e o \
\ ,f
PRO*^
Final Regulatory Impact Analysis for the
Reconsideration of the National Ambient Air
Quality Standards for Particulate Matter
-------
ii
-------
EPA-452/R-24-006
January 2024
Final Regulatory Impact Analysis for the Reconsideration of the National Ambient Air
Quality Standards for Particulate Matter
U.S. Environmental Protection Agency
Office of Air Quality Planning and Standards
Health and Environmental Impacts Division
Research Triangle Park, NC
iii
-------
CONTACT INFORMATION
This document has been prepared by staff from the Office of Air Quality Planning and
Standards, U.S. Environmental Protection Agency. Questions related to this document
should be addressed to U.S. Environmental Protection Agency, Office of Air Quality
Planning and Standards, C439-02, Research Triangle Park, North Carolina 27711 (email:
oaqpseconomics@epa.gov). The docket for the final regulatory impact analysis is EPA-HQ-
OAR-2019-0587. The docket for the notice of final rulemaking is EPA-HQ-OAR-2015-0072.
ACKNOWLEDGEMENTS
In addition to EPA staff from the Office of Air Quality Planning and Standards, personnel
from the Office of Policy's National Center for Environmental Economics contributed data
and analysis to this document.
iv
-------
TABLE OF CONTENTS
LIST OF TABLES xi
LIST OF FIGURES xvii
EXECUTIVE SUMMARY 1
Overview of the Final Rule 1
Overview of the Regulatory Impact Analysis 2
ES.l Design of the Regulatory Impact Analysis 3
ES.1.1 Establishing the Analytical Baseline 6
ES.l.2 Estimating PM2.5 Emissions Reductions Needed for Annual and 24-hour Revised and
Alternative Standard Levels Analyzed 8
ES.l.3 Control Strategies and PM2.5 Emissions Reductions 11
ES.l.4 Estimates of PM2.5 Emissions Reductions Still Needed after Applying Controls 12
ES.l.5 Engineering Costs 13
ES.l.6 Human Health Benefits 15
ES.l.7 Welfare Benefits of Meeting the Primary and Secondary Standards 19
ES.l.8 Environmental Justice 20
ES.2 Qualitative Assessment of the Remaining Air Quality Challenges 22
ES.3 Changes in Data Used and Methods Between Proposal and Final RIAs 24
ES.4 Results of Benefit-Cost Analysis 26
ES.5 References 29
CHAPTER 1: OVERVIEW AND BACKGROUND 31
Overview of the Final Rule 31
Overview of the Regulatory Impact Analysis 31
1.1 Background 33
1.1.1 National Ambient Air Quality Standards 3 4
1.1.2 Role of Executive Orders in the Regulatory Impact Analysis 35
1.1.3 Nature of the Analysis 36
1.2 The Need for National Ambient Air Quality Standards 36
1.3 Design ofthe Regulatory Impact Analysis 37
1.3.1 Establishing the Baseline for Evaluating Revised and Alternative Standard
Levels 38
1.3.2 Cost Analysis Approach 40
1.3.3 Benefits Analysis Approach 40
1.3.4 Welfare Benefits of Meeting the Primary and Secondary Standards 41
1.4 Organization of the Regulatory Impact Analysis 41
1.5 References 43
CHAPTER 2: AIR QUALITY MODELING AND METHODS 45
Overview 45
2.1 PM2.5 Characteristics 46
2.1.1 PM2.5 Size and Composition 46
2.1.2 PM2.5 Regional Characteristics 49
2.1.3 PM2.5 Trends 51
2.2 Modeling PM2.5 in the Future 56
2.2.1 Air Quality Modeling Platform 57
2.2.1.1 Model Configuration 57
v
-------
2.2.1.2 Emissions Inventory 59
2.2.1.3 Model Evaluation 61
2.2.2 Future-Year PM2.5 Design Values 62
2.3 Calculating Emission Reductions for Meeting the Existing, Revised, and Alternative
Standard Levels 64
2.3.1 Developing Air Quality Ratios 64
2.3.2 Emission Reductions to Meet 12/35 68
2.3.3 Emission Reductions to Meet Revised and Alternative Standards 70
2.3.4 Limitations of Using Air Quality Ratios 72
2.4 Description of Air Quality Challenges in Select Areas 73
2.4.1 Cincinnati, OH and Fort Lee, NJ Near-Road Sites 73
2.4.2 Border Areas 75
2.4.2.1 Imperial County, CA 75
2.4.2.2 Cameron and Hidalgo County, TX 77
2.4.3 Small Mountain Valleys in the West 79
2.4.4 California Areas 83
2.4.4.1 San Joaquin Valley, CA 83
2.4.4.2 South Coast Air Basin, CA 87
2.4.4.3 Additional California Areas 90
2.4.5 Additional Considerations 92
2.5 Calculating PM2.5 Concentration Fields for Standard Combinations 93
2.5.1 Creating the PM2.5 Concentration Field for 2032 93
2.5.2 Creating Spatial Fields Corresponding to Meeting Standards 94
2.6 References 96
APPENDIX 2A: ADDITIONAL AIR QUALITY MODELING INFORMATION 101
Overview 101
2A.1 2018 CMAQ Modeling 102
2A.1.1 Model Configuration 103
2A.1.2 Model Performance Evaluation 104
2A.2 Projecting PM2.5 DVs to 2032 119
2A.2.1 Monitoring Data for PM2.5 Projections 120
2A.2.2 Future-Year PM2.5 Design Values 135
2A.3 Developing Air Quality Ratios and Estimating Emission Reductions 141
2A.3.1 Developing Air Quality Ratios for Primary PM2.5 Emissions 141
2A.3.2 Developing Air Quality Ratios for NOx in Southern California 145
2A.3.3 Developing Air Quality Ratios for NOx in SJV, CA 147
2A.3.4 Applying Air Quality Ratios to Estimate Emission Reductions 149
2A.3.4.1 Emission Reductions Needed to Meet 12/35 150
2A.3.4.2 Emission Reductions Needed to Meet 10/35, 9/35, 8/35, and 10/30 154
2A.4 Calculating PM2.5 Concentration Fields for Standard Combinations 158
2A.4.1 Creating the PM2.5 Concentration Field for 2032 158
2A.4.2 Creating Spatial Fields Corresponding to Meeting Standards 160
2A.5 References 163
CHAPTER 3: CONTROL STRATEGIES AND PM2.5 EMISSIONS REDUCTIONS 168
Overview 168
3.1 Preparing the 12/35 |a,g/m3 Analytical Baseline 171
3.2 Illustrative Control Strategies and PM2.5 Emissions Reductions from the Analytical
Baseline 172
vi
-------
3.2.1 Estimating PM2.5 Emissions Reductions Needed for Annual and 24-hour Revised and
Alternative Standard Levels Analyzed 173
3.2.2 Applying End-of-Pipe and Area Source Controls 177
3.2.3 Estimates of PM2.5 Emissions Reductions Resulting from Applying Control
Technologies 182
3.2.4 Estimates of PM2.5 Emissions Reductions Still Needed after Applying End-of-Pipe and
Area Source Controls 190
3.2.5 Qualitative Assessment of the Remaining Air Quality Challenges and Emissions
Reductions Potentially Still Needed 201
3.2.5.1 Near-Road Monitors (Northeast) 202
3.2.5.2 Border Areas (Southeast, California) 205
3.2.5.3 Small Mountain Valleys (West) 208
3.2.5.4 California Areas 210
3.3 Limitations and Uncertainties 214
3.4 References 217
APPENDIX 3 A: CONTROL STRATEGIES AND PM2.5 EMISSIONS REDUCTIONS 218
Overview 218
3A.1 Types of Control Technologies 218
3A.1.1 PM Controls for Non-EGU Point Sources 218
3A.1.2 PM Controls for Non-point (Area) Sources 219
3 A.2 EGU Trends Reflected in EPA's Integrated Planning Model (IPM) v6 Platform, Post-IRA
2022 Reference Case Projections 220
3A.3 Applying End-of-Pipe and Area Source Controls 222
CHAPTER 4: ENGINEERING COST ANALYSIS AND QUALITATIVE DISCUSSION OF
SOCIAL COSTS 250
Overview 250
4.1 Estimating Engineering Costs 251
4.1.1 Methods, Tools, and Data 252
4.1.2 Cost Estimates for the Control Strategies 253
4.2 Limitations and Uncertainties in Engineering Cost Estimates 260
4.3 Social Costs 261
4.4 References 267
APPENDIX 4A: ENGINEERING COST ANALYSIS 269
Overview 269
4A.1 Estimated Costs by County for Revised and Alternative Standard Levels 269
CHAPTER 5: BENEFITS ANALYSIS APPROACH AND RESULTS 275
Overview 275
5.1 Human Health Benefits Analysis Methods 279
5.1.1 Selecting Air Pollution Health Endpoints to Quantify 279
5.1.2 Calculating Counts of Air Pollution Effects Using the Health Impact Function 281
5.1.3 Calculating the Economic Valuation of Health Impacts 283
5.2 Benefits Analysis Data Inputs 283
5.2.1 Demographic Data 284
5.2.2 Baseline Incidence and Prevalence Estimates 285
5.2.3 Effect Coefficients 288
5.2.3.1 PM2.5 Premature Mortality Effect Coefficients for Adults 288
vii
-------
5.2.4 Unquantified Human Health Benefits 290
5.2.5 Unquantified Welfare Benefits 293
5.2.5.1 Visibility Impairment Benefits 295
5.2.6 Climate Effects of PM2.5 296
5.2.6.1 Climate Effects of Carbonaceous Particles 297
5.2.6.2 Climate Effects: Summary and Conclusions 298
5.2.7 Economic Valuation Estimates 298
5.3 Characterizing Uncertainty 299
5.3.1 Monte Carlo Assessment 300
5.3.2 Sources of Uncertainty Treated Qualitatively 301
5.4 Benefits Results 302
5.4.1 Benefits of the Applied Control Strategies for the Revised and Alternative
Combinations of Primary PM2.5 Standard Levels 302
5.5 Discussion 308
5.6 References 311
APPENDIX 5A: BENEFITS OF THE REVISED AND ALTERNATIVE STANDARD LEVELS
318
Overview 318
5A.1 Benefits of the Revised, Less, and More Stringent Alternative Standard Levels of Primary
PM2.5 Standards 319
5A.2 References 325
CHAPTER 6: ENVIRONMENTAL JUSTICE 326
6.1 Analyzing EJ Impacts in This Final Action 326
6.2 Introduction 329
6.3 EJ Analysis of Exposures Under Current, Revised, and Alternative Standard Levels 331
6.3.1 National 333
6.3.1.1 Absolute Exposures Under Current and Alternative Standard Levels and Exposure
Changes When Moving from Current to Revised and Alternative Standard
Levels 334
6.3.1.2 Proportional Exposure Changes When Moving from Currentto Revised and
Alternative Standard Levels 344
6.3.2 Regional 346
6.3.2.1 Absolute Exposures Under Current, Revised, and Alternative Standard Levels 346
6.3.2.2 Absolute Exposure Changes When Moving from Currentto Revised and Alternative
Standard Levels 352
6.3.2.3 Proportional Exposure Changes When Moving from Currentto Revised and
Alternative Standard Levels 355
6.4 EJ Analysis of Health Effects under Current, Revised, and Alternative Standard
Levels 355
6.4.1 National 358
6.4.1.1 Absolute Mortality Rates Under Current, Revised, and Alternative Standard Levels
and Mortality Rate Changes When Moving from Current to Revised and Alternative
Standard Levels 358
6.4.1.2 Proportional Mortality Rate Changes When Moving from Current to Revised and
Alternative Standard Levels 361
6.4.2 Regional 363
6.4.2.1 Absolute Mortality Rates Under Current, Revised, and Alternative Standard
Levels 363
viii
-------
6.4.2.2 Absolute Mortality Rate Changes When Moving from Current to Revised and
Alternative Standard Levels 366
6.4.2.3 Proportional Mortality Rate Changes When Moving from Current to Revised and
Alternative Standard Levels 368
6.5 Summary 369
6.6 Environmental Justice Appendix 372
6.6.1 EJ Exposure Analysis Input Data 372
6.6.1.1 Educational Attainment 373
6.6.1.2 Poverty Status 374
6.6.1.3 Unemployment Status 374
6.6.1.4 Health Insurance Status 374
6.6.1.5 Linguistic Isolation 375
6.6.1.6 Redlined Areas 375
6.6.2 EJ Health Effects Analysis Input Data 375
6.6.3 National EJ Analysis of Total Exposures and Exposure Changes Associated with
Meeting the Revised and Alternative Standard Levels 379
6.6.3.1 Absolute National Exposures Under Current, Revised, and Alternative Standard
Levels and Exposure Changes When Moving from Current Standard to Revised and
Alternative Standard Levels 379
6.6.3.2 Proportional Regional Exposure Changes When Moving from Current to Revised and
Alternative Standard Levels 384
6.6.4 Regional EJ Analysis of Total Exposures and Exposure Changes Associated with
Meeting the Standards 386
6.6.4.1 Absolute Regional Exposures Under Current, Revised, and Alternative Standard
Levels 386
6.6.4.2 Absolute Regional Exposure Changes When Moving from Current Standard to
Revised and Alternative Standard Levels 388
6.6.4.3 Proportional Regional Exposure Changes When Moving from Current to Revised and
Alternative Standard Levels 390
6.6.5 National EJ Analysis of Total Mortality Rates and Rate Changes Associated with
Meeting the Standards 392
6.6.5.1 Absolute Mortality Rates Under Current and Alternative Standard Levels and
Mortality Rate Reductions When Moving from Current to Revised and Alternative
Standard Levels 392
6.6.5.2 Proportional Regional Exposure Changes When Moving from Current to Revised and
Alternative Standard Levels 395
6.6.6 Regional National EJ Analysis of Total Mortality Rates and Mortality Rate Changes
Associated with Meeting the Standards 396
6.6.6.1 Absolute Regional Mortality Rates Under Current, Revised, and Alternative Standard
Levels 396
6.6.6.2 Absolute Regional Exposure Changes When Moving from Current to Revised and
Alternative Standard Levels 399
6.6.6.3 Proportional Regional Exposure Changes When Moving from Current to Revised and
Alternative Standard Levels 400
6.7 References 402
CHAPTER 7: LABOR IMPACTS 404
Overview 404
7.1 Labor Impacts 404
7.2 References 409
ix
-------
CHAPTER 8: COMPARISON OF BENEFITS AND COSTS 410
Overview 410
8.1 Results 411
8.2 Limitations of Present Value Estimates 418
8.3 References 421
x
-------
LIST OF TABLES
Table ES-1 Summary of PM2.5 Emissions Reductions Needed by Area in 2032 to Meet Current
Primary Annual and 24-hour Standards of 12/35 |a,g/m3 (tons/year) 7
Table ES-2 By Area, Summary of PM2.5 Emissions Reductions Needed, InTons/Year and as
Percent of Total Reduction Needed Nationwide, for the Revised and Alternative
Primary Standard Levels of 10/35 |a,g/m3,10/30 |a,g/m3, 9/35 |a,g/m3, and 8/35
|a,g/m3 in 2032 9
Table ES-3 Summary of PM2.5 Estimated Emissions Reductions from CoST by Area for the
Revised and Alternative Primary Standard Levels of 10/35 |a,g/m3,10/30 |a,g/m3,
9/35 |a,g/m3, and 8/35 |a,g/m3 in 2032 (tons/year) 12
Table ES-4 Summary of PM2.5 Emissions Reductions Still Needed by Area for the Revised and
Alternative Primary Standard Levels of 10/35 |a,g/m3,10/30 |a,g/m3, 9/35 |a,g/m3,
and 8/35 |a,g/m3 in 2032 (tons/year) 13
Table ES-5 By Area, Summary of Annualized Control Costs for the Revised and Alternative
Primary Standard Levels of 10/35 |a,g/m3,10/30 |a,g/m3, 9/35 |a,g/m3, and 8/35
|a,g/m3 for 2032 (millions of 2017$) 14
Table ES-6 Estimated Avoided PM-Related Premature Mortalities and Illnesses of the Control
Strategies for the Revised and Alternative Primary PM2.5 Standard Levels for 2032
(95% Confidence Interval) 17
Table ES-7 Estimated Monetized Benefits of the Control Strategies for the Revised and
Alternative Primary PM2.5 Standard Levels in 2032, Incremental to Attainment of
12/35 |J.g/m3 (billions of 2017$) 18
Table ES-8 Estimated Monetized Benefits by Area of the Control Strategies for the Revised and
Alternative Primary PM2.5 Standard Levels in 2032, Incremental to Attainment of
12/35 |J.g/m3 (billions of 2017$) 19
Table ES-9 Summary of Counties by Bin that Still Need Emissions Reductions for the Revised
Primary Standard Levels of 9/35 |a,g/m3 23
Table ES-10 Estimated Monetized Benefits, Costs, and Net Benefits of the Control Strategies
Applied Toward the Primary Revised and Alternative Standard Levels of 10/35
|a,g/m3,10/30 |a,g/m3, 9/35 |a,g/m3, and 8/35 |a,g/m3 in 2032 for the U.S. (millions of
2017$) 27
Table ES-11 Summary of Present Values and Equivalent Annualized Values for Estimated
Monetized Compliance Costs, Benefits, and Net Benefits of the Control Strategies
Applied Toward the Revised Primary Standard Levels of 9/35 |a,g/m3 (millions of
2017$, 2032-2051, discounted to 2023 using 3 and 7 percent discount rates) 28
Table 2-1 Annual and 24-Hour Air Quality Ratios for Primary PM2.5 Emissions 67
Table 2-2 Information on Areas with Challenging Residential Wood Combustion Issues 83
Table 2-3 Design Value Information for Additional California Areas 92
Table 2A-1 Definition of Statistics Used in the CMAQ Model Performance Evaluation 107
Table 2A-2 CMAQ Performance Statistics for PM2.5 at AQS Sites in 2018 109
XI
-------
Table 2A-3
CMAQ Performance Statistics for PM2.5 Sulfate at CSN and IMPROVE Sites in 2018
Ill
Table 2A-5
Table 2A-6
Table 2A-7
Table 2A-8
Table 2A-9
Table 2A-10
Table 2A-11
Table 2A-12
Table 2A-13
Table 2A-14
Table 3-1
Table 3-2
Table 3-3
Table 3-4
Table 3-5
Table 3-6
Table 2A-4 CMAQ Performance Statistics for PM2.5 Nitrate at CSN and IMPROVE Sites in 2018
113
CMAQ Performance Statistics for PM2.5 EC at CSN and IMPROVE Sites in 2018 115
CMAQ Performance Statistics for PM2.5 OC at CSN and IMPROVE Sites in 2018 117
November Wildfire Episodes and Counties Where Data Were Excluded if PM2.5
Concentrations Exceeded the Extreme Value Threshold of 64 [ig nr3 122
PM2.5 DVs for 2032 Projection and 12/35 Analytical Baseline for the Highest DVs in
the County for Counties with Annual 2032 DVs Greater 8 jug nr3 or 24-hour 2032
DVs Greater than 30 jug nr3 138
Annual and 24-Hour Air Quality Ratios for Primary PM2.5 Emissions 143
County Groups for Calculating Air Quality Ratios for NOx Emission Changes in
Southern California 146
2032 PM2.5 DVs and NOx-adjusted PM2.5 DVs for the Highest Annual and 24-Hour DV
Monitors in South Coast Counties 147
2032 PM2.5 DVs and NOx-adjusted PM2.5 DVs for the Highest Annual and 24-Hour DV
Monitors in SJV Counties 149
Summary of Primary PM2.5 Emissions Reductions by County Needed to Meet the
Existing Standards (12/35) for Counties with 2032a Annual DVs greater than 8 jug
m-3 or 24-Hour DVs Greater than 30 jug nr3 151
Primary PM2.5 Emission Reductions Needed to Meet the Revised and Alternative
Standard Levels of 10/35,10/30, 9/35, and 8/35 Relative to the 12/35 Analytical
Baseline 155
Summary of PM2.5 Emissions Reductions Needed by Area in 2032 to Meet Current
Primary Annual and 24-hour Standards of 12/35 |a,g/m3 (tons/year) 172
By Area, Summary of PM2.5 Emissions Reductions Needed, inTons/Year and as
Percent of Total Reductions Needed Nationwide, for Revised and Alternative
Primary Standard Levels of 10/35 |a,g/m3,10/30 |a,g/m3, 9/35 |a,g/m3, and 8/35
|a,g/m3 in 2032 175
By Inventory Sector, Controls Applied in Analyses of the Current Standards and the
Revised and Alternative Primary Standard Levels 182
Summary of PM2.5 Estimated Emissions Reductions from CoST by Area for the
Revised and Alternative Primary Standard Levels of 10/35 |a,g/m3,10/30 |a,g/m3,
9/35 |a,g/m3, and 8/35 |a,g/m3 in 2032 (tons/year) 183
Summary of PM2.5 Emissions and Estimated Emissions Reductions from CoST by
Inventory Sector for Revised and Alternative Primary Standard Levels of 10/35
|a,g/m3,10/30 |a,g/m3, 9/35 |a,g/m3, and 8/35 |a,g/m3 in 2032 (tons/year) 185
Summary of Estimated Emissions Reductions from CoST by Inventory Sector and
Control Technology for Revised and Alternative Primary Standard Levels of 10/35
|a,g/m3,10/30 |a,g/m3, 9/35 |a,g/m3, and 8/35 |a,g/m3 in 2032 (tons/year) 187
-------
Table 3-7 Summary of Estimated PM2.5 Emissions Reductions from CoST by Inventory Source
Classification Code Sectors for Revised and Alternative Primary Standard Levels of
10/35 |a,g/m3,10/30 |a,g/m3, 9/35 ng/m3, and 8/35 |a,g/m3 in 2032 (tons/year)... 188
Table 3-8 Summary of PM2.5 Emissions Reductions Still Needed by Area for the Revised and
Alternative Primary Standard Levels of 10/35 |a,g/m3,10/30 |a,g/m3, 9/35 |a,g/m3,
and 8/35 |a,g/m3 in 2032 (tons/year) 192
Table 3-9 Summary of PM2.5 Emissions Reductions Still Needed by Area and by County for the
Revised and Alternative Primary Standard Levels of 10/35 |a,g/m3,10/30 |a,g/m3,
9/35 |a,g/m3, and 8/35 |a,g/m3 in 2032 (tons/year) 192
Table 3-10 Current Rules from Several California Air Districts and City of Portola for Area
Fugitive Dust Emissions, Non-point Source Emissions, and Residential Wood
Combustion Emissions 197
Table 3-11 Voluntary Measures from the Houston-Galveston Area Advance Plan for PM 200
Table 3-12 Summary of Counties by Bin that Still Need Emissions Reductions for Revised
Primary Standard Levels of 9/35 |a,g/m3 202
Table 3-13 Summary of Estimated PM2.5 Emissions Reductions Needed and Emissions
Reductions Identified by CoST for the West for the Revised Primary Standard Levels
of 9/35 |a,g/m3 in 2032 (tons/year) 208
Table 3 A-1 By Area and Emissions Inventory Sector, Controls Applied in Analyses of the Current
Standards, Revised, and Alternative Primary Standard Levels 225
Table 3A-2 Summary of PM2.5 Estimated Emissions Reductions from CoST for the Northeast (31
counties) for Revised and Alternative Primary Standard Levels of 10/35 |a,g/m3,
10/30 |a,g/m3, 9/35 |a,g/m3, and 8/35 |a,g/m3 in 2032 (tons/year) 227
Table 3A-3 Summary of PM2.5 Estimated Emissions Reductions from CoST for the Adjacent
Counties in the Northeast (51 counties) for Revised and Alternative Primary
Standard Levels of 10/35 |a,g/m3,10/30 |a,g/m3, 9/35 |a,g/m3, and 8/35 |a,g/m3 in
2032 (tons/year) 229
Table 3A-4 Summary of PM2.5 Estimated Emissions Reductions from CoST for the Southeast (33
counties) for Revised and Alternative Primary Standard Levels of 10/35 |a,g/m3,
10/30 |a,g/m3, 9/35 |a,g/m3, and 8/35 |a,g/m3 in 2032 (tons/year) 231
Table 3A-5 Summary of PM2.5 Estimated Emissions Reductions from CoST for the Adjacent
Counties in the Southeast (34 counties) for Revised Alternative Primary Standard
Levels of 10/35 |a,g/m3,10/30 |a,g/m3, 9/35 |a,g/m3, and 8/35 |a,g/m3 in 2032
(tons/year) 232
Table 3A-6 Summary of PM2.5 Estimated Emissions Reductions from CoST for the West (29
counties) for Revised and Alternative Primary Standard Levels of 10/35 |a,g/m3,
10/30 |a,g/m3, 9/35 |a,g/m3, and 8/35 |a,g/m3 in 2032 (tons/year) 232
Table 3A-7 Summary of PM2.5 Estimated Emissions Reductions from CoST for California (36
counties) for Revised and Alternative Primary Standard Levels of 10/35 |a,g/m3,
10/30 |a,g/m3, 9/35 |a,g/m3, and 8/35 |a,g/m3 in 2032 (tons/year) 234
Table 3A-8 Remaining PM2.5 Emissions and Potential Additional Reduction Opportunities 235
xiii
-------
Table 4-1 By Area, Summary of Annualized Control Costs for Revised and Alternative Primary
Standard Levels of 10/35 ng/m3,10/30 ng/m3, 9/35 ng/m3, and 8/35 |a,g/m3 for
2032 (millions of 2017$) 254
Table 4-2 By Emissions Inventory Sector, Summary of Annualized Control Costs for Revised
and Alternative Primary Standard Levels of 10/35 |a,g/m3,10/30 |a,g/m3, 9/35
|a,g/m3, and 8/35 |a,g/m3 for 2032 (millions of 2017$) 256
Table 4-3 By Area and by Emissions Inventory Sector, Summary of Annualized Control Costs
for Revised and Alternative Primary Standard Levels of 10/35 |a,g/m3,10/30 |a,g/m3,
9/35 |a,g/m3, and 8/35 |a,g/m3 for 2032 (millions of 2017$) 256
Table 4-4 By Control Technology, Summary of Annualized Control Costs for Revised and
Alternative Primary Standard Levels of 10/35 |a,g/m3,10/30 |a,g/m3, 9/35 |a,g/m3,
and 8/35 |a,g/m3 for 2032 (millions of 2017$) 258
Table 4-5 By Emissions Inventory Sector and Control Technology, Summary of Annualized
Control Costs for Revised and Alternative Primary Standard Levels of 10/35 |a,g/m3,
10/30 |a,g/m3, 9/35 |a,g/m3, and 8/35 |a,g/m3 for 2032 (millions of 2017$) 259
Table 4A-1 Summary of Estimated Annual Control Costs for the Northeast (31 counties) for
Revised and Alternative Primary Standard Levels of 10/35 |a,g/m3,10/30 |a,g/m3,
9/35 |a,g/m3, and 8/35 |a,g/m3 for 2032 (millions of 2017$) 269
Table 4A-2 Summary of Estimated Annual Control Costs for Adjacent Counties in the Northeast
(51 counties) for Revised and Alternative Primary Standard Levels of 10/35 |a,g/m3,
10/30 |a,g/m3, 9/35 |a,g/m3, and 8/35 |a,g/m3 for 2032 (millions of 2017$) 270
Table 4A-3 Summary of Estimated Annual Control Costs for the Southeast (33 counties) for
Revised and Alternative Primary Standard Levels of 10/35 |a,g/m3,10/30 |a,g/m3,
9/35 |a,g/m3, and 8/35 |a,g/m3 for 2032 (millions of 2017$) 271
Table 4A-4 Summary of Estimated Annual Control Costs for Adjacent Counties in the Southeast
(34 counties) for Revised and Alternative Primary Standard Levels of 10/35 |a,g/m3,
10/30 |a,g/m3, 9/35 |a,g/m3, and 8/35 |a,g/m3 for 2032 (millions of 2017$) 272
Table 4A-5 Summary of Estimated Annual Control Costs for the West (29 counties) for Revised
and Alternative Primary Standard Levels of 10/35 |a,g/m3,10/30 |a,g/m3, 9/35
|a,g/m3, and 8/35 |a,g/m3 for 2032 (millions of 2017$) 273
Table 4A-6 Summary of Estimated Annual Control Costs for California (36 counties) for Revised
and Alternative Primary Standard Levels of 10/35 |a,g/m3,10/30 |a,g/m3, 9/35
|a,g/m3, and 8/35 |a,g/m3 for 2032 (millions of 2017$) 274
Table 5-1 Estimated Monetized Benefits of the Applied Control Strategies for the Revised and
Alternative Primary PM2.5 Standard Levels in 2032, Incremental to Attainment of
12/35 (billions of 2017$) 278
Table 5-2 Human Health Effects of Pollutants Potentially Affected by Attainment of the
Primary PM2.5 NAAQS 281
Table 5-3 Baseline Incidence Rates for Use in Impact Functions 286
Table 5-4 Causal Determinations Identified in Integrated Science Assessment for Oxides of
Nitrogen, Oxides of Sulfur, and Particulate Matter — Ecological Criteria 2020b 294
xiv
-------
Table 5-5 Estimated Avoided PM-Related Premature Mortalities and Illnesses of the Applied
Control Strategies for the Revised and Alternative Primary Standard Levels of 10/35
Hg/m3,10/30 |a,g/m3, 9/35 ng/m3, and 8/35 |a,g/m3 for 2032 (95% Confidence
Interval) 304
Table 5-6 Monetized PM-Related Premature Mortalities and Illnesses of the Applied Control
Strategies for the Revised and Alternative Primary Standard Levels of 10/35 |a,g/m3,
10/30 |a,g/m3, 9/35 |a,g/m3, and 8/35 |a,g/m3 for 2032 (Millions of 2017$, 3%
discount rate; 95% Confidence Interval) 305
Table 5-7 Monetized PM-Related Premature Mortalities and Illnesses of the Applied Control
Strategies for the Revised and Alternative Primary Standard Levels of 10/35 |a,g/m3,
10/30 |a,g/m3, 9/35 |a,g/m3, and 8/35 |a,g/m3 for 2032 (Millions of 2017$, 7%
discount rate; 95% Confidence Interval) 306
Table 5-8 Estimated Monetized Benefits of the Applied Control Strategies for the Revised and
Alternative Primary Standard Levels of 10/35 |a,g/m3,10/30 |a,g/m3, 9/35 |a,g/m3,
and 8/35 |a,g/m3 for 2032, Incremental to Attainment of 12/35 (billions of 2017$)
307
Table 5-9 Estimated Monetized Benefits by Area of the Applied Control Strategies for the
Revised and Alternative Primary Standard Levels of 10/35 |a,g/m3,10/30 |a,g/m3,
9/35 |a,g/m3, and 8/35 |a,g/m3 for 2032, Incremental to Attainment of 12/35 (billions
of 2017$) 308
Table 5 A-1 Estimated Avoided PM-Related Premature Mortalities and Illnesses of Meeting the
Revised and Alternative Primary PM2.5 Standard Levels for 2032 (95% Confidence
Interval) 320
Table 5 A-2 Monetized Avoided PM-Related Premature Mortalities and Illnesses of Meeting the
Revised and Alternative Primary PM2.5 Standard Levels for 2032 (Millions of 2017$,
3% discount rate; 95% Confidence Interval) 321
Table 5 A-3 Monetized Avoided PM-Related Premature Mortalities and Illnesses of Meeting the
Revised and Alternative Primary PM2.5 Standard Levels for 2032 (Millions of 2017$,
7% discount rate; 95% Confidence Interval) 322
Table 5A-4 Total Estimated Monetized Benefits of Meeting the Revised and Alternative Primary
Standard Levels in 2032, Incremental to Attainment of 12/35 (billions of 2017$)
323
Table 5 A-5 Total Estimated Monetized Benefits by Area of Meeting the Revised and Alternative
Primary Standard Levels in 2032, Incremental to Attainment of 12/35 (billions of
2017$) 324
Table 6-1 Populations Included in the PM2.5 Exposure Analysis 333
Table 6-2 Hazard Ratios, Beta Coefficients, and Standard Errors (SE) from (Di et al., 2017)
377
Table 7-1 Baseline Industry Employment 406
Table 7-2 Employment per $1 Million Output (2017$) by Industry (4-digit NAICS) 408
Table 8-1 Estimated Monetized Benefits, Costs, and Net Benefits of the Control Strategies
Applied Toward Primary Revised and Alternative Standard Levels of 10/35 |a,g/m3,
xv
-------
10/30 |a,g/m3, 9/35 |a,g/m3, and 8/35 |a,g/m3 in 2032 for the U.S. (millions of 2017$)
412
Table 8-2 Summary of Present Values and Equivalent Annualized Values for Estimated
Monetized Compliance Costs of the Control Strategies Applied Toward the Primary
Revised and Alternative Standard Levels of 10/35 ng/m3,10/30 |a,g/m3, 9/35 |a,g/m3
8/35 |a,g/m3 (millions of 2017$, 2032-2051, discounted to 2023, 3 percent discount
rate) 414
Table 8-3 Summary of Present Values and Equivalent Annualized Values for Estimated
Monetized Compliance Costs of the Control Strategies Applied Toward the Primary
Revised and Alternative Standard Levels of 10/35 |a,g/m3,10/30 |a,g/m3, 9/35 |a,g/m3
8/35 |a,g/m3 (millions of 2017$, 2032-2051, discounted to 2023, 7 percent discount
rate) 415
Table 8-4 Summary of Present Values and Equivalent Annualized Values for Estimated
Monetized Benefits of the Control Strategies Applied Toward the Primary Revised
and Alternative Standard Levels of 10/35 |a,g/m3,10/30 |a,g/m3, 9/35 |a,g/m3 8/35
|a,g/m3 (millions of 2017$, 2032-2051, discounted to 2023, 3 percent discount rate)
416
Table 8-5 Summary of Present Values and Equivalent Annualized Values for Estimated
Monetized Benefits of the Control Strategies Applied Toward the Primary Revised
and Alternative Standard Levels of 10/35 |a,g/m3,10/30 |a,g/m3, 9/35 |a,g/m3 8/35
|a,g/m3 (millions of 2017$, 2032-2051, discounted to 2023, 7 percent discount rate)
417
Table 8-6 Summary of Present Values and Equivalent Annualized Values for Estimated
Monetized Compliance Costs, Benefits, and Net Benefits of the Control Strategies
Applied Toward the Revised Primary Alternative Standard Levels of 9/3 5 |a,g/m3
(millions of 2017$, 2032-2051, discounted to 2023 using 3 and 7 percent discount
rates) 418
xvi
-------
LIST OF FIGURES
Figure ES-1 Geographic Areas Used in Analysis 7
Figure ES-2 Counties Projected to Exceed in Analytical Baseline for the Revised and Alternative
Standard Levels of 10/35 ng/m3, 9/35 ng/m3, and 8/35 |a,g/m3 10
Figure ES-3 Counties that Still Need PM2.5 Emissions Reductions for the Revised Standard Levels
of 9/35 |a,g/m3 13
Figure 2-1 Annual Average PM2.5 Concentrations over the U.S. in 2019 Based on the Hybrid
Satellite Modeling Approach of van Donkelaar etal. (2021) 49
Figure 2-2 Seasonally Weighted Annual Average PM2.5 Concentrations in the U.S. from 2000 to
2019 (406 sites) 52
Figure 2-3 National Emission Trends of PM2.5, PM10, and Precursor Gases from 1990 to 2017
52
Figure 2-4 Annual Anthropogenic Source Sector Emission Totals (1000 tons per year) for NOx,
SO2, and PM2.5 for 2018 and 2032 54
Figure 2-5 Gridded PM2.5 Concentrations over Selected Urban Areas Based on the 2032
Modeling Case Described Below with the Enhanced Voronoi Neighbor Averaging
Approach 56
Figure 2-6 Map of the 12US2 (12 x 12 km Horizontal Resolution) Modeling Domain 58
Figure 2-7 Regional Groupings for Calculating Air Quality Ratios 66
Figure 2-8 Counties with Projected 2032 PM2.5 DVs that Exceed the 24-Hour (Daily Only),
Annual (Annual Only) or Both (Both) Existing Standards (12/35 jug nr3) 69
Figure 2-9 Counties with PM2.5 DVs that Exceed Alternative Annual (Annual Only), 24-Hour
(Daily Only), or Both (Both) Standards in the 12/35 Analytical Baseline 71
Figure 2-10 Total Primary PM2.5 Emission Reductions Needed to Meet the Revised and
Alternative Standard Levels of 10/35,10/30, 9/35, and 8/35 Relative to the 12/35
Analytical Baseline in the Eastern and Western U.S 72
Figure 2-11 Cincinnati Near-Road Site (Left) and Storage Building Construction Near the Site
(Right) 74
Figure 2-12 Fort Lee, NJ Near-Road Site (Red Balloon) and Roadway Exchange Leading to George
Washington Bridge 75
Figure 2-13 Imperial County and the Nonattainment Area 76
Figure 2-14 Nighttime Aerial View of Calexico, CA and Mexicali, MX 77
Figure 2-15 Annual Source Sector Emission Totals for PM2.5 for 2018 and 2032 in Imperial
County 77
Figure 2-16 Location of Mission and Brownsville Monitors in Hidalgo and Cameron County,
respectively, with Annual Wind Patterns from Meteorological Measurements 79
Figure 2-17 Annual Source Sector Emission Totals for PM2.5 for 2018 and 2032 in Cameron and
Hidalgo County Combined 79
XVll
-------
Figure 2-18 Air Pollution Layer Associated with a Temperature Inversion in Missoula, MT in
November 2018 80
Figure 2-19 Plumas County, CA (Grey), Portola Nonattainment Area (Red), and City of Portola
(Purple) 81
Figure 2-20 Lincoln County, MT (Grey), Libby Nonattainment Area (Red), and City of Libby
(Purple) 81
Figure 2-21 San Joaquin Valley Nonattainment Area and Location of Highest PM2.5 Monitor in
Bakersfield (06-029-0016) 84
Figure 2-22 Recent Annual PM2.5 DVs at the Highest SJV Monitor for Design Value Periods (e.g.,
11-13: 2011-2013). Dashed line is the 2012 Annual PM2.5 NAAQS Level (12 jug nr3)
85
Figure 2-23 Decrease in the Number of Days SJV Exceeded the 24-hr NAAQS Level (35 jug nr3)
85
Figure 2-24 Annual Source Sector PM2.5 Emission Totals in SJV Counties for 2032 Modeling Case
87
Figure 2-25 South Coast Air Basin Nonattainment Area and Locations of Highest PM2.5 Monitors
in Los Angeles (06-037-1302), Riverside (06-065-8005), and San Bernardino (06-
071-0027) 88
Figure 2-26 Recent Annual PM2.5 DVs at the Highest South Coast Monitor for Design Value
Periods (e.g., 11-13: 2011-2013). Dashed line is the 2012 Annual PM2.5 NAAQS Level
(12 |a,gm-3) 89
Figure 2-27 Annual Source Sector PM2.5 Emission Totals in the SoCAB Counties for 2032
Modeling Case 90
Figure 2-28 PM2.5 Concentration for 2032 based on eVNA Method 94
Figure 2-29 PM2.5 Concentration Improvement Associated with Meeting 9/35 Relative to the
12/35 Analytical Baseline 95
Figure 2A-1 Map of the 12US2 (12x12 km Horizontal Resolution) Modeling Domain Used for
the PM NAAQS RIA 104
Figure 2A-2 U.S. Climate Regions (Karl and Koss, 1984) Used in the CMAQ Model Performance
Evaluation 107
Figure 2A-3 Comparison of CMAQ Predictions of PM2.5 and Observations at AQS Sites for County
Highest PM2.5 Monitors with PM2.5 DVs Greater than 8/30 108
Figure 2A-4 NMB in 2018 CMAQ Predictions of PM2.5 Components at CSN and IMPROVE Sites
110
Figure 2A-5 NMB in 2016 CMAQ Predictions of PM2.5 Components at CSN and IMPROVE Sites for
Monitors in Counties with PM2.5 DVs Greater than 8/30 110
Figure 2A-6 Visible Satellite Imagery from NASA's Worldview Platform Showing Smoke from the
Camp Fire on 11/10/2018 123
Figure 2A-7 Visible Satellite Imagery from NASA's Worldview Platform Showing Smoke from the
North Bay/Wine Country Fires on 10/09/2017 124
xviii
-------
2 A-8
2A-9
2 A-10
2 A-11
2A-12
2A-13
2A-14
2 A-15
2 A-16
2A-17
2A-18
2A-19
2 A-20
2A-21
2A-22
2A-23
2A-24
2A-25
2A-26
Visible Satellite Imagery from NASA's Worldview Platform Showing Smoke from
Fires Across the Pacific Northwest/Northern California on 08/29/2017 124
Visible Satellite Imagery from NASA's Worldview Platform Showing Smoke from
Fires in Washington and Oregon on 08/09/2018 125
Visible Satellite Imagery from NASA's Worldview Platform Showing Smoke from
Fires in Montana on 08/19/2018 125
Visible Satellite Imagery from NASA's Worldview Platform Showing Smoke from
Several Fires in eastern California including the Empire Fire on 9/1/2017 126
Visible Satellite Imagery from NASA's Worldview Platform Showing Smoke from the
416/Burro Complex Fires on 06/10/2018 126
Visible Satellite Imagery from NASA's Worldview Platform Showing Smoke from the
Creek Fire on 10/26/2020 127
Visible Satellite Imagery from NASA's Worldview Platform Showing Smoke from the
Carr/Mendocino/Ferguson Fires on 08/04/2018 127
Visible Satellite Imagery from NASA's Worldview Platform Showing Smoke from
Fires in the Appalachians on 11/10/2016 128
Daily PM2.5 (in [ig nr3) from a Subset of Monitors Impacted by the Camp Fire in
November 2018 128
Daily PM2.5 (in [ig nr3) from a Subset of Monitors Impacted by the North Bay/Wine
Country Fires in October 2017 129
Daily PM2.5 (in [ig nr3) from a Subset of Monitors Impacted by Fires in the Pacific
Northwest/Northern California in August-September 2017 130
Daily PM2.5 (in [ig m 3) from a Subset of Monitors Impacted by Fires in Washington
and Oregon in July-August 2018 131
Daily PM2.5 (in [ig m 3) from the Monitors Impacted by Fires and Smoke in Montana
in August 2018 132
Daily PM2.5 (in [ig m 3) from a Subset of Monitors Impacted by Fires in Montana,
Washington and Idaho in August 2015 133
Daily PM2.5 (in [ig m 3) from the Monitor in Plata, CO Impacted by the 416/Burro Fire
Complex in June 2018 133
Daily PM2.5 (in [ig m 3) from the Two monitors Impacted by the Butte Fire in
September 2015 134
Daily PM2.5 (in [ig m 3) from a Subset of Monitors Impacted by the
Carr/Mendocino/Ferguson Fires in August 2018 134
Daily PM2.5 (in [ig m 3) from a Subset of Monitors Impacted by Fires in the
Appalachians in November 2016 135
Counties with Projected 2032 PM2.5 DVs that Exceed the 24-Hour (Daily Only),
Annual (Annual Only) or Both the 24-Hour and Annual (Both) Standards for the
Combination of Existing Standards (12/35) 136
xix
-------
Figure 2A-27 Counties with PM2.5 DVs in the 12/35 Analytical Baseline that Exceed the 24-Hour
(24-hr Only), Annual (Annual Only) or Both the 24-Hour and Annual (Both)
Standards for Combinations of Alternative Standards 137
Figure 2A-28 Counties with 50% Reduction in Anthropogenic Primary PM2.5 Emissions in 2028
Sensitivity Modeling 142
Figure 2A-29 Regional Groupings for Calculating Air Quality Ratios 143
Figure 2A-30 Counties Used in Estimating the Relative Impact of Emissions in Core and
Neighboring Counties 144
Figure 2A-31 Counties with 50% Reduction in Anthropogenic NOx Emissions in 2028 Sensitivity
Modeling 146
Figure 2A-32 Total Primary PM2.5 Emission Reductions Needed to Meet the Revised and
Alternative Standard Levels of 10/35,10/30, 9/35, and 8/35 Relative to the 12/35
Analytical Baseline in the East and West 154
Figure 2A-33 PM2.5 Concentration for 2032 based on eVNA Method 160
Figure 2A-34 PM2.5 Concentration Improvement Associated with Meeting 9/35 Relative to the
12/35 Analytical Baseline 162
Figure 3-1 Geographic Areas Used in Analysis 171
Figure 3-2 Counties Projected to Exceed in Analytical Baseline for Revised and Alternative
Standard Levels of 10/35 |a,g/m3, 9/35 |a,g/m3, and 8/35 |a,g/m3 175
Figure 3-3 Counties Projected to Exceed in Analytical Baseline for Alternative Standard Levels
of 10/30 |a,g/m3 176
Figure 3-4 PM2.5 Emissions Reductions and Costs Per Ton (CPT) in 2032 (tons, 2017$) 180
Figure 3-5 Counties that Still Need PM2.5 Emissions Reductions for Less Stringent Alternative
Standard Levels of 10/35 |a,g/m3 194
Figure 3-6 Counties that Still Need PM2.5 Emissions Reductions for Revised Standard Levels of
9/35 |a,g/m3 194
Figure 3-7 Counties that Still Need PM2.5 Emissions Reductions for More Stringent Alternative
Standard Levels of 8/35 |a,g/m3 195
Figure 3-8 Counties that Still Need PM2.5 Emissions Reductions for More Stringent Alternative
Standard Levels of 10/30 |a,g/m3 195
Figure 5-1 Data Inputs and Outputs for the BenMAP-CE Model 284
Figure 6-1 Heat Map of National Average Annual PM2.5 Concentrations and Concentration
Reductions (|ig/m3) by Demographic for Current, Revised, and Alternative PM
NAAQS Levels (annual/24-hr) After Application of Controls in 2032 336
Figure 6-2 National Distributions of Annual PM2.5 Concentrations by Demographic for Current,
Revised, and Alternative PM NAAQS Levels After Application of Controls in 2032
339
Figure 6-3 National Distributions of High Annual PM2.5 Concentrations by Demographic for
Current, Revised, and Alternative PM NAAQS Levels After Application of Controls in
2032 341
xx
-------
Figure 6-4 National Distributions of Annual PM2.5 Concentration Reductions by Demographic
from Current to Revised and Alternative PM NAAQS Levels After Application of
Controls in 2032 343
Figure 6-5 Heat Map of National Percent Reductions (%) in Average Annual PM2.5
Concentrations for Demographic Groups When Moving from Current to Revised and
Alternative PM NAAQS Levels After Application of Controls in 2032 346
Figure 6-6 Heat Map of Regional Average Annual PM2.5 Concentrations ([ig/m3) by
Demographic for Current, Revised, and Alternative PM NAAQS Levels After
Application of Controls in 2032 348
Figure 6-7 Regional Distributions of Annual PM2.5 Concentration Reductions for Demographic
Groups for Current PM NAAQS Levels and the Revised 9/35 Standard Scenario After
Application of Controls in 2032 349
Figure 6-8 Regional Distributions of High Annual PM2.5 Concentration Reductions for
Demographic Groups for Current PM NAAQS Levels and the 9/35 Revised Standard
Scenario After Application of Controls in 2032 (Revised Scale) 351
Figure 6-9 Heat Map of Regional Reductions in Average Annual PM2.5 Concentrations (|ig/m3)
for Demographic Groups When Moving from Current to Revised and Alternative PM
NAAQS Levels After Application of Controls in 2032 353
Figure 6-10 Regional Distributions of Annual PM2.5 Concentration Reductions for Demographic
Groups When Moving from Current PM NAAQS Levels to 9/35 After Application of
Controls in 2032 354
Figure 6-11 Heat Map of Regional Percent Reductions (%) in Average Annual PM2.5
Concentrations for Demographic Groups When Moving from Current to Revised and
Alternative PM NAAQS Levels After Application of Controls in 2032 355
Figure 6-12 Heat Map of National Average Annual Total Mortality Rates and Rate Reductions
(per 100K) for Demographic Groups for Current, Revised, and Alternative PM
NAAQS Levels After Application of Controls in 2032 (NH, Non-Hispanic) 359
Figure 6-13 National Distributions of Total Annual Mortality Rates (per 100k) for Demographic
Groups for Current, Revised, and Alternative PM NAAQS Levels After Application of
Controls in 2032 (NH, Non-Hispanic) 360
Figure 6-14 National Distributions of Annual Mortality Rate Reductions for Demographic Groups
When Moving from Current to Revised and Alternative PM NAAQS Levels After
Application of Controls in 2032 (NH, Non-Hispanic) 361
Figure 6-15 Heat Map of National Average Percent Mortality Rate Reductions (per 100k People)
for Demographic Groups When Moving from Current to Revised and Alternative PM
NAAQS Levels After Application of Controls in 2032 (NH, Non-Hispanic) 363
Figure 6-16 Heat Map of Regional Average Annual Total Mortality Rates (per 100K) for
Demographic Groups for Current, Revised, and Alternative PM NAAQS Levels After
Application of Controls in 2032 (NH, Non-Hispanic) 364
Figure 6-17 Regional Distributions of Total Annual Mortality Rates (per 100k) for Demographic
Groups for Current, Revised, and Alternative PM NAAQS Levels After Application of
Controls in 2032 (NH, Non-Hispanic) 365
xxi
-------
Figure 6-18 Heat Map of Regional Average Annual Mortality Rate Reductions (per 100k) for
Demographic Groups When Moving from Current to Revised and Alternative PM
NAAQS Levels After Application of Controls in 2032 (NH, Non-Hispanic) 366
Figure 6-19 Regional Distributions of Annual Mortality Rate Reductions (per 100k) for
Demographic Groups When Moving from Current to Revised and Alternative PM
NAAQS Levels After Application of Controls in 2032 (NH, Non-Hispanic) 367
Figure 6-20 Heat Map of Regional Average Proportional Mortality Rate Reductions (per 100k)
for Demographic Groups When Moving from Current to Revised and Alternative PM
NAAQS Levels After Application of Controls in 2032 (NH, Non-Hispanic) 368
Figure 6-21 Heat Map of National Average Annual PM2.5 Concentrations and Concentration
Reductions (|ig/m:i) Associated Either with Control Strategies (Controls) or with
Meeting the Standards (Standards) by Demographic for Current (12/35) and
Revised and Alternative PM NAAQS Levels (10/35,10/30, 9/35, and 8/35) in 2032
380
Figure 6-22 National Distributions of Annual PM2.5 Concentrations Associated Either with
Control Strategies or with Meeting the Standards by Demographic for Current,
Revised, and Alternative PM NAAQS Levels in 2032 381
Figure 6-23 National Distributions of Annual Reductions in PM2.5 Concentrations Associated
Either with Control Strategies or with Meeting the Standards by Demographic When
Moving from Current to Revised and Alternative PM NAAQS Levels in 2032 382
Figure 6-24 National Distributions of Annual Concentrations Experienced by the Reference
Population Associated Either with Control Strategies or with Meeting the Standards
for Current PM NAAQS of 12/35 in 2032 383
Figure 6-25 Heat Map of National Percent Reductions (%) in Average Annual PM2.5
Concentrations Associated Either with Control Strategies or with Meeting the
Standard Levels by Demographic When Moving from Current to Revised and
Alternative PM NAAQS Standard Levels in 2032 385
Figure 6-26 Heat Map of Regional Average Annual PM2.5 Concentrations ([ig/m3) Associated
Either with Control Strategies or with Meeting the Standards by Demographic for
Current, Revised, and Alternative PM NAAQS Standard Levels in 2032 387
Figure 6-27 Regional Distributions of Annual PM2.5 Concentrations Associated Either with
Control Strategies or with Meeting the 12/35 and Revised 9/35 Standard Levels in
2032 388
Figure 6-28 Heat Map of National Average Annual Reductions in PM2.5 Concentrations (|ig/m:i)
Associated Either with Control Strategies or with Meeting the Standards by
Demographic When Moving from Current to Revised and Alternative PM NAAQS
Levels in 2032 389
Figure 6-29 Regional Distributions of Annual PM2.5 Concentration Reductions When Moving
From 12/35-9/35 Associated Either with Control Strategies or Meeting the
Standards in 2032 390
Figure 6-30 Heat Map of Regional Percent Reductions (%) in Average Annual PM2.5
Concentrations Associated Either with Control Strategies or with Meeting the
Standards by Demographic When Moving from Current to Revised and Alternative
PM NAAQS Standard Levels in 2032 391
xxii
-------
Figure 6-31 Heat Map of National Average Annual Total Mortality Rates (per 100K People)
Associated Either with Control Strategies or with Meeting the Standards by
Demographic for Current, Revised, and Alternative PM NAAQS Levels in 2032 393
Figure 6-32 National Distributions of Total Mortality Rates (per 100k) Associated Either with
Control Strategies or with Meeting the Standards by Demographic for Current,
Revised, and Alternative PM NAAQS Levels in 2032 (NH, Non-Hispanic) 394
Figure 6-33 National Distributions of Annual T otal Mortality Rate Reductions (per 100k)
Associated Either with Control Strategies or with Meeting the Standards by
Demographic When Moving from Current to Revised and Alternative PM NAAQS
Levels in 2032 395
Figure 6-34 Heat Map of National Percent Reductions (%) in Average Mortality Rate Reductions
Associated Either with Control Strategies or with Meeting the Standards by
Demographic When Moving from Current to Revised and Alternative PM NAAQS
Levels in 2032 396
Figure 6-35 Heat Map of Regional Average Annual Total Mortality Rates (per 100K People)
Associated Either with Control Strategies or with Meeting the Standards by
Demographic for Current, Revised, and Alternative PM NAAQS Levels in 2032 (NH,
Non-Hispanic) 397
Figure 6-36 Regional Distributions of Total Mortality Rates (per 100k) Associated Either with
Control Strategies or with Meeting the Standards by Demographic for Current,
Revised, and Alternative PM NAAQS Levels in 2032 398
Figure 6-37 Heat Map of Regional Average Annual Total Mortality Rate Reductions (per 100K
People) Associated Either with Control Strategies or with Meeting the Standards by
Demographic When Moving from Current to Revised and Alternative PM NAAQS
Levels in 2032 (NH, Non-Hispanic) 399
Figure 6-38 Regional Distributions of Average Annual T otal Mortality Rate Reductions (per
100k) Associated Either with Control Strategies or with Meeting the Standards by
Demographic for When Moving from Current to Revised and Alternative PM NAAQS
Levels in 2032 400
Figure 6-39 Heat Map of Regional Percent Reductions (%) in Average Mortality Rate Reductions
Associated Either with Control Strategies or with Meeting the Standards by
Demographic When Moving from Current to Revised and Alternative PM NAAQS
Levels in 2032 (NH, Non-Hispanic) 401
xxiii
-------
EXECUTIVE SUMMARY
Overview of the Final Rule
In setting primary and secondary national ambient air quality standards (NAAQS),
the Environmental Protection Agency's (EPA) responsibility under the law is to establish
standards that protect public health and welfare. The Clean Air Act (CAA) requires the EPA,
for each criteria pollutant, to set standards that protect public health with "an adequate
margin of safety" and public welfare from "any known or anticipated adverse effects." As
interpreted by the Agency and the courts, the CAA requires the EPA to base the decisions
for primary standards on health considerations only; economic factors cannot be
considered. The prohibition against considering cost in the setting of the primary air
quality standards does not mean that costs, benefits, or other economic consequences are
unimportant. The Agency believes that consideration of costs and benefits is an essential
decision-making tool for the efficient implementation of these standards. The impacts of
costs, benefits, and efficiency are considered by the States when they make decisions
regarding what timelines, strategies, and policies are appropriate for their circumstances.
On June 10, 2021, the EPA announced its decision to reconsider the 2020 Particulate
Matter (PM) NAAQS final action. The EPA is reconsidering the December 2020 decision
because the available scientific evidence and technical information indicate that the current
standards may not be adequate to protect public health and welfare, as required by the
CAA. In general, the Administrator recognizes that the primary annual PM2.5 standard is
most effective at controlling exposures to "typical" daily PM2.5 concentrations that are
experienced over the year, while the primary 24-hour PM2.5 standard, with its 98th
percentile form, is most effective at limiting daily "peak" PM2.5 concentrations. The EPA has
concluded that the existing primary annual PM2.5 standard for PM, set at a level of 12.0
|ig/m3, is not requisite to protect public health with an adequate margin of safety. The EPA
Administrator is revising the existing annual standard to provide increased public health
protection. Specifically, the EPA Administrator is revising the level of the annual standard
to 9 |ig/m3. The EPA Administrator is retaining the primary 24-hour PM2.5 standard at its
current level of 35 |ig/m3. The Administrator is also retaining the primary 24-hour PM10
1
-------
standard, which provides public health protection against PMio-2.5-related health effects, at
its current level of 150 |ig/m3.
The EPA also concluded that the existing secondary PM standards are requisite to
protect public welfare from known or anticipated effects and is not changing the secondary
standards for PM at this time. Specifically, for the secondary annual PM2.5 standard, the EPA
Administrator is retaining the existing standard of 15.0 |ig/m3. For the secondary 24-hour
PM2.5 standard, the EPA Administrator is retaining the existing standard of 35 |ig/m3. For
the secondary 24-hour PM10 standard, the EPA Administrator is retaining the existing
standard of 150 |ig/m3.
Overview of the Regulatory Impact Analysis
Per Executive Orders 12866,13563, and 14094 and the guidelines of the Office of
Management and Budget's (OMB) Circular A-4, in this Regulatory Impact Analysis (RIA) we
analyze the revised annual and current 24-hour alternative standard levels of 9/35 |~ig/m3,
as well as the following less and more stringent alternative standard levels: (1) a less
stringent alternative annual standard level of 10 |~ig/m3 in combination with the current
24-hour standard (i.e.,10/35 |j,g/m3), (2) a more stringent alternative annual standard level
of 8 ng/m3 in combination with the current 24-hour standard (i.e., 8/35 |j,g/m3), and (2) a
more stringent alternative 24-hour standard level of 30 |~ig/m3 in combination with an
annual standard level of 10 ng/m3 (i.e., 10/30 |j,g/m3). Because the EPA is not changing the
current secondary PM2.5 standards at this time, as well as retaining the primary and
secondary PM10 standards, we did not evaluate alternative levels of these standards. The
RIA includes the following chapters: Chapter 2: Air Quality Modeling and Methods; Chapter
3: Control Strategies and PM2.5 Emissions Reductions; Chapter 4: Engineering Cost Analysis
and Qualitative Discussion of Social Costs; Chapter 5: Benefits Analysis Approach and
Results; Chapter 6: Environmental Justice; Chapter 7: Labor Impacts; and Chapter 8:
Comparison of Benefits and Costs.
The RIA presents estimates of the costs and benefits of applying illustrative national
control strategies in 2032 after implementing existing and expected regulations and
assessing emissions reductions to meet the current annual and 24-hour fine particulate
2
-------
matter NAAQS (12/35 |j,g/m3). The selection of 2032 as the analysis year in the RIA does
not predict or prejudge attainment dates that will ultimately be assigned to individual
areas under the CAA. The CAA contains a variety of potential attainment dates and
flexibility to move to later dates, provided that the date is as expeditious as practicable. For
the purposes of this analysis, the EPA assumes that it would likely finalize designations for
the revised fine particulate matter NAAQS in late 2025. Furthermore, also for the purposes
of this analysis and depending on the precise timing of the effective date of those
designations, the EPA assumes that nonattainment areas classified as Moderate would
likely have to attain in late 2032. As such, we selected 2032 as the primary year of analysis.
The analyses in this RIA rely on national-level data (emissions inventory and control
measure information) for use in national-level assessments (air quality modeling, control
strategies, environmental justice, and benefits estimation). However, the ambient air
quality issues being analyzed are highly complex and local in nature, and the results of
these national-level assessments therefore contain uncertainty. It is beyond the scope of
this RIA to develop detailed local information for the areas being analyzed, including
populating the local emissions inventory, obtaining local information to increase the
resolution of the air quality modeling, and obtaining local information on emissions
controls, all of which would reduce some of the uncertainty in these national-level
assessments. For example, having more refined data would be ideal for agricultural dust
and burning, prescribed burning, and non-point (area) sources due to their large
contribution to primary PM2.5 emissions and the limited availability of emissions controls.1
In addition, for residential wood combustion emissions, people will respond differently to
the various regulations and incentives offered for controlling PM2.5 emissions from wood
burning, making it important to identify the right balance of controls for each area.
ES.l Design of the Regulatory Impact Analysis
The goal of this RIA is to provide estimates of the potential costs and benefits of the
illustrative national control strategies in 2032. Because States are ultimately responsible
for implementing strategies to meet alternative standard levels, this RIA provides insights
1 Examples of area source emissions include area fugitive dust, residential wood combustion, and commercial
cooking emissions.
3
-------
and analysis of a limited number of illustrative control strategies that states might adopt to
implement a revised standard level.
We developed our projected baselines for emissions and air quality for 2032. To
estimate the costs and benefits of the illustrative national control strategies for the revised
and less and more stringent annual and 24-hour PM2.5 alternative standard levels, we first
prepared an analytical baseline for 2032 that assumes full compliance with the current
standards of 12/35 |~ig/m3. From that analytical baseline, we estimate PM2.5 emissions
reductions needed to reach the revised and alternative annual and 24-hour PM2.5 standard
levels and then analyze illustrative control strategies that areas might employ.
Because PM2.5 concentrations are most responsive to primary PM emissions
reductions, for the illustrative control strategies we analyze direct, local PM2.5 emissions
reductions by individual counties.2 For the eastern U.S. where counties are relatively small
and terrain is relatively flat, we identified potential PM2.5 emissions reductions within each
county and in adjacent counties within the same state, where needed. As discussed in
Chapter 3, Section 3.2.2, when we applied the emissions reductions from adjacent counties,
we used a |~ig/m3 per ton PM2.5 air quality ratio that was four times less responsive than the
ratio used when applying in-county emissions reductions. Because the counties in the
western U.S. are generally large and the terrain is more complex, we only identified
potential PM2.5 emissions reductions within each county.
We then prepared illustrative control strategies. We apply end-of-pipe control
technologies to non-electric generating unit (non-EGU) stationary sources (e.g., fabric
filters, electrostatic precipitators, venturi scrubbers) and area source controls to non-point
(area) sources (e.g., installing controls on charbroilers), to residential wood combustion
2 As discussed in Chapter 2, Section 2.1.3, the spatial distributions of PM2.5 concentrations in the U.S. are
characterized by an "urban increment" of consistently higher PM2.5 concentrations over urban areas than in
surrounding areas. Monitored concentrations are highest in urban areas and relatively low in rural areas.
Conceptually, PM2.5 concentrations in urban areas can be viewed as the superposition of the urban
increment and the contributions from regional and natural background sources. The decreases in
anthropogenic SO2 and NOx emissions in recent decades have reduced regional background concentrations
and increased the relative importance of the urban increment. The projections of additional large reductions
in SO2 and NOx emissions in the 2032 case further motivate the need for control of local primary PM2.5
sources to address the highest PM2.5 concentrations in urban areas. The 2032 projections include wildfire
emissions at their 2018 levels, but these emissions were not targeted for control.
4
-------
sources (e.g., converting woodstoves to gas logs), and for area fugitive dust emissions (e.g.,
paving unpaved roads) in analyzing PM2.5 emissions reductions. Below we discuss the SO2
and NOx emissions reductions from mobile sources and EGUs reflected in the projections
between 2018 and 2032. We analyze PM2.5 emissions reductions in and near counties with
projected exceedances because this is the most efficient approach for assessing reductions
in future PM2.5 concentrations after accounting for the large projected SO2 and NOx
emissions reductions. The estimated PM2.5 emissions reductions from these control
applications do not fully account for all the emissions reductions needed to reach the
revised and less and more stringent alternative standard levels in some counties in the
northeast, southeast, west, and California. In Chapter 2, Section 2.4 and Chapter 3, Section
3.2.5, we discuss the remaining air quality challenges for areas in the northeast and
southeast, as well as in the west and California for the revised standard levels of 9/35 |Lxg/-
m3; the areas include counties with near-road monitors, counties in border areas, counties
in small western mountain valleys, and counties in California's air basins and districts. The
characteristics of the air quality challenges for these areas include features of certain near-
road sites with challenging local conditions, cross-border transport, effects of complex
terrain in the west, and identifying wildfire influence on projected PM2.5 DVs that could
potentially qualify for exclusion as atypical, extreme, or unrepresentative events (U.S. EPA,
2019a). Lastly, we estimate the engineering costs and human health benefits associated
with the illustrative control strategies, as well as assess environmental justice
considerations.
Chapter 2, Section 2.1.3, includes discussions of historical and projected emissions
trends for direct PM2.5 and precursor emissions (i.e., SO2, NOx, VOC, and ammonia), as well
as of the "urban increment" of consistently higher PM2.5 concentrations over urban areas.
We did not apply controls to EGUs or mobile sources beyond what is reflected in the
projections between 2018 and 2032. The projections reflect SO2 and NOx emissions
decreases between 2018 and 2032 — over this period (1) NOx emissions are projected to
decrease by 3.6 million tons (41 percent), with the greatest reductions from mobile source
and EGU emissions inventory sectors, and (2) SO2 emissions are projected to decrease by
5
-------
1.1 million tons (48 percent), with the greatest reductions from the EGU emissions
inventory sector.
ES.1.1 Establishing the Analytical Baseline
To project air quality to the future, the Community Multiscale Air Quality Modeling
System (CMAQ) model was applied to simulate air quality over the U.S. during 2018 and for
a case with emissions representative of 2032. In the 2032 projections, PM2.5 design values
(DVs) exceeded the current standards for some counties in the west and California.3 As
described in Chapter 2, Section 2.3.2, we adjusted the PM2.5 DVs for 2032 to account for
emissions reductions needed to attain the current annual and 24-hour PM2.5 standards of
12/35 |j,g/m3to form the 12/35 |~ig/m3 analytical baseline; it is from this baseline that we
estimate the incremental costs and benefits associated with control strategies for the
revised and less and more stringent alternative standard levels relative to the current
standards. For EGUs, the analytical baseline reflects, among other existing regulations, the
Final Good Neighbor Plan for the 2015 Ozone NAAQS (2023), the Revised Cross-State Air
Pollution Rule Update (2021), the Standards of Performance for Greenhouse Gas Emissions
from New, Modified, and Reconstructed Stationary Sources (2015), and the Mercury and
Air Toxics Rule (2011). The baseline also reflects provisions of tax incentives in the
Inflation Reduction Act of 2022 (IRA). For mobile sources, the baseline reflects the Control
of Air Pollution from New Motor Vehicles: Heavy-Duty Engine and Vehicle Standards
(2022), the Final Rule to Revise Existing National GHG Emissions Standards for Passenger
Cars and Light Trucks Through Model Year 2026 (2021), the GHG Emissions Standards and
Fuel Efficiency Standards for Medium- and Heavy-Duty Engines and Vehicles, Phase 2
(2016), and the Tier 3 Motor Vehicle Emission and Fuel Standards (2014). For non-EGUs,
the baseline reflects the Final Good Neighbor Plan for the 2015 Ozone NAAQS (2023), the
New Source Performance Standards (NSPS) for oil and natural gas sources (2016), the
NSPS for process heaters (2013), the NSPS for natural gas turbines and reciprocating
3 PM2.5 DVs were projected to 2032 using the air quality model results in a relative sense, as recommended by
the EPA modeling guidance, by projecting monitoring data with relative response factors [RRFs] developed
from the 2018 and 2032 CMAQ modeling.
6
-------
internal combustion engines (2012), and the NSPS for residential wood combustion (2015).
For a more complete list of regulations, please see Chapter 2, Section 2.2.1.
We present results throughout the RIA by northeast, southeast, west, and California,
and Figure ES-1 includes a map of the U.S. with these areas identified. Table ES-1 presents a
summary of the PM2.5 emissions reductions needed by area to meet the current standards
to form the 12/35 |.ig/m3 analytical baseline.
West
Northeast
1 ca \
.California
Southeast
Figure ES-1 Geographic Areas Used in Analysis
Table ES-1 Summary of PM2.5 Emissions Reductions Needed by Area in 2032 to
Meet Current Primary Annual and 24-hour Standards of 12/35 (ig/rn3
(tons/year)
Area
12/35
Northeast
0
Southeast
0
West
1,494
CA
6,032
Total
7,526
7
-------
Sixteen counties need PM2.5 emissions reductions to meet the current standards in
2032 - 11 counties in California and five counties in the west.4 The counties in California
include several counties in the San Joaquin Valley Air Pollution Control District and the
South Coast Air Quality Management District, as well as Plumas County, Colusa County, and
Siskiyou County in Northern California, Mono County in Eastern California, and Imperial
County in Southern California. No counties in the northeast or southeast U.S. need PM2.5
emissions reductions to meet the current annual and 24-hour standards.
ES.1.2 Estimating PM2.5 Emissions Reductions Needed for Annual and 24-hour
Revised and Alternative Standard Levels Analyzed
We apply regional PM2.sair quality ratios to estimate the emissions reductions
needed to reach the revised and less and more stringent annual and 24-hour alternative
standard levels analyzed. To develop air quality ratios that relate the change in DV in a
county to the change in primary PM2.5 emissions in that county, we performed air quality
sensitivity modeling with reductions in primary PM2.5 emissions in selected counties. More
specifically, we conducted a 2028 CMAQ sensitivity modeling simulation with 50 percent
reductions in primary PM2.5 emissions from anthropogenic sources in counties with annual
2028 DVs greater than 8 |~ig/m3. We divided the change in annual and 24-hour PM2.5 DVs in
these counties by the change in emissions in the respective counties to determine the air
quality ratio at individual monitors.
We developed representative air quality ratios for regions of the U.S. from the ratios
at individual monitors as in the 2012 PM2.5 NAAQS review (U.S. EPA, 2012). These regions
are shown in Chapter 2, Figure 2-7, and the air quality ratios for primary PM2.5 emissions
used in estimating the emission reductions needed to just meet the revised and alternative
standard levels analyzed are listed in Chapter 2, Table 2-1. We estimated the emissions
reductions needed to just meet the revised and alternative standard levels analyzed using
the primary PM2.5 air quality ratios in combination with the required incremental change in
concentration. Chapter 2, Section 2.3.1 includes a brief discussion of developing air quality
ratios and estimated emissions reductions needed to just meet the revised and alternative
4 The 16 counties require primary PM emissions reductions to meet the current standards of 12/35 |J.g/m3
following application of the NOx emission reductions in San Joaquin Valley and the South Coast to adjust the
2032 DVs. For additional discussion, see Appendix 2A, Section 2A.3.2 and Section 2A.3.3.
8
-------
standard levels analyzed, and Appendix 2A, Section 2A.3 includes more detailed
discussions.
Table ES-2 presents a summary of the estimated emissions reductions needed by
area to reach the annual and 24-hour revised and alternative standard levels. For each set
of standard levels, Table ES-2 also includes an area's percent of the total estimated
emissions reductions needed nationwide to reach those standard levels in all locations. For
example, for the less stringent alternative standard levels of 10/35 |j,g/m3, California's
10,753 estimated tons needed is 81 percent of the total estimated emissions reductions
needed nationwide to meet 10/35 |~ig/m3. See Appendix 2A, Table 2A-14 for the estimated
PM2.5 emissions reductions, from the analytical baseline, needed by county for the revised
and alternative standard levels analyzed. Figure ES-2 shows the counties projected to
exceed the annual and 24-hour revised and alternative standard levels of 10/35 |~ig/m3,
9/35 ng/m3, and 8/35 |~ig/m3 in the analytical baseline. Additional information on the air
quality modeling, as well as information about projected future DVs, DV targets, and air
quality ratios is provided in Chapter 2 and Appendix 2A.
Table ES-2 By Area, Summary of PM2.5 Emissions Reductions Needed, In
Tons/Year and as Percent of Total Reduction Needed Nationwide, for
the Revised and Alternative Primary Standard Levels of 10/35 ng/m3,
10/30 |ig/m3, 9/35 (j,g/m3, and 8/35 jig/m3 in 2032
Area
10/35
10/30
9/35
8/35
Northeast
1,032
1,073
6,974
20,620
Southeast
531
531
3,279
18,658
West
987
6,673
3,132
10,277
CA
10,753
16,660
19,402
31,518
Total
13,303
24,938
32,786
81,073
Area
10/35
10/30
9/35
8/35
Northeast
8%
4%
21%
25%
Southeast
4%
2%
10%
23%
West
7%
27%
10%
13%
CA
81%
67%
59%
39%
9
-------
H Reductions required for 10/35, 9/35, and 8/35
Reductions required for 9/35 and 8/35
I Reductions required for 8/35
Figure ES-2 Counties Projected to Exceed in Analytical Baseline for the Revised
and Alternative Standard Levels of 10/35 (ig/m3, 9/35 (ig/m3, and
8/35 |ig/m3
For each set of alternative standard levels, Chapter 2, Section 2.3.3 includes a
discussion of the number of counties that are projected to exceed in 2032, and Figure 2-9
includes maps of counties projected to exceed along with the number of counties. The
following summarizes the number of counties, by revised and alternative standard levels,
in each geographic area that need PM2.5 emissions reductions from the analytical baseline.
• 10/35 |_ig/m3 — 20 counties need PM2.5 emissions reductions. This includes 4
counties in the northeast, 1 county in the southeast, 2 counties in the west,
and 13 counties in California.
• 10/30 !~ig/m3 — 49 counties need PM2.5 emissions reductions. This includes 4
counties in the northeast, 1 county in the southeast, 19 counties in the west,
and 25 counties in California.
• 9/35 ng/m3 — 52 counties need PM2.5 emissions reductions. This includes 12
counties in the northeast, 7 counties in the southeast, 10 counties in the west,
and 23 counties in California.
10
-------
• 8/35 ng/m3 — 117 counties need PM2.5 emissions reductions. This includes
31 counties in the northeast, 33 counties in the southeast, 21 counties in the
west, and 32 counties in California.
ES.1.3 Control Strategies and PM2.5 Emissions Reductions
We identified controls using the EPA's Control Strategy Tool (CoST) (U.S. EPA,
2019b) and the control measures database.5 CoST estimates emissions reductions and
engineering costs associated with end-of-pipe control technologies or area source controls
applied to non-electric generating unit (non-EGU) point, non-point (area), residential wood
combustion, and area fugitive dust sources of air pollutant emissions by matching end-of-
pipe or area source controls to emissions sources by source classification code (SCC). For
these control strategy analyses, to maximize the number of emissions sources we included
a lower emissions source size threshold (5 tons per year) and a higher marginal cost per
ton threshold ($160,000/ton) than reflected in prior NAAQS RIAs (25-50 tpy, $15,000-
$20,000/ton). In Chapter 3, Figure 3-4 shows estimated PM2.5 emissions reductions for
several emissions source sizes and cost thresholds up to the $160,000/ton marginal cost
threshold. We selected the $160,000/ton marginal cost threshold because it is around that
cost level that (i) road paving controls get selected and applied, and (ii) opportunities for
additional emissions reductions diminish. In addition, in the northeast and southeast we
applied emissions reductions from adjacent counties, using a ratio of 4:1. That is, four tons
of PM2.5 emissions reductions would be required from an adjacent county to reduce one ton
of emissions reduction needed in a given county. For additional discussion, see Chapter 2,
Section 2.3.1 and Chapter 3, Section 3.2.2.
By area, Table ES-3 includes a summary of the estimated emissions reductions from
control applications for the revised and alternative standard levels analyzed. These
emissions reductions were used to create the PM2.5 spatial surfaces described in Appendix
2A, Section 2A.4.2 for the human health benefits assessments presented in Chapter 5. See
5 More information about CoST and the control measures database can be found at the following link:
https://www.epa.gov/economic-and-cost-analysis-air-pollution-regulations/cost-analysis-modelstools-air-
pollution.
11
-------
Chapter 3, Table 3-5 through Table 3-7 for additional summaries of estimated PM2.5
emissions reductions from CoST.
Table ES-3 Summary of PM2.5 Estimated Emissions Reductions from CoST by Area
for the Revised and Alternative Primary Standard Levels of 10/35 (j,g/-
m3,10/30 |ig/m3,9/35 |ig/m3, and 8/35 jig/m3 in 2032 (tons/year)
PM2.5 Emissions Reductions
Area
10/35
10/30
9/35
8/35
Northeast
1,032
1,074
7,226
14,036
Northeast (Adjacent Counties)
0
0
2,599
11,911
Southeast
521
521
1,959
13,995
Southeast (Adjacent Counties)
45
45
354
3,086
West
470
2,715
1,386
5,323
CA
3,010
4,652
5,069
7,181
Total
5,078
9,006
18,592
55,532
Note: Totals may not match related tables due to independent rounding. In the northeast and southeast
when we applied the emissions reductions from adjacent counties, we used a ppb/ton PM2.5 air quality
ratio that was four times less responsive than the ratio used when applying in-county emissions
reductions.
ES.1.4 Estimates of PM2.5 Emissions Reductions Still Needed after Applying Controls
The estimated PM2.5 emissions reductions from the control strategies do not fully
account for all the emissions reductions needed to reach the revised and less and more
stringent alternative standard levels in some counties in the northeast, southeast, west, and
California. By area, Table ES-4 includes a summary of the estimated emissions reductions
still needed after control applications for the revised and alternative standard levels
analyzed. See Chapter 3, Table 3-9 for an additional summary of estimated emissions
reductions still needed. Figure ES-3 shows the counties that still need emissions reductions
after control applications for the revised standard levels of 9/35 |~ig/m3. Section ES.2 below
includes a qualitative discussion of the remaining air quality challenges. In addition,
Chapter 2, Section 2.4 and Chapter 3, Section 3.2.5 provide more detailed discussions of
these air quality challenges.
12
-------
Table ES-4
Summary of PM2.5 Emissions Reductions Still Needed by Area for the
Revised and Alternative Primary Standard Levels of 10/35 jig/m3,
10/30 jig/m3,9/35 ng/m3, and 8/35 (ig/m3 in 2032 (tons/year)
Area
10/35
10/30
9/35
8/35
Northeast
0
0
130
3,285
Southeast
0
0
1,038
3,519
West
516
3,959
1,747
4,982
CA
7,739
11,986
14,411
24,366
Total
8,255
15,945
17,327
36,152
| Counties with Sufficient Identified Reductions to Meet 9/35
I Counties Still Needing Reductions to Meet 9/35
Figure ES-3 Counties that Still Need PM2.5 Emissions Reductions for the Revised
Standard Levels of 9/35 ng/m3
ES.1.5 Engineering Costs
The EPA also used CoST and the control measures database to estimate engineering
control costs. We estimated costs for non-EGU point, non-point (area), residential wood
combustion, and area fugitive dust sources of air pollutant emissions. CoST calculates
engineering costs using one of two different methods: (1) an equation that incorporates
key operating unit information, such as unit design capacity or stack flow rate, or (2) an
average annualized cost-per-ton factor multiplied by the total tons of reduction of a
pollutant. The engineering cost analysis uses the equivalent uniform annual costs (EUAC)
13
-------
method, in which annualized costs are calculated based on the equipment life for the
control and the interest rate incorporated into a capital recovery factor. Annualized costs
represent an equal stream of yearly costs over the period the control is expected to
operate. The cost estimates reflect the engineering costs annualized using a 7 percent
interest rate.
By area, Table ES-5 includes a summary of estimated control costs from control
applications for the revised and alternative standard levels analyzed. See Chapter 4, Table
4-2 through Table 4-5 for additional summaries of estimated control costs associated with
the control strategies.
Table ES-5 By Area, Summary of Annualized Control Costs for the Revised and
Alternative Primary Standard Levels of 10/35 |ig/m3,10/30 |ig/m3,
9/35 (ig/m3, and 8/35 (ig/m3 for 2032 (millions of 2017$)
Area
10/35
10/30
9/35
8/35
Northeast
$5.3
$5.5
$203.6
$371.1
Northeast (Adjacent Counties)
$0
$0
$62.1
$364.2
Southeast
$35.8
$35.8
$60.4
$299.7
Southeast (Adjacent Counties)
$0.02
$0.02
$25.5
$69.2
West
$39.7
$112.4
$57.7
$140.6
CA
$121.8
$186.1
$184.4
$256.7
Total
$202.5
$339.8
$593.8
$1,501.5
Note: The estimated PM2.5 emissions reductions from the control strategies do not fully account for all the
emissions reductions needed to reach the revised and less and more stringent alternative standard levels
in some counties in the northeast, southeast, west, and California.
For the alternative standard levels of 10/35 |~ig/m3, the majority of the estimated
costs are incurred in California because 13 of the 20 counties that need emissions
reductions are located in California. Looking at the more stringent alternative standard
levels of 10/30 |~ig/m3, in the west an additional 17 counties need emissions reductions and
estimated costs increase significantly; also, in the west estimated costs for the revised
standard levels of 9/35 |~ig/m3 are higher than for 10/35 |~ig/m3 but lower than for 10/30
Hg/m3.
For revised and more stringent alternative standard levels of 9/35 |~ig/m3 and 8/35
Hg/m3, more controls are available to apply in the northeast and the southeast as compared
to California and the west. Therefore, the estimated costs for the northeast and southeast
are higher for 9/35 ng/m3 and 8/35 ng/m3. In addition, in the northeast and southeast
14
-------
when we applied the emissions reductions from adjacent counties, we applied a ratio of
4:1. Application of this ratio also contributes to the higher estimated cost estimates for
alternative standard levels of 9/35 |~ig/m3 and 8/35 |~ig/m3 in those areas.
ES.1.6 Human Health Benefits
We estimate the quantity and economic value of air pollution-related effects using a
"damage-function." This approach quantifies counts of air pollution-attributable cases of
adverse health outcomes and assigns dollar values to those counts, while assuming that
each outcome is independent of one another. We construct this damage function by
adapting primary research—specifically, air pollution epidemiology studies and economic
value studies—from similar contexts. This approach is sometimes referred to as "benefits
transfer."
We use the environmental Benefits Mapping and Analysis Program—Community
Edition (BenMAP-CE) software program to quantify counts of premature deaths and
illnesses attributable to photochemical modeled changes in annual mean PM2.5 for the year
2032 using health impact functions (Sacks et al., 2018). A health impact function combines
information regarding: the concentration-response relationship between air quality
changes and the risk of a given adverse outcome; the population exposed to the air quality
change; the baseline rate of death or disease in that population; and the air pollution
concentration to which the population is exposed.
After quantifying the change in adverse health impacts, the final step is to estimate
the economic value of these avoided impacts. The appropriate economic value for a change
in a health effect depends on whether the health effect is viewed ex ante (before the effect
has occurred) or ex post (after the effect has occurred). Reductions in ambient
concentrations of air pollution generally lower the risk of future adverse health effects by a
small amount for a large population. The appropriate economic measure is therefore ex
ante willingness-to-pay (WTP) for changes in risk. However, epidemiological studies
generally provide estimates of the relative risks of a particular health effect avoided due to
a reduction in air pollution. A convenient way to use this data in a consistent framework is
15
-------
to convert probabilities to units of avoided statistical incidences. This measure is calculated
by dividing individual WTP for a risk reduction by the related observed change in risk.
Applying the impact and valuation functions to the estimated changes in PM2.5 yields
estimates of the changes in physical damages (e.g., premature mortalities, cases of hospital
admissions and emergency department visits) and the associated monetary values for
those changes. Table ES-6 presents the estimated avoided incidences of PM-related
illnesses and premature mortality resulting from emissions reductions associated with the
application of the illustrative control strategies for the revised and alternative standard
levels in 2032. Table ES-7 and Table ES-8 present a summary of the monetized benefits
associated with emissions reductions from the application of the illustrative control
strategies for the revised and alternative standard levels, both nationally and by area,
thereby allowing the comparison of cost and benefits of the application of the illustrative
controls. As mentioned above and discussed in Chapter 3, Section 3.2.4, the estimated PM2.5
emissions reductions from control applications do not fully account for all the emissions
reductions needed to reach the revised and less and more stringent alternative standard
levels in some counties in the northeast, southeast, west, and California. In Chapter 2,
Section 2.4 and Chapter 3, Section 3.2.5, we discuss the remaining air quality challenges for
areas in the northeast and southeast, as well as in the west and California for the revised
standard levels of 9/35 |~ig/m3. In Appendix 5A, a set of tables summarizes the benefits
associated with identifying all of the emissions reductions needed to reach the revised and
less and more stringent alternative standard levels. For Table ES-7 and Table ES-8, the
monetized value of unquantified effects is represented by adding an unknown "B" to the
aggregate total. This B represents both uncertainty and a bias in this analysis, as it reflects
health and welfare benefits that we are unable to quantify. Note that not all known PM
health effects could be quantified or monetized.
16
-------
Table ES-6 Estimated Avoided PM-Related Premature Mortalities and Illnesses of
the Control Strategies for the Revised and Alternative Primary PM2.5
Standard Levels for 2032 (95% Confidence Interval)
Avoided Mortality3
10/35
10/30
9/35
8/35
(Pope Illetal., 2019)
(adult mortality ages 18-
99 years)
1,700
(1,200 to 2,100)
2,000
(1,400 to 2,600)
4,500
(3,200 to 5,700)
9,500
(6,800 to 12,000)
(Wu et al., 2020) (adult
mortality ages 65-99
years)
810
(710 to 900)
970
(850 to 1,100)
2,100
(1,900 to 2,400)
4,500
(4,000 to 5,100)
(Woodruff et al., 2008)
(infant mortality)
1.7
(-1.0 to 4.3)
2.0
(-1.2 to 5.1)
5.0
(-3.1 to 13)
11
(-7.2 to 29)
Avoided Morbidity
10/35
10/30
9/35
8/35
Hospital admissions—
cardiovascular (age > 18)
140
(100 to 180)
160
(120 to 200)
330
(240 to 420)
690
(500 to 870)
Hospital admissions—
respiratory
90
(31 to 150)
100
(35 to 170)
230
(79 to 370)
480
(170 to 780)
ED visits—cardiovascular
260
(-98 to 600)
300
(-110 to 690)
660
(-250 to 1,500)
1,400
(-550 to 3,300)
ED visits—respiratory
470
(93 to 990)
560
(110 to 1,200)
1,300
(250 to 2,700)
2,900
(570 to 6,000)
Acute Myocardial
Infarction
30
(17 to 42)
35
(20 to 49)
72
(42 to 100)
150
(86 to 210)
Cardiac arrest
14
(-5.9 to 33)
17
(-6.9 to 38)
36
(-15 to 81)
75
(-31 to 170)
Hospital admissions—
Alzheimer's Disease
360
(270 to 440)
400
(300 to 500)
910
(690 to 1,100)
2,000
(1,500 to 2,400)
Hospital admissions—
Parkinson's Disease
47
(24 to 69)
56
(29 to 81)
130
(67 to 190)
280
(140 to 400)
Stroke
54
(14 to 93)
65
(17 to 110)
140
(35 to 230)
290
(74 to 490)
Lung cancer
65
(20 to 110)
77
(23 to 130)
160
(49 to 270)
340
(100 to 560)
Hay Fever/Rhinitis
15,000
(3,600 to 26,000)
17,000
(4,200 to 30,000)
38,000
(9,100 to 65,000)
79,000
(19,000 to 140,000)
Asthma Onset
2,300
(2,200 to 2,300)
2,600
(2,500 to 2,700)
5,700
(5,500 to 6,000)
12,000
(12,000 to 13,000)
Asthma symptoms -
Albuterol use
310,000
(-150,000 to
760,000)
370,000
(-180,000 to
900,000)
800,000
(-390,000 to
1,900,000)
1,700,000
(-820,000 to
4,100,000)
Lost work days
110,000
(96,000 to
130,000)
130,000
(110,000 to
150,000)
290,000
(240,000 to
330,000)
610,000
(510,000 to
700,000)
Minor restricted-activity
days
670,000
(540,000 to
790,000)
780,000
(630,000 to
920,000)
1,700,000
(1,400,000 to
2,000,000)
3,500,000
(2,900,000 to
4,200,000)
Note: Values rounded to two significant figures.
a Reported here are two alternative estimates of the number of premature deaths among adults due to long-term
exposure to PM2.5. These values should not be added to one another.
17
-------
Table ES-7 Estimated Monetized Benefits of the Control Strategies for the Revised
and Alternative Primary PM2.5 Standard Levels in 2032, Incremental to
Attainment of 12/35 iig/m3 (billions of 2017$)
Benefits Estimate 10/35 10/30 9/35 8/35
Economic value of avoided PIVh.s-related morbidities and premature deaths using PM2.5 mortality
estimate from (Pope III et al., 2019)
3% discount $17+ B $21 + B $46 + B $99 + B
rate
7% discount $16+ B $19 + B $42 + B $89 + B
rate
Economic value of avoided PIVh.s-related morbidities and premature deaths using PM2.5 mortality
estimate from (Wu etal., 2020)
3% discount 5 + g $10 + B $22 + B $48 + B
rate
7% discount $7.6+ B $9.2 + B $20+ B $43 + B
rate
Notes: Rounded to two significant figures.
The estimated PM2.5 emissions reductions from the control strategies do not fully account for all the emissions
reductions needed to reach the revised and less and more stringent alternative standard levels in some counties
in the northeast, southeast, west, and California.
Avoided premature deaths account for over 98% of monetized benefits here, which are discounted over the SAB-
recommended 20-year segmented lag.
It was not possible to quantify all benefits due to data limitations in this analysis. "B" is the sum of all
unquantified health and welfare benefits.
18
-------
Table ES-8
Benefits
Estimate
Estimated Monetized Benefits by Area of the Control Strategies for the
Revised and Alternative Primary PM2.5 Standard Levels in 2032,
Incremental to Attainment of 12/35 iig/m3 (billions of 2017$)
Area
10/35
10/30
9/35
8/35
Economic value of avoided PIVh.s-related morbidities and premature deaths using PM2.5 mortality estimate
from (Pope III etal., 2019)
Northeast $2.4 + B $2.5 + B $18 + B $37 + B
Southeast $0.51 + B $0.51 + B $5.3 + B $25 + B
West $0,059 + B $1.3 + B $2.3 + B $10 + B
California $15 + B $17+ B $21+ B $27+ B
3%
discount
rate
7%
discount
rate
Northeast
Southeast
West
California
$2.1+ B
$0.46 + B
$0,053 + B
$13+ B
$2.2 + B
$0.46 + B
$1.2 + B
$15+ B
$16+ B
$4.8 + B
$2.1+ B
$19+ B
$34+ B
$22+ B
$9.3 + B
$24+ B
Economic value of avoided PIVh.s-related morbidities and premature deaths using PM2.5 mortality estimate
from (Wu et al., 2020)
30/0
discount
rate
7%
discount
rate
Northeast
Southeast
West
California
$1.2 + B
$0.24+ B
$0.03+ B
$7.0+ B
$1.2+ B
$0.24+ B
$0.64+ B
$8.1 + B
$8.7 + B
$2.5+ B
$1.1+ B
$10+ B
$18+ B
$12+ B
$5.0+ B
$13+ B
Northeast
Southeast
West
California
$1.0+ B
$0.21+ B
$0.027+ B
$6.3 + B
$1.1+ B
$0.21+ B
$0.58 + B
$7.3+ B
$7.8 + B
$2.2 + B
$1.0+ B
$9.1 + B
$16+ B
$10+ B
$4.5 + B
$12+ B
Notes: Rounded to two significant figures.
The estimated PM2.5 emissions reductions from the control strategies do not fully account for all the emissions
reductions needed to reach the revised and less and more stringent alternative standard levels in some counties in
the northeast, southeast, west, and California.
Avoided premature deaths account for over 98% of monetized benefits here, which are discounted over the SAB-
recommended 20-year segmented lag.
It was not possible to quantify all benefits due to data limitations in this analysis. "B" is the sum of all unquantified
health and welfare benefits.
ES.1.7 Welfare Benefits of Meeting the Primary and Secondary Standards
Even though the primary standards are designed to protect against adverse effects
to human health, the emissions reductions would have welfare benefits in addition to the
direct health benefits. The term welfare benefits covers both environmental and societal
benefits of reducing pollution. Welfare benefits of the primary PM standard include
improved visibility, reduced climate effects, reduced materials damage, reduced vegetation
effects resulting from PM exposure, and reduced ecological effects from particulate matter
deposition and from nitrogen emissions. This RIA does not assess welfare effects
quantitatively; this is discussed further in Chapter 5.
19
-------
ES.1.8 Environmental Justice
Environmental justice (EJ) concerns for each rulemaking are unique and should be
considered on a case-by-case basis, and EPA's EJ Technical Guidance (U.S. EPA, 2015) states
that "[t]he analysis of potential EJ concerns for regulatory actions should address three
questions:
1. Are there potential EJ concerns associated with environmental stressors
affected by the regulatory action for population groups of concern in the
baseline?
2. Are there potential EJ concerns associated with environmental stressors
affected by the regulatory action for population groups of concern for the
regulatory option(s) under consideration?
3. For the regulatory option(s) under consideration, are potential EJ concerns
created or mitigated compared to the baseline?"
To address these questions, EPA developed an analytical approach that considers
the purpose and specifics of the rulemaking, as well as the nature of known and potential
exposures and impacts. For the rule, we quantitatively evaluate the potential for disparities
in PM2.5 exposures and mortality rates across different demographic populations under
illustrative control strategies associated with implementation of the current standards
(12/35 ng/m3, i.e., the baseline) and lower alternative PM2.5 standard levels (10/35 mg/m3,
10/30 |Lxg /m3, 9/35 |Lxg /m3, and 8/35 |Lxg /m3) at the national and regional levels.
Specifically, we provide information on total burden, absolute changes, and proportional
changes in 1) exposures, in terms of annual PM2.5 concentrations and 2) premature
mortality, in terms of rates per 100,000 individuals across and within various demographic
populations. Each type of analysis has strengths and weaknesses, but when taken together,
can respond to the above three questions from EPA's Environmental Justice (EJ) Technical
Guidance.
Beginning with the first question, under the 12/35 |~ig/m3 analytical baseline, some
populations are predicted to experience disproportionately higher annual PM2.5 exposures
nationally than the reference (overall) population, both in terms of aggregated average
20
-------
exposure and across the distribution of air quality. Specifically, populations who are
linguistically isolated, Hispanic, Asian, Black, less educated, unemployed, uninsured, and
living below the poverty line live in areas with higher national annual PM2.5 concentrations,
on average and across the distributions, than both the overall reference population or
other populations (e.g., non-Hispanic, White, and more educated) (Section 6.3.1). In
addition, those living in urban areas that received Home Owners' Loan Corporation (HOLC)
neighborhood quality grades for mortgage lending purposes have higher national annual
PM2.5 concentrations, both for urban areas designated as "redlined" (i.e., 'Grade D' or
"hazardous") and those not redlined (i.e., Grades A, B, and C) as compared to everywhere
else (identified here as 'Ungraded by HOLC') (Mitchell et al., 2018, Swope et al., 2022, Lee et
al., 2022). Those living in urban areas that received a grade of D are estimated to
experience the highest concentrations, both on average and across PM2.5 concentration
distributions, of all demographic groups analyzed. These disproportionalities are also
observed at the regional level, to different extents.
In response to the second question, while lower standard levels would be predicted
to reduce PM2.5 exposures and mortality rates across all demographic groups, disparities
seen in the baseline persist under lower alternative standard levels. However, as to the
third question, for most populations assessed, PM2.5 exposure disparities are mitigated in
the illustrative air quality scenarios reflecting control strategies (10/35 |~ig/m3,10/30
Hg/m3, 9/35 ng/m3, and 8/35 |j,g/m3) as compared to the baseline (12/35 |j,g/m3), and
more so as the alternative standard levels become more stringent. At the national scale,
populations that are linguistically isolated, Hispanic, Asian, those less educated, and
unemployed populations are estimated to see greater proportional reductions in PM2.5
concentrations than reference populations under all lower standard levels evaluated, with
proportional reductions increasing as the standard levels decrease. This is also the case for
urban areas that received HOLC neighborhood quality grades and for those living in areas
that were historically redlined within those urban areas. In addition, exposure disparities
in baseline Black and uninsured populations are estimated to be mitigated when moving to
alternative standard levels of 8/35 |~ig/m3. Considering the four geographic areas
(northeast, southeast, west, and California), proportionally greater reductions in PM2.5
21
-------
concentrations experienced by various populations with baseline exposure disparities are
most notable in California, whereas PM2.5 concentration reductions are greatest. In
Appendix Section 6.6.3 we provide insight into exposures in areas with remaining air
quality challenges (i.e., without sufficient emissions control strategies to reach alternative
standard levels).
In terms of health effects, some populations are also predicted to experience
disproportionately higher rates of premature mortality than the reference population
under the baseline scenario. Black populations over the age of 64 are predicted to
experience substantially greater mortality rate burdens as compared to White populations
over the age of 64. When moving to more stringent standard levels, Black and non-Hispanic
Black populations are predicted to experience proportionally similar mortality rate
reductions as compared to the reference populations under control strategies associated
with 12/35-10/35 or 12/35-10/30, but greater reductions in mortality rates under control
strategies associated with 12/35-9/35 or 12/35-8/35. Disparities in national PM2.5
mortality rates across demographic groups are mitigated for Hispanics in all the alternative
PM standard levels (10/35,10/30, 9/35, and 8/35), as compared to the baseline.
ES.2 Qualitative Assessment of the Remaining Air Quality Challenges
For the revised standard levels of 9/35 |~ig/m3, the analysis indicates that some
areas in the northeast and southeast, as well as in the west and California may still need
emissions reductions (Figure ES-3). As discussed in Chapters 2 and 3, the remaining air
quality challenges for the revised standard levels can be grouped into the following "bins":
counties with near-road monitors, border areas, small mountain valleys, and California
areas. By bin, Table ES-9 below summarizes the counties that may need additional
emissions reductions for the revised standard levels.
22
-------
Table ES-9 Summary of Counties by Bin that Still Need Emissions Reductions for
the Revised Primary Standard Levels of 9/35 (ig/m3
PM2.5 Emissions
Reductions
Bin
Area
Counties3 for 9/35 mg/m3
Still Needed
Near-Road Monitors
Northeast
Bergen County, NJ
Hamilton County, OH
75
55
Border Areas
Southeast
Cameron County, TX
Hidalgo County, TX
351
687
California
Imperial County, CA
2,516
Small Mountain Valleys
West
Plumas County, CA
Lemhi County, ID
Shoshone County, ID
Lincoln County, MT
Klamath County, OR
626
235
558
894
60
California Areas
Fresno County, CA (SJVAPCD)
Kern County, CA (SJVAPCD)
Kings County, CA (SJVAPCD)
Los Angeles County, CA (SCAQMD)
Madera County, CA (SJVAPCD)
Merced County, CA (SJVAPCD)
Orange County, CA (SCAQMD)
Riverside County, CA (SCAQMD)
San Bernardino County, CA (SCAQMD)
Stanislaus County, CA (SJVAPCD)
Tulare County, CA (SJVAPCD)
San Joaquin County, CA (SJVAPCD)
Alameda County, CA (BAAQMD)
Calaveras County, CA
Sutter County, CA
441
726
730
1,450
38
339
1,179
2,551
2,475
414
763
48
34
50
31
Note: For California counties that are part of multi-county air districts, the relevant district is indicated in parentheses;
BAAQMD = Bay Area Air Quality Management District, SCAQMD = South Coast Air Quality Management District, and
SJVAPCD= San Joaquin Valley Air Pollution Control District.
a The following counties had no identified PM2.5 emissions reductions because available controls were applied for the
current standard of 12/35 |ig/m3 and additional controls were not available to apply for analyses of the revised and
alternative standards: Colusa, Mono, Plumas, and Riverside, CA, Lake, OR, and Yakima, WA.
The characteristics of the air quality challenges for these areas include features of
certain near-road sites with challenging local conditions, cross-border transport, effects of
complex terrain in the west and California, and identifying wildfire influence on projected
PM2.5 DVs that could potentially qualify for exclusion as atypical, extreme, or
unrepresentative events (U.S. EPA, 2019a). For bin-specific detailed discussions of these air
quality challenges, see Chapter 2, Section 2.4. Further, for each bin for discussions of the
estimated PM2.5 emissions reductions needed, the control strategy analyses and controls
applied, the estimated PM2.5 emissions reductions still needed after the application of
controls, and the bin-specific air quality challenges, see Chapter 3, Section 3.2.5.
23
-------
For counties with near-road monitors, understanding the nature of the local
contributions under complex conditions would require detailed local studies beyond the
scope of the RIA. For the border areas that may be influenced by cross-border emissions,
more detailed analyses of international transport emissions will be needed to assess the
relevance of Section 179B of the Clean Air Act. For the small mountain valleys in the west
that are influenced by the temperature inversions, residential wood combustion, and
wildfire smoke additional detailed analyses that reflect local PM2.5 response factors,
emissions inventory information, and control measure information will be needed. In
addition, more detailed analyses will be needed to characterize the influence of wildfires on
PM2.5 concentrations and the potential for some wildfires to qualify for exclusion as
atypical, extreme, or unrepresentative events.
The air quality in the SJVAPCD and SCAQMD is influenced by complex terrain and
meteorological conditions that are best characterized with a high-resolution air quality
modeling platform developed for the specific conditions of the air basins. Specific, local
information on area source controls to reduce emissions from agricultural dust and
burning, prescribed burning, and many of the non-point (area) emissions sources (e.g.,
commercial and residential cooking) will be needed given the magnitude of emissions from
these sources in these areas. In addition, Alameda, Calaveras, and Sutter Counties were
influenced by wildfires during the monitoring period used for the air quality projections.
Further, more detailed analyses will be needed to characterize the influence of wildfires on
PM2.5 concentrations and the potential for some wildfires to qualify for exclusion as
atypical, extreme, or unrepresentative events.
ES.3 Changes in Data Used and Methods Between Proposal and Final RIAs
Between the proposal and final RIAs, we made some minor updates to the data used,
methods applied, and results presented; overall the results of the benefit-cost analysis
presented in this final rule RIA are similar to the results presented in the proposed rule
RIA.
For the proposal RIA, to project air quality to the future the CMAQ model was
applied to simulate air quality over the U.S. during 2016 and for a case with emissions
representative of 2032. For the final RIA, the CMAQ model was used to simulate air quality
24
-------
over the U.S. during 2018 and for a case with emissions representative of 2032. Some of the
additional policies and final rules reflected in the updated modeling include provisions of
tax incentives in the Inflation Reduction Act of 2022, the 2023 Final Good Neighbor Plan for
the 2015 Ozone NAAQS, the 2022 Control of Air Pollution from New Motor Vehicles, and
the 2021 Final Rule to Revise Existing National GHG Emissions Standards for Passenger
Cars and Light Trucks Through Model Year 2026. For additional information on the air
quality modeling platform, see Chapter 2, Section 2.2.1.
In addition, for the control strategy analyses, with the exception of two area source
controls, the non-point (area) source, residential wood combustion, and area source
fugitive dust controls were applied at different rule penetration (RP) rates depending on
the reductions needed in particular areas at different standard levels.6 The controls were
applied at between 5 percent and 35 percent RP at 5 percent increments. In the proposal
RIA, these controls were applied at 10 percent and 25 percent RP.
When accounting for reductions from neighbor counties in the northeast and
southeast, we identified controls and reductions from adjacent or neighboring counties in 2
rounds. Note that a county can be both a core/home county and an adjacent/neighboring
county. In round 1, we identified controls and reductions in the home counties for
application in the home counties. In round 2, we identified controls and reductions in
neighboring counties for application in home counties that still needed reductions. If a
county was both a home county and a neighboring county in round 1, in the final RIA
before round 2 we adjusted any potential remaining reductions needed for a home county
to account for reductions that were applied in round 1 for application in its neighboring
county. In the proposal RIA, we did not make these adjustments before applying controls in
round 2. For additional information on the control strategy analyses, see Chapter 3, Section
3.2.2.
Lastly, for the EJ analyses we added several population groups to the exposure EJ
assessment, including employment status, health insurance status, linguistic isolation, and
redlined areas and added a second epidemiologic study that stratifies PM2.5-attributable
6 RP is the percent of the area source inventory emissions that the control is applied to at a specified percent
control efficiency.
25
-------
mortality by race/ethnicity, similar to the approach used by benefits assessment. In
addition, we replaced the case study of areas changing when moving to a revised standard
level of 9/35 |~ig/m3 with new figures showing nuanced disparities at high exposures under
various standard levels. For additional information on the EJ analyses, see Chapter 6.
ES.4 Results of Benefit-Cost Analysis
As discussed above and in Chapter 3, Section 3.2.4, the estimated PM2.5 emissions
reductions from control applications do not fully account for all the emissions reductions
needed to reach the revised and less and more stringent alternative standard levels in
some counties in the northeast, southeast, west, and California. In Chapter 2, Section 2.4
and Chapter 3, Section 3.2.5, we discuss the remaining air quality challenges for areas in
the northeast and southeast, as well as in the west and California for the revised standard
levels of 9/35 |~ig/m3. The EPA calculates the monetized net benefits of the revised and
alternative standard levels by subtracting the estimated monetized compliance costs from
the estimated monetized benefits in 2032. The estimates of costs and benefits do not fully
account for all of the emissions reductions needed to reach the revised and less and more
stringent alternative standard levels. In 2032, the monetized net benefits of the revised
standard levels of 9/35 |~ig/m3 are approximately $22 billion and $46 billion using a 3
percent real discount rate for the benefits estimates (in 2017$). The benefits are associated
with two point estimates from two different epidemiologic studies discussed in more detail
in Chapter 5, Section 5.2.3. Table ES-10 presents a summary of these impacts for the
revised standard levels and the less and more stringent alternative standard levels for
2032.
26
-------
Table ES-10 Estimated Monetized Benefits, Costs, and Net Benefits of the Control
Strategies Applied Toward the Primary Revised and Alternative
Standard Levels of 10/35 |ig/m3,10/30 |ig/m3, 9/35 |ig/m3, and 8/35
(ig/m3 in 2032 for the U.S. (millions of 2017$)
10/35
10/30
9/35
8/35
Benefits3
$8,500 and $17,000
$10,000 and $21,000
$22,000 and $46,000
$48,000 and $99,000
Costsb
$200
$340
$590
$1,500
Net Benefits
$8,300 and $17,000
$9,900 and $21,000
$22,000 and $46,000
$46,000 and $97,000
Notes: Rows may not appear to add correctly due to rounding.
We focus results to provide a snapshot of costs and benefits in 2032, using the best available information to
approximate social costs and social benefits recognizing uncertainties and limitations in those estimates.
The estimated PM2.5 emissions reductions from the control strategies do not fully account for all the emissions
reductions needed to reach the revised and less and more stringent alternative standard levels in some
counties in the northeast, southeast, west, and California.
a We assume that there is a cessation lag between the change in PM exposures and the total realization of
changes in mortality effects. Specifically, we assume that some of the incidences of premature mortality related
to PM2.5 exposures occur in a distributed fashion over the 20 years following exposure, which affects the
valuation of mortality benefits at different discount rates. Similarly, we assume there is a cessation lag between
the change in PM exposures and both the development and diagnosis of lung cancer. The benefits are associated
with two point estimates from two different epidemiologic studies, and we present the benefits calculated at a
real discount rate of 3 percent. The benefits exclude additional health and welfare benefits that could not be
quantified (see Chapter 5, Sections 5.2.4 and 5.2.5).
b The costs are annualized using a 7 percent interest rate.
As part of fulfilling analytical guidance with respect to E.0.12866, the EPA presents
estimates of the present value (PV) of the monetized benefits and costs over the twenty-
year period 2032 to 2051. To calculate the present value of the social net benefits of the
revised standard levels, annual benefits and costs are discounted to 2023 at 3 percent and
7 percent discount rates as recommended by OMB's Circular A-4. The EPA also presents the
equivalent annualized value (EAV), which represents a flow of constant annual values that,
had they occurred in each year from 2032 to 2051, would yield a sum equivalent to the PV.
The EAV represents the value of a typical cost or benefit for each year of the analysis, in
contrast to the 2032-specific estimates.
For the twenty-year period of 2032 to 2051, for the revised standard levels of 9/35
Hg/m3the PV of the net benefits, in 2017$ and discounted to 2023, is $540 billion when
using a 3 percent discount rate and $280 billion when using a 7 percent discount rate. The
EAV is $36 billion per year when using a 3 percent discount rate and $27 billion when
using a 7 percent discount rate. The comparison of benefits and costs in PV and EAV terms
for the revised standard levels can be found in Table ES-11. Estimates in the table are
presented as rounded values.
27
-------
Table ES-11 Summary of Present Values and Equivalent Annualized Values for
Estimated Monetized Compliance Costs, Benefits, and Net Benefits of
the Control Strategies Applied Toward the Revised Primary Standard
Levels of 9/35 ng/m3 (millions of 2017$, 2032-2051, discounted to
2023 using 3 and 7 percent discount rates)
Benefits3
Costsb
Net Benefits
Year
3%
7%
3%
7%
3%
7%
2032
$35,000
$25,000
$460
$320
$35,000
$25,000
2033
$34,000
$24,000
$440
$300
$34,000
$23,000
2034
$33,000
$22,000
$430
$280
$33,000
$22,000
2035
$32,000
$21,000
$420
$260
$32,000
$20,000
2036
$31,000
$19,000
$400
$250
$31,000
$19,000
2037
$31,000
$18,000
$390
$230
$30,000
$18,000
2038
$30,000
$17,000
$380
$220
$29,000
$17,000
2039
$29,000
$16,000
$370
$200
$28,000
$15,000
2040
$28,000
$15,000
$360
$190
$28,000
$14,000
2041
$27,000
$14,000
$350
$180
$27,000
$14,000
2042
$26,000
$13,000
$340
$160
$26,000
$13,000
2043
$26,000
$12,000
$330
$150
$25,000
$12,000
2044
$25,000
$11,000
$320
$140
$25,000
$11,000
2045
$24,000
$10,000
$310
$130
$24,000
$10,000
2046
$23,000
$9,800
$300
$130
$23,000
$9,600
2047
$23,000
$9,100
$290
$120
$22,000
$9,000
2048
$22,000
$8,500
$280
$110
$22,000
$8,400
2049
$21,000
$8,000
$280
$100
$21,000
$7,900
2050
$21,000
$7,400
$270
$96
$21,000
$7,300
2051
$20,000
$7,000
$260
-ee-
00
vo
$20,000
$6,900
Present Value
$540,000
$290,000
$7,000
$3,700
$540,000
$280,000
Equivalent
Annualized Value
$36,000
$27,000
$470
$350
$36,000
$27,000
Notes: Rows may not appear to add correctly due to rounding. The annualized present value of costs and
benefits are calculated over a 20-year period from 2032 to 2051.
¦'The benefits values use the larger of the two avoided premature deaths estimates presented in Chapter 5,
Table 5-5, and are discounted at a rate of 3 percent over the SAB-recommended 20-year segmented lag. The
benefits exclude additional health and welfare benefits that could not be quantified (see Chapter 5, Sections
5.2.4 and 5.2.5).
b The costs are annualized using a 7 percent interest rate.
28
-------
ES.5 References
Mitchell, B and Franco J. (2018). HOLC "redlining" maps: The persistent structure of
segregation and economic inequality. National Community Reinvestment Coalition,
https://ncrc.org/?s=redlining+maps, https://ncrc.org/wp-
content/uploads/dlm_uploads/2018/02/NCRC-Research-HOLC-10.pdf.
Lee, EK, Donley, G, Cielielski, T, Gill, I, Yamoah, 0, Roche, A, Martinez, R, and Freedman, D.
Health outcomes in redlined versus non-redlined neighborhoods: A systematic review
and meta-analysis. Social Science & Medicine. 294 (2022) 114696.
Pope III, CA, Lefler, JS, Ezzati, M, Higbee, JD, Marshall, JD, Kim, S-Y, Bechle, M, Gilliat, KS,
Vernon, SE and Robinson, AL (2019). Mortality risk and fine particulate air pollution in
a large, representative cohort of US adults. Environmental health perspectives 127(7):
077007.
Sacks, JD, Lloyd, JM, Zhu, Y, Anderton, J, Jang, CJ, Hubbell, B and Fann, N (2018). The
Environmental Benefits Mapping and Analysis Program-Community Edition (BenMAP-
CE): A tool to estimate the health and economic benefits of reducing air pollution.
Environmental Modelling Software 104: 118-129
Swope, C, Hernandez, D, and Cushing, L. (2022). The Relationship of Historical Redlining
with Present-Day Neighborhood Environmental and Health Outcomes: A Scoping
Review and Conceptual Model. Journal of Urban Health 99: 959-983.
U.S. EPA (2012). Regulatory Impact Analysis for the Final Revisions to the National
Ambient Air Quality Standards for Particulate Matter, U.S. Environmental Protection
Agency, Office of Air Quality Planning and Standards, Research Triangle Park, NC. EPA-
452/R-12-005. Available at: https://www.epa.gov/sites/default/files/2020-
07/documents/naaqs-pm_ria_final_2012-12.pdf.
U.S. EPA (2015). Guidance on Considering Environmental Justice During the Development
of Regulatory Actions. Available at https://www.epa.gov/sites/default/files/2015-
06/documents/considering-ej-in-rulemaking-guide-final.pdf.
U.S. EPA (2019a). Additional Methods, Determinations, and Analyses to Modify Air Quality
Data Beyond Exceptional Events. U.S. Environmental Protection Agency, Office of Air
Quality Planning and Standards, Research Triangle Park, NC, EPA-457/B-19-002.
Available: https://www.epa.gov/sites/default/files/2019-
04/documents/clarification_memo_on_data_modification_methods.pdf
U.S. EPA (2019b). CoST v3.7 User's Guide. Office of Air Quality Planning and Standards,
Research Triangle Park, NC. November 2019. Available at <
https://www.cmascenter.org/help/documentation. cfm?model=cost&version=3.7>.
29
-------
Woodruff, TJ, Darrow, LA and Parker, JD (2008). Air pollution and postneonatal infant
mortality in the United States, 1999-2002. Environmental Health Perspectives 116(1):
110-115.
Wu, X, Braun, D, Schwartz, J, Kioumourtzoglou, M and Dominici, F (2020). Evaluating the
impact of long-term exposure to fine particulate matter on mortality among the elderly.
Science advances 6(29): eaba5692.
30
-------
CHAPTER 1: OVERVIEW AND BACKGROUND
Overview of the Final Rule
On June 10, 2021, the EPA announced its decision to reconsider the 2020 Particulate
Matter (PM) NAAQS final action. The EPA is reconsidering the December 2020 decision
because the available scientific evidence and technical information indicate that the current
standards may not be adequate to protect public health and welfare, as required by the
CAA. In general, the Administrator recognizes that the primary annual PM2.5 standard is
most effective at controlling exposures to "typical" daily PM2.5 concentrations that are
experienced over the year, while the primary 24-hour PM2.5 standard, with its 98th
percentile form, is most effective at limiting daily "peak" PM2.5 concentrations. The EPA has
concluded that the existing primary annual PM2.5 standard for PM, set at a level of 12.0
|ig/m3, is not requisite to protect public health with an adequate margin of safety. The EPA
Administrator is revising the existing annual standard to provide increased public health
protection. Specifically, the EPA Administrator is revising the level of the annual standard
to 9 |ig/m3. The EPA Administrator is retaining the primary 24-hour PM2.5 standard at its
current level of 35 |ig/m3. The Administrator is also retaining the primary 24-hour PM10
standard, which provides public health protection against PMio-2.5-related health effects, at
its current level of 150 |ig/m3.
The EPA also concluded that the existing secondary PM standards are requisite to
protect public welfare from known or anticipated effects and is not changing the secondary
standards for PM at this time. Specifically, for the secondary annual PM2.5 standard, the EPA
Administrator is retaining the existing standard of 15.0 |ig/m3. For the secondary 24-hour
PM2.5 standard, the EPA Administrator is retaining the existing standard of 35 |ig/m3. For
the secondary 24-hour PM10 standard, the EPA Administrator is retaining the existing
standard of 150 |ig/m3. The docket for the rulemaking is EPA-HQ-OAR-2015-0072.
Overview of the Regulatory Impact Analysis
This chapter summarizes the purpose and background of this Regulatory Impact
Analysis (RIA). In this RIA, we are analyzing the revised annual and current 24-hour
alternative standard levels of 9/35 |~ig/m3, as well as the following less and more stringent
31
-------
alternative standard levels: (1) a less stringent alternative annual standard level of 10 |Lxg/-
m3 in combination with the current 24-hour standard (i.e.,10/35 |j,g/m3), (2) a more
stringent alternative annual standard level of 8 |~ig/m3 in combination with the current 24-
hour standard (i.e., 8/35 |j,g/m3), and (3) a more stringent alternative 24-hour standard
level of 30 |~ig/m3 in combination with the an annual standard level of 10 ng/m3 (i.e., 10/30
Hg/m3). The RIA presents estimated costs and benefits of the control strategies analyzed
for the revised and less and more stringent alternative standard levels. According to the
Clean Air Act ("the Act"), the Environmental Protection Agency (EPA) must use health-
based criteria in setting the NAAQS and cannot consider estimates of compliance cost.
The analyses in this RIA rely on national-level data (emissions inventory and control
measure information) for use in national-level assessments (air quality modeling, control
strategies, environmental justice, and benefits estimation). However, the ambient air
quality issues being analyzed are highly complex and local in nature, and the results of
these national-level assessments therefore contain uncertainty. It is beyond the scope of
this RIA to develop detailed local information for the areas being analyzed, including
populating the local emissions inventory, obtaining local information to increase the
resolution of the air quality modeling, and obtaining local information on emissions
controls, all of which would reduce some of the uncertainty in these national-level
assessments. For example, having more refined data would be ideal for agricultural dust
and burning, prescribed burning, and non-point (area) sources due to their large
contribution to primary PM2.5 emissions and the limited availability of emissions controls.1
In addition, for residential wood combustion emissions, people will respond differently to
the various regulations and incentives offered for controlling PM2.5 emissions from wood
burning, making it important to identify the right balance of controls for each area.
To maximize the number of emissions sources included and controls analyzed in the
analyses, we included a lower emissions source size threshold (5 tons per year) and a
higher marginal cost per ton threshold ($160,000/ton) than reflected in prior NAAQS RIAs
(25-50 tpy, $15,000-$20,000/ton). As discussed in Chapter 2, Section 2.1.3, given historical
1 Examples of area source emissions include area fugitive dust, residential wood combustion, and commercial
cooking emissions.
32
-------
and projected trends in NOx and SO2 emissions reductions (reducing background PM
concentrations and increasing the importance of urban PM concentrations), we analyze
direct PM emissions reductions because our modeling indicates that these reductions will
be the most effective at reducing PM concentrations in counties projected to exceed the
revised and alternative standard levels. The spatial distributions of PM2.5 concentrations in
the U.S. are characterized by an "urban increment" of consistently higher PM2.5
concentrations over urban than surrounding areas. Monitored concentrations are highest
in urban areas and relatively low in rural areas. Conceptually, PM2.5 concentrations in
urban areas can be viewed as the superposition of the urban increment and the
contributions from regional and natural background sources. The decreases in
anthropogenic SO2 and NOx emissions in recent decades have reduced regional background
concentrations and increased the relative importance of the urban increment. The
projections of additional large reductions in SO2 and NOx emissions in the 2032 case
further motivate the need for control of local primary PM2.5 sources to address the highest
PM2.5 concentrations in urban areas. Lastly, Chapter 2, Section 2.4 and Chapter 3, Section
3.2.5 discuss the remaining air quality challenges for areas in the northeast and southeast,
as well as in the west and California for the revised standard levels of 9/35 |j,g/m3; the
areas include counties with near-road monitors, border areas, counties in small western
mountain valleys, and counties in California's air basins and districts. The characteristics of
the air quality challenges for these areas include features of certain near-road sites with
challenging local conditions, cross-border transport, effects of complex terrain in the west,
and identifying wildfire influence on projected PM2.5 DVs that could potentially qualify for
exclusion as atypical, extreme, or unrepresentative events (U.S. EPA, 2019).
The remainder of this chapter provides a brief background on the NAAQS, the need
for the NAAQS, and an overview of structure of this RIA. The EPA prepared this RIA both to
provide the public with information on the benefits and costs of meeting a revised PM2.5
NAAQS and to meet the requirements of Executive Orders 12866,13563, and 14094.
1.1 Background
In setting primary ambient air quality standards, the EPA's responsibility under the
law is to establish standards that protect public health, without consideration of the costs
33
-------
of implementing those standards. As interpreted by the Agency and the courts, the CAA
requires the EPA to create standards based on health considerations only. The prohibition
against the consideration of cost in the setting of the primary air quality standards,
however, does not mean that costs or other economic consequences are unimportant or
should be ignored. The Agency believes that consideration of costs and benefits is essential
to making efficient, cost-effective decisions for implementing these standards. The impact
of cost and efficiency is considered by states during the implementation process, as they
decide what timelines, strategies, and policies are appropriate for their circumstances. This
RIA is not part of the standard setting and is intended to inform the public about the
potential costs and benefits that may result when new standards are implemented.
1.1.1 National Ambient Air Quality Standards
Sections 108 and 109 of the CAA govern the establishment and revision of the
NAAQS. Section 108 (42 U.S.C. 7408) directs the Administrator to identify pollutants that
"may reasonably be anticipated to endanger public health or welfare" and to issue air
quality criteria for them. These air quality criteria are intended to "accurately reflect the
latest scientific knowledge useful in indicating the kind and extent of all identifiable effects
on public health or welfare which may be expected from the presence of [a] pollutant in the
ambient air." PM is one of six pollutants for which the EPA has developed air quality
criteria.
Section 109 (42 U.S.C. 7409) directs the Administrator to propose and promulgate
"primary" and "secondary" NAAQS for pollutants identified under section 108. Section
109(b)(1) defines a primary standard as an ambient air quality standard "the attainment
and maintenance of which in the judgment of the Administrator, based on [the] criteria and
allowing an adequate margin of safety, [is] requisite to protect the public health." A
secondary standard, as defined in section 109(b)(2), must "specify a level of air quality the
attainment and maintenance of which in the judgment of the Administrator, based on [the]
criteria, is requisite to protect the public welfare from any known or anticipated adverse
effects associated with the presence of [the] pollutant in the ambient air." Welfare effects as
defined in section 302(h) [42 U.S.C. 7602(h)] include but are not limited to "effects on soils,
water, crops, vegetation, manmade materials, animals, wildlife, weather, visibility and
34
-------
climate, damage to and deterioration of property, and hazards to transportation, as well as
effects on economic values and on personal comfort and well-being."
Section 109(d) of the CAA directs the Administrator to review existing criteria and
standards at 5-year intervals. When warranted by such review, the Administrator is to
retain or revise the NAAQS. After promulgation or revision of the NAAQS, the standards are
implemented by the states.
1.1.2 Role of Executive Orders in the Regulatory Impact Analysis
While this RIA is separate from the NAAQS decision-making process, several
statutes and executive orders still apply to any public documentation. The analyses
required by these statutes and executive orders are presented in the final rule preamble,
and below we briefly discuss requirements of Orders 12866,13563, and 14094 and the
guidelines of the Office of Management and Budget (OMB) Circular A-4 (U.S. OMB, 2003).
In accordance with Executive Orders 12866,13563, and 14094 and the guidelines of
OMB Circular A-4, the RIA presents the estimated benefits and costs associated with
control strategies for a range of annual and 24-hour PM2.5 standard levels. The estimated
benefits and costs associated with emissions controls are incremental to a baseline of
attaining the current standards (annual and 24-hour PM2.5 standards of 12/35 |~ig/m3 in
ambient air). OMB Circular A-4 requires analysis of one potential alternative standard level
more stringent than the final standard and one less stringent than the final standard. The
Agency is revising the current annual PM2.5 standard to a level of 9 |~ig/m3. The Agency is
also retaining the current 24-hour standard of 35 |~ig/m3. In this RIA, we are analyzing the
revised annual and current 24-hour alternative standard levels of 9/35 ng/m3, as well as
the following less and more stringent alternative standard levels: (1) a less stringent
alternative annual standard level of 10 ng/m3 in combination with the current 24-hour
standard (i.e.,10/35 |j,g/m3), (2) a more stringent alternative annual standard level of 8
Hg/m3 in combination with the current 24-hour standard (i.e., 8/35 |j,g/m3), and (2) a more
stringent alternative 24-hour standard level of 30 ng/m3 in combination with an annual
standard level of 10 ng/m3 (i.e., 10/30 |j,g/m3).
35
-------
1.1.3 Nature of the Analysis
The control strategies presented in this RIA are an illustration of one possible set of
control strategies states might choose to implement in response to the revised standards.
States—not the EPA—will implement the revised NAAQS and will ultimately determine
appropriate emissions control strategies and measures. State Implementation Plans (SIPs)
will likely vary from the EPA's estimates provided in this analysis due to differences in the
data and assumptions that states use to develop these plans. Because states are ultimately
responsible for implementing strategies to meet the revised standards, the control
strategies in this RIA are considered hypothetical. The hypothetical strategies were
constructed with the understanding that there are inherent uncertainties in estimating and
projecting emissions and applying controls to specific emissions or emissions sources.
Additional important uncertainties and limitations are documented in the relevant
chapters of the RIA.
The EPA's national program rules require technology application or emissions limits
for a specific set of sources or source groups. In contrast, a NAAQS establishes a standard
level and requires states to identify and secure emissions reductions to meet the standard
level from any set of sources or source groups. To avoid double counting the impacts of
NAAQS and other national program rules, the EPA includes previously promulgated federal
regulations and enforcement actions in its baseline for this analysis (See Section 1.3.1
below for additional discussion of the baseline). The benefits and costs of the revised
standards will not be realized until specific controls are mandated by SIPs or other federal
regulations.
1.2 The Need for National Ambient Air Quality Standards
OMB Circular A-4 indicates that one of the reasons a regulation such as the NAAQS
may be issued is to address a market failure. The major types of market failure include
externality, market power, and inadequate or asymmetric information. Correcting market
failures is one reason for regulation, but it is not the only reason. Other possible
justifications include improving the function of government, removing distributional
unfairness, or promoting privacy and personal freedom.
36
-------
Environmental problems are classic examples of externalities -- uncompensated
benefits or costs imposed on another party as a result of one's actions. For example, the
smoke from a factory may adversely affect the health of local residents and soil the
property in nearby neighborhoods. If bargaining was costless and all property rights were
well defined, people would eliminate externalities through bargaining without the need for
government regulation.
From an economics perspective, setting an air quality standard is a straightforward
remedy to address an externality in which firms emit pollutants, resulting in health and
environmental problems without compensation for those incurring the problems. Setting a
standard with an adequate margin of safety attempts to place the cost of control on those
who emit the pollutants and lessens the impact on those who suffer the health and
environmental problems from higher levels of pollution. For additional discussion on the
PM2.5 air quality problem, see Chapter 2 of the Policy Assessment for the Reconsideration
of the National Ambient Air Quality Standards for Particulate Matter (U.S. EPA, 2022a).
1.3 Design of the Regulatory Impact Analysis
The RIA presents the estimates of costs and benefits of applying hypothetical
national control strategies for the revised and less and more stringent alternative annual
and 24-hour standard levels of 10/35 |~ig/m3,10/30 |~ig/m3, 9/35 |~ig/m3, and 8/35 |~ig/m3,
incremental to attaining the current PM2.5 standards and implementing existing and
expected regulations. We assume that potential nonattainment areas everywhere in the
U.S. will be designated such that they are required to attain the revised standards by 2032.
The selection of 2032 as the analysis year in the RIA does not predict or prejudge
attainment dates that will ultimately be assigned to individual areas under the CAA. The
CAA contains a variety of potential attainment dates and flexibility to move to later dates,
provided that the date is as expeditious as practicable. For the purposes of this analysis, the
EPA assumes that it would likely finalize designations for the revised particulate matter
NAAQS in late 2025. Furthermore, also for the purposes of this analysis and depending on
the precise timing of the effective date of those designations, the EPA assumes that
nonattainment areas classified as Moderate would likely have to attain in late 2032. As
such, we selected 2032 as the primary year of analysis. States with areas classified as
37
-------
Moderate and higher are required to develop attainment demonstration plans for those
nonattainment areas.
The EPA recognizes that areas designated nonattainment for the final PM2.5 NAAQS
and classified as Moderate will likely incur some costs prior to the 2032 analysis year.
States with nonattainment areas designated as Moderate are required by the CAA to
develop SIPs demonstrating attainment by no later than the assigned attainment date. The
CAA also requires these states to address Reasonably Available Control Technologies
(RACT) for sources in the Moderate nonattainment area, which would lead to additional
point source controls in an area beyond existing federal emissions control requirements.
Additionally, the CAA requires some Moderate areas with larger populations to implement
basic vehicle inspection and maintenance in the area. Should these federal programs and
CAA required programs prove inadequate for the area to attain the revised standards by
the attainment date, the state would need to identify additional emissions controls in its
SIP to meet attainment requirements.
1.3.1 Establishing the Baseline for Evaluating Revised and Alternative Standard
Levels
To develop and evaluate control strategies, it is important to estimate PM2.5 levels in
the future after attaining the current standards of 12/35 |~ig/m3, taking into account
projections of future air quality reflecting on-the-books Federal regulations, enforcement
actions, state regulations, population and where possible, economic growth. Establishing
this baseline for the analysis then allows us to estimate the incremental costs and benefits
associated with the revised and alternative standard levels. For the purposes of this
analysis and depending on the precise timing of the effective date of designations, the EPA
assumes that areas will be designated such that they are required to reach attainment by
2032, and we developed our projected baselines for emissions and air quality for 2032.2
Attaining the current standards of 12/35 |~ig/m3 reflects emissions reductions (i)
already achieved as a result of national regulations, (ii) expected prior to 2032 from
2 Because of the complex nature of air quality in California, we adjusted baseline air quality in 2032 to reflect
mobile source NOx emissions reductions for California that would occur between 2032 and 2035. These
emissions reductions are the result of mobile source regulations expected to be fully implemented by 2035.
California provided the mobile source inventory data for 2035.
38
-------
recently promulgated national regulations (i.e., reductions that were not realized before
promulgation of the previous standard but are expected prior to attainment of the current
PM2.5 standards), and (iii) from additional controls that the EPA estimates need to be
included to reach the current standards. Additional emissions reductions achieved as a
result of state and local agency regulations and voluntary programs are reflected to the
extent that they are represented in emissions inventory information submitted to the EPA
by state and local agencies.
We took two steps to develop the baseline for this analysis. First, national PM2.5
concentrations were projected to the analysis year (2032) based on forecasts of population
and where possible, economic growth and the application of emissions controls resulting
from national rules promulgated prior to this analysis, as well as state programs and
enforcement actions. Second, we estimated additional emissions reductions needed to
meet the current standards of 12/35 |~ig/m3. Below is a list of some of the national rules
reflected in the baseline. For a more complete list, please see Chapter 2, Section 2.2.1 (Air
Quality Modeling Platform) and the technical support document (TSD) for the 2018v2
emissions modeling platform titled Preparation of Emissions Inventories for the 2018v2
North American Emissions Modeling Platform (U.S. EPA, 2023a). If the national rules
reflected in the baseline result in changes in PM2.5 concentrations or actual emissions
reductions that are lower or higher than those estimated, the costs and benefits estimated
in this RIA would be higher or lower, respectively.
For EGUs, rules in the baseline include:
• Final Good Neighbor Plan for the 2015 Ozone NAAQS (2023),
• Revised Cross-State Air Pollution Rule Update (2021),
• Standards of Performance for Greenhouse Gas Emissions from New, Modified, and
Reconstructed Stationary Sources (2015),
• Mercury and Air Toxics Rule (2011), and
• Provisions of tax incentives in the Inflation Reduction Act of 2022 (IRA).
For mobile sources, rules in the baseline include:
• Control of Air Pollution from New Motor Vehicles: Heavy-Duty Engine and Vehicle
Standards (2022),
39
-------
• Final Rule to Revise Existing National GHG Emissions Standards for Passenger Cars
and Light Trucks Through Model Year 2026 (2021),
• GHG Emissions Standards and Fuel Efficiency Standards for Medium- and Heavy-
Duty Engines and Vehicles, Phase 2 (2016), and
• Tier 3 Motor Vehicle Emission and Fuel Standards (2014).
For non-EGUs, rules in the baseline include:
• Final Good Neighbor Plan for the 2015 Ozone NAAQS (2023),
• New Source Performance Standards (NSPS) for oil and natural gas sources (2016),
• NSPS for process heaters (2013),
• NSPS for natural gas turbines and reciprocating internal combustion engines
(2012), and
• NSPS for residential wood combustion (2015).
We did not conduct this analysis incremental to controls applied as part of previous
NAAQS analyses because the data and modeling on which these previous analyses were
based are now considered outdated and are not compatible with this PM2.5 NAAQS analysis.
1.3.2 Cost Analysis Approach
The EPA estimated the costs of applying hypothetical national control strategies.
Where available, we apply end-of-pipe and area source controls to achieve emissions
reductions and present the costs associated with these PM2.5 emissions reductions. These
cost estimates reflect only engineering costs, which generally include the costs of
purchasing, installing, and operating the referenced controls. The end-of-pipe and area
source control strategies selected for analysis illustrate one way in which nonattainment
areas could reduce emissions. As mentioned above, the air quality issues being analyzed
are highly complex and local in nature, and the results of these national-level assessments
contain uncertainty. The EPA anticipates that state and local governments will consider
programs that are best suited for local conditions.
1.3.3 Benefits Analysis Approach
The EPA estimated the number and economic value of the avoided PM2.5-
attributable premature deaths and illnesses associated with the control strategies analyzed
40
-------
for the revised alternative standard levels. We quantified an array of mortality and
morbidity effects using the BenMAP-CE tool (U.S. EPA 2023b), which has been used in
recent RIAs. As compared to the 2012 PM NAAQS RIA (U.S. EPA, 2012), the Agency applied
concentration-response relationships from newer epidemiologic studies, assessed a wider
array of human health endpoints and updated other economic and demographic input
parameters. Each of these updates is fully described in Chapter 5, the benefits analysis
approach and results chapter. Unquantified health benefits, welfare benefits, and climate
benefits are also discussed in Chapter 5.
1.3.4 Welfare Benefits of Meeting the Primary and Secondary Standards
Even though the primary standards are designed to protect against adverse effects
to human health, the emissions reductions would have welfare benefits in addition to the
direct health benefits. The term welfare benefits covers both environmental and societal
benefits of reducing pollution. Welfare benefits of the primary PM standard include
reduced vegetation effects resulting from PM exposure, reduced ecological effects from
particulate matter deposition and from nitrogen emissions, reduced climate effects, and
changes in visibility. This RIA does not assess welfare effects quantitatively; this is
discussed further in Chapter 5.
1.4 Organization of the Regulatory Impact Analysis
This RIA is organized into the following remaining chapters:
• Chapter 2: Air Quality Modeling and Methods. The data, tools, and methods
used for the air quality modeling are described in this chapter, as well as the
post-processing techniques used to produce a number of air quality metrics
for input into the analysis of benefits and costs.
• Chapter 3: Control Strategies and PM2.5Emissions Reductions. The chapter
presents the hypothetical control strategies and estimated emissions
reductions in 2032 after applying the control strategies.
• Chapter 4: Engineering Cost Analysis and Qualitative Discussion of Social Costs.
The chapter summarizes the methods, tools, and data used to estimate the
41
-------
engineering costs of the revised and alternative standard levels analyzed.
The chapter also provides a qualitative discussion of social costs.
• Chapter 5: Benefits Analysis Approach and Results. The chapter quantifies the
estimated health-related benefits of the PM-related air quality improvements
associated with the control strategies for the revised and alternative
standard levels analyzed. The chapter also presents qualitative discussions of
welfare benefits and climate benefits.
• Chapter 6: Environmental Justice. This chapter includes an assessment of
environmental justice impacts associated with the control strategies for the
revised and alternative standard levels analyzed.
• Chapter 7: Labor Impacts. This chapter provides a qualitative discussion of
potential labor impacts.
• Chapter 8: Comparison of Benefits and Costs. The chapter compares estimates
of the benefits with costs and summarizes the net benefits of the revised and
alternative standard levels analyzed.
42
-------
1.5 References
U.S. EPA (2011). Regulatory Impact Analysis for the Final Mercury and Air Toxics
Standards. Research Triangle Park, NC. U.S. Environmental Protection Agency, Office of
Air Quality Planning and Standards, Health and Environmental Impact Division. U.S.
EPA. EPA-452/R-11-011. December 2011. Available at:
https://www.epa.gOv/sites/default/files/2020-07/documents/utilities_ria_final-
mats_2011-12.pdf.
U.S. EPA (2012). Regulatory Impact Analysis for the Final Revisions to the National
Ambient Air Quality Standards for Particulate Matter. Research Triangle Park, NC. U.S.
Environmental Protection Agency, Office of Air Quality Planning and Standards, Health
and Environmental Impact Division. U.S. EPA. EPA-452/R-12-005. December 2012.
Available at: https://www3.epa.gov/ttnecasl/regdata/RIAs/finalria.pdf.
U.S. EPA (2014). Control of Air Pollution from Motor Vehicles: Tier 3 Motor Vehicle
Emission and Fuel Standards Final Rule Regulatory Impact Analysis. U.S. Environmental
Protection Agency, Office Transportation and Air Quality. EPA-420/R-14-005. March
2014. Available at:
https://nepis.epa.gov/Exe/ZyPDF.cgi/P100ISWM.PDF?Dockey=P100ISWM.PDF.
U.S. EPA (2015). Regulatory Impact Analysis for the Final Standards of Performance for
Greenhouse Gas Emissions from New, Modified, and Reconstructed Stationary Sources:
Electric Utility Generating Units. Research Triangle Park, NC. U.S. Environmental
Protection Agency, Office of Air Quality Planning and Standards, Health and
Environmental Impact Division. U.S. EPA. EPA-452/R-15-005. August 2015.
https://www.epa.gov/sites/default/files/2020-07/documents/utilities_ria_final-nsps-
egus_2015-08.pdf.
U.S. EPA., U.S. DOT (2016). Greenhouse Gas Emissions and Fuel Efficiency Standards for
Medium- and Heavy-Duty Engines and Vehicles - Phase 2 Regulatory Impact Analysis.
U.S. Environmental Protection Agency, Office Transportation and Air Quality. U.S.
Department of Transportation, National Highway Traffic Safety Administration. EPA-
420/R-16-900. August 2016. Available at:
https://nepis.epa.gov/Exe/ZyPDF.cgi/P100P7NS.PDF?Dockey=P100P7NS.PDF
U.S. EPA (2019). Additional Methods, Determinations, and Analyses to Modify Air Quality
Data Beyond Exceptional Events. U.S. Environmental Protection Agency, Office of Air
Quality Planning and Standards, Research Triangle Park, NC, EPA-457/B-19-002.
Available: https://www.epa.gov/sites/default/files/2019-
04/documents/clarification_memo_on_data_modification_methods.pdf
U.S. EPA., U.S. DOT (2020). Final Regulatory Impact Analysis, The Safer Affordable Fuel-
Efficient (SAFE) Vehicles Rule for Model Year 2021 - 2026 Passenger Cars and Light
Trucks. U.S. Environmental Protection Agency, Office of Air Quality Planning and
Standards, Health and Environmental Impact Division. U.S. Department of
Transportation, National Highway Traffic Safety Administration. March 2020. Available
43
-------
at:
https://www.nhtsa.gov/sites/nhtsa.gov/files/documents/final_safe_fria_web_version_
200701.pdf
U.S. EPA (2021). Regulatory Impact Analysis for the Final Revisions Revised Cross-State Air
Pollution Rule (CSAPR) Update for the 2008 Ozone NAAQS. Research Triangle Park, NC.
U.S. Environmental Protection Agency, Office of Air Quality Planning and Standards,
Health and Environmental Impact Division. U.S. EPA. EPA-452/R-21-002. March 2021.
Available at: https://www.epa.gov/sites/default/files/2021-
03/documents/revised_csapr_update_ria_final.pdf.
U.S. EPA (2022a). Policy Assessment for the Reconsideration of the National Ambient Air
Quality Standards for Particulate Matter. Office of Air Quality Planning and Standards,
Health and Environmental Impacts Division. Research Triangle Park, NC. U.S. EPA. EPA-
452/R-22-004. May 2022. Available at: https://www.epa.gov/naaqs/particulate-
matter-pm-standards-policy-assessments-current-review-O.
U.S. EPA (2023a). Technical Support Document (TSD): Preparation of Emissions
Inventories for the 2018v2 North American Emissions Modeling Platform. Research
Triangle Park, NC. Office of Air Quality Planning and Standards, Air Quality Assessment
Division. U.S. EPA. EPA-454/B-23-003. September 2023. Available at:
https://www.epa.gov/air-emissions-modeling/2018v3-emissions-modeling-platform-
technical-support-document.
U.S. EPA (2023b). Environmental Benefits Mapping and Analysis Program - Community
Edition User's Manual. Office of Air Quality Planning and Standards. Research Triangle
Park, NC. U.S. EPA. Available at: https://www.epa.gov/sites/production/files/2015-
04/documents/benmap-ce_user_manual_march_2015.pdf.
U.S. 0MB (2003). Circular A-4, September 17, 2003, Available at:
https: //www. whitehouse.gov/wp-
content/uploads/legacy_drupal_files / omb / circulars/A4/a-4.pdf.
44
-------
CHAPTER 2: AIR QUALITY MODELING AND METHODS
Overview
To evaluate the incremental costs and benefits of meeting the revised and
alternative PM2.5 standard levels relative to meeting the existing standards, models were
used to predict PM2.5 concentrations and emissions associated with the standard levels. Air
quality was simulated using a 2018-based modeling platform with the Community
Multiscale Air Quality (CMAQ) model. The modeling platform paired a 2018 CMAQ
simulation with a corresponding CMAQ simulation with emissions representative of 2032
that reflects effects of finalized rules and other factors.
Air quality ratios, which relate a change in PM2.5 design values (DVs) to a change in
emissions, were used to estimate the emission reductions needed to meet the existing,
revised, and alternative NAAQS in areas projected to exceed the standards in 2032. These
emission estimates are used in identifying controls and associated costs of meeting the
revised and alternative standard levels relative to meeting the existing standards. A PM2.5
concentration field was developed using the 2032 CMAQ modeling and was adjusted
according to the required change in PM2.5 concentrations to create PM2.5 fields associated
with meeting standard levels. These PM2.5 concentration fields are used in calculating the
health benefits associated with meeting the standard levels.
The overall steps in the process are as follows:
Step 1. Project annual and 24-hour PM2.5 DVs to 2032 using a CMAQ simulation for 2018
and a corresponding CMAQ simulation with emissions representative of 2032
that reflects effects of finalized rules and other factors.
Step 2. Develop air quality ratios that relate a change in PM2.5 DV to a change in
emissions for use in estimating the emissions reductions needed to just meet the
existing, revised, and alternative NAAQS. The air quality ratios are developed
using CMAQ sensitivity modeling with reductions in anthropogenic emissions in
select counties.
Step 3. Using the air quality ratios from Step 2, estimate the emission reductions beyond
the 2032 modeling case that are needed to meet the existing standards and
45
-------
adjust PM2.5 DVs accordingly. The resulting PM2.5 DVs define the 12/35 analytical
baseline that is used as the reference case in estimating the incremental costs
and benefits of meeting revised and alternative standard levels relative to
existing standards. Note that emission reductions applied to meet the existing
standards do not contribute to incremental costs and benefits in the Regulatory
Impact Analysis (RIA).
Step 4. Using the air quality ratios from Step 2, estimate the primary PM2.5 emission
reductions needed to meet the alternative standard levels beyond the 12/35
analytical baseline. These emission reduction estimates are used in developing
controls to meet the alternative standard levels.
Step 5. Develop a gridded national PM2.5 concentration field associated with the 2032
case by fusing the 2032 CMAQ modeling with projected monitor concentrations.
Adjust the 2032 concentration field according to the changes in PM2.5 DVs
needed to meet standard levels to create PM2.5 fields associated with each
standard level. These PM2.5 concentration fields are used in calculating the
health benefits associated with meeting revised and alternative standard levels.
In the remainder of this chapter, contextual information on PM2.5 and its
characteristics in the U.S. is first provided in Section 2.1. The projection of air quality from
2018 to 2032 is then described in Section 2.2. In Section 2.3, the development of air quality
ratios and their application to estimating emission reductions is described. In Section 2.4,
the air quality challenges in select areas are described in terms of highly local influences on
PM2.5 concentrations. Finally, the development of the PM2.5 concentration fields associated
with meeting the existing, revised, and alternative standards is described in Section 2.5.
2.1 PM2.5 Characteristics
2.1.1 PM2.5 Size and Composition
As described in the Integrated Science Assessment (US EPA, 2019a) and Policy
Assessment (US EPA, 2022a), PM (particulate matter) refers to the mass concentration of
suspended particles in the atmosphere. Atmospheric particles range in size from less than
1 nanometer (10-9 meter) to over 100 micrometers (|im, or 10"6 meter) in diameter. For
reference, a typical strand of human hair is 70 [im in diameter and a grain of salt is about
46
-------
100 [im. Atmospheric particles are often classified into size ranges associated with the
three distinct modes evident in measured ambient particle size distributions. The size
ranges include ultrafine particles (<0.1 [im), accumulation mode or fine particles (0.1 to ~3
[im), and coarse particles (>1 [im). For regulatory purposes, fine particles are measured as
PM2.5, which refers to the total mass concentration of particles with aerodynamic diameter
less than 2.5 [im.
PM is made up of many different chemical components. The major components
include carbonaceous matter (elemental and organic carbon) and inorganic species such as
sulfate, nitrate, ammonium, and crustal species. PM includes solid and liquid particles as
well as multiphase particles (e.g., particles with a solid core surrounded by an inorganic
aqueous solution with an organic coating). The phase state and composition of an
atmospheric particle can vary with atmospheric conditions. For example, the aqueous
phase of a particle may effloresce (i.e., crystallize) when the atmospheric relative humidity
falls below a threshold. Similarly, as gas-phase concentrations and meteorological
conditions (e.g., temperature and relative humidity) change, chemical species can condense
and evaporate from particles to maintain or approach equilibrium with their gas-phase
counterparts (Seinfeld and Pandis, 2016).
PM can be directly emitted into the atmosphere or formed in the atmosphere
through chemical and physical processes. PM that is directly emitted into the atmosphere
by sources is referred to as primary PM. Elemental carbon and crustal species are examples
of primary PM components. PM that is formed in situ through atmospheric processes is
referred to as secondary PM. Secondary PM is formed through pathways including new
particle nucleation, condensation and reactive uptake of gas-phase species, and cloud and
fog evaporation (Seinfeld and Pandis, 2016). Nucleation of new particles occurs when
molecular clusters formed from gas-phase species grow into stable particles. Condensation
of atmospheric gases onto preexisting particles occurs when gas-phase concentrations
exceed the equilibrium vapor concentrations of the particle constituents. PM formation
from cloud and fog processes occurs when semi- and non-volatile chemical species formed
via aqueous chemistry in cloud and fog remain suspended in ambient particles following
cloud/fog evaporation.
47
-------
Gaseous SO2 emissions lead to PM2.5 formation following SO2 oxidation to sulfuric
acid in the gas and aqueous phases (Seinfeld and Pandis, 2016). Sulfuric acid is essentially
non-volatile under atmospheric conditions and leads to PM2.5 sulfate formation by
contributing to new particle formation, condensation onto preexisting particles, and
remaining in particles following cloud/fog evaporation. Enhanced particle acidity due to
PM2.5 sulfate formation reduces the equilibrium vapor concentration of ammonia (the
primary atmospheric base) and promotes condensation of ammonia onto particles, thereby
forming PM2.5 ammonium. PM2.5 sulfate and associated water and acidity also influence
chemical pathways for the formation of secondary organic aerosol (SOA).
Gaseous NOx emissions lead to PM2.5 formation following NOx oxidation to nitric
acid, which is semi-volatile under atmospheric conditions (Seinfeld and Pandis, 2016).
Condensation of nitric acid onto particles tends to be favorable under cool, humid
conditions with abundant ammonia, and results in PM2.5 nitrate formation. Due to effects of
nitric acid on particle acidity, ammonia often co-condenses with nitric acid to yield PM2.5
ammonium. NOx emissions also influence secondary PM concentrations by modulating
many atmospheric oxidation processes and by contributing to the production of organic
nitrates. Monoterpene nitrates and isoprene nitrates are examples of PM2.5 species that can
be formed from products of anthropogenic NOx emissions and biogenic volatile organic
compound (VOC) emissions. SOA formation occurs following the oxidation ofVOC
emissions in the atmosphere. SOA formation is an active area of research and involves
myriad species and reactions occurring in the gas, particle, and aqueous phases. Gaseous
ammonia emissions can influence PM concentrations by affecting cloud and aerosol acidity
in addition to condensing on particles to form PM2.5 ammonium.
The emission sources of primary PM2.5 and the gaseous precursors of PM2.5 have
recently been summarized in the PM NAAQS Policy Assessment (USEPA, 2022a). EGUs
make up the largest emissions source sector for SO2. The largest NOx emissions sectors
include mobile sources (on-road and non-road) and EGUs. Ammonia emissions are greatest
from the agricultural sector (fertilizer and livestock waste) and from fires. VOC emissions
are largest from mobile sources, industrial processes, fires, and biogenic sources. Primary
PM2.5 emissions are largest from fires, fugitive dust (paved/unpaved road dust and
48
-------
construction dust), and area sources (e.g., residential wood combustion). Fires are an
important source of particulate organic matter. Note that some PM2.5 components (e.g.,
elemental carbon and crustal species) occur due to direct emissions alone while other PM2.5
components (e.g., organic carbon and sulfate) occur due to a combination of direct
emissions and secondary formation in the atmosphere.
2.1.2 PM2.5 Regional Characteristics
PM2.5 concentrations vary in magnitude and composition over the U.S. with distinct
regional and seasonal features. The characteristics of PM2.5 concentrations in the U.S. have
recently been summarized in the Integrated Science Assessment (USEPA, 2019a), and the
spatial distribution of PM2.5 over the U.S. is shown in Figure 2-1 based on a hybrid satellite
modeling method (van Donkelaar et al., 2021). In the Eastern U.S., organic carbon and
sulfate have the highest contribution to total PM2.5 concentrations in most locations. In the
Upper Midwest and Ohio Valley, nitrate can also be an important contributor to PM2.5, due
to the cool, humid conditions in winter and influence of ammonia that promotes
ammonium nitrate formation. In the Southeastern U.S., organic carbon concentrations are
relatively high due to the abundance of biogenic VOC emissions that contribute to SOA
formation following oxidation in the presence of anthropogenic emissions. Areas of
relatively high PM2.5 concentrations within the Eastern U.S. are associated with urban
centers.
-120 -110 -100 -90 -80 -70
Longitude
Figure 2-1 Annual Average PM2.5 Concentrations over the U.S. in 2019 Based on
the Hybrid Satellite Modeling Approach of van Donkelaar et al. (2021)
4-9
-------
The Western U.S. is characterized by some of the lowest and highest PM2.5
concentrations in the country, with relatively sharp spatial gradients in PM2.5 compared to
the east. The complex terrain of the Western U.S. has an important influence on air
pollution processes as does the relative abundance of wildfires (and prescribed burning).
In the Northwest, meteorological temperature inversions often occur in small mountain
valleys in winter and trap pollution emissions in a shallow atmospheric layer at the surface.
Emissions from home heating with residential wood combustion can build up in the surface
layer and produce episodically high PM2.5 concentrations in winter. Elevated wintertime
PM2.5 in these mountain valleys can approach or sometimes exceed the 24-hour PM2.5
standard, which is based on a 98th percentile form.
In large western air basins (e.g., San Joaquin Valley, CA; South Coast Air Basin, CA;
and Salt Lake Valley, UT), emission sources are more diverse than in the small mountain
valleys and include NOx emissions from urban centers and ammonia from agriculture.
Meteorological conditions are also more complex than in the smaller valleys and can
include a persistent aloft temperature inversion from high-pressure-driven air subsidence
in addition to a near-surface temperature inversion from nighttime radiative cooling. The
near-surface inversion has the effect of concentrating primary PM2.5 emissions near the
ground, whereas the aloft inversion caps the nighttime residual air layer, in which NOx is
converted to nitrate through heterogeneous aerosol chemistry. In the morning, when the
near-surface inversion breaks and the surface mixed layer grows due to surface heating,
the PM2.5 nitrate and ammonium formed overnight in the residual layer are entrained to
the surface. This entrainment has the effect of diluting primary PM2.5 concentrations near
the surface and enhancing surface concentrations of secondary PM2.5. PM2.5 concentrations
in the South Coast Air Basin are also affected by the land-sea breeze circulation and a semi-
permanent high-pressure cell. Due to the large populations, diverse emission sources, and
terrain-driven meteorological features, the San Joaquin Valley and South Coast Air Basin
experience elevated annual-average PM2.5 concentrations as well as short-term PM2.5
enhancements. These characteristics can create challenges for meeting both the annual and
24-hour PM2.5 standards.
50
-------
PM2.5 concentrations in the Western U.S. are also strongly influenced by emissions
from wildfires, which are relatively common in summer but increasingly occur year-round.
In the Southwest, dust emission sources are prevalent, and windblown dust makes
substantial contributions to PM2.5 concentrations under dry, windy conditions. Organic
carbon is often the largest PM2.5 contributor in the west due to the influence of combustion
sources such as wildfire and residential wood combustion. Crustal species are also
important contributors in dust-prone areas, and ammonium nitrate is a major PM2.5
component in large air basins during meteorological stagnation periods in fall and winter.
Along the border with Mexico, western areas also experience important cross-border
transport contributions to PM2.5 (e.g., Calexico, CA experiences contributions from the
much the larger city of Mexicali, MX, which is in the same airshed just across the border).
2.1.3 PM2.5 Trends
Over the last several decades, PM2.5 concentrations have decreased on average over
the U.S. (Figure 2-2). As described in the recent PM NAAQS Policy Assessment (USEPA,
2022a), the reductions in PM2.5 concentrations correspond to the reductions in PM2.5
precursor emissions illustrated in Figure 2-3. Among the PM2.5 precursors (i.e., SO2, NOx,
VOC, and ammonia), the largest emission reductions occurred for SO2 and NOx. SO2
emissions decreased by 84% between 2002 and 2017, and NOx emissions decreased by
60%. Reductions in SO2 emissions were relatively large from stationary sources such as
EGUs in the Eastern U.S. NOx emission reductions were driven by reduced emissions from
mobile sources and EGUs. Compared with SO2 and NOx, emissions of primary PM2.5 and
ammonia have been relatively flat in recent decades. The small changes in primary PM2.5
emissions in Figure 2-3 are likely due to changes in emission estimation methods for
source sectors over time. Wildfire emissions are not included in the data for Figure 2-3, but
an upward trend in PM2.5 emissions is evident in estimates generated for National Emission
Inventory years (i.e., 2005, 2008, 2011, 2014, and 2017).1 Studies have also predicted that
climate change presents increased potential for very large fires in the contiguous U.S. in the
future (e.g., Barbero et al., 2015).
1 https://www.epa.gov/air-emissions-inventories/air-pollutant-emissions-trends-data
51
-------
00000000001111111111
012345678901 23456789
Figure 2-2 Seasonally Weighted Annual Average PM2.5 Concentrations in the U.S.
from 2000 to 2019 (406 sites)
Note: The white line indicates the mean concentration while the gray shading denotes the 10th and
90th percentile concentrations.
30000
Year
_#_NH3 -*-NOx -«-PM2.5 -*-PM10 -*-S02 -*-VOCs
Figure 2-3 National Emission Trends of PM2.5, PM10, and Precursor Gases from
1990 to 2017
Note: Data do not include wildfire emissions.
As described in the PM NAAQS Policy Assessment (USEPA, 2022a), PM2.5 precursor
emission reductions have altered the seasonal variation in PM2.5 concentrations over the
52
-------
U.S. Through 2008, the peak in the national average PM2.5 concentration occurred during
summer, largely due to sulfate formation from summertime increases in EGU SO2 emissions
in the Eastern U.S. and wildfires in the West. However, starting in 2009, the summertime
peaks in PM2.5 concentrations have been smaller than those in winter as PM2.5 sulfate
concentrations have decreased (Chan et al., 2018). The decrease in sulfate in the Eastern
U.S. has increased the relative contribution of organic carbon and sources of primary PM2.5,
whose emissions have remained flat as SO2 emissions have decreased. Primary PM2.5
sources in urban centers contribute to the "urban increment" of consistently higher PM2.5
concentrations in urban than surrounding areas (Chan et al., 2018).
To explore how emission trends may persist into the future, models are applied to
project emission inventories accounting for expected future emission changes from
finalized rules and other factors. Air quality models are then used to simulate pollutant
concentrations under conditions of the projected future emissions. For the purposes of the
RIA, model projections from 2018 to 2032 were developed for air quality analyses as
described in Section 2.2. As shown in Figure 2-4, the trends in NOx, SO2, and primary PM2.5
emissions from the recent past (Figure 2-3) are projected to continue into the near future.
From 2018 to 2032, anthropogenic NOx emissions are projected to decrease by 3.6 million
tons (41%), with the greatest reductions from mobile-source sectors (nonroad and onroad)
and EGUs. SO2 emissions are projected to decrease by 1.1 million tons (48%), with the
greatest reductions from the EGU sector. For primary PM2.5, emissions are relatively flat
from 2018 to 2032, with a decrease of 116k tons (4%) mainly due to reductions from
mobile sources and EGUs. Primary PM2.5 emissions from the largest emitting sectors (e.g.,
dust, agricultural and prescribed fires, residential wood combustion, and areas sources)
are essentially constant or slightly increasing (e.g., dust) (Figure 2-4).2 This projected
behavior is consistent with past trends, in which NOx and SO2 emissions declined steadily
while primary PM2.5 emissions were relatively constant (Figure 2-3).
2 Prescribed burning emissions were held constant at 2018 levels in the model projections, although these
emissions could potentially change in the future. Available evidence indicates that wildfire acres burned
have increased over time, which, in turn, has drawn attention to prescribed fires as an important tool for
reducing wildfire risk and the severity of wildfires and wildfire smoke (88 FR, 54118, 54126, August 9,
2023). A few agencies have efforts in place to increase fuel load minimization efforts in areas at high risk of
wildfire (as noted in the PM NAAQS proposal 88 FR 5570, January 27, 2023).
53
-------
2000
1000
0 ........ .
AgPrFire Area Nonroad Onroad EGU O&G Other NonEGU ResWood
S02
>*
o 1000-
f- I H 2018
c 500" I ¦ 2032
'c/5
W o- —
E AgPrFire Area Nonroad Onroad EGU O&G Other NonEGU ResWood
^ P PM25
1000-
750-
500-
250-
0- ......... .
Dust AgPrFire Area Nonroad Onroad EGU O&G Other NonEGlResWood
Figure 2-4 Annual Anthropogenic Source Sector Emission Totals (1000 tons per
year) for NOx, SO2, and PM2.5 for 2018 and 2032
Note that AgPrFire: agricultural and prescribed fire; Area: non-point area sources; O&G: oil and gas;
Other: airports, commercial marine vehicles, rail, and solvents; NonEGU: remaining non-EGU point
sources; RWC: residential wood combustion.
As mentioned above, spatial distributions of PM2.5 concentrations in the U.S. are
characterized by an "urban increment" of consistently higher PM2.5 concentrations over
urban than surrounding areas. Monitored concentrations are highest in urban areas and
relatively low in rural areas. Conceptually, PM2.5 concentrations in urban areas can be
viewed as the superposition of the urban increment and the contributions from regional
and natural background sources. The decreases in anthropogenic SO2 and NOx emissions in
recent decades have reduced regional background concentrations and increased the
relative importance of the urban increment. The projections of additional large reductions
54
-------
in SO2 and NOx emissions in the 2032 case further motivates the need for control of local
primary PM2.5 sources to address the highest PM2.5 concentrations in urban areas.
In Figure 2-5, PM2.5 concentrations are shown over four urban areas in the Eastern
U.S. based on the 2032 modeling case described in Section 2.2. A common feature of these
diverse locations is the relatively high PM2.5 concentrations over the urban area and lower
concentrations just outside of the urban core. PM2.5 concentrations in the urban core of
these Eastern U.S. areas exceed revised and alternative standards levels considered in the
RIA, whereas concentrations surrounding the urban core are below the revised and
alternative standard levels. In the illustrative control strategy analysis of the RIA, the urban
exceedances are addressed by focusing on primary PM2.5 emission controls in the local
county. This approach is consistent with the exceedances being driven by the urban PM2.5
increment, the relatively high responsiveness of PM2.5 concentrations to primary PM2.5
emission reductions, and the reductions in regional PM2.5 concentrations from the large SO2
and NOx emission reductions in recent decades and in the 2032 projection. Patterns may
vary in the Western U.S. where the spatial extent of the PM2.5 increment may be influenced
by complex terrain that defines distinct air basins.
55
-------
Longitude
ug/m3
I
7
-85.0 -84.5 -84.0
Longitude
Figure 2-5 Gridded PM2.5 Concentrations over Selected Urban Areas Based on the
2032 Modeling Case Described Below with the Enhanced Voronoi
Neighbor Averaging Approach
ug/m3
I
30.5
ug/m3
Is
8 30.0
7
6
29.5
29.0
-96.5
Houston
-96.0 -95.5 -95.0 -94.5
Longitude
ug/m3
34.0
33.5
2.2 Modeling PM2.5 in the Future
To evaluate the incremental costs and benefits of meeting the revised and
alternative PM2.5 standard levels in this RIA relative to meeting the existing standards,
models were used to predict PM2.5 concentrations associated with emissions
representative of a 2032 future year to inform subsequent analyses. The projections were
performed using a 2018-based modeling platform with the Community Multiscale Air
Birmingham
34.0
33.6
32.8
-87.5 -87.0 -86.5
Chicago
42.5
41.5
41.0
-88.5 -88.0 -87.5 -87.0
Longitude
56
-------
Quality (CMAQ) model (www.epa.gov/cmaq). The modeling platform paired a 2018 CMAQ
simulation with a corresponding CMAQ simulation based on emissions representative of
2032. The 2032 emission projections account for numerous factors including the effects of
finalized rules. This modeling platform was chosen because it represents the most recent,
complete set of emissions information currently available for national-scale modeling. The
approach used for projecting future-year air quality with the platform is described in this
section.
2.2.1 Air Quality Modeling Platform
To project air quality to the future, the CMAQ model was applied to simulate air
quality over the U.S. during 2018 and for a case with emissions representative of 2032.
Other than the differences in emissions inventories for the 2018 and 2032 CMAQ
simulations, all other model inputs specified for the 2018 base year remained unchanged in
the 2032 modeling case. Inputs for CMAQ simulations include files with emissions,
meteorology, and initial and boundary condition data.
2.2.1.1 Model Configuration
CMAQ is a three-dimensional grid-based Eulerian air quality model designed to
estimate the formation and fate of oxidant precursors, primary and secondary PM2.5
concentrations, and deposition over regional spatial scales (e.g., over the contiguous U.S.)
(Appel et al., 2021; Appel et al., 2018; Appel et al., 2017). CMAQ simulates the key processes
(e.g., emissions, transport, chemistry, and deposition) that affect primary (directly emitted)
and secondary (formed by atmospheric processes) PM2.5 using state-of-the-science process
parameterizations and input data for emissions, meteorology, and initial and boundary
conditions. CMAQ's representation of the chemical and physical mechanisms that govern
the formation and fate of air pollution enable simulations of the impacts of emission
controls on PM2.5 concentrations.
CMAQ version 5.3.2 (www.epa.gov/cmaq) was used to simulate air quality for 2018
to provide a reference simulation for the 2032 air quality projection. The geographic extent
of the air quality modeling domain is shown in in Figure 2-6. The modeling domain covers
the 48 contiguous states along with parts of Canada and Mexico with a horizontal
resolution of 12 x 12 km. Air quality modeling for a larger 12-km domain (USEPA, 2021)
57
-------
was used to provide chemical boundary conditions for the 12US2 domain simulation used
in air quality analyses in the RIA.
Gas-phase chemistry in the CMAQ simulations was based on the Carbon Bond 2006
mechanism (CB6r3) (Emery et al., 2015), and deposition was modeled with the M3DRY
parameterization. Aerosol processes were parameterized with the AER07 module using
ISORROPIAII for inorganic aerosol thermodynamics (Fountoukis and Nenes, 2007) and the
non-volatile treatment for primary organic aerosol (Appel et al., 2017; Simon and Bhave,
2012). Emissions of biogenic compounds were modeled with the Biogenic Emission
Inventory System (BEIS) (Bash et al., 2016). Anthropogenic emissions were based on 2018
version 2 emissions modeling platform (USEPA, 2023a), which included emissions for 2018
and the projected 2032 case. Meteorological data were based on a 2018 simulation with
version 3.8 of the Weather Research Forecasting (WRF) model (Skamarock et al., 2008).
The meteorological fields include hourly-varying horizontal wind components (i.e., speed
and direction), temperature, moisture, vertical diffusion rates, and rainfall rates for each
grid cell in each vertical layer. Additional details on the model configuration are available in
section 2A.1.1 of Appendix 2A.
Figure 2-6 Map of the 12US2 (12 x 12 km Horizontal Resolution) Modeling
Domain
58
-------
2.2.1.2 Emissions Inventory
The future-year emissions inventory is projected from the 2018 version 2 emissions
modeling platform. The projected emission case is labeled 2032, although the emission
projections are based on a combination of projection years, with some sources held
constant at base year levels.3 The development of the 2018 base-year inventory, the
projection methodology, and the controls applied to create the projected inventory are
described in detail in the emissions Technical Support Document (TSD): Preparation of
Emissions Inventories for the 2018v2 North American Emissions Modeling Platform (USEPA,
2023a). The types of sources included in the emission inventory include stationary point
sources such as EGUs and non-EGUs; non-point emissions sources including those from oil
and gas production and distribution, agriculture, residential wood combustion, fugitive
dust, and residential and commercial heating and cooking; mobile source emissions from
onroad and nonroad vehicles, aircraft, commercial marine vessels, and locomotives; wild,
prescribed, and agricultural fires; and biogenic emissions from vegetation and soils.4
The EGU emissions were developed using the Post-IRA 2022 Reference Case of the
EPA's Power Sector Platform v6 using Integrated Planning Model (IPM), (USEPA, 2023b)
where the 2023 Federal Good Neighbor Plan Addressing Regional Ozone Transport for the
2015 Ozone National Ambient Air Quality Standards (Final GNP) was also reflected. The
IPM is a multiregional, dynamic, deterministic linear programming model of the U.S.
electric power sector. The EGU projected inventory represents demand growth, fuel
resource availability, generating technology cost and performance, and other economic
factors affecting power sector behavior. It also reflects environmental rules and
regulations, consent decrees and settlements, plant closures, and newly built units for the
calendar year 2030. In this analysis, the projected EGU emissions include provisions of tax
3 2032: nonroad, onroad, airports, non-EGU point (except for biorefineries / ethanol plants), paved-road dust,
oil and gas (except in WRAP states), residential wood combustion (except held constant from base year in
CA, OR, and WA), locomotives, livestock, solvents, other U.S. nonpoint sources, Canadian onroad and
nonroad emissions, Mexico onroad emissions; 2030: US EGUs and commercial marine vessels; 2028: Canada
and Mexico nonpoint and point emissions; 2018: fertilizer, fires, biogenic, and US fugitive dust (other than
paved road) emissions
4 Emissions reductions from the Federal Implementation Plan for Managing Emissions from Oil and Natural
Gas Sources on Indian Country Lands within the Uintah and Ouray Indian Reservation in Utah (2022) are
not reflected in the baseline for this analysis. Given the focus of this rule, any potential impacts are likely to
be small. Also, the impacts of any proposed rules, such as those for onroad mobile sources, are not reflected.
59
-------
incentives impacting electricity supply in the Inflation Reduction Act of 2022 (IRA), Final
GNP, 2021 Revised Cross-State Air Pollution Rule Update (RCU), the 2016 Standards of
Performance for Greenhouse Gas Emissions from New, Modified, and Reconstructed
Stationary Sources, the Mercury and Air Toxics Rule (MATS) finalized in 2011, and other
finalized rules. Documentation and results of the Post-IRA 2022 Reference Case, where the
Final GNP was also included for EGUs, are available at (https://www.epa.gov/power-
sector-modeling/final-pm-naaqs).
Regulations for non-EGU point sources and non-point sources reflected in the
inventories include:
• Good Neighbor Plan for the 2015 ozone NAAQS;
• New Source Performance Standards (NSPS) for oil and natural gas sources
(2016), process heaters (2013), natural gas turbines (2012), and
reciprocating internal combustion engines;
• NSPS for residential wood combustion (2015);
• Fuel sulfur rules in mid-Atlantic and northeast states (current through
2019);
• NSPS and Emission Guidelines for Commercial and Industrial Solid Waste
Incineration (CISWI) from March 2011;
• NSPS Subpart JA for Standards of Performance for Petroleum Refineries
from June 2008;
• Specific consent decrees; and
• Ozone Transport Commission controls for Portable Fuel Containers,
consumer products, architectural and industrial maintenance coatings, and
various other solvents.
Known closures are also implemented for non-EGU point sources.
Onroad and nonroad mobile source emissions were developed using the Motor
Vehicle Emission Simulator version 3 (MOVES3). The SMOKE-MOVES emissions modeling
60
-------
framework was used that leverages MOVES-generated emission factors, county and SCC-
specific activity data, and hourly meteorological data. MOVES3 was run in emission rate
mode to create emission factor tables for the 2032 future modeling year for all
representative counties and fuel months. These emissions represent the effects the Control
of Air Pollution from New Motor Vehicles: Heavy-Duty Engine and Vehicle Standards
(2022); the Final Rule to Revise Existing National GHG Emissions Standards for Passenger
Cars and Light Trucks Through Model Year 2026 (2021); the Safer Affordable Fuel Efficient
(SAFE) Vehicles Final Rule for Model Years 2021-2026 ( 2020); the Greenhouse Gas
Emissions Standards and Fuel Efficiency Standards for Medium- and Heavy-Duty Engines
and Vehicles - Phase 2 (2016); the Tier 3 Vehicle Emission and Fuel Standards Program
(2014); and other finalized rules. A full discussion of the future year base inventory is
provided in USEPA (2023a). Nonroad emissions rules related to nonroad spark-ignition
engines, equipment, and vessels from October 2008 are reflected.
Emissions for commercial marine vessels and locomotive engines reflect the rules
finalized in 2010 and 2008:
• Growth and control from Locomotives and Marine Compression-Ignition
Engines Less than 30 Liters per Cylinder: March 2008
• Category 3 marine diesel engines Clean Air Act and International Maritime
Organization standards: April 2010
• Growth and control from Locomotives and Marine Compression-Ignition
Engines Less than 30 Liters per Cylinder: March 2008
2.2.1.3 Model Evaluation
An operational model performance evaluation for PM2.5 and its speciated
components (e.g., sulfate, nitrate, elemental carbon, and organic carbon) was performed to
estimate the ability of the CMAQ modeling system to replicate the 2018 base year
concentrations. This evaluation includes statistical assessments of model predictions
versus observations from national monitoring networks paired in time and space. Details
on the evaluation methodology and the calculation of performance statistics are provided
in section 2A.1.2 of Appendix 2A. Overall, the performance statistics for PM2.5 and its
61
-------
components from the CMAQ 2018 simulation are within or close to the ranges found in
other recent applications. These model performance results provide confidence that our
use of the 2018 modeling platform is a scientifically credible approach for assessing PM2.5
concentrations for the purposes of the RIA.
2.2.2 Future-Year PM2.5 Design Values
To evaluate the incremental costs and benefits associated with meeting revised and
alternative standard levels relative to the existing standard, PM2.5 DVs were first projected
to 2032 accounting for emission reductions expected from finalized rules. The air quality
and emission changes associated with meeting the existing, revised, and alternative
standard levels were then estimated as described below in Section 2.3. PM2.5 DVs were
projected to 2032 using the air quality model results in a relative sense, as recommended
by the EPA modeling guidance (USEPA, 2018), by projecting monitoring data with relative
response factors (RRFs) developed from the 2018 and 2032 CMAQ modeling.
PM2.5 RRFs were calculated as the ratios of modeled PM2.5 species concentrations in
the future year (2032) to the base year (2018) for each PM2.5 component (i.e., sulfate,
nitrate, organic carbon, elemental carbon, crustal material, and ammonium). The 2032
PM2.5 DVs were calculated by applying the species-specific RRFs to ambient PM2.5 species
concentrations from the PM2.5 monitoring network. Observed PM2.5 concentrations were
disaggregated into species concentrations by applying the SANDWICH method (Frank,
2006) and through interpolation of PM2.5 species data from the Chemical Speciation
Network (CSN) and the Interagency Monitoring of Protected Visual Environments
(IMPROVE) monitoring network. The RRF method for projecting PM2.5 DVs was
implemented using EPA's Software for Modeled Attainment Test-Community Edition
(SMAT-CE) (USEPA, 2018; Wang etal., 2015). More details on the PM2.5 projection method
using RRFs are provided in the user's guide for the predecessor to the SMAT-CE software
(Abt, 2014).
Ambient PM2.5 measurements from the 2016-2020 period centered on the 2018
CMAQ modeling period were used in projecting PM2.5 DVs. PM2.5 species measurements
from the IMPROVE and CSN networks during 2017-2019 were used to disaggregate the
measured total PM2.5 concentrations into components. In addition to exclusion of EPA-
62
-------
concurred exceptional events, limited exclusion of wildfire and fireworks influence on
PM2.5 concentrations was applied to the 2016-2020 PM2.5 monitoring data. Monitoring data
were evaluated (i.e., screened) for potential wildfire and fireworks influence because PM2.5
concentrations may be influenced by atypical, extreme, or unrepresentative events such as
wildfires or fireworks that may be appropriate for exclusion as described in EPA's
memorandum Additional Methods, Determinations, and Analyses to Modify Air Quality Data
Beyond Exceptional Events (USEPA, 2019b). The steps in implementing the limited
screening of major wildfire and fireworks influence on PM2.5 concentrations are as follows.
Step 1. An extreme value cutoff of 64 [ig nr3 was identified based the 99.9th percentile
value from all daily PM2.5 concentrations across all sites in the U.S. EPA's Air
Quality System database of observations (2002-2020).
Step 2. Specific months were evaluated for instances of monitors exceeding the extreme
value cutoff. Months included were June-October (although November can be a
high fire month for parts of the western U.S., it becomes more difficult to
distinguish wildfire PM2.5 from residential wood smoke and other anthropogenic
sources during the late fall).
Step 3. The presence of visible wildfire smoke was corroborated using satellite imagery
from NASA's Worldview platform (https://worldview.earthdata.nasa.gov) for
the time periods and geographic locations identified through Steps 1 and 2.
Timeseries for individual sites were also examined to confirm PM2.5
enhancements temporally consistent with the wildfire events identified.
Step 4. For extreme wildfire smoke periods identified through Steps 1-3 above, all
concentrations above the extreme value cutoff of 64 [ig nr3 at impacted sites
were excluded.
Step 5. In addition to the evaluation criteria above, data corresponding to the Creek Fire
(eastern CA during September-November 2020), the Camp Fire (northern CA
during November 2018), and the Appalachian Fires (NC, TN, GA during
November 2016) were evaluated for exclusion if concentrations exceeded the
extreme value threshold of 64 [ig nr3. These large fire episodes show obvious
63
-------
impacts across multiple monitors and were clearly documented with satellite
imagery.
Step 6. In addition to the limited exclusion of major wildfire influence, data were
evaluated to identify days for potential exclusion due to the influence of isolated
fireworks events on PM2.5 concentrations. The 99.9th percentile value of 64 [ig nr
3 was applied as the cutoff across all sites for New Year's Eve and the Fourth of
July.
The percentage of days excluded from the 2016-2020 dataset was 0.9% at affected
sites; the total percentage of days excluded overall from the dataset was 0.09%. Since the
cutoff value (64 |Lxg nr3) is much greater than the 24-hour and annual standard levels,
wildfire contributions to PM2.5 concentrations above the standard levels likely persists in
the data following screening. Comprehensive identification and exclusion of such wildfire
impacts would require detailed analyses that are beyond the scope of this national
assessment. More information on the wildfire and fireworks screening are provided in
section 2A.2.1 of Appendix 2A.
2.3 Calculating Emission Reductions for Meeting the Existing, Revised, and
Alternative Standard Levels
To estimate the tons of emissions reductions needed to reach attainment of the
existing, revised, and alternative standard levels, we calculated air quality ratios based on
how modeled concentrations changed with changes in emissions in CMAQ sensitivity
modeling. Air quality ratios represent an estimate of how the DVs at a monitor would
change in response to emissions reductions and have been used in prior PM NAAQS RIAs
(USEPA, 2012a, 2012b). Air quality ratios have units of |Lxg nr3 per 1000 tons of emissions.
The remainder of this section describes the development of air quality ratios and their
application to estimating emission reductions for meeting the existing, revised, and
alternative standards.
2.3.1 Developing Air Quality Ratios
In the illustrative control strategy analysis in the RIA, the revised and alternative
standard level exceedances are addressed by focusing on primary PM2.5 emission controls
in the local county. This approach is consistent with the exceedances generally being driven
64
-------
by the urban PM2.5 increment, the relatively high responsiveness of PM2.5 concentrations to
primary PM2.5 emission reductions, and the reductions in regional PM2.5 concentrations
from the large SO2 and NOx emission reductions in recent decades and in the 2032
projection (Section 2.1.3). To develop air quality ratios that relate the change in DV in a
county to the change in primary PM2.5 emissions in that county, CMAQ sensitivity modeling
was performed with reductions in primary PM2.5 emissions in selected counties. The
modeling was conducted using CMAQ version 5.2.1 for a 2028 modeling case similar to that
of recent regional haze modeling (USEPA, 2019c) due to the availability of the 2028 (but
not 2032) modeling at the time of the work. Since air quality ratios reflect the sensitivity of
air quality to emission changes (rather than absolute concentrations), the air quality ratios
based on the 2028 modeling are suitable for application to our 2032 modeling case.
To develop air quality ratios for primary PM2.5 emissions, we used the following
method:
Step 1. A CMAQ sensitivity simulation was conducted with 50% reductions in primary
PM2.5 emissions from anthropogenic sources in counties with annual 2028 DVs
greater than 8 |Lxg rrr3.
Step 2. The change in annual and 24-hour PM2.5 DVs at monitors in counties where
emission reductions were applied was calculated using projected DVs from the
2028 modeling with the SMAT-CE software.
Step 3. The change in DVs at individual monitors was divided by the change in
emissions in the respective county to determine the air quality ratio (|Lxg nr3 per
1000 tons) for the individual monitors.
Step 4. The responsiveness of air quality at a specific monitor location to primary PM2.5
emission reductions depends on several factors including the specific
meteorology and topography in an area and the nearness of the emissions
source to the monitor. As described in a previous PM NAAQS RIA (USEPA,
2012a), the strong local influence of changes in directly emitted PM2.5 on air
quality produces large variability in air quality ratios that can result in non-
representative values for general application. To address this issue,
65
-------
representative air quality ratios for regions of the U.S. were developed from the
ratios at individual monitors. The five regions are illustrated in Figure 2-7. The
Northeast region was defined by combining the Upper Midwest, Ohio Valley, and
Northeast U.S. climate regions (Karl and Koss, 1984); the Southeast region was
defined by combining the Southeast and South U.S. climate regions (Karl and
Koss, 1984); and California was separated into southern and northern regions as
done previously (USEPA, 2012a) due to differences in PM2.5 responsiveness in
those areas. For each region, representative air quality ratios were calculated as
the 75th percentile of air quality ratios for individual monitors within the region.
The 75th percentile was selected to avoid use of extreme values while accounting
for the relatively high responsiveness of the highest-DV monitors that are most
relevant to our application.
45°N -
40°IM-
35°N -
30°N -
25°N -
Figure 2-7
The air quality ratios for primary PM2.5 emissions used in estimating the emission
reductions needed to meet standard levels at monitors in the five regions are shown in
Table 2-1. These data give an estimate of how PM2.5 DVs at a monitor would change if 1000
tons of primary PM2.5 emissions were reduced in the county in which the monitor is
located. Additional details on the development of the air quality ratios are available in
section 2A.3.1 of Appendix 2A.
Northeast
Southeast
West
NoCal
SoCal
120°W
110°W 100°W
90°W
80°W
70°W
Regional Groupings for Calculating Air Quality Ratios
66
-------
Table 2-1
Annual and 24-Hour Air Quality Ratios for Primary PM2.5 Emissions
Region Annual Air Quality Ratio 24-hour Air Quality Ratio
fug m ! per kton) (p.g m ! per kton)
Northeast 1.37 4.33
Southeast 1.22 3.51
West 2.14 8.70
Northern California 3.15 9.97
Southern California 1.18 2.56
The air quality ratios in Table 2-1 relate the change in DV in a county to a change in
emissions in that county. The ratios are developed for county-level spatial scales because
concentrations are most responsive to changes in local emissions. However, emission
controls may not always be identified in the local county, and emission reductions in
neighboring counties may sometimes be appropriate, such as in the Eastern U.S. where
counties are relatively small and terrain is relatively flat. To apply emission reductions in
the neighboring counties in the Eastern U.S., the responsiveness of annual PM2.5 DVs for
emission reductions within the county was compared to the responsiveness of DVs in the
neighboring counties using the 2028 sensitivity modeling. Annual DVs were estimated to
be 4 times more responsive on average for emission reductions in the county containing
the monitor than for emission reductions in a neighboring county in the Eastern U.S.
Primary PM2.5 emission reductions were not applied in neighboring counties in the
Western U.S. (including California) due to the large size of the counties and the complex
terrain that often isolates the influence of primary PM2.5 emissions to the local air basin.
Additional information related to air quality ratios for neighboring counties is available in
section 2A.3.1 of Appendix 2A.
At monitors in the South Coast Air Basin and San Joaquin Valley (SJV) of California,
PM2.5 DVs exceeded the existing standards in the 2032 modeling case. Air quality
management plans apply reductions in NOx emissions in addition to reductions in primary
PM2.5 emissions to meet the existing NAAQS in these air basins (SCAQMD, 2017; SJVAPCD,
2018). The NOx emission reductions help in meeting the existing standards by reducing
concentrations of PM2.5 ammonium nitrate in the air basins as described in Section 2.1.2. In
creating the 12/35 analytical baseline of DVs associated with meeting existing standards,
67
-------
we applied 75% reductions in NOx emissions in SJV and South Coast in addition to primary
PM2.5 emission reductions. To apply the NOx emission reductions, air quality ratios for NOx
emissions were developed for South Coast and SJV monitors. Air quality ratios for South
Coast were developed using 2028 sensitivity modeling for NOx emissions similar to the
approach described above for the primary PM2.5 air quality ratios. For SJV, air quality ratios
were developed from sensitivity modeling results presented in the SJV air quality
management plan (SJVAPCD, 2018), which was based on a fine-scale CMAQ modeling
platform. Additional details on the South Coast and SJV air quality ratios for NOx are
available in section 2A.3.2 and 2A.3.3 of Appendix 2A. Note that the NOx emission
reductions were applied in attaining the existing standards and therefore do not contribute
to the incremental costs and benefits of meeting revised and alternative standard levels
relative to meeting the existing standards.
2.3.2 Emission Reductions to Meet 12/35
PM2.5 DVs from the 2032 projection were adjusted using air quality ratios to
correspond with just meeting the existing standard level to create the 12/35 analytical
baseline. The 12/35 analytical baseline is used as the reference case for estimating the
incremental costs and benefits of meeting the revised and alternative standard levels
relative to the existing 12/35 standard combination.
The counties with projected 2032 PM2.5 DVs that exceed the existing standard levels
and require air quality adjustments to meet 12/35 are shown in Figure 2-8. Counties that
exceed only the 24-hour standard are in northern California, Oregon, Washington, Idaho,
and Montana. Elevated PM2.5 episodically occurs in winter in these areas due to
meteorological temperature inversions that concentrate PM2.5 in shallow layers near the
ground in complex terrain. In California, multiple counties exceed both the annual and 24-
hour standards, and two counties (San Bernardino and Imperial) exceed only the annual
standard.
68
-------
12/35
Figure 2-8 Counties with Projected 2032 PM2.5 DVs that Exceed the 24-Hour
(Daily Only), Annual (Annual Only) or Both (Both) Existing Standards
(12/35 |Jg m 3)
40
cs
-1 35
Both
Annual Only
Daily Only
-100 -90
Longitude
To create the PM2.5 DVs for the 12/35 analytical baseline, the reductions in primary
PM2.5 emissions needed to just meet 12/35 at the highest DV monitor by county were
calculated using the air quality ratios in Table 2-1. The emission reductions were calculated
as follows:
A Emission^ = w^e;,std-wTarflet,std x
sta AQratiostd 1 J
where AEmissioristd is the emission reduction required to meet the annual or 24-hour
standard; DVTarget,std is the level of the annual or 24-hour standard to be met; DVModeistd is the
modeled PM2.5 design value for the annual or 24-hour standard at the county highest
monitor; Aqratiostd is the air quality ratio for that standard; and the factor of 1000 converts
units from kton to ton.
For example, the highest 2032 annual PM2.5 DV in Kern County is 13.70 |j,g m-3 at site
06-029-0016 after applying the 75% NOx emission reduction to the 2032 DVs. The annual
air quality ratio for primary PM2.5 emissions in Northern California is 3.15 |j,g m-3 per kton.
Therefore, to meet an annual standard of 12 |j,g m-3, a total of 527 tons of primary PM2.5
emissions would be needed (i.e., (13.70-12.04)/3.15 x 1000). The highest 2032 24-hour
PM2.5 DV in Kern County is 36.4 |j,g m-3 at site 06-029-0016 after applying the 75% NOx
69
-------
emission reduction to the 2032 DVs. The 24-hour air quality ratio for primary PM2.5
emissions in Northern California is 9.97 |Lxg nr3 per kton. Therefore, to meet a 24-hour
standard of 35 |Lxg nr3, a total of 100 tons of primary PM2.5 emissions would be needed (i.e.,
(36.4-35.4)/9.97 x 1000). To determine the overall emission reductions needed to meet the
combination of annual and 24-hour standards, the maximum needed reduction across
standards is calculated. For the Kern County example, a total 527 tons of primary PM2.5
emission reductions are needed to meet the 12/35 standard combination (i.e., the
maximum of 527 tons and 100 tons).
After the emission reductions needed to meet a standard combination are identified,
the PM2.5 DVs are adjusted to correspond with the emission reductions. The PM2.5 DVs
associated with meeting a standard combination at the highest monitor in a county are
calculated as follows:
DVstd.combo DVinitial AETTlissiOTlg^d combo ^ AQvutio/1000 (2-2)
In the Kern County example, the adjusted annual DV for the 12/35 case is 12.04 ngnr3
(13.70-527*3.15/1000) and the adjusted 24-hour DV is 31.1 [ig nr3 (36.4-527*9.97/1000).
2.3.3 Emission Reductions to Meet Revised and Alternative Standards
PM2.5 DVs in the 12/35 analytical baseline exceed the levels of the revised and
alternative standards in some areas of the country. The emission reductions needed to
resolve these exceedances and the associated air quality improvements contribute to the
incremental costs and benefits of the revised and alternative standard levels.
Exceedances of the revised and alternative standard levels in the 12/35 analytical
baseline are shown by county in Figure 2-9. Since the PM2.5 DVs have been adjusted to meet
the 24-hour standard level of 35 |Lxg nr3 in the analytical baseline, there are no exceedances
of the 24-hour standard for the cases of 10/35, 9/35, and 8/35. For the 10/35 case, five
counties in the east, two in the northwest, and thirteen in California have annual PM2.5 DVs
greater than 10 |Lxg nr3 in the 12/35 analytical baseline. For the 10/30 case, twenty-nine
counties have 24-hr DVs greater than 30 |Lxg nr3 with annual DVs less than 10 |Lxg nr3, and
nine counties exceed both the 24-hr and annual standards. For the 9/35 case, nineteen
counties exceed the annual standard in the Eastern U.S., compared with five for the 10/35
70
-------
and 10/30 cases. The total number of counties exceeding the standards increases from 52
to 117 when moving from 9/35 to 8/35. Additional information on PM2.5 DVs for the 2032
projection and 12/35 analytical baseline are available in section 2A.2.2 of Appendix 2A.
10/35
10/30
B
Both
Annual Only
24-hr Only
120°W 110°W 100°W 90°W 80 °W 70°W 120°W 110°W 100°W 90°W 80°W 70°W
Figure 2-9 Counties with PM2.5 DVs that Exceed Alternative Annual (Annual Only),
24-Hour (Daily Only), or Both (Both) Standards in the 12/35
Analytical Baseline
The primary PM2.5 emission reductions needed to meet the revised and alternative
standard levels of 10/35,10/30, 9/35, and 8/35 relative to the 12/35 analytical baseline
were calculated using Equation 2-1 and the air quality ratios in the Table 2-1, The emission
reductions needed to meet the standard levels in the Eastern and Western U.S. are shown
in Figure 2-10. These emission estimates are used to inform identification of emission
controls for meeting the standard levels analyzed. Additional information on estimating the
emission reductions needed to meet revised and alternative standards is available in
section 2A.3.4.2 of Appendix 2A.
71
-------
40,000-
> 30,000-
c 20,000- East
I II
E 10,000- H
. JJII
12/35 10/35 10/30 9/35 8/35
Standard Level
Figure 2-10 Total Primary PM2.5 Emission Reductions Needed to Meet the Revised
and Alternative Standard Levels of 10/35,10/30,9/35, and 8/35
Relative to the 12/35 Analytical Baseline in the Eastern and Western
U.S.
2.3.4 Limitations of Using Air Quality Ratios
There are important limitations to the methodology of calculating and using air
quality ratios to predict the response of air quality to emissions changes. The air quality
ratios are calculated with results from only two CMAQ model runs and assume that the
monitor DVs would decrease with additional reductions in the future similar to how the
CMAQ model runs predicted changes in air quality concentrations. In addition, the model
response to emissions changes is analyzed at the county-level and air quality
concentrations at a monitor are assumed to decrease linearly with emission reductions in a
county. Due to the strong local influence of changes in primary PM2.5 emissions on air
quality, the generalized air quality ratio approach may not capture the specific features of
how the DV at a monitor in a county would respond to changes in specific primary PM2.5
emissions in the county. Ideally, direct modeling would be applied to account for the
location of the source relative to the location of the monitor using a model configuration
designed to capture the local features near the source. Such source-specific, high-resolution
modeling is beyond the scope of this national assessment.
The exact impact of using the air quality ratio methodology to estimate the emission
reductions needed for attainment and the associated effect on the cost and benefits is
uncertain and may vary from monitor-to-monitor. We do not believe that this methodology
¦
12/35 10/35 10/30 9/35 8/35
Standard Level
72
-------
tends towards any general trend or results systematically in either an underestimation or
overestimation of the costs and benefits of attaining the revised and alternative standard
levels.
2.4 Description of Air Quality Challenges in Select Areas
Several groups of areas have air quality characteristics that limit our ability to
characterize how standard levels might be met given highly local influences that require
more specific information beyond what is available for this type of national analysis. The
challenging air quality characteristics include features of certain near-road sites, cross-
border transport, effects of complex terrain in the west, and identifying wildfire influence
on projected PM2.5 DVs that could potentially qualify for exclusion as atypical, extreme, or
unrepresentative events (USEPA, 2019b). In particular, we note that our analysis is limited
in its ability to evaluate potential air quality improvements in border counties, major
California air basins, small western mountain valleys, and two near-road sites with
challenging local features. As a result, we have treated these areas differently in the control
strategy analysis as described in Chapter 3. In this section, we describe the nature of the air
quality conditions in these areas and the challenges they present for our national
assessment.
2.4.1 Cincinnati, OH and Fort Lee, NJ Near-Road Sites
Of the 47 near-road sites included in the model projections, two (Cincinnati, OH, 39-
061-0048 and Fort Lee, NJ, 34-003-0010) were challenging for identifying sufficient
emission controls to meet the revised and alternative standard levels. The challenges in
Cincinnati (Hamilton County) appear to be related to construction of a storage facility near
the site during the 2016-2020 monitoring period used in our projections. In Figure 2-11,
the near-road monitor and neighboring storage facility are shown on the left, and images
from before, during, and after construction are shown on the right. The RIA modeling
projected a 2016-2020 DV of 11.65 |~ig/m3 to decrease by 1.21 |~ig/m3 to 10.44 |~ig/m3 in
2032. The most recent ambient PM2.5 DV at this site is 10.3 |~ig/m3 for the 2020-2022
period, which is less influenced by construction than the prior years used in model
projections. The construction activity near the Cincinnati site may have contributed to the
high base-period DV of 11.65 |~ig/m3 that led to the high projected DV, which resulted in the
73
-------
challenges for meeting the revised standard at the site. However, a detailed local analysis
beyond the scope of the national RIA would be needed to determine the full contribution of
the construction and other local influences on the PM2.5 DV at this site.
Figure 2-11 Cincinnati Near-Road Site (Left) and Storage Building Construction
Near the Site (Right)
Source: Map Data ©2023 Google,
The Fort Lee, NJ near-road site also presents challenges for identifying sufficient
emission controls to meet revised and alternative standard levels in the RIA. The Fort Lee
site (Bergen County) is close to the roadway interchange that leads to the George
Washington Bridge (Figure 2-12). Six major highways converge in the area leading to the
bridge, and the location has been reported to be the most congested freight-significant
highway location in the nation (ART1, 2023). Additionally, the site is located near the urban
activity of downtown Fort Lee (Figure 2-12). The projected 2032 DV at the site is 9.78
ug/m3. This value is higher than the other Bergen County site (6.90 jj.g/m3) and monitors in
Manhattan (maximum: 8.94 p,g/m3) demonstrating the importance of local contributions to
the concentrations. However, understanding the nature of the local contributions and
developing an approach for meeting the standard under the complex conditions at this site
would require a detailed study that goes beyond the scope of the RIA.
74
-------
Figure 2-12 Fort Lee, NJ Near-Road Site (Red Balloon) and Roadway Exchange
Leading to George Washington Bridge.
Source: Map Data ©2023 Google.
2.4.2 Border Areas
2.4.2.1 Imperial County, CA
As described in the Clean Air Act Section 179B5 Technical Demonstration by the
California Air Resources Board (CARB, 2018b), the Imperial County PM2.5 nonattainment
area is an agricultural community located in the southeast corner of California that shares a
southern border with Mexicali, Mexico. Imperial County includes three PM2.5 monitoring
sites, located in the cities of Calexico (site ID: 06-025-0005), El Centro (site ID: 06-025-
1003), and Brawley (site ID: 06-025-0007) (Figure 2-13). Although these three cities are of
similar size and have similar emission sources, the PM2.5 DV at the Calexico monitor closest
to the IJ.S.-Mexico border is much greater than the other two monitors. The projected 2032
annual PM2.5 DV is 12.36 \xg nr3 in Calexico, 9.70 (.ig nr3 in Brawley, and 8.69 |j.g nr3 in El
Centro. The Calexico monitor is in an airshed that includes both Calexico and Mexicali and
is less than one mile from the international border. Previous analysis has demonstrated
that Mexicali emissions have a daily influence on PM2.5 concentrations in Calexico and can
contribute to PM2.5 NAAQS exceedances there (CARB, 2018a, 2018b).
5179B refers the section of the Clean Air Act that addresses situations where a nonattainment area would be
able to attain and maintain, or would have attained, the NAAQS but for emissions emanating from outside of
the U.S.
75
-------
The city of Mexicali has a population of about 700,000 (CARB, 2018a) and Calexico
has a population of 38,633 (2020 U.S. Census). The nighttime aerial view of Calexico and
Mexicali in Figure 2-14 illustrates the much larger scale of urban activity in Mexicali than
Calexico. Substantially greater emissions have been estimated for Mexicali than Calexico
(i.e., 3,4x greater for NOx, 13.7x greater for combined SO2 and sulfate, and 57% greater for
primary PM2.5, (CARB, 2018b). PM2.5 emissions in Imperial County are dominated by dust
with limited contribution from other controllable sectors (Figure 2-15). Considering the
influence of Mexicali emissions on PM2.5 concentrations in Calexico, the limited emissions
available for control in Imperial County, and the relatively lower concentrations predicted
at the two Imperial County monitors away from the border, EPA believes it is reasonable to
assume that a significant portion of the emissions affecting this area cannot be controlled in
California. However, a detailed local analysis beyond the scope of the RIA would be needed
to evaluate this possibility.
Figure 2-13 Imperial County and the Nonattainment Area
Source: (CARB, 2018a)
76
-------
Mexicali
'3*$
jm*> 4 * N- Cla ,,. -| ... ]T -r
' • -rsfcigftflr ««Macgvv ' '^JT^d
' - •"¦ '
" *
V - V^mv~~
- Calexico " *¦ ~ £y
Figure 2-14 Nighttime Aerial View of Calexico, CA and Mexicali, MX
Source: (CARB, 2018b)
CO
©
2000-
w
O 1000-
co
w
LU
2018
2032
Dust AgPrFire Area Nonroad Onroad EGU Other NonEGU ResWood
Figure 2-15 Annual Source Sector Emission Totals for PM2.5 for 2018 and 2032 in
Imperial County
Note: Sector names defined in Figure 2-4
2.4.2.2 Cameron and Hidalgo County, TX
The Brownsville monitor in Cameron County, TX (site ID: 48-061-0006) and the
Mission monitor in Hidalgo County, TX (site ID: 48-215-0043) are in the Lower Rio Grande
Valley, which includes the northern portion of the state of Tamaulipas, Mexico. Addressing
the exceedances of the 9/35 standard level at the monitors in Cameron (2032 annual DV:
9.90 j.ig nr3) and Hidalgo (2032 annual DV: 10.69 ng nr3) is challenging due to the location
77
-------
of these areas along the U.S.-Mexico border. The Brownsville monitor is within one mile of
the Mexican metropolitan area of Matamoros (population: 540,000; datamexico.org) and
the Mission monitor is about nine miles from the Mexican metropolitan area of Reynosa
(population: 700,000; datamexico.org). Due to the southeast to northwest wind pattern
(Figure 2-16), emissions from these local metropolitan areas in Mexico might influence
PM2.5 concentrations at the Brownsville and Mission monitors. Studies have also identified
long-range transport of emissions from agricultural burning and wildfire in the
southwestern states of Mexico and Central America as major regional sources that
influence air quality along the U.S.-Mexico border (Karnae and John, 2019; TCEQ, 2015).
Long-range transport of Saharan dust also episodically influences concentrations in this
area based on speciation data, satellite imagery, and wind-flow back trajectories (TCEQ,
2015).
Dust makes up the largest fraction of primary PM2.5 emissions in Hidalgo and
Cameron County in the 2018 and 2032 modeling cases (Figure 2-17). Paved-road dust
emissions are projected to increase in these counties between 2018 and 2032 due to
projected increases in the vehicle miles travelled. Area source emissions are also projected
to increase due to population-based emission projection factors. Increases in dust and area
source emissions from 2018 to 2032 offset the decreases in primary PM2.5 emissions
projected for EGUs and mobile (onroad/nonroad) sources in Cameron and Hidalgo County
(Figure 2-17). A local area analysis would be better suited than the national RIA to
understand the potential growth in dust and area source emissions as well as the potential
contributions of cross-border transport to projected exceedances in this area.
78
-------
Hidalgo
Cameron
Mission"
'Isla Blanca Park
Mexico
Figure 2-16 Location of Mission and Brownsville Monitors in Hidalgo and Cameron
County, respectively, with Annual Wind Patterns from Meteorological
Measurements
Source: (TCEQ, 2015)
Figure 2-17 Annual Source Sector Emission Totals for PM2.5 for 2018 and 2032 in
Cameron and Hidalgo County Combined
Note: Sector names defined in Figure 2-4
2.4.3 Small Mountain Valleys in the West
As described in Section 2.1.2, meteorological temperature inversions often occur in
small northwestern mountain valleys in winter and trap pollution emissions in a shallow
atmospheric layer at the surface. Primary PM2.5 emissions, particularly from home heating
with residential wood combustion, can build up in the surface layer and produce high PM2.5
concentrations in winter (e.g., Figure 2-18). The mountain valleys are often very small in
size relative to the area of the surrounding county and the scales resolved by
79
-------
photochemical air quality models. For instance, the Portola nonattainment area for the
2012 PM2.5 NAAQS and the city of Portola are shown within Plumas County, CA in Figure 2-
19, The Lib by nonattainment area for the 1997 PM2.5 NAAQS and the city of Libby are
shown within Lincoln County, MT in Figure 2-20,
Figure 2-18 Air Pollution Layer Associated with a Temperature Inversion in
Missoula, MT in November 2018
Source: Tommy Martino, Missoulian6
6 Missoula health official: Air quality likely to worsen over next few days. David Erickson Missoulian, updated
Jan. 14,2019. Available at — https://missoulian.com/news/local/missoula-health-official-air-quality-likely-
to-worsen-over-next-few-days/article_clf00499-8al0-562 5-8af8-043d7ba02a4d.amp.html,
80
-------
.Susanville
Vestwood-V¦
.Chester
Herlong
G '11V I i I'
^CresMnt Mills:
Chilcool
3Sierra Brooks*
- \i rf'fv- *1
40.4°N
nera^fl.'
40.2°N -i
Standish
J -10.0"'.- (
39.8°N -agaJiaS'l _
Ijise^jj
39.6°N J
121.6°W 121.4°W 121.2 °W 121.0°W 120.8°W 120.6°W 120.4°W 120.2°W
Longitude
Figure 2-19 Plumas County, CA (Grey), Porto la Nonattainment Area (Red), and City
of Portola (Purple)
Source: Map Data ©2022 Google.
49.0°N
48.8°N
48.6°N
48.4°N
48.2°N
48.0°N
116.0°W 115.8°W 115.6°W 115.4°W 115.2°W 115.0°W 114.8°W 114.6°W
Longitude
Figure 2-20 Lincoln County, MT (Grey), Libby Nonattainment Area (Red), and City
of Libby (Purple)
Source: Map Data ©2022 Google.
81
-------
Due to the small size of the urban areas within the western mountain valleys, air
quality planning is commonly based on linear rollback methods (rather than air quality
process modeling) for these areas (e.g., LRAPA, 2012; NSAQMD, 2017). The linear rollback
approach relates wood-smoke contribution estimates at the exceeding monitor to the local
(sub-county) wood combustion emission totals to estimate the tons of emission reductions
needed to meet the standard. Due to the high effectiveness of reducing PM2.5 emissions
near monitors under stagnant meteorological conditions, the PM2.5 response factors from
linear rollback methods estimate that relatively small emission reductions can greatly
influence PM2.5 concentrations in the mountain valleys. For instance, based on the linear
rollback analysis in the Portola, CA state implementation plan (NSAQMD, 2017), a
reduction of 100 tons of primary PM2.5 emissions would reduce the annual DV by about 6.6
|Lxg nr3. This responsiveness is about 30x more efficient than photochemical modeling
estimates of PM2.5 responsiveness for county-wide emission reductions under typical
meteorological conditions (i.e., outside of mountain valley stagnation conditions). Our
national RIA analysis did not apply linear rollback-based response factors for the mountain
valleys because emission and control information are available only at the county level, and
therefore controls cannot be targeted to the local communities in our analysis. To address
standard exceedances in the small mountain valleys, a detailed analysis would be necessary
that considers local PM2.5 response factors and applies controls in the local community.
Challenges due to the wood-smoke issues just described occur in five western
counties including Plumas, CA; Lincoln, MT; Shoshone, ID; Lemhi, ID; and Klamath, OR. The
populations of the relevant cities within these counties are less than 3,200, except for
Klamath Falls (22,000) (Table 2-2). In addition to challenges related to residential wood
combustion and meteorological temperature inversions, PM2.5 concentrations in these
areas may also be influenced by wildfire smoke that could potentially qualify as atypical,
extreme, or unrepresentative events. Some wildfire influence likely persists in the
projected 2032 PM2.5 DVs despite the removal of EPA-concurred exceptional events and the
wildfire screening described in Section 2.2.2. Sensitivity projections using an extreme value
cutoff concentration of 34 |Lxg nr3 (99.0th percentile) rather than the default 64 |Lxg nr3
(99.9th percentile) were performed to explore the potential for wildfire impacts to affect
82
-------
attainment of the standards. The alternative cutoff value led to DV projections (Table 2-2)
that are 0.27 to 1.08 |Lxg nr3 lower than the default approach at the five sites. Therefore,
projections for these sites may include an important contribution from wildfire. However,
a detailed local analysis would be needed to fully characterize the wildfire influence on
attainment in these areas as well as the wood-smoke issues discussed above.
Table 2-2 Information on Areas with Challenging Residential Wood Combustion
Issues
County,
City
Annual 2032 DV
Annual 2032 DV
State
(Population3)
(M-g nr3)
Alternative Fire
Screeningb
fug m-3)
Plumas, CA
Portola (1,913)
14.02
13.75
Lincoln, MT
Libby (2,845)
11.63
11.16
Shoshone, ID
Pinehurst (1,620)
10.56
10.08
Lemhi, ID
Salmon (3,182)
9.85
9.44
Klamath, OR
Klamath Falls
10.69
9.61
(22,002)
a Population from Census.gov (https://www.census.gov/programs-surveys/popest/technical-
documentation/research/evaluation-estimates/2020-evaluation-estimates/2010s-cities-and-towns-
total.html]
b Screening based on a 99.0 percentile cutoff concentration of 34 |_ig nr3 (rather than the default 99.9
percentile value of 64 |ag nr3)
2.4.4 Califo rnia Areas
Several areas in California present challenges in the RIA analysis in addition to the
Imperial and Plumas County areas discussed above. The additional areas, described in this
section, are SJV, South Coast Air Basin, and three relatively isolated counties.
2.4.4.1 San Joaquin Valley, CA
SJV is a large inter-mountain air basin covering approximately 25,000 square miles
(SJVAPCD, 2018) that makes up the southern portion of California's Central Valley. SJV is
formed by the Sierra Nevada mountains in the east, the coastal mountain ranges in the
west, and the convergence of mountain ranges at the Tehachapi mountains in the south.
The SJV nonattainment area (Figure 2-21) includes eight counties with a combined
population of about 4.3 million. Due to the typical north to south wind pattern (Ying and
Kleeman, 2009) and wintertime meteorological inversions, PM2.5 concentrations tend to be
highest in the south near Bakersfield and the convergence of the mountain ranges.
83
-------
Stockton
ammoth
lakes"**
Fresno
Visalia
Bakersfield
San Luis^
Obispo
Santa Maria
°
Figure 2-21 San Joaquin Valley Nonattainnient Area and Location of Highest PM2.5
Monitor in Bakersfield (06-029-0016)
Source: Map Data ©2022 Google.
SJV is currently in nonattainment of the 1997 and 2012 annual PM2.5 NAAQS and the
2006 24-hr PM2.5 NAAQS. The ambient DVs at the highest SJV monitor for the 2009-2011 to
2019-2021 DV periods are shown in Figure 2-22. Discerning progress from the SJV DVs
over this period is complicated by the year-to-year variability in wildfire activity and
meteorological conditions that strongly influence PM2.5 concentrations. However, the
effectiveness of SJV control strategies has previously been demonstrated in terms of
reductions in the annual number of days that exceed the 24-hr standard level of 35 jig nr3
(Figure 2-23; (SJVAPCD, 2018)). SJV control strategies focus on reducing NOx emissions to
lower ammonium nitrate concentrations and reducing primary PM2.5 emissions to lower
carbonaceous and crustal PM2.5 concentrations (SJVAPCD, 2018). These strategies are
based on decades of modeling research and multiple intensive field measurement
campaigns such as the 1995 Integrated Monitoring Study (IMS), the 2000/2001 California
Regional PM10/PM2.5 Air Quality Study (CRPAQS), the 2010 California Research at the
84
-------
Nexus of Air Quality and Climate Change (CalNex) study, and the 2013 Deriving
Information on Surface Conditions from Column and Vertically Resolved Observations
Relevant to Air Quality (DISCOVER-AQ) study The effectiveness of NOx reduction for
control of ammonium nitrate in SJV has also been demonstrated using data from the long-
term ambient monitoring record (Pusede et al., 2016).
25-
00""^
20-
F
15-
TO
LT)
C\j
10-
CL
5-
0-
SJV
09-11 10-12 11-13 12-14 13-15 14-16 15-17 16-18 17-19 18-20 19-21
Design Value Period
Figure 2-22 Recent Annual PM2.5 DVs at the Highest SJV Monitor for Design Value
Periods (e.g., 11-13: 2011-2013). Dashed line is the 2012 Annual PM2.5
NAAQS Level (12 |ag m 3)
Figure 2-23 Decrease in the Number of Days SJV Exceeded the 24-hr NAAQS Level
(35 ^g m 3)
Source: (SJVAPCD, 2018)
85
-------
SJV air quality is influenced by emissions from large cities such as Bakersfleld
(population: 400,000) and Fresno (population: 540,000), an extremely productive
agricultural region, dust exacerbated by drought, major goods transport corridors (i.e.,
Interstate-5 and Highway 99), and wildfire. Primary PM2.5 emission totals are shown for
SJV in our 2032 modeling case in Figure 2-24. PM2.5 emissions are largest from agricultural
dust from the production of crops and livestock, paved and unpaved road dust, prescribed
burning, and cooking. Wildfire also contributed 13,100 tons of PM2.5 emissions to SJV based
on 2018 levels.
The highest projected 2032 annual DV in SJV Is 15.34 |Lxg nr3 in Bakersfleld (site ID:
06-029-0016). To address standard exceedances in SJV in the RIA, we applied 75% NOx
emissions reductions beyond the 2032 modeling case and pursued emission reductions of
primary PM2.5. However, the RIA is not well suited to identifying the specific measures
needed to meet standards in SJV given the nature and magnitude of the air quality
challenges. Challenges include air quality influenced by complex terrain and meteorological
conditions that would be best characterized with a high-resolution modeling platform
developed for the specific conditions of the valley. Also, specific local information on
measures for reducing emissions from agricultural dust and burning and prescribed
burning would be valuable given the magnitude of those emissions in SJV. Characterizing
the influence of wildfire on PM2.5 concentrations and potential atypical, extreme, or
unrepresentative events in SJV would also benefit from a local analysis. Wildfire screening
is particularly complex in California because different parts of the state have different
wildfire seasonality (e.g., Barbero et al., 2014), and severe wildfire episodes can occur
during periods where anthropogenic PM2.5 concentrations may also be high. Progress
toward meeting the alternative standards in SJV will likely occur as an outgrowth of
existing efforts to meet the 1997, 2006, and 2012 PM2.5 NAAQS.
86
-------
4000i
Figure 2-24 Annual Source Sector PM2.5 Emission Totals in SJV Counties for 2032
Modeling Case
Note that DustAgProd: Dust from Agricultural Production; AgBurn: Open Agricultural Burning;
DustRoad: Paved and Unpaved Road Dust; NonEGUpt: Non-EGU Point Sources; Onroad: Onroad
Mobile Sources; ResWoodComb: Residential Wood Combustion; Cooking: Commercial Cooking and
Residential Grilling; Other: Airports, Commercial Marine Vehicles, Rail, Solvents, and Other Non-
Point Area Sources; Nonroad: Nonroad Mobile Sources; WasteBurn: Open Waste Burning;
DustConstruct: Construction Dust; GasComb: Gas Combustion; and DustMineQrry: Dust from
Mining and Quarrying. Wildfire emissions (Not Shown) are 13,100 tons. Point Source Emissions for
NonlPM, EGU, and Oil&Gas Reflect Levels in the Nonattainment Area.
2.4.4.2 South Coast Air Basin, CA
The South Coast Air Basin (SoCAB) is formed by mountain ranges on three sides and
the Pacific Ocean in the west (Figure 2-25). SoCAB includes all or part of four counties (LA,
Riverside, San Bernardino, and Orange) with a combined population of over 17 million and
diverse emission sources associated with the large population, the ports of LA and Long
Beach, wildfire, and transportation of goods. The semi-permanent Pacific high-pressure
system leads to subsidence temperature inversions over SoCAB that can influence air
pollution processes by capping vertical mixing over the basin (Jacobson, 2002; Lu and
Turco, 1995). The sea-breeze circulation transports emissions from coastal ports and Los
Angeles to inland areas such as Riverside and San Bernardino (Lu and Turco, 1995;
Neuman et al., 2003; Pilinis et al., 2000). This transport, along with concurrent formation of
secondary PM2.5 and limited ventilation due to terrain blocking and temperature
87
-------
inversions, causes the highest PM2.5 concentrations to occur downwind of LA in Riverside
and San Bernardino. For instance, the projected 2032 annual DV at the highest site in LA is
12.87 |_ignr3 (site ID: 06-037-1302) and is 13.79 (ignr3 in Riverside (site ID: 06-065-8005)
and 14.33 jjgnr3 in San Bernardino (site ID: 06-071-0027).
Site ID
• 060371302
* 060658005
¦ 060710027
' Santa Clarita
Thousand'
Oaks f
San Bernardino^
Ontario^ FJjverside
° Vo
Anaheim MorenoValley
Los Angeles
o
Santa Mpnica
Irvine
Huntington,
Beach
Temecula
34.8°N
34.6°N -
Victorvi e
o> 34.2°N -
34.0°N -
33.8"N -
33.6°N -
33.4°N -
119.0°W
118.5°W
118.0°W 117.5°W
Longitude
-J 'v>
117.0°W 116.5°W
Figure 2-25 South Coast Air Basin Nonattainment Area and Locations of Highest
PM2.5 Monitors in Los Angeles (06-037-1302), Riverside (06-065-
8005), and San Bernardino (06-071-0027)
Source: Map Data ©2022 Google,
PM2.5 DVs in SoCAB exceed the 2012 annual PM2.5 NAAQS and the 2006 24-hr PM2.5
NAAQS. As in SJV, limited progress is evident in the trend of recent annual DVs in SoCAB
(Figure 2-26). However, year-to-year variability in wildfire emissions and meteorology
might mask air quality management progress. The 2016 Air Quality Management Plan
demonstrates the effectiveness of control programs during the 1999 to 2015 period in
which SoCAB experienced significant population growth (SCAQMD, 2017). Emission
control programs for SoCAB focus on reducing NOx emissions to lower ammonium nitrate
concentrations and primary PM2.5 emissions to lower carbonaceous PM2.5 concentrations.
Ammonium nitrate tends to be elevated in Riverside and San Bernardino due to the mixing
88
-------
of NOx emissions from LA with ammonia emissions from dairy facilities near Chino during
transport inland (Neuman et al., 2003; Nowak et al., 2012). The largest primary PM2.5
emission sources in our 2032 modeling include commercial and residential cooking,
onroad mobile sources, and paved and unpaved road dust (Figure 2-27). PM2.5 control
strategies in SoCAB are based on decades of study including intensive measurement and
modeling campaigns such as the 1987 Southern California Air Quality Study (SQAQS) and
the 2010 CalNex campaign.
O)
3
LO
C\j
25-
20-
15-
10-
5-
0-
SOCAB
09-11 10-12 11-13 12-14 13-15 14-16 15-17 16-18 17-19 18-20 19-21
Design Value Period
Figure 2-26 Recent Annual PM2.5 DVs at the Highest South Coast Monitor for Design
Value Periods (e.g., 11-13: 2011-2013). Dashed line is the 2012
Annual PM2.5 NAAQS Level (12 (j,g m 3)
89
-------
I 400°-
c 3000-
R Hln.
1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
§ 5 | I I I
-------
been influenced by wildfire during the 2016-2020 monitoring period used for model
projections. To explore the possibility that wildfire influence persists in the 2032
projections, the 2032 DVs were recalculated in Sutter, Calaveras, and Alameda Counties
using a threshold of 34 |~ig/m3 (i.e., the 99.0th percentile across 2002-2020 data) rather
than 64 |~ig/m3 (i.e., the 99.9th percentile across 2002-2020 data) to exclude days with
potential wildfire influence. The recalculated projected DVs are shown in Table 2-3 along
with the default projected 2032 DVs. At the Sutter and Calaveras sites, the 2032 DVs
decrease from above to below the annual standard of 9 |~ig/m3 when the lower threshold
(i.e., 34 |j,g/m3) is applied in a place of the default threshold (i.e., 64 |j,g/m3). These results
suggest the important influence of fires on the projected DVs, especially considering that
the lower threshold is close to 4x the revised annual standard level of 9 |~ig/m3, and
additional wildfire influence likely persists even using the more stringent screening
threshold.
For the Alameda site, the 2032 DV decreases from 10.37 |~ig/m3 to 10.07 |~ig/m3
when using the 99.0th percentile (i.e., 34 |j,g/m3) rather than the 99.9th percentile (i.e., 64
Hg/m3) value to screen wildfire influence. The ambient PM2.5 DV for the 2020-2022 period
at this site is 8.6 |~ig/m3, and the 2019-2021 DV is 8.5 |~ig/m3. For the 2016-2018, 2017-
2019, and 2018-2020 DV periods that overlap the 2016-2020 monitoring period of the RIA
projections, higher ambient DVs of 12.0 |~ig/m3, H-7 Hg/m3, and 10.8 |~ig/m3 were
measured. Wildfire influence during this period may explain the much higher projected
2032 DV than the most recent two ambient DVs (i.e., 2019-2021 and 2020-2022) at the
Alameda site. For instance, CalFire reports (https://www.fire.ca.gov/incidents/) that 2017
was the most destructive wildfire year on record in California at the time in terms of
property damage; the 2018 wildfire year included a total of over 7,500 fires that burned an
area of over 1.7 million acres; and the 2020 wildfire year had nearly 10,000 fires that
burned over 4.2 million acres, making 2020 the largest wildfire season recorded in
California's modern history. Although detailed analysis of wildfire influence would be
needed to determine the full extent of the fire impacts at the Sutter, Calaveras, and Alameda
sites, the existing evidence suggests that wildfire has an important contribution to the
projected exceedances at these sites in the RIA.
91
-------
Table 2-3 Design Value Information for Additional California Areas
County
Site ID
Annual 2032 DV
m-3)
Annual 2032 DV
Alternative Fire
Screening3
m-3}
Sutter
06-101-0003
9.41
8.99
Calaveras
06-009-0001
9.60
8.88
Alameda
06-001-0011
10.37
10.07
a Screening based on a 99.0 percentile cutoff concentration of 34 |_ig m :i (rather than the default 99.9
percentile value of 64 |ag nr3].
2.4.5 Additional Considerations
The 2020-2022 annual PM2.5 DV at the North Pole Fire Station site (02-090-0035) in
Fairbanks, Alaska is 12.2 |~ig/m3 and the 24-hour PM2.5 DV is 70 |j,g/m3. Although the annual
DV is greater than the standard levels considered in the RIA, emission reductions to meet
the existing 24-hour standard could potentially reduce the annual DV to a concentration
below the standard levels considered. The North Pole site experiences local air quality
challenges that have led to persistent exceedances of the 24-hour standard set in 2006.
Elevated PM2.5 concentrations occur in the local area under extreme temperature inversion
conditions during winter when residential wood combustion emissions are prevalent. A
detailed local analysis that develops information on local emissions, modeling, and controls
would be necessary to explore this area further and is beyond the scope of the RIA.
The Hidden Valley site (04-021-3015) in Pinal, AZ was not compared with the
annual NAAQS in this assessment because Hidden Valley is a replacement for the Cowtown
Road site that was not comparable to the annual standard. The Pinal County Air Quality
Control District (PCAQCD) is currently excluding the Hidden Valley monitor in determining
minimum monitoring requirements while EPA determines its comparability to the annual
NAAQS (PCAQCD, 2020). PM2.5 concentrations at the site are influenced by local emissions
from a dairy feedlot, unpaved roads, and agricultural cropland. Projected annual PM2.5 DVs
are more than 4 |~ig/m3 lower at the 04-021-0001 site two CMAQ grid cells away in Pinal
County due to the very local nature of the emissions and air quality impacts at the Hidden
Valley site.
92
-------
2.5 Calculating PM2.5 Concentration Fields for Standard Combinations
National PM2.5 concentration fields corresponding to meeting the existing, revised,
and alternative standard levels were developed to inform health benefit calculations. First,
a gridded PM2.5 concentration field for the 2032 CMAQ modeling case was developed using
the enhanced Voronoi Neighbor Average (eVNA) method. Next, the incremental difference
in annual PM2.5 DVs between the 2032 case and case of meeting standard combinations was
calculated at monitors and interpolated to the spatial grid. The resulting field of
incremental PM2.5 concentration was then subtracted from the 2032 eVNA field to create
the gridded field for the standard combinations. The steps in developing the PM2.5
concentration fields are described further below.
2.5.1 Creating the PM2.5 Concentration Field for 2032
The gridded field of annual average PM2.5 concentrations for 2032 was developed
using the eVNA method that combines information from the model and monitors to predict
PM2.5 concentrations. The eVNA approach was applied using SMAT-CE and has been
previously described in EPA's modeling guidance document (USEPA, 2018) and the user's
guide for the predecessor software to SMAT-CE (Abt, 2014). Briefly, the steps in developing
the eVNA PM2.5 concentration field for 2032 are as follows:
Step 1. Quarterly average PM2.5 component concentrations measured during the 2017-
2019 period were interpolated to the spatial grid using inverse distance-
squared-weighting of monitored concentrations that were further weighted by
the ratio of the 2018 CMAQ value in the prediction grid cell to CMAQ value in the
monitor-containing grid cell. The weighting by CMAQ predictions adjusts the
interpolation of monitor data to account for spatial gradients in the CMAQ fields.
This step results in an interpolated spatial field of gradient-adjusted observed
concentrations for each PM2.5 component and each quarter representative of
2018.
Step 2. The 2018 eVNA component concentration in each grid cell is multiplied by the
corresponding ratio (i.e., RRF) of the quarterly-average CMAQ concentration
predictions in 2032 and 2018. This step results in spatial concentration fields for
each PM2.5 component in each quarter of 2032.
93
-------
Step 3. The 2032 PM2.5 component concentrations are summed to give the total PM2.5
concentration for each quarter in 2032. The quarterly PM2.5 concentrations are
then averaged to create the 2032 PM2.5 concentration field. The resulting PM2.5
concentration field for 2032 is shown in Figure 2-28.
ug/m3
>15
Q3 A A
T3 40
3
-120 -110 -100 -90 -80 -70
Longitude
Figure 2-28 PM2.5 Concentration for 2032 based on eVNA Method
2.5.2 Creating Spatial Fields Corresponding to Meeting Standards
To create spatial fields corresponding to meeting standard levels, the 2032
concentration field was adjusted according to the change in PM2.5 concentrations
associated with the difference in annual PM2.5 DVs between the 2032 case and the cases
where standards are met. To implement this adjustment, the following steps were applied:
Step 1. The difference in annual PM2.5 DVs was calculated at the county highest monitor
between the 2032 case and cases of meeting the 12/35,10/30,10/35, 9/35, and
8/35 standard combinations for counties with monitors exceeding the standard
levels. The PM2.5 DVs for the cases where alternative standard levels are met
were developed by applying the air quality ratios to the emission reductions for
the county (i.e., Eqn. 2-2). For the county non-highest monitors, the difference in
PM2.5 DVs was estimated by proportionally adjusting the DVs according to the
percent change in PM2.5 DV at the highest monitor.
94
-------
Step 2.
Step 3.
The difference in DVs between the 2032 case and the cases of meeting the
standard combinations were then interpolated to the spatial grid using inverse-
distance-squared VNA interpolation (Abt, 2014; Gold et al., 1997). The
interpolated field was clipped to grid cells within 50 km of monitors whose DVs
changed in meeting the standard level (USEPA, 2012a).
National PM2.5 concentration fields were developed for each standard
combination by subtracting the corresponding spatial field of PM2.5
concentration differences from Step 2 from the 2032 eVNA concentration field.
In Figure 2-29, the spatial field for the incremental change in PM2.5 concentration
between the 12/35 analytical baseline and the case of meeting the 9/35 standard
combination is shown. Additional details on the method for developing PM2.5 concentration
fields are available in section 2A.4 of Appendix 2A.
50-
CD
"§ 40-
05
30-
jig m
¦
-120 -110 -100 -90 -80 -70
Longitude
Figure 2-29 PM2.5 Concentration Improvement Associated with Meeting 9/35
Relative to the 12/35 Analytical Baseline
95
-------
2.6 References
Abt (2014). User's Guide: Modeled Attainment Test Software. Abt Associates, Prepared for
US EPA Office of Air Quality Planning and Standards
Appel, KW, Bash, JO, Fahey, KM, Foley, KM, Gilliam, RC, Hogrefe, C, Hutzell, WT, Kang, D,
Mathur, R, Murphy, BN, Napelenok, SL, Nolte, CG, Pleim, JE, Pouliot, GA, Pye, HOT, Ran, L,
Roselle, SJ, Sarwar, G, Schwede, DB, Sidi, FI, Spero, TL and Wong, DC (2021). The
Community Multiscale Air Quality (CMAQ) model versions 5.3 and 5.3.1: system
updates and evaluation. Geosci. Model Dev. 14(5): 2867-2897.
Appel, KW, Napelenok, S, Hogrefe, C, Pouliot, G, Foley, KM, Roselle, SJ, Pleim, JE, Bash, J, Pye,
HOT, Heath, N, Murphy, B and Mathur, R (2018). Overview and Evaluation of the
Community Multiscale Air Quality (CMAQ) Modeling System Version 5.2. Air Pollution
Modeling and its Application XXV, Cham, Springer International Publishing.
Appel, KW, Napelenok, SL, Foley, KM, Pye, HOT, Hogrefe, C, Luecken, DJ, Bash, JO, Roselle, SJ,
Pleim, JE, Foroutan, H, Hutzell, WT, Pouliot, GA, Sarwar, G, Fahey, KM, Gantt, B, Gilliam,
RC, Heath, NK, Kang, DW, Mathur, R, Schwede, DB, Spero, TL, Wong, DC and Young, JO
(2017). Description and evaluation of the Community Multiscale Air Quality (CMAQ)
modeling system version 5.1. Geoscientific Model Development 10(4): 1703-1732.
American Transportation Research Institute (2023) The Nation's Top Truck Bottlenecks
2023. https://truckingresearch.org/2023/02/top-100-truck-bottlenecks-2023/
BAAQMD (2009). Fine Particulate Matter Data Analysis and Modeling in the Bay Area. Bay
Area Air Quality Managemet District. San Francisco, CA. Available:
https://www.baaqmd.gOv/~/media/files/planning-and-research/research-and"
modeling/pm-data-analysis-and-modeling-report.pdf?la=en.
Barbero, R, Abatzoglou, JT, Steel, EA and K Larkin, N (2014). Modeling very large-fire
occurrences over the continental United States from weather and climate forcing.
Environmental Research Letters 9(12): 124009.
Barbero R., Abatzoglou J. T., Larkin N. K., Kolden C. A., Stocks B. (2015) Climate change
presents increased potential for very large fires in the contiguous United States.
International Journal ofWildland Fire 24, 892-899. https://doi.org/10.1071/WF15083
Bash, JO, Baker, KR and Beaver, MR (2016). Evaluation of improved land use and canopy
representation in BEIS v3.61 with biogenic VOC measurements in California. Geosci.
Model Dev. 9(6): 2191-2207.
CARB (2018a). Staff Report. 2018 State Implementation Plan for the Imperial County 12
|ig/m3 PM2.5 Annual Standard. California Air Resources Board. Available:
https://ww3.arb.ca.gov/planning/sip/planarea/imperial/staffreport.pdf.
96
-------
CARB (2018b). Clean Air Act Section 179B Technical Demonstration: Imperial County
PM2.5 Nonattainment Area. California Air Resources Board. Available:
https://ww3.arb.ca.gov/planning/sip/planarea/imperial/finall79bwoe.pdf
Chan, EAW, Gantt, B and McDow, S (2018). The reduction of summer sulfate and switch
from summertime to wintertime PM2.5 concentration maxima in the United States.
Atmospheric Environment 175: 25-32.
Emery, C, Jung, J, Koo, B and Yarwood, G (2015). Improvements to CAMx snow cover
treatments and Carbon Bond chemical mechanism for winter ozone. Final Report,
prepared for Utah Department of Environmental Quality, Salt Lake City, UT. Prepared
by Ramboll Environ, Novato, CA, August 2015.
Fountoukis, C and Nenes, A (2007). ISORROPIA II: a computationally efficient
thermodynamic equilibrium model for K+, Ca2+, Mg2+, NH4+, Na+, S042-, N03-, C1-,
H20 aerosols. Atmos. Chem. Phys. 7(17): 4639-4659.
Frank, NH (2006). Retained Nitrate, Hydrated Sulfates, and Carbonaceous Mass in Federal
Reference Method Fine Particulate Matter for Six Eastern U.S. Cities. Journal of the Air &
Waste Management Association 56(4): 500-511.
Gold, C, Remmele, PR and Roos T. (1997). In: Algorithmic Foundation of Geographic
Information Systems. In: Lecture Notes in Computer Science, Vol. 1340 (van Kereveld
M, Nievergelt J, Roos T, Widmayer P, eds) Berlin, Germany: Springer-Verlag. Voronoi
methods in GIS. pp. 21-35.
Jacobson, MZ (2002). Atmospheric Pollution: History, Science, and Regulation. Cambridge
University Press. Cambridge, https://www.cambridge.org/core/books/atmospheric-
pollution/36CC6F9EE8EB7242031F3599FEE56961.
Karl, TR and Koss, WJ (1984). Regional and National Monthly, Seasonal, and Annual
Temperature Weighted by Area, 1895-1983. Historical Climatology Series 4-3, National
Climatic Data Center, Asheville, NC, 38 pp. https://www.ncdc.noaa.gov/monitoring-
references/maps/us-climate-regions.php
Karnae, S and John, K (2019). Source apportionment of PM2.5 measured in South Texas
near U.S.A. - Mexico border. Atmospheric Pollution Research 10(5): 1663-1676.
LRAPA (2012). Oakridge Fine Particulate Matter (PM2.5) Attainment Plan. Lane Regional
Air Quality Protection Agency, Springfield, Oregon.
Lu, R and Turco, RP (1995). Air pollutant transport in a coastal environment—II. Three-
dimensional simulations over Los Angeles basin. Atmospheric Environment 29(13):
1499-1518.
97
-------
Neuman, JA, Nowak, JB, Brock, CA, Trainer, M, Fehsenfeld, FC, Holloway, JS, Hiibler, G,
Hudson, PK, Murphy, DM, Nicks Jr, DK, Orsini, D, Parrish, DD, Ryerson, TB, Sueper, DT,
Sullivan, A and Weber, R (2003). Variability in ammonium nitrate formation and nitric
acid depletion with altitude and location over California. Journal of Geophysical
Research: Atmospheres 108(D17).
Nowak, JB, Neuman, JA, Bahreini, R, Middlebrook, AM, Holloway, JS, McKeen, SA, Parrish,
DD, Ryerson, TB and Trainer, M (2012). Ammonia sources in the California South Coast
Air Basin and their impact on ammonium nitrate formation. Geophysical Research
Letters 39(7).
NSAQMD (2017). Portola Fine Particulate Matter (PM2.5) Attainment Plan. Northern Sierra
Air Quality Management District. Available: http://myairdistrict.com/wp-
content/uploads/2016/12/2017_sip.pdf.
PADEP (2014). Commonwealth of Pennsylvania Department of Environmental Protection
2014 Annual Ambient Air Monitoring Network Plan. Available:
https: //www.dep.state.pa.us/dep/deputate / airwaste / aq/aqm/docs/2 014_PA_Air_Mon
itoring_Network_Plan.pdf.
Pilinis, C, Capaldo, KP, Nenes, A and Pandis, SN (2000). MADM-A New Multicomponent
Aerosol Dynamics Model. Aerosol Science and Technology 32(5): 482-502.
Pinal County Air Quality Control District (PCAQCD) (2020). 2020 Ambient Monitoring
Network Plan and 2019 Data Summary.
https://www.pinal.gov/ArchiveCenter/ViewFile/Item/217
Pusede, SE, Duffey, KC, Shusterman, AA, Saleh, A, Laughner, JL, Wooldridge, PJ, Zhang, Q,
Parworth, CL, Kim, H, Capps, SL, Valin, LC, Cappa, CD, Fried, A, Walega, J, Nowak, JB,
Weinheimer, AJ, Hoff, RM, Berkoff, TA, Beyersdorf, AJ, Olson, J, Crawford, JH and Cohen,
RC (2016). On the effectiveness of nitrogen oxide reductions as a control over
ammonium nitrate aerosol. Atmos. Chem. Phys. 16(4): 2575-2596.
SCAQMD (2017). Final 2016 Air Quality Management Plan, South Coast Air Quality
Management District. Available: http://www.aqmd.gov/docs/default-source/clean-air-
plans/air-quality-management-plans/2016-air-quality-management-plan/final-2016-
aqmp/final2016aqmp.pdf?sfvrsn=15.
Seinfeld, J.H. and Pandis, S.N. (2016). Atmospheric Chemistry and Physics: From Air
Pollution to Climate Change. John Wiley & Sons, Hoboken.
Simon, H and Bhave, PV (2012). Simulating the Degree of Oxidation in Atmospheric Organic
Particles. Environmental Science & Technology 46(1): 331-339.
98
-------
SJVAPCD (2018). San Joaquin Valley Air Pollution Control District, 2018 Plan for the 1997,
2006, and 2012 PM2.5 Standards, Available:
http://valleyair.org/pmplans/documents/2018/pm-plan-adopted/2018-Plan-for-the-
1997-2006-and-2012-PM2.5-Standards.pdf.
Skamarock, WC, Klemp, JB, Dudhia, J, Gill, DO, Barker, DM, Duda, MG, Huang, X, Wang, W
and Powers, JG (2008). A description of the Advanced Reserch WRF version 3. NCAR
Technical Note NCAR/TN-475+STR.
TCEQ (2015). Texas Five-Year Ambient Monitoring Network Assessment. Texas
Commission on Environmental Quality. Available:
https://www.epa.gov/sites/default/files/2020-01/documents/txassess2015.pdf.
USEPA (2012a). Regulatory Impact Analysis for the Proposed Revisions to the National
Ambient Air Quality Standards for Particulate Matter. U.S. Environmental Protection
Agency, Office of Air Quality Planning and Standards, Research Triangle Park, NC. EPA-
452/R-12-003. Available: https://www.epa.gov/sites/default/files/2020-
07/documents/naaqs-pm_ria_proposed_2012-06.pdf
USEPA (2012b). Regulatory Impact Analysis for the Final Revisions to the National Ambient
Air Quality Standards for Particulate Matter, U.S. Environmental Protection Agency,
Office of Air Quality Planning and Standards, Research Triangle Park, NC. EPA-452/R-
12-005. Available: https://www.epa.gov/sites/default/files/2020-
07/documents/naaqs-pm_ria_final_2012-12.pdf.
USEPA (2018). Modeling Guidance for Demonstrating Attainment of Air Quality Goals for
Ozone, PM2.5, and Regional Haze. EPA -454/B-07-002 U.S. EPA, Office of Air Quality
Planning and Standards. Research Triangle Park, NC. EPA454/R-18-009. Available:
https://www3.epa.gov/ttn/scram/guidance/guide/03-PM-RH-Modeling_Guidance-
2018.pdf. http://www.epa.gov/ttn/scram/guidance/guide/final-03-pm-rh-
guidance.pdf.
USEPA (2019a). Integrated Science Assessment (ISA) for Particulate Matter (Final Report,
2019). U.S. Environmental Protection Agency, Washington, DC, EPA/600/R-19/188.
USEPA (2019b). Additional Methods, Determinations, and Analyses to Modify Air Quality
Data Beyond Exceptional Events. U.S. Environmental Protection Agency, Office of Air
Quality Planning and Standards, Research Triangle Park, NC, EPA-457/B-19-002.
Available: https://www.epa.gov/sites/default/files/2019-
04/documents/clarification_memo_on_data_modification_methods.pdf.
USEPA (2019c). Technical Support Document for EPA's Updated 2028 Regional Haze
Modeling. U.S. Environmental Protection Agency, Office of Air Quality Planning and
Standards, Research Triangle Park, NC. Available:
https://www.epa.gov/sites/default/files/2019-
10/documents/updated_2028_regional_haze_modeling-tsd-2019_0.pdf
99
-------
USEPA (2021). Bayesian Space-time Downscaling Fusion Model (Downscaler) Derived
Estimates of Air Quality for 2018. U.S. Environmental Protection Agency, Office of Air
Quality Planning and Standards, Air Quality Assessment Division, Research Triangle
Park, NC. EPA-454/R-21-003. Available:
https://www.epa.gov/system/files/documents/2021-
ll/2018_ds_annual_report_final.pdf
USEPA (2022a). Policy Assessment for the Reconsideration of the National Ambient Air
Quality Standards for Particulate Matter. US Environmental Protection Agency, Office of
Air Quality Planning and Standards. EPA-452/R-22-004. Available:
https://www.epa.gov/naaqs/particulate-matter-pm-standards-policy-assessments-
current-review-O.
USEPA (2023a). Technical Support Document (TSD): Preparation of Emissions Inventories
for the 2018v2 North American Emissions Modeling Platform. U.S. Environmental
Protection Agency, EPA-454/B-23-003 Available: https://www.epa.gov/air-emissions-
modeling/2018v2-platform.
USEPA (2023b). EPA's Power Sector Modeling Platform v6 using IPM Post-IRA 2022
Reference Case, https://www.epa.gov/power-sector-modeling/post-ira-2022-
reference-case.
van Donkelaar, A, Hammer, MS, Bindle, L, Brauer, M, Brook, JR, Garay, MJ, Hsu, NC,
Kalashnikova, OV, Kahn, RA, Lee, C, Levy, RC, Lyapustin, A, Sayer, AM and Martin, RV
(2021). Monthly Global Estimates of Fine Particulate Matter and Their Uncertainty.
Environmental Science & Technology 55(22): 15287-15300.
Wang, H, Zhu, Y, Jang, C, Lin, C-J, Wang, S, Fu, JS, Gao, J, Deng, S, Xie, J, Ding, D, Qiu, X and
Long, S (2015). Design and demonstration of a next-generation air quality attainment
assessment system for PM2.5 and 03. Journal of Environmental Sciences 29: 178-188.
Ying, Q and Kleeman, M (2009). Regional contributions to airborne particulate matter in
central California during a severe pollution episode. Atmospheric Environment 43(6):
1218-1228.
100
-------
APPENDIX 2A: ADDITIONAL AIR QUALITY MODELING INFORMATION
Overview
A 2018-based modeling platform was used to project future-year air quality for
2032 to identify areas that would exceed the existing, revised, and alternative PM NAAQS
after accounting for expected emission reductions from 'on-the-books' rules. This platform
uses the Community Multiscale Air Quality (CMAQ; www.epa.gov/cmaq) model for air
quality simulation and incorporates the most recent, complete set of base year emissions
information available for national modeling. PM2.5 design values (DVs) were projected to
2032 using relative response factors (RRFs) developed from CMAQ simulations based on
emissions estimated for 2018 and projected to 2032.
Air quality ratios, which relate a change in PM2.5 DVs to a change in emissions, were
used to estimate the emission reductions needed to just meet the existing, revised, and
alternative NAAQS in areas projected to exceed the standards in 2032. The emission
reduction estimates are used in identifying controls and associated costs of meeting the
standards. To inform calculations of the health benefits of meeting standards, annual-mean
PM2.5 concentration fields corresponding to cases where the existing, revised, and
alternative NAAQS are just met were developed. The PM2.5 concentration fields were
created by adjusting the 2032 field based on the CMAQ modeling using the incremental
change in annual PM2.5 DV needed to meet the standards.
The overall steps in the air quality analysis are:
1. Project annual and 24-hour PM2.5 DVs to 2032 using a CMAQ simulation for
2018 and a corresponding CMAQ simulation with emissions representative
of 2032.
2. Develop air quality ratios that relate a change in PM2.5 DVs to a change in
emissions for use in estimating the emission reductions needed to just meet
the existing and alternative NAAQS. The air quality ratios are developed
using the change in DVs associated with CMAQ sensitivity modeling where
50% reductions in anthropogenic emissions were applied in targeted
counties relative to previous CMAQ modeling for 2028.
101
-------
3. Using the air quality ratios from Step 2, estimate the emission reductions
needed to meet the existing standards (12/35) beyond the 2032 modeling
case. For counties in the San Joaquin Valley (SJV) and South Coast Air Basin of
California, 75% reductions in anthropogenic NOx emissions are applied in
addition to reductions in primary PM2.5 emissions in this step.
Concentrations of ammonium nitrate are elevated in SJV and South Coast,
and these areas are pursuing both NOx and primary PM2.5 emission
reductions to meet the existing standards. For other counties, primary PM2.5
emission reductions alone are applied. The resulting PM2.5 DVs define the
12/35 analytical baseline that is used as the reference case in estimating the
incremental costs and benefits of meeting revised and alternative standards
relative to existing standards.
4. Using the air quality ratios from Step 2, estimate the primary PM2.5 emission
reductions needed to meet the revised and alternative standards beyond the
12/35 analytical baseline.
5. Develop a gridded national PM2.5 concentration field associated with the
2032 case by fusing the 2032 CMAQ modeling with projected monitor
concentrations. Adjust the 2032 concentration field according to the changes
in PM2.5 DVs needed to meet standard levels to create national PM2.5
concentration fields associated with meeting the existing and alternative
standard levels.
In the remainder of this Appendix, the 2018 air quality model configuration and
simulation are described and evaluated in Section 2A.1. The projection of air quality from
2018 to 2032 is described in Section 2A.2. The development of air quality ratios and their
application to estimating emission reductions is described in Section 2A.3. Finally, the
development of the PM2.5 concentration fields is described in Section 2A.4.
2A.1 2018 CMAQ Modeling
CMAQ modeling was performed for 2018 to provide a reference simulation for the
PM2.5 DV projections to 2032 that are described in section 2A.2.
102
-------
2A.1.1 Model Configuration
CMAQ is a three-dimensional grid-based Eulerian air quality model designed to
estimate the formation and fate of oxidant precursors, primary and secondary PM2.5
concentrations, and deposition over regional spatial scales (e.g., over the contiguous U.S.)
(Appel et al., 2021; Appel et al., 2018; Appel et al., 2017). CMAQ simulates the key processes
(e.g., emissions, transport, chemistry, and deposition) that affect primary (directly emitted)
and secondary (formed by atmospheric processes) PM using state-of-the-science process
parameterizations and input data for emissions, meteorology, and initial and boundary
conditions. CMAQ's representation of the chemical and physical mechanisms that govern
the formation and fate of air pollution enable simulations of the impacts of emission
controls on PM2.5 concentrations.
CMAQ version 5.3.2 (doi: 10.5281/zenodo.4081737) was used to simulate air
quality for 2018 to provide a reference simulation for the 2032 air quality projection. The
geographic extent of the air quality modeling domain (12US2) is shown in Figure 2A-1. The
12US2 modeling domain covers the 48 contiguous states along with parts of Canada and
Mexico with a horizontal resolution of 12 x 12 km. Air quality modeling for a larger 12-km
domain (US EPA 2021) was used to provide chemical boundary conditions for the 12US2
domain simulation used in projecting air quality to the future. The modeling domains have
35 vertical layers with a top at about 17.6 km (50 millibars). The CMAQ simulation
included 10 days of model spin-up in December 2017 and produced hourly pollutant
concentrations for each grid cell across each modeling domain.
Gas-phase chemistry in the CMAQ simulations was based on the Carbon Bond 2006
mechanism (CB6r3) (Emery et al., 2015), and deposition was modeled with the M3DRY
parameterization. Aerosol processes were parameterized with the AER07 module using
ISORROPIAII for inorganic aerosol thermodynamics (Fountoukis and Nenes, 2007) and the
non-volatile treatment for primary organic aerosol (Appel et al., 2017; Simon and Bhave,
2012). Emissions used were based on version 2 of the 2018 emissions modeling platform
as described in detail previously (USEPA, 2023a). Emissions of anthropogenic precursors
for secondary organic aerosol (SOA) (Murphy et al., 2017) were not added to the
simulation beyond what was captured in the National Emissions Inventory. Emissions of
103
-------
biogenic compounds were modeled with the Biogenic Emission Inventory System (BEIS)
(Bash et al., 2016). Emissions of sea-spray aerosol (Gantt et al., 2015) were simulated
online within CMAQ using 2018 meteorology.
The 2018 meteorological data were derived from running Version 3.8 of the
Weather Research Forecasting Model (WRF) (Skamarock et al., 2008). The meteorological
outputs from WRF include hourly-varying horizontal wind components (i.e., speed and
direction), temperature, moisture, vertical diffusion rates, and rainfall rates for each grid
cell in each vertical layer. Selected physics options used in the WRF simulations include
Pleim-Xiu land surface model (Pleim et al., 2001; Xiu and Pleim, 2001), Asymmetric
Convective Model version 2 planetary boundary layer scheme (Pleim, 2007), Kain-Fritsch
cumulus parameterization (Kain, 2004) utilizing the moisture-advection trigger (Ma and
Tan, 2009), Morrison double moment microphysics (H. Morrison et al., 2005; Hugh
Morrison and Gettelman, 2008), and RRTMG longwave and shortwave radiation schemes
(Iacono et al., 2008). The meteorological model configuration and evaluation have been
described previously (USEPA, 2023b).
Figure 2A-1 Map of the 12US2 (12 x 12 km Horizontal Resolution) Modeling
Domain Used for the PM NAAQS RIA
2A.1.2 Model Performance Evaluation
CMAQ predictions were evaluated by comparison with observations from U.S.
monitoring networks in 2018. Modeled PM2.5 concentrations were compared with available
observations from U.S. EPA's Air Quality System (AQS) database (www.epa.gov/aqs).
104
-------
Modeled concentrations of PM2.5 components (nitrate; sulfate; elemental carbon, EC; and
organic carbon, OC) were compared with observations from the Chemical Speciation
Network (CSN) and Interagency Monitoring of Protected Visual Environments (IMPROVE)
network (USEPA, 2019d). CSN sites tend to be in relatively urban areas and IMPROVE sites
in relatively rural areas. Model predictions were paired with observations in space and
time by averaging predictions to the observation sampling period and matching
predictions with monitors in a model grid cell. Regional performance statistics were
summarized according to the U.S. climate regions defined in Figure 2A-2. The absolute and
normalized bias and error statistics and Pearson correlation coefficient used in evaluating
model performance are defined in Table 2A-1. As described below, performance statistics
for this application are generally within the range of model performance statistics reported
in previous applications (Kelly etal., 2019; Simon etal., 2012) and suggest that the
simulations are suitable for use in our application.
In Figure 2A-3, PM2.5 model performance is shown for the AQS sites having the
highest PM2.5 DVs in the county for counties with projected annual PM2.5 DVs greater than 8
|Lxg nr3 or 24-hour DVs greater than 30 |Lxg nr3. For regions in the eastern U.S., normalized
mean biases (NMBs) are within 12% and Pearson correlation coefficients are 0.57 or
greater for all regions, except for the South (r=0.53). In western regions, the model is
generally biased low compared with observations, with NMBs ranging from -35% in the
Northwest to -21% in the West. Underpredictions in western regions could be related to
challenges in representing the influence of complex terrain in the 12-km modeling,
challenges in simulating wildfire impacts, and underestimates of windblown dust influence.
PM2.5 performance statistics by region and season across all sites are provided in Table 2A-
2. For the annual period, NMB is within 20% in eastern regions and correlation coefficients
are 0.55 or greater. In the western regions, NMB ranged from -23.4% in the West to -5.5%
in the Northwest and correlation coefficients ranged from 0.37 to 0.60.
Model performance statistics for PM2.5 sulfate by region and season for sites in the
CSN and IMPROVE networks are provided in Table 2A-3. The annual NMBs in sulfate
predictions are within ±22% for all regions except the Northwest (NMB: 30.6%) and West
(NMB: -32.9%) at CSN sites and within ±26% for all regions except the Southwest (NMB: -
105
-------
34.1%) and West (NMB: -30.5%) at IMPROVE sites. Overpredictions of PM2.5 species
concentrations in the Northwest have been previously attributed to challenges in
simulating the atmospheric mixing height near the Puget Sound and at coastal sites and in
simulating wildfire influence on concentrations (Kelly et al., 2019). Concentrations are
relatively low in the Northwest compared with the eastern U.S., and mean biases (MBs) in
sulfate predictions are <0.15 |Lxg nr3 for both networks in the Northwest. Correlation
coefficients over the annual period for sulfate predictions and observations were greater
than 0.55 in six of the nine regions for CSN sites and seven of the nine regions at IMPROVE
sites. Spatially, sulfate predictions tend to be biased slightly low in the southern and
eastern parts of the domain and biased slightly high toward the Northwestern part of the
domain (Figure 2A-4 and 2A-5).
Model performance statistics for PM2.5 nitrate by region and season for sites in the
CSN and IMPROVE networks are provided in Table 2A-4. In five of the nine regions, the
annual NMB in nitrate predictions is within ±13% at CSN sites and within ±22% at
IMPROVE sites. Nitrate predictions are biased low in the West at CSN (NMB: -60.6%) and
IMPROVE (NMB: -45.6%) sites. Underpredictions of nitrate during meteorological
inversion episodes in western mountain basins have been identified in the past due to
challenges in resolving the influence of complex terrain and chemical and meteorological
coupling in 12-km modeling (Baker et al., 2011; Kelly et al., 2019). Outside of the
Northwest, correlation coefficients for the annual period ranged from 0.54 to 0.85 at CSN
sites and 0.48 to 0.82 at IMPROVE sites. Spatially, nitrate predictions tend to be biased high
in the eastern US and low in western U.S. (Figure 2A-4 and 2A-5).
Model performance statistics for PM2.5 OC by region and season for sites in the CSN
and IMPROVE networks are provided in Table 2A-5. The annual NMB in OC predictions is
within ±40% for six of the nine regions at CSN sites and five of the nine regions at IMPROVE
sites. PM2.5 OC predictions are biased high (positive NMB) in eight of the nine regions at
CSN sites and seven of the nine regions at IMPROVE sites. Correlation coefficients over the
annual period for OC predictions and observations were greater than 0.5 in five of the nine
regions for CSN sites and six of the nine regions at IMPROVE sites. Spatially, OC predictions
tend to be biased high in the eastern U.S. and low in the western U.S., although spatial
106
-------
variability exists (Figure 2A-4 and 2A-5). Modeling of the emissions, volatility and
atmospheric chemistry related to organic aerosol formation is an active area of research
(USEPA, 2019d).
Model performance statistics for PM2.5 EC by region and season for sites in the CSN
and IMPROVE networks are provided in Table 2A-6. The annual NMB in EC predictions
ranges from -17.2% to -52.5% at CSN sites and from -19.2% to -40.0% at IMPROVE sites.
Correlation coefficients for the EC predictions and observations over the annual period
were greater than 0.5 in four of the nine regions for CSN sites and seven of nine regions for
IMPROVE sites. Spatially, EC predictions tend to be biased slightly low throughout much of
the US (Figure 2A-4 and 2A-5).
Northeast
Northern Rockies & Plains
Northwest
| Ohio Valley
South
Southeast
| Southwest
Upper Midwest
West
40-
T3
35-
30-
25-,
Google (Mapdata©20i8GooaleilNEGl ( Mp*irn
120 -100
Longitude
Figure 2A-2 U.S. Climate Regions (Karl and Koss, 1984) Used in the CMAQ Model
Performance Evaluation
Table 2A-1
Statistic
Definition of Statistics Used in the CMAQ Model Performance
Evaluation
Description
MB (|-Lgm :i] = -TIM - O0
RMSE Cngm¦3) = V2"=1(Pi-Oi)2/n
ZftPj-Qj)
NMB (%} = ' x 100
°i
Mean bias (MB] is defined as the average difference between
predicted (P) and observed (0) concentrations for the total
number of samples (n)
Root mean-squared error (RMSE)
The normalized mean bias (NMB] is defined as the sum of the
difference between predictions and observations divided by
the sum of observed values
107
-------
Statistic Description
NME (%) = SJ| °il x 100
Zj Oi
Normalized mean error (NME] is defined as the sum of the
absolute value of the difference between predictions and
observations divided by the sum of observed values
r _ 2"=1(P;-P)(0;-0)
Pearson correlation coefficient
JzlUCPi-P^jE'^COi-oyz
75
50
25
0
cf 75
E
0
75
50
25
Northeast
Ohio Valley
Upper Midwest
NMB: 6 %
MB: 0.58
RMSE;.5.08
r:0J3l
NMB: 10%
MB: 1.05
RMSE;.5.1
r.0.57
'.< • '
NMB: 12%
MB: 1.18
RMSE;,4.51
r: 0.68
Southeast
South
Southwest
NMB: 4 %
MB: 0.35 ,
RMSE^.97
r: 0.68
NMB: 8 %
MB: 0.76 , '
RMSE;.5.56
r: 0^3
NMB: -30 % .
MB: -2.81, '
RMSE;.8.08
r: 0,3^
Northern Rockies & Plains
Northwest
West
NMB: -32 % .
MB:-3.3 ,
RMSE;,9'.1
r: 0.50
NMB: -35 %
MB: -4.02y '
RMSE;f5.72
r: 0.37
.'•V . .' *. • .
% . Nldg:.-pi % .
* MB:-2.74 ' ..
w * . *.PMSE;.T1.27 .
"...•;: cpstk
0 25 50 75 0 25 50 75 0 25 50 75
Observed (ug rrf3)
Figure 2A-3 Comparison of CMAQ Predictions of PM2.5 and Observations at AQS
Sites for County Highest PM2.5 Monitors with PM2.5 DVs Greater than
8/30
108
-------
Table 2A-2 CMAQ Performance Statistics for PM2.5 at AQS Sites in 2018
Region
Season
N
Avg.
Obs.
(M.gm-3)
Avg.
Mod.
C^gm-3)
MB
(M.gm-3)
NMB
(%)
RMSE
(^gm-3)
NME
(%)
r
Northeast
Winter
13417
8.57
11.38
2.81
32.8
6.47
51.1
0.62
Spring
13593
6.28
7.79
1.51
24.0
4.21
46.8
0.58
Summer
13544
8.22
7.56
-0.65
-7.9
3.99
35.4
0.66
Fall
13255
5.80
7.23
1.43
24.6
4.72
52.9
0.57
Annual
53809
7.22
8.49
1.27
17.6
4.94
46.0
0.61
Southeast
Winter
10798
7.39
8.88
1.49
20.2
4.66
44.0
0.66
Spring
11263
7.97
7.64
-0.33
-4.2
3.40
31.3
0.69
Summer
11346
8.63
7.69
-0.94
-10.9
3.38
29.5
0.62
Fall
11026
6.92
7.84
0.92
13.3
3.76
37.6
0.65
Annual
44433
7.74
8.00
0.27
3.4
3.83
35.1
0.63
Ohio Valley
Winter
11826
8.92
11.18
2.25
25.2
5.51
44.5
0.62
Spring
12274
7.86
9.18
1.33
16.9
4.54
41.8
0.51
Summer
12228
9.96
8.95
-1.01
-10.2
4.10
30.6
0.57
Fall
11837
7.42
9.06
1.64
22.0
4.49
41.9
0.67
Annual
48165
8.55
9.58
1.04
12.1
4.69
39.2
0.58
Upper Midwest
Winter
6646
9.16
12.20
3.04
33.2
6.47
47.6
0.63
Spring
6674
6.88
9.12
2.24
32.5
5.90
52.2
0.48
Summer
6542
8.37
7.28
-1.09
-13.0
4.78
37.3
0.49
Fall
6681
5.85
7.65
1.80
30.9
3.94
46.9
0.72
Annual
26543
7.56
9.07
1.51
20.0
5.36
45.7
0.58
South
Winter
8418
7.28
9.27
2.00
27.5
5.28
51.2
0.49
Spring
8786
8.98
8.96
-0.02
-0.2
4.30
33.1
0.54
Summer
8919
11.41
9.70
-1.72
-15.0
5.18
32.5
0.66
Fall
8861
6.52
7.94
1.42
21.8
4.93
49.2
0.56
Annual
34984
8.57
8.97
0.40
4.6
4.93
39.7
0.55
Winter
6379
7.33
6.81
-0.51
-7.0
6.97
54.4
0.44
Southwest
Spring
6567
5.39
4.99
-0.40
-7.5
3.97
47.0
0.28
Summer
6868
8.31
4.64
-3.66
-44.1
7.13
54.6
0.49
Fall
6716
5.72
5.93
0.21
3.7
5.14
53.8
0.52
Annual
26530
6.69
5.57
-1.12
-16.7
5.95
52.8
0.42
N. Rockies &
Winter
5319
5.07
4.64
-0.44
-8.6
8.41
65.8
0.18
Plains
Spring
5331
4.98
4.61
-0.37
-7.5
3.41
46.0
0.51
Summer
5235
8.94
5.19
-3.75
-41.9
8.14
53.6
0.65
Fall
5450
4.96
5.30
0.33
6.7
4.50
57.6
0.46
Annual
21335
5.97
4.93
-1.04
-17.3
6.48
55.5
0.37
Northwest
Winter
10537
6.51
6.50
-0.01
-0.2
8.39
77.3
0.27
Spring
10953
4.31
5.30
0.99
23.0
4.36
63.0
0.31
Summer
11279
12.57
8.22
-4.35
-34.6
18.21
57.8
0.55
Fall
10988
8.31
10.06
1.74
21.0
10.16
66.9
0.51
Annual
43757
7.98
7.54
-0.44
-5.5
11.54
64.7
0.43
West
Winter
11744
10.05
7.95
-2.10
-20.9
7.79
46.2
0.61
Spring
12181
6.81
5.60
-1.21
-17.7
4.24
41.1
0.49
Summer
12288
12.97
9.79
-3.18
-24.5
13.86
45.7
0.46
Fall
12123
12.78
9.31
-3.47
-27.1
15.03
45.1
0.75
Annual
48336
10.66
8.17
-2.49
-23.4
11.17
44.9
0.60
109
-------
PM25EC
A '"^0 ' " A
V^» 4°afe«
Wx
D
fh
3 A
PM25 OC
-120
-100
PM25_N03
® o-b -}$0>
PM25 S04
GWli
-80 -120
Longitude
-100
-80
I
> 120
100
80
60
40
20
0
-20
-40
60
-80
-100
-120
network
• CSN
A IMPROVE
Figure 2A-4 NMB in 2018 CMAQ Predictions of PM2.5 Components at CSN and
IMPROVE Sites
PM25 EC
PM25 OC
PM25 N03
PM25 S04
-120 -100 -80 -120 -100 -80
Longitude
Figure 2A-5 NMB in 2016 CMAQ Predictions of PM2.5 Components at CSN and
IMPROVE Sites for Monitors in Counties with PM2.5 DVs Greater than
8/30
> 120
100
80
60
40
20
0
-20
-40
-60
-80
-100
-120
network
• CSN
110
-------
Table 2A-3 CMAQ Performance Statistics for PM2.5 Sulfate at CSN and IMPROVE
Sites in 2018
Region Network
Season
N
Avg
Obs.
(.US m1)
Avg.
Mod.
(.Lig m1)
MB
(Hgm3)
NMB
(%)
RMSE
(Hgm3)
NME
(%)
r
Northeast CSN
Winter
717
0.97
0.99
0.02
2.3
0.76
45.5
0.29
Spring
741
0.98
1.00
0.02
2.3
0.45
30.8
0.68
Summer
765
1.06
0.88
-0.18
-17.0
0.54
32.2
0.78
Fall
719
0.80
0.78
-0.02
-1.9
0.37
31.4
0.67
Annual
2942
0.96
0.92
-0.04
-4.2
0.55
35.0
0.60
IMPROVE
Winter
365
0.71
0.69
-0.02
-2.2
0.28
28.7
0.72
Spring
371
0.67
0.66
-0.01
-1.3
0.31
29.5
0.74
Summer
381
0.80
0.57
-0.23
-29.0
0.44
37.6
0.83
Fall
376
0.51
0.51
0.00
0.1
0.26
36.8
0.71
Annual
1493
0.67
0.61
-0.07
-9.7
0.33
33.1
0.75
Southeast CSN
Winter
515
0.80
0.80
0.00
0.0
0.40
35.4
0.49
Spring
505
0.99
0.91
-0.07
-7.5
0.41
29.7
0.61
Summer
524
0.99
0.88
-0.11
-11.4
0.54
34.3
0.47
Fall
471
0.90
0.84
-0.05
-6.1
0.34
28.7
0.66
Annual
2015
0.92
0.86
-0.06
-6.6
0.43
32.0
0.55
IMPROVE
Winter
361
0.80
0.69
-0.11
-13.6
0.34
33.7
0.60
Spring
376
1.09
0.84
-0.25
-22.6
0.45
30.8
0.68
Summer
385
1.19
0.77
-0.42
-35.4
0.60
39.5
0.68
Fall
358
0.91
0.72
-0.20
-21.4
0.39
33.0
0.73
Annual
1480
1.00
0.76
-0.25
-24.5
0.46
34.5
0.67
Ohio Valley CSN
Winter
531
1.23
1.11
-0.13
-10.2
0.73
38.8
0.43
Spring
549
1.18
1.18
0.00
0.0
0.56
33.8
0.55
Summer
576
1.37
1.18
-0.19
-13.5
0.70
32.1
0.52
Fall
530
1.09
1.02
-0.07
-6.5
0.50
31.2
0.71
Annual
2186
1.22
1.12
-0.10
-7.9
0.63
34.0
0.55
IMPROVE
Winter
197
0.93
0.78
-0.16
-16.6
0.45
33.7
0.69
Spring
210
1.18
0.99
-0.19
-15.8
0.44
27.1
0.74
Summer
208
1.37
0.95
-0.41
-30.3
0.66
37.9
0.61
Fall
207
0.98
0.83
-0.15
-15.1
0.48
36.3
0.71
Annual
822
1.12
0.89
-0.23
-20.3
0.51
33.8
0.69
Upper Midwest CSN
Winter
310
1.00
1.06
0.06
5.6
0.51
33.7
0.69
Spring
305
0.86
1.01
0.15
17.6
0.42
37.6
0.69
Summer
315
0.93
0.94
0.01
1.0
0.73
39.3
0.55
Fall
311
0.69
0.89
0.21
30.1
0.42
47.6
0.77
Annual
1241
0.87
0.98
0.11
12.1
0.54
38.9
0.65
IMPROVE
Winter
202
0.76
0.84
0.08
10.6
0.49
36.1
0.61
Spring
210
0.78
0.82
0.04
5.0
0.35
31.9
0.67
Summer
207
0.68
0.54
-0.14
-20.5
0.32
32.3
0.84
Fall
207
0.47
0.59
0.12
25.3
0.30
49.2
0.80
Annual
826
0.67
0.70
0.02
3.7
0.37
36.2
0.69
South CSN
Winter
325
1.08
1.15
0.06
5.8
0.71
40.6
0.61
Spring
338
1.40
1.13
-0.27
-19.2
0.71
32.4
0.74
Summer
353
1.51
1.14
-0.37
-24.6
0.80
36.0
0.57
Fall
315
0.99
1.05
0.06
6.0
0.68
43.2
0.59
Annual
1331
1.25
1.12
-0.14
-10.9
0.73
37.3
0.60
IMPROVE
Winter
222
0.74
0.80
0.06
7.9
0.45
41.4
0.64
Spring
228
1.17
0.87
-0.30
-25.3
0.60
33.6
0.62
Summer
213
1.33
0.81
-0.52
-39.2
0.71
41.9
0.69
Fall
230
0.89
0.76
-0.13
-14.6
0.50
37.8
0.71
Annual
893
1.03
0.81
-0.22
-21.3
0.57
38.5
0.63
Southwest CSN
Winter
213
0.44
0.39
-0.04
-10.2
0.36
50.7
0.37
Spring
207
0.43
0.44
0.01
2.7
0.19
32.4
0.48
Summer
210
0.64
0.32
-0.33
-50.9
0.45
53.2
0.21
Fall
207
0.46
0.39
-0.06
-13.9
0.20
34.0
0.42
111
-------
Region Network
Season
N
Avg
Obs.
tam3)
Avg.
Mod.
tam3)
MB
(M-g m')
NMB
(%)
RMSE
(M-g m')
NME
(%)
r
Annual
837
0.49
0.38
-0.11
-21.6
0.32
43.7
0.24
IMPROVE
Winter
833
0.23
0.24
0.01
5.1
0.17
43.6
0.62
Spring
876
0.43
0.36
-0.07
-17.2
0.22
32.4
0.57
Summer
841
0.65
0.25
-0.40
-62.2
0.49
63.1
0.56
Fall
851
0.38
0.27
-0.11
-29.1
0.23
41.5
0.73
Annual
3401
0.42
0.28
-0.14
-34.1
0.30
47.6
0.53
N. Rockies & CSN
Winter
160
0.53
0.71
0.19
35.2
0.52
59.9
0.61
Plains
Spring
158
0.75
0.82
0.07
9.1
0.46
40.4
0.74
Summer
164
0.58
0.62
0.04
6.6
0.40
42.8
0.63
Fall
161
0.53
0.67
0.14
26.5
0.42
52.1
0.66
Annual
643
0.60
0.70
0.11
18.1
0.45
47.9
0.67
IMPROVE
Winter
520
0.29
0.32
0.03
10.1
0.25
47.3
0.82
Spring
576
0.53
0.52
-0.01
-1.8
0.35
35.9
0.75
Summer
595
0.46
0.32
-0.14
-30.9
0.24
37.8
0.74
Fall
584
0.37
0.38
0.01
2.1
0.27
39.4
0.68
Annual
2275
0.42
0.39
-0.03
-7.5
0.28
39.1
0.74
Northwest CSN
Winter
144
0.28
0.52
0.24
83.9
0.41
106.8
0.23
Spring
133
0.41
0.60
0.19
45.3
0.27
53.0
0.73
Summer
144
0.61
0.71
0.09
15.1
0.69
59.5
0.18
Fall
165
0.54
0.61
0.07
12.6
0.29
38.8
0.56
Annual
586
0.47
0.61
0.14
30.6
0.44
58.4
0.36
IMPROVE
Winter
437
0.13
0.24
0.11
84.5
0.19
104.0
0.43
Spring
476
0.30
0.38
0.08
28.4
0.19
51.5
0.70
Summer
475
0.42
0.45
0.03
6.6
0.32
50.8
0.41
Fall
446
0.26
0.33
0.07
26.2
0.21
52.4
0.51
Annual
1834
0.28
0.35
0.07
25.7
0.23
57.3
0.57
West CSN
Winter
270
0.58
0.50
-0.08
-13.4
0.60
58.4
0.31
Spring
270
0.78
0.62
-0.16
-20.8
0.52
43.7
0.61
Summer
270
1.31
0.68
-0.63
-48.4
0.97
54.5
0.30
Fall
266
0.92
0.61
-0.31
-33.5
0.60
44.5
0.61
Annual
1076
0.90
0.60
-0.30
-32.9
0.69
50.2
0.48
IMPROVE
Winter
542
0.25
0.26
0.01
5.1
0.29
61.4
0.50
Spring
553
0.52
0.40
-0.12
-22.9
0.36
45.4
0.55
Summer
565
0.85
0.45
-0.40
-47.0
0.63
54.4
0.26
Fall
562
0.46
0.33
-0.12
-26.6
0.30
43.0
0.59
Annual
2222
0.52
0.36
-0.16
-30.5
0.42
50.5
0.51
112
-------
Table 2A-4 CMAQ Performance Statistics for PM2.5 Nitrate at CSN and IMPROVE
Sites in 2018
Region Network
Season
N
Avg.
Obs.
km3)
Avg.
Mod.
fug m1)
MB
(ugm')
NMB
(%)
RMSE
(M-gm3)
NME
(%)
r
Northeast CSN
Winter
717
1.81
2.11
0.30
16.4
1.23
48.2
0.70
Spring
741
0.89
0.92
0.03
3.5
0.69
51.5
0.71
Summer
765
0.35
0.24
-0.10
-30.0
0.42
65.6
0.49
Fall
719
0.62
0.79
0.17
27.5
0.71
66.6
0.74
Annual
2942
0.91
1.00
0.09
10.5
0.81
53.8
0.77
IMPROVE
Winter
365
0.66
1.19
0.53
81.3
0.91
102.6
0.62
Spring
371
0.29
0.33
0.05
16.2
0.31
66.3
0.62
Summer
381
0.15
0.15
-0.00
-1.2
0.15
66.0
0.55
Fall
376
0.21
0.34
0.13
61.1
0.42
98.3
0.58
Annual
1493
0.32
0.50
0.17
53.8
0.53
89.5
0.70
Southeast CSN
Winter
515
0.79
1.36
0.57
71.9
1.42
99.6
0.62
Spring
505
0.43
0.34
-0.08
-19.8
0.34
51.4
0.60
Summer
525
0.21
0.21
0.01
2.6
0.27
76.5
0.33
Fall
474
0.39
0.64
0.26
66.5
0.79
98.3
0.65
Annual
2019
0.45
0.64
0.19
41.0
0.84
85.3
0.66
IMPROVE
Winter
361
0.55
0.75
0.20
35.7
0.85
85.9
0.48
Spring
376
0.41
0.30
-0.11
-26.5
0.35
56.2
0.61
Summer
385
0.18
0.19
0.01
7.3
0.19
70.9
0.52
Fall
358
0.26
0.38
0.12
48.2
0.46
84.9
0.65
Annual
1480
0.35
0.40
0.05
15.5
0.52
74.8
0.56
Ohio Valley CSN
Winter
531
3.01
2.61
-0.40
-13.2
2.01
41.1
0.72
Spring
549
1.34
1.20
-0.14
-10.6
0.85
42.3
0.73
Summer
576
0.40
0.27
-0.13
-31.7
0.42
67.8
0.34
Fall
530
1.10
1.18
0.09
8.0
0.91
47.9
0.82
Annual
2186
1.44
1.29
-0.14
-10.0
1.19
44.6
0.80
IMPROVE
Winter
197
1.51
1.31
-0.21
-13.6
1.29
53.4
0.67
Spring
210
0.81
0.60
-0.21
-26.0
0.67
51.1
0.65
Summer
208
0.20
0.18
-0.02
-9.6
0.18
59.9
0.58
Fall
207
0.57
0.59
0.02
3.9
0.66
62.1
0.73
Annual
822
0.76
0.66
-0.10
-13.4
0.79
54.8
0.73
Upper Midwest CSN
Winter
310
3.08
2.92
-0.16
-5.3
1.54
34.1
0.81
Spring
305
1.44
1.17
-0.27
-18.6
1.06
43.6
0.76
Summer
315
0.36
0.22
-0.14
-39.6
0.34
64.1
0.48
Fall
311
0.89
1.00
0.11
12.6
0.77
50.7
0.79
Annual
1241
1.44
1.32
-0.11
-7.9
1.02
40.9
0.85
IMPROVE
Winter
202
2.06
1.73
-0.33
-16.2
1.40
43.6
0.77
Spring
210
0.80
0.52
-0.29
-35.7
0.75
53.3
0.74
Summer
207
0.14
0.11
-0.02
-17.8
0.14
61.3
0.58
Fall
207
0.41
0.55
0.14
34.5
0.50
64.6
0.83
Annual
826
0.84
0.72
-0.13
-14.9
0.83
49.2
0.82
South CSN
Winter
325
1.24
1.24
-0.00
-0.1
1.05
56.2
0.67
Spring
338
0.62
0.34
-0.28
-44.6
0.66
60.4
0.63
Summer
353
0.30
0.26
-0.04
-12.4
0.39
79.0
0.24
Fall
315
0.44
0.69
0.25
55.3
1.09
102.5
0.72
Annual
1331
0.65
0.62
-0.02
-3.4
0.84
67.6
0.64
IMPROVE
Winter
222
0.97
0.84
-0.12
-12.8
0.81
53.3
0.69
Spring
228
0.63
0.32
-0.31
-49.3
0.66
57.5
0.69
Summer
213
0.25
0.14
-0.11
-45.6
0.22
66.1
0.44
Fall
230
0.31
0.38
0.07
22.1
0.45
79.6
0.68
Annual
893
0.54
0.42
-0.12
-22.2
0.58
59.8
0.68
Southwest CSN
Winter
213
2.19
0.94
-1.25
-57.2
2.69
66.2
0.49
Spring
207
0.58
0.24
-0.33
-58.0
0.63
71.4
0.49
Summer
209
0.31
0.20
-0.11
-35.7
0.46
99.3
-0.14
Fall
207
0.86
0.52
-0.34
-39.9
1.34
75.2
0.38
113
-------
Region Network
Season
N
Avg.
Obs.
(M-gm3)
Avg
Mod.
tam3)
MB
(M-g m')
NMB
(%)
RMSE
(M-gm3)
NME
(%)
r
Annual
836
0.99
0.48
-0.51
-51.9
1.56
71.5
0.54
IMPROVE
Winter
833
0.28
0.12
-0.16
-55.6
0.52
75.3
0.50
Spring
876
0.20
0.06
-0.14
-70.1
0.20
76.9
0.29
Summer
841
0.18
0.03
-0.15
-82.9
0.20
83.4
0.44
Fall
851
0.15
0.08
-0.08
-50.4
0.26
77.6
0.51
Annual
3401
0.20
0.07
-0.13
-64.3
0.32
77.9
0.48
N. Rockies & CSN
Winter
160
1.51
1.18
-0.33
-22.0
1.38
50.9
0.66
Plains
Spring
158
1.28
0.65
-0.63
-49.3
1.15
54.4
0.75
Summer
164
0.25
0.14
-0.11
-44.3
0.23
68.3
0.48
Fall
161
0.57
0.67
0.10
18.5
0.57
62.8
0.82
Annual
643
0.90
0.66
-0.24
-26.7
0.95
55.2
0.71
IMPROVE
Winter
520
0.39
0.34
-0.05
-13.9
0.65
75.4
0.53
Spring
576
0.48
0.19
-0.30
-61.2
0.75
72.9
0.49
Summer
595
0.14
0.03
-0.11
-78.3
0.18
80.6
0.50
Fall
584
0.20
0.21
0.01
3.2
0.35
89.4
0.50
Annual
2275
0.30
0.19
-0.11
-37.9
0.53
77.4
0.50
Northwest CSN
Winter
144
0.93
1.04
0.11
11.8
1.48
88.6
0.26
Spring
133
0.43
0.53
0.10
23.9
0.57
79.3
0.62
Summer
144
0.36
0.27
-0.08
-23.5
0.46
89.5
0.17
Fall
165
0.96
1.18
0.21
21.9
1.36
82.0
0.50
Annual
586
0.69
0.78
0.09
13.0
1.09
84.8
0.44
IMPROVE
Winter
437
0.20
0.25
0.05
24.9
0.47
109.4
0.41
Spring
476
0.14
0.11
-0.03
-23.9
0.22
74.3
0.53
Summer
475
0.19
0.07
-0.12
-63.4
0.25
75.0
0.38
Fall
446
0.22
0.19
-0.03
-12.3
0.51
95.0
0.40
Annual
1834
0.19
0.15
-0.03
-18.4
0.38
89.3
0.40
West CSN
Winter
270
3.82
1.54
-2.28
-59.6
4.85
68.1
0.62
Spring
270
1.52
0.59
-0.93
-61.3
1.95
66.9
0.64
Summer
270
1.14
0.50
-0.64
-56.2
1.25
60.3
0.61
Fall
266
2.47
0.90
-1.57
-63.7
3.62
70.1
0.59
Annual
1076
2.24
0.88
-1.35
-60.6
3.24
67.4
0.64
IMPROVE
Winter
542
0.71
0.44
-0.27
-38.5
1.48
59.0
0.88
Spring
553
0.40
0.23
-0.17
-41.8
0.38
56.6
0.73
Summer
565
0.47
0.22
-0.25
-54.1
0.58
66.1
0.41
Fall
562
0.54
0.27
-0.27
-50.0
1.15
70.7
0.70
Annual
2222
0.53
0.29
-0.24
-45.6
0.99
63.2
0.81
114
-------
Table 2A-5 CMAQ Performance Statistics for PM2.5 EC at CSN and IMPROVE Sites in
2018
Region Network
Season
N
Avg
Obs.
(.US m1)
Avg.
Mod.
(.Lig m1)
MB
(M-gm3)
NMB
(%)
RMSE
(Hgm3)
NME
(%)
r
Northeast CSN
Winter
725
0.68
0.62
-0.06
-9.2
0.48
44.0
0.54
Spring
726
0.58
0.44
-0.14
-23.7
0.66
48.4
0.35
Summer
723
0.66
0.43
-0.24
-35.6
0.43
46.0
0.60
Fall
719
0.64
0.53
-0.10
-16.2
0.45
45.6
0.59
Annual
2893
0.64
0.51
-0.13
-21.1
0.51
45.9
0.48
IMPROVE
Winter
393
0.24
0.26
0.02
10.4
0.13
37.4
0.81
Spring
399
0.18
0.14
-0.04
-22.4
0.13
42.4
0.69
Summer
422
0.25
0.13
-0.12
-48.1
0.19
50.6
0.82
Fall
395
0.19
0.16
-0.02
-11.2
0.10
35.1
0.84
Annual
1609
0.22
0.17
-0.04
-19.2
0.14
42.0
0.74
Southeast CSN
Winter
409
0.57
0.48
-0.10
-16.6
0.30
34.4
0.76
Spring
423
0.52
0.35
-0.17
-32.7
0.35
43.1
0.67
Summer
471
0.47
0.35
-0.12
-25.1
0.32
46.6
0.45
Fall
410
0.64
0.44
-0.20
-31.0
0.43
40.4
0.69
Annual
1713
0.55
0.40
-0.14
-26.4
0.35
41.0
0.66
IMPROVE
Winter
386
0.30
0.24
-0.06
-20.0
0.18
37.8
0.83
Spring
406
0.33
0.19
-0.14
-43.3
0.27
48.5
0.78
Summer
417
0.29
0.15
-0.15
-50.0
0.24
53.7
0.83
Fall
386
0.33
0.20
-0.13
-39.8
0.32
45.9
0.84
Annual
1595
0.31
0.19
-0.12
-38.7
0.26
46.6
0.80
Ohio Valley CSN
Winter
537
0.57
0.50
-0.07
-12.9
0.30
34.7
0.58
Spring
544
0.59
0.40
-0.19
-31.8
0.36
42.1
0.54
Summer
549
0.67
0.36
-0.30
-45.6
0.45
48.3
0.41
Fall
537
0.63
0.44
-0.18
-29.3
0.41
39.8
0.49
Annual
2167
0.62
0.43
-0.19
-30.6
0.38
41.5
0.47
IMPROVE
Winter
192
0.24
0.20
-0.05
-18.8
0.11
33.2
0.72
Spring
209
0.28
0.16
-0.11
-41.2
0.16
46.9
0.66
Summer
213
0.29
0.12
-0.17
-58.7
0.21
60.3
0.57
Fall
208
0.24
0.18
-0.06
-26.5
0.13
40.0
0.63
Annual
822
0.26
0.16
-0.10
-37.9
0.16
46.1
0.57
Upper Midwest CSN
Winter
308
0.45
0.51
0.06
14.3
0.36
49.7
0.52
Spring
303
0.43
0.38
-0.04
-9.9
0.32
46.6
0.55
Summer
327
0.53
0.31
-0.22
-41.7
0.43
52.6
0.47
Fall
304
0.45
0.34
-0.11
-25.0
0.34
43.4
0.49
Annual
1242
0.47
0.38
-0.08
-17.2
0.37
48.4
0.46
IMPROVE
Winter
231
0.21
0.24
0.03
12.5
0.15
39.9
0.72
Spring
236
0.22
0.18
-0.04
-18.9
0.15
47.2
0.69
Summer
242
0.30
0.14
-0.16
-53.6
0.41
68.6
0.52
Fall
227
0.19
0.14
-0.05
-26.2
0.18
46.4
0.65
Annual
936
0.23
0.17
-0.06
-25.0
0.25
52.6
0.53
South CSN
Winter
303
0.55
0.48
-0.07
-12.9
0.36
38.7
0.58
Spring
297
0.51
0.36
-0.15
-29.2
0.28
37.1
0.69
Summer
316
0.39
0.32
-0.07
-18.9
0.25
46.6
0.53
Fall
313
0.57
0.46
-0.10
-18.3
0.31
37.8
0.72
Annual
1229
0.51
0.41
-0.10
-19.6
0.30
39.6
0.65
IMPROVE
Winter
219
0.17
0.13
-0.04
-22.8
0.11
44.0
0.61
Spring
228
0.26
0.15
-0.11
-42.5
0.18
51.3
0.70
Summer
215
0.20
0.08
-0.11
-58.0
0.18
62.2
0.58
Fall
230
0.18
0.12
-0.06
-33.7
0.11
43.5
0.75
Annual
892
0.20
0.12
-0.08
-40.0
0.15
50.6
0.64
Southwest CSN
Winter
214
1.27
0.71
-0.57
-44.5
2.35
59.5
0.39
Spring
180
0.46
0.38
-0.08
-16.5
0.54
45.0
0.41
Summer
200
0.51
0.34
-0.17
-32.7
0.31
43.8
0.49
Fall
195
0.79
0.60
-0.19
-23.8
0.44
39.2
0.61
115
-------
Region Network
Season
N
Avg
Obs.
tam3)
Avg.
Mod.
tom1)
MB
(M-gm3)
NMB
(%)
RMSE
(M-g m')
NME
(%)
r
Annual
789
0.77
0.51
-0.26
-33.5
1.28
49.8
0.42
IMPROVE
Winter
819
0.16
0.10
-0.05
-34.2
0.20
46.4
0.87
Spring
885
0.10
0.06
-0.03
-33.8
0.08
46.7
0.83
Summer
838
0.16
0.08
-0.08
-50.0
0.15
59.3
0.64
Fall
849
0.15
0.10
-0.05
-33.8
0.17
50.8
0.75
Annual
3391
0.14
0.09
-0.05
-38.5
0.15
51.3
0.79
N. Rockies & CSN
Winter
140
0.39
0.23
-0.16
-40.2
0.61
73.4
-0.01
Plains
Spring
139
0.28
0.16
-0.12
-42.4
0.26
53.3
0.45
Summer
155
0.56
0.17
-0.39
-69.6
0.98
72.2
0.20
Fall
140
0.41
0.23
-0.19
-45.0
0.43
57.8
0.21
Annual
574
0.41
0.20
-0.22
-52.5
0.64
65.9
0.14
IMPROVE
Winter
532
0.07
0.07
-0.00
-3.8
0.12
72.2
0.33
Spring
570
0.11
0.06
-0.05
-48.5
0.12
56.5
0.52
Summer
594
0.28
0.11
-0.17
-59.4
0.27
63.3
0.73
Fall
582
0.17
0.14
-0.03
-16.4
0.24
61.6
0.51
Annual
2278
0.16
0.10
-0.06
-40.0
0.20
62.6
0.56
Northwest CSN
Winter
143
0.89
0.87
-0.02
-2.0
0.77
58.9
0.41
Spring
135
0.51
0.56
0.05
10.6
0.36
49.1
0.63
Summer
145
0.69
0.36
-0.33
-47.5
0.86
55.1
0.62
Fall
158
1.14
0.86
-0.28
-24.3
0.70
43.1
0.62
Annual
581
0.82
0.67
-0.15
-18.1
0.70
50.7
0.50
IMPROVE
Winter
440
0.10
0.14
0.04
36.1
0.20
92.5
0.78
Spring
473
0.11
0.09
-0.02
-18.8
0.14
64.3
0.70
Summer
475
0.48
0.21
-0.27
-55.8
1.48
74.3
0.34
Fall
457
0.31
0.26
-0.05
-15.3
0.49
66.5
0.71
Annual
1845
0.25
0.17
-0.08
-30.8
0.80
72.6
0.39
West CSN
Winter
265
1.24
0.82
-0.42
-33.7
0.76
43.6
0.55
Spring
253
0.42
0.41
-0.01
-2.4
0.21
34.3
0.70
Summer
273
0.55
0.50
-0.05
-8.9
0.38
37.9
0.65
Fall
246
1.09
0.72
-0.38
-34.3
1.12
40.4
0.78
Annual
1037
0.82
0.61
-0.21
-25.7
0.70
40.4
0.68
IMPROVE
Winter
539
0.15
0.11
-0.03
-22.2
0.18
56.3
0.81
Spring
555
0.11
0.07
-0.04
-37.4
0.09
53.7
0.70
Summer
563
0.53
0.35
-0.17
-32.7
1.48
64.4
0.36
Fall
555
0.32
0.24
-0.08
-23.8
1.03
70.1
0.37
Annual
2212
0.28
0.20
-0.08
-29.3
0.91
63.9
0.39
116
-------
Table 2A-6 CMAQ Performance Statistics for PM2.5 OC at CSN and IMPROVE Sites in
2018
Region Network
Season
N
Avg.
Obs.
km3)
Avg.
Mod.
fug m1)
MB
(ugm')
NMB
(%)
RMSE
(ugm')
NME
(%)
r
Northeast CSN
Winter
725
1.57
3.35
1.78
113.7
2.55
119.2
0.60
Spring
726
1.50
2.63
1.12
74.5
1.68
85.6
0.62
Summer
723
2.37
2.70
0.34
14.2
1.39
37.4
0.72
Fall
719
1.33
2.56
1.23
92.2
2.13
98.9
0.63
Annual
2893
1.69
2.81
1.12
66.0
1.99
79.1
0.57
IMPROVE
Winter
393
0.84
2.04
1.20
142.3
1.46
145.6
0.73
Spring
399
0.77
1.45
0.69
89.5
1.02
104.1
0.60
Summer
422
1.49
1.63
0.14
9.4
1.06
43.8
0.58
Fall
395
0.77
1.26
0.50
65.0
0.87
77.4
0.74
Annual
1609
0.97
1.59
0.62
63.8
1.12
83.5
0.57
Southeast CSN
Winter
409
1.91
2.94
1.04
54.4
1.74
63.0
0.77
Spring
423
2.00
2.75
0.75
37.6
1.49
50.6
0.77
Summer
471
2.33
2.81
0.49
20.9
1.26
41.1
0.64
Fall
410
1.89
2.88
0.98
51.9
1.59
59.5
0.78
Annual
1713
2.04
2.84
0.80
39.3
1.52
52.4
0.73
IMPROVE
Winter
386
1.19
1.82
0.63
52.6
1.29
65.7
0.81
Spring
406
1.39
1.99
0.59
42.6
1.57
63.5
0.69
Summer
417
1.54
1.76
0.23
14.9
0.91
40.3
0.70
Fall
386
1.20
1.81
0.62
51.6
2.44
67.7
0.42
Annual
1595
1.33
1.85
0.51
38.4
1.64
57.9
0.60
Ohio Valley CSN
Winter
537
1.57
2.81
1.24
79.4
1.80
89.1
0.57
Spring
544
1.74
2.66
0.92
53.0
1.73
66.3
0.69
Summer
549
2.63
2.37
-0.26
-9.9
1.07
31.4
0.62
Fall
537
1.52
2.20
0.68
45.1
1.37
61.8
0.60
Annual
2167
1.87
2.51
0.64
34.5
1.52
57.7
0.53
IMPROVE
Winter
192
0.92
1.68
0.76
82.7
1.44
93.7
0.59
Spring
209
1.20
1.85
0.65
54.5
1.46
74.9
0.72
Summer
213
1.64
1.72
0.07
4.5
0.89
40.1
0.57
Fall
208
0.99
1.53
0.55
55.3
1.21
71.8
0.65
Annual
822
1.19
1.69
0.50
42.0
1.27
65.2
0.58
Upper Midwest CSN
Winter
308
1.33
3.22
1.89
141.6
2.66
144.9
0.53
Spring
303
1.38
2.65
1.28
92.6
2.20
104.5
0.42
Summer
327
2.77
2.60
-0.17
-6.1
5.44
47.4
0.45
Fall
304
1.21
1.87
0.66
54.7
1.24
73.9
0.41
Annual
1242
1.69
2.59
0.90
53.0
3.33
82.4
0.38
IMPROVE
Winter
231
0.72
2.01
1.29
178.7
1.97
180.7
0.52
Spring
236
0.86
2.00
1.14
133.1
1.81
143.8
0.49
Summer
242
1.97
2.10
0.12
6.2
6.23
67.1
0.51
Fall
227
0.65
1.02
0.37
56.7
0.79
82.2
0.50
Annual
936
1.06
1.79
0.73
68.4
3.46
104.0
0.45
South CSN
Winter
303
1.66
2.76
1.10
66.3
2.07
87.4
0.49
Spring
297
2.03
2.74
0.71
35.2
1.60
55.1
0.67
Summer
316
2.08
2.63
0.56
26.7
1.75
60.9
0.47
Fall
313
1.60
2.58
0.98
61.2
2.07
77.6
0.74
Annual
1229
1.84
2.68
0.84
45.4
1.88
68.9
0.56
IMPROVE
Winter
219
0.66
1.27
0.61
92.9
1.05
100.0
0.64
Spring
228
1.31
1.98
0.67
51.5
2.06
65.9
0.63
Summer
215
1.39
1.43
0.04
2.8
1.07
52.4
0.51
Fall
230
0.77
1.17
0.40
52.2
0.97
76.2
0.71
Annual
892
1.03
1.46
0.43
42.2
1.37
68.8
0.57
Southwest CSN
Winter
214
2.30
3.02
0.72
31.4
2.63
69.6
0.43
Spring
180
1.18
1.74
0.56
47.1
1.07
63.6
0.50
Summer
200
2.27
1.63
-0.64
-28.3
1.68
46.0
0.51
Fall
195
1.61
2.51
0.90
56.3
1.75
73.7
0.44
117
-------
Region Network
Season
N
Avg.
Obs.
(M-gm3)
Avg
Mod.
tom1)
MB
(M-g m')
NMB
(%)
RMSE
(M-g m')
NME
(%)
r
Annual
789
1.87
2.25
0.38
20.5
1.90
62.3
0.41
IMPROVE
Winter
819
0.53
0.68
0.15
28.5
0.55
59.9
0.84
Spring
885
0.46
0.72
0.26
56.0
0.43
70.9
0.71
Summer
838
1.22
0.84
-0.38
-31.2
1.18
52.7
0.55
Fall
849
0.60
0.80
0.20
32.5
1.93
69.5
0.33
Annual
3391
0.70
0.76
0.06
8.4
1.18
60.8
0.43
N. Rockies & CSN
Winter
140
1.07
1.28
0.20
18.8
1.63
96.3
0.04
Plains
Spring
139
0.98
1.19
0.20
20.7
0.86
60.3
0.62
Summer
155
2.92
1.36
-1.57
-53.6
3.15
61.3
0.45
Fall
140
1.12
1.23
0.11
10.2
1.13
62.4
0.26
Annual
574
1.56
1.27
-0.30
-19.0
1.95
67.2
0.35
IMPROVE
Winter
532
0.33
0.67
0.34
100.5
0.81
124.4
0.32
Spring
570
0.57
0.75
0.18
30.8
0.88
67.5
0.40
Summer
594
2.78
1.38
-1.40
-50.5
2.82
55.9
0.81
Fall
582
1.12
1.18
0.06
5.8
1.75
68.4
0.44
Annual
2278
1.23
1.01
-0.23
-18.4
1.79
64.5
0.64
Northwest CSN
Winter
143
2.32
4.24
1.93
83.1
3.68
99.0
0.58
Spring
135
1.53
2.94
1.41
92.1
2.08
96.8
0.76
Summer
145
3.51
2.44
-1.07
-30.5
4.21
52.5
0.64
Fall
158
2.90
4.46
1.56
53.7
3.07
73.0
0.57
Annual
581
2.59
3.55
0.96
36.9
3.36
75.0
0.41
IMPROVE
Winter
440
0.43
0.95
0.51
118.0
1.29
147.2
0.65
Spring
473
0.54
0.85
0.30
55.9
0.81
79.8
0.57
Summer
475
3.95
2.59
-1.37
-34.6
6.64
68.1
0.59
Fall
457
1.93
2.55
0.63
32.5
6.44
91.6
0.69
Annual
1845
1.74
1.74
0.00
0.2
4.71
80.2
0.57
West CSN
Winter
265
3.61
4.32
0.72
19.8
2.96
48.9
0.48
Spring
253
1.63
2.28
0.65
39.6
1.64
52.3
0.57
Summer
273
3.44
3.37
-0.07
-2.0
3.28
42.9
0.63
Fall
246
3.77
4.00
0.23
6.2
3.35
41.3
0.85
Annual
1037
3.12
3.50
0.38
12.1
2.90
45.4
0.67
IMPROVE
Winter
539
0.70
0.95
0.25
35.9
1.11
70.3
0.68
Spring
555
0.67
0.79
0.12
17.8
0.47
49.6
0.68
Summer
563
4.28
3.84
-0.44
-10.2
18.69
70.9
0.26
Fall
555
2.26
2.08
-0.17
-7.6
8.93
73.7
0.26
Annual
2212
1.99
1.93
-0.06
-3.2
10.45
69.8
0.24
118
-------
2A.2 Projecting PM2.5 DVs to 2032
PM2.5 DVs were projected to 2032 using air quality modeling to inform estimates of
the emission reductions needed to meet standards beyond the reductions expected to
occur due to finalized rules. The projections were performed by pairing the 2018 CMAQ
simulation with a corresponding CMAQ simulation based on emissions representative of
2032. The 2032 emissions case accounts for factors including emission reductions between
2018 and 2032 from 'on-the-books' rules and has been described in detail previously
(USEPA, 2023a). Other than differences in the emissions inputs, all aspects of the 2032
CMAQ modeling were specified identical to the 2018 modeling. These aspects include the
meteorology, boundary conditions, the 12-km modeling domain, and the model
configuration.
To predict the influence of the emission reductions between 2018 and 2032 on
PM2.5 DVs, PM2.5 relative response factors (RRFs) were calculated using the CMAQ results to
project monitoring data to 2032. RRFs are the ratios of modeled PM2.5 species
concentrations in the future year (2032) to the base year (2018). RRFs are used in
projecting air quality to help mitigate the influence of systematic biases in model
predictions (e.g., systematic biases in the 2018 and 2032 modeling may partially cancel in
the ratio) (Cohan and Chen, 2014; NRC, 2004; USEPA, 2018). RRFs are calculated for each
PM2.5 component (i.e., sulfate, nitrate, organic carbon, elemental carbon, crustal material,
and ammonium). The annual and 24-hour PM2.5 DVs for the future year are calculated by
applying the species-specific RRFs to ambient PM2.5 concentrations from the PM2.5
monitoring network, which are disaggregated into species concentrations by applying the
SANDWICH method (Frank, 2006) and through interpolation of PM2.5 species data from the
CSN and IMPROVE monitoring networks. Details on the PM2.5 projection method using
RRFs are provided in the user's guide for the predecessor to the SMAT-CE software (Abt,
2014). The RRF method for calculating future-year PM2.5 annual and 24-hour PM2.5 DVs
was implemented here using the Software for Modeled Attainment Test-Community
Edition (SMAT-CE) (USEPA, 2018; Wangetal., 2015).
119
-------
2A.2.1 Monitoring Data for PM2.5 Projections
PM2.5 DVs were projected using ambient PM2.5 measurements from the 2016-2020
period centered on the 2018 CMAQ modeling period. PM2.5 species measurements from the
IMPROVE and CSN networks during 2017-2019 were used to disaggregate the measured
total PM2.5 concentrations into components for the RRF calculations. As in the 2012 PM2.5
NAAQS RIA (USEPA, 2012a), limited exclusion of wildfire and fireworks influence on PM2.5
concentrations was applied to the 2016-2020 PM2.5 monitoring data in addition to
exclusion of EPA-concurred exceptional events. Monitoring data were evaluated (i.e.,
screened) for potential wildfire and fireworks influence because PM2.5 concentrations may
be influenced by atypical, extreme, or unrepresentative events such as wildfires or
fireworks that may be appropriate for exclusion as described in EPA's memorandum
Additional Methods, Determinations, and Analyses to Modify Air Quality Data Beyond
Exceptional Events (USEPA, 2019a). Due to the challenges in identifying wildfire influence
on monitored concentrations, only limited screening of extreme values is possible here, and
wildfire impacts on concentrations likely persists in the screened data.
The steps in implementing the limited screening of major wildfire and fireworks
influence on PM2.5 concentrations are as follows:
1. An extreme value cutoff of 64 [ig nr3 was identified based the 99.9th
percentile value from all daily PM2.5 concentrations across all sites in the
long-term AQS observation record (2002-2020).
2. Specific months were screened for instances of monitors exceeding the
extreme value cutoff. Months included were June-October (while November
can be a high fire month for parts of the western U.S., it becomes more
difficult to distinguish wildfire PM2.5 from residential wood smoke and other
anthropogenic sources during the late fall).
3. The presence of visible wildfire smoke was corroborated using satellite
imagery from NASA's Worldview platform
(https://worldview.earthdata.nasa.gov) for the time periods and geographic
locations identified through steps 1 and 2. Timeseries for individual sites
120
-------
flagged were also examined to confirm PM2.5 enhancements temporally
consistent with the wildfire events identified (Figures 2A-16 to 2A-25).
4. For extreme wildfire smoke periods identified through steps 1-3 above, all
concentrations above the extreme value cutoff of 64 [ig nr3 at impacted sites
were removed.
5. In addition to the evaluation criteria above, data corresponding to the Creek
Fire (eastern CA during September-November 2020), the Camp Fire
(northern CA during November 2018), and the Appalachian Fires (NC, TN, GA
during November 2016) were evaluated for exclusion if concentrations
exceeded the extreme value threshold of 64 |Lxg nr3. These large fire episodes
show obvious impacts across multiple monitors and were clearly
documented with satellite imagery (Figures 2A-6 to and 2A-15).
6. In addition to the limited exclusion of major wildfire influence, data were
evaluated to identify days for potential exclusion due to the influence of
isolated fireworks events on PM2.5 concentrations. The 99.9th percentile value
of 64 [ig nr3 was applied as the cutoff across all sites for New Year's Eve and
the Fourth of July.
The list of counties that were evaluated for fire impacts for the November episodes
are shown in Table 2A-7. Example satellite imagery and timeseries of PM2.5 at impacted
monitors for episodes are shown in Figures 2A-6 to 2A-15. The percentage of days
excluded from the 2016-2020 dataset was 0.9% at affected sites; the total percentage of
days excluded overall from the dataset was 0.09%.
121
-------
Table 2A-7 November Wildfire Episodes and Counties Where Data Were Excluded
if PM2.5 Concentrations Exceeded the Extreme Value Threshold of 64
Hg m3
Episode
Dates
Impacted County
State
Creek Fire
Nov. 1-10, 2020
Mono
CA
Camp Fire
Nov. 8-20, 2018
Alameda
CA
Stanislaus
CA
San Joaquin
CA
Sonoma
CA
Butte
CA
Contra Costa
CA
Colusa
CA
Fresno
CA
Mendocino
CA
Sacramento
CA
Napa
CA
Solano
CA
Placer
CA
San Francisco
CA
Marin
CA
Yolo
CA
T ehama
CA
Lake
CA
Santa Clara
CA
Santa Cruz
CA
Nevada
CA
Kings
CA
Merced
CA
San Mateo
CA
Madera
CA
Monterey
CA
Appalachian Fires
Nov. 7-24, 2016
Hamilton
TN
Knox
TN
Loudon
TN
Roane
TN
Blount
TN
Swain
NC
Mitchell
NC
Buncombe
NC
Jackson
NC
Walker
GA
Clarke
GA
Richmond
GA
Hall
GA
122
-------
Episode
Dates
Impacted County State
Greenville SC
Richland SC
Edgefield SC
Lexington SC
Charleston SC
Figure 2A-6 Visible Satellite Imagery from NASA's Worldview Platform Showing
Smoke from the Camp Fire on 11/10/2018
123
-------
Figure 2A-7 Visible Satellite Imagery from NASA's Worldview Platform Showing
Smoke from the North Bay/Wine Country Fires on 10/09/2017
Figure 2A-8 Visible Satellite Imagery from NASA's Worldview Platform Showing
Smoke from Fires Across the Pacific Northwest/Northern California on
08/29/2017
124
-------
Figure 2A-9 Visible Satellite Imagery from NASA's Worldview Platform Showing
Smoke from Fires in Washington and Oregon on 08/09/2018
Figure 2A-10 Visible Satellite Imagery from NASA's Worldview Platform Showing
Smoke from Fires in Montana on 08/19/2018
125
-------
Figure 2A-11 Visible Satellite Imagery from NASA's Worldview Platform Showing
Smoke from Several Fires in eastern California including the Empire
Fire on 9/1/2017
Figure 2A-12 Visible Satellite Imagery from NASA's Worldview Platform Showing
Smoke from the 416/Burro Complex Fires on 06/10/2018
126
-------
Figure 2A-13 Visible Satellite Imagery from NASA's Worldview Platform Showing
Smoke from the Creek Fire on 10/26/2020
Figure 2A-14 Visible Satellite Imagery from NASA's Worldview Platform Showing
Smoke from the Carr/Mendocino/Ferguson Fires on 08/04/2018
127
-------
Figure 2A-15 Visible Satellite Imagery from NASA's Worldview Platform Showing
Smoke from Fires in the Appalachians on 11/10/2016
100 200
100 200
100 200
Figure 2A-16 Daily PM2.5 (in ^g m3) from a Subset of Monitors Impacted by the
Camp Fire in November 2018
Note: Bottom axis shows day in 2018. Red line indicates extreme value threshold of 64 |ig m 3 used for
screening.
128
-------
o
o —
0 100 200 300 0 100 200 300 0 100 200 300
Figure 2A-17 Daily PM2.5 (in [xg m3) from a Subset of Monitors Impacted by the
North Bay/Wine Country Fires in October 2017
Note: Bottom axis shows day in 2017. Red line indicates extreme value threshold of 64 |ig m 1 used for
screening.
O O O QQ-, cfftxC&PcP
129
-------
Missoula
fr mi ifa-ifrtftffi
Jackson
O
O
O
O
O
<§>
O
0 °°o
s%8&
Spokane
WW
Figure 2A-18 Daily PM2.5 (in ng m3) from a Subset of Monitors Impacted by Fires in
the Pacific Northwest/Northern California in August-September 2017
Note: Bottom axis shows day in 2017. Red line indicates extreme value threshold of 64 jig m 3 used for
screening.
130
-------
Spokane
1
—I—
300
King
Okanogan ,
§ -
1 1
0 100
200
I
300
Snohomish
I I
0 100
200
i
300
Kitsap
I I
0 100
200
i
300
King
§ "
i r
0 100
§ -
Skagit
<
I I i
0 100 200
I
300
Snohomish
<
J
1
0 an n- **
I I I
0 100 200
I
300
King
I
Mm
I I I
0 100 200
I
300
King
~r
100
T
200
Figure 2A-19 Daily PM2.5 (in ng m3) from a Subset of Monitors Impacted by Fires in
Washington and Oregon in July-August 2018
Note: Bottom axis shows day in 2018. Red line indicates extreme value threshold of 64 jig nr3 used for
screening.
131
-------
Lincoln
0 100
200
300
Lewis and Clark
nfftt §£ti iffi Tmi
0 100
200
300
Fergus
Figure 2A-20 Daily PM2.5 (in ng m3) from the Monitors Impacted by Fires and Smoke
in Montana in August 2018
Note: Bottom axis shows day in 2018. Red line indicates extreme value threshold of 64 jig m 3 used for
screening.
132
-------
Figure 2A-21 Daily PM2.5 (in ng m3) from a Subset of Monitors Impacted by Fires in
Montana, Washington and Idaho in August 2015
Note: Bottom axis shows day in 2015. Red line indicates extreme value threshold of 64 jig m3 used for
screening.
0
LO -
La Plata
<
3
)
O
O -
'1
O
LO
<
A Jk
O -
oOooooooof
0 100 200 300
Figure 2A-22 Daily PM2.5 (in ng m3) from the Monitor in Plata, CO Impacted by the
416/Burro Fire Complex in June 2018
Note: Bottom axis shows day in 2018. Red line indicates extreme value threshold of 64 jig m3 used for
screening.
133
-------
Figure 2A-23 Daily PM2.5 (in ng m3) from the Two monitors Impacted by the Butte
Fire in September 2015
Note: Bottom axis shows day in 2015. Red line indicates extreme value threshold of 64 jig m3 used for
screening.
Figure 2A-24 Daily PM2.5 (in ng m3) from a Subset of Monitors Impacted by the
Carr/Mendocino/Ferguson Fires in August 2018
Note: Bottom axis shows day in 2018. Red line indicates extreme value threshold of 64 jig m3 used for
screening.
134
-------
0 100 200 300 0 100 200 300
Figure 2A-25 Daily PM2.5 (in ng m3) from a Subset of Monitors Impacted by Fires in
the Appalachians in November 2016
Note: Bottom axis shows day in 2016. Red line indicates extreme value threshold of 64 jig nr3 used for
screening.
2A.2.2 Future-Year PM2.5 Design Values
PM2.5 DVs were projected to 2032 using air quality modeling as described above and
compared with the existing standard combination, 12/35. Counties with projected 2032
PM2.5 DVs exceeding the existing standards are shown in Figure 2A-26. Counties that
exceed only the 24-hour standard are in northern California, Oregon, Washington, Idaho,
and Montana. Elevated PM2.5 episodically occurs in winter in these areas due to
meteorological temperature inversions that concentrate PM2.5 in shallow layers, especially
in mountainous terrain. In California, multiple counties exceed both the annual and 24-
135
-------
hour standards and two counties (San Bernardino and Imperial) exceed only the annual
standard.
12/35
Figure 2A-26 Counties with Projected 2032 PM2.5 DVs that Exceed the 24-Hour
(Daily Only), Annual (Annual Only) or Both the 24-Hour and Annual
(Both) Standards for the Combination of Existing Standards (12/35)
40
cs
-1 35
Both
Annual Only
Daily Only
-120 -110 -100 -90 -80 -70
Longitude
As described below in section 2A.3.4, PM2.5 DVs for 2032 were adjusted to
correspond with just meeting the existing standard level to form the 12/35 analytical
baseline used in estimating the incremental costs and benefits of meeting the revised and
alternative standards relative to the existing standards. The county exceedances of the
alternative standards in the 12/35 analytical baseline are shown in Figure 2A-27. Since the
PM2.5 DVs have been adjusted to meet the 24-hour standard level of 35 |j,g m-3 in the
analytical baseline, there are no exceedances of the 24-hour standard for the cases of
10/35, 9/35, and 8/35. For the 10/35 case, five counties in the east, two in the northwest,
and thirteen in California have annual PM2.5 DVs greater than 10 |j,g m-3 in the 12/35
analytical baseline. For the 10/30 case, twenty-nine counties have 24-hr DVs greater than
30 |j,g m-3 and annual DVs less than 10 |j,g m-3, and nine counties exceed both the 24-hr and
annual standards. For the 9/35 case, nineteen counties exceed the annual standard in the
eastern US, compared with five for the 10/35 and 10/30 cases. The total number of
counties exceeding the standards increases from 52 to 117 when moving from 9/35 to
8/35. In Table 2A-8, PM2.5 DVs are shown for the 2032 projections and 12/35 analytical
136
-------
baseline for sites with the highest annual and 24-hour PM2.5 DVs in counties with 2032 DVs
that exceed an annual standard 8 j.xg nr3 or a 24-hour standard of 30 (ig nr3.
10/35
10/30
30°N -Annual Only: 20
24-hr Only: 0
25°N -Tota': 20
30°N -Annual Only: 52
24-hr Only: 0
25°N -To,al: P
9/35
8/35
Both
Annual Only
24-hr Only
120°W 110°W 100°W 90°W 80°W 70°W 120°W 110°W 100°W 90°W 80°W 70°W
Figure 2A-27 Counties with PM2.5 DVs in the 12/35 Analytical Baseline that Exceed
the 24-Hour (24-hr Only), Annual (Annual Only) or Both the 24-Hour
and Annual (Both) Standards for Combinations of Alternative
Standards
137
-------
Table 2A-8 PM2.5 DVs for 2032 Projection and 12/35 Analytical Baseline for the
Highest DVs in the County for Counties with Annual 2032 DVs Greater
8 |ig m3 or 24-hour 2032 DVs Greater than 30 (j,g m3
State
County
Annual
2032 DV
fug m-')
24-hour
2032 DV
fug m-')
Annual
12/35 DV
ftigm-3)
24-hour
12/35 DV
ftigm-3)
AL
Jefferson
9.00
19.4
9.00
19.4
AL
Russell
8.44
17.1
8.44
17.1
AZ
Maricopa
9.59
29.2
9.59
29.2
AZ
Pinal3
7.95
34.6
7.95
34.6
AZ
Santa Cruz
8.71
24.9
8.71
24.9
AZ
Yuma
8.30
20.8
8.30
20.8
AR
Pulaski
8.92
21.3
8.92
21.3
CA
Alameda
10.37
27.6
10.37
27.6
CA
Butte
8.82
33.4
8.82
33.4
CA
Calaveras
9.60
30.0
9.60
30.0
CA
Colusa
8.24
37.4
7.61
35.4
CA
Contra Costa
9.44
25.2
9.44
25.2
CA
Fresno
13.66
48.2
12.04
33.8
CA
Imperial
12.36
33.5
12.04
32.8
CA
Kern
15.34
53.0
12.04
31.2
CA
Kings
14.51
44.0
11.64
25.5
CA
Los Angeles
12.87
37.7
12.04
33.5
CA
Madera
11.47
37.3
9.96
28.7
CA
Mendocino
8.07
32.7
8.07
32.7
CA
Merced
11.85
36.7
10.77
29.6
CA
Mono
9.63
38.7
8.59
35.4
CA
Napa
8.90
25.6
8.90
25.6
CA
Orange
11.07
32.7
10.74
29.6
CA
Placer
8.07
29.9
8.07
29.9
CA
Plumas
14.02
44.9
11.01
35.4
CA
Riverside
13.79
36.9
12.04
30.7
CA
Sacramento
9.55
32.4
9.55
32.4
CA
San Bernardino
14.33
35.0
12.04
27.6
CA
San Diego
9.32
24.8
9.32
24.8
CA
San Francisco
8.62
24.3
8.62
24.3
CA
San Joaquin
11.91
37.2
10.00
30.7
CA
San Luis Obispo
8.08
24.8
8.08
24.8
CA
San Mateo
8.40
24.1
8.40
24.1
CA
Santa Barbara
8.12
33.5
8.12
33.5
CA
Santa Clara
9.79
28.3
9.79
28.3
CA
Shasta
7.86
30.7
7.86
30.7
CA
Siskiyou
8.25
37.3
7.65
35.4
CA
Solano
9.55
23.7
9.55
23.7
CA
Stanislaus
12.32
38.5
10.99
29.7
138
-------
State
County
Annual
2032 DV
fag m-3}
24-hour
2032 DV
fag m-3}
Annual
12/35 DV
fugm-3!
24-hour
12/35 DV
fugm-3!
CA
Sutter
9.41
32.3
9.41
32.3
CA
Tehama
7.51
33.3
7.51
33.3
CA
Tulare
14.43
44.0
12.04
23.7
CA
Ventura
9.09
34.8
9.09
34.8
CO
Adams
9.13
24.2
9.13
24.2
CO
Denver
9.06
25.6
9.06
25.6
CO
Weld
8.76
24.9
8.76
24.9
DC
District of Columbia
8.37
20.1
8.37
20.1
FL
Broward
9.06
19.1
9.06
19.1
GA
Bibb
8.18
16.5
8.18
16.5
GA
Chatham
8.14
18.2
8.14
18.2
GA
Dougherty
8.29
21.5
8.29
21.5
GA
Fulton
8.86
18.5
8.86
18.5
GA
Gwinnett
8.78
21.2
8.78
21.2
GA
Muscogee
8.29
27.6
8.29
27.6
GA
Richmond
9.10
20.4
9.10
20.4
ID
Benewah
-
34.1
-
34.1
ID
Canyon
8.66
32.7
8.66
32.7
ID
Lemhi
9.85
36.5
9.58
35.4
ID
Shoshone
10.56
35.3
10.56
35.3
IL
Cook
9.77
22.5
9.77
22.5
IL
DuPage
8.05
18.7
8.05
18.7
IL
McLean
8.08
17.1
8.08
17.1
IL
Macon
8.34
17.8
8.34
17.8
IL
Madison
9.15
19.1
9.15
19.1
IL
Saint Clair
8.44
18.2
8.44
18.2
IN
Lake
9.18
21.9
9.18
21.9
IN
Marion
10.00
23.1
10.00
23.1
KS
Shawnee
8.38
20.5
8.38
20.5
KS
Wyandotte
8.41
22.7
8.41
22.7
KY
Jefferson
8.68
20.8
8.68
20.8
LA
Caddo
9.53
20.5
9.53
20.5
LA
East Baton Rouge
8.44
21.1
8.44
21.1
LA
West Baton Rouge
8.58
19.7
8.58
19.7
MI
Wayne
10.34
26.3
10.34
26.3
MS
Hinds
8.42
17.6
8.42
17.6
MT
Flathead
7.72
31.7
7.72
31.7
MT
Lewis and Clark
8.79
36.5
8.52
35.4
MT
Lincoln
11.63
31.6
11.63
31.6
MT
Missoula
9.09
28.2
9.09
28.2
NV
Clark
9.04
26.3
9.04
26.3
NJ
Bergen
9.78
21.6
9.78
21.6
139
-------
State
County
Annual
24-hour
Annual
24-hour
2032 DV
2032 DV
12/35 DV
12/35 DV
fag m-3}
fag m-3}
fugm-3!
fugm-3!
NJ
Camden
9.27
22.2
9.27
22.2
NJ
Union
8.35
20.5
8.35
20.5
NY
New York
8.94
22.4
8.94
22.4
NC
Forsyth
8.07
22.1
8.07
22.1
NC
Mecklenburg
8.35
18.1
8.35
18.1
OH
Butler
9.42
21.1
9.42
21.1
OH
Cuyahoga
9.87
21.6
9.87
21.6
OH
Franklin
8.11
20.2
8.11
20.2
OH
Hamilton
10.44
21.8
10.44
21.8
OH
Jefferson
8.06
19.4
8.06
19.4
OH
Stark
8.29
18.8
8.29
18.8
OK
Cleveland
8.63
19.1
8.63
19.1
OK
Oklahoma
8.71
19.2
8.71
19.2
OK
Tulsa
8.66
21.9
8.66
21.9
OR
Crook
8.23
32.8
8.23
32.8
OR
Harney
9.47
31.4
9.47
31.4
OR
Jackson
9.55
30.5
9.55
30.5
OR
Josephine
8.62
26.0
8.62
26.0
OR
Klamath
10.69
40.5
9.44
35.4
OR
Lake
8.68
38.1
8.02
35.4
OR
Lane
8.48
32.1
8.48
32.1
PA
Allegheny
10.71
32.7
10.71
32.7
PA
Beaver
8.20
19.2
8.20
19.2
PA
Cambria
8.46
20.7
8.46
20.7
PA
Chester
8.69
21.8
8.69
21.8
PA
Delaware
10.08
24.9
10.08
24.9
PA
Lancaster
8.71
24.6
8.71
24.6
PA
Lebanon
8.27
24.9
8.27
24.9
PA
Philadelphia
8.74
21.4
8.74
21.4
PA
York
8.42
19.6
8.42
19.6
RI
Providence
8.12
17.6
8.12
17.6
TN
Davidson
8.46
17.2
8.46
17.2
TX
Bowie
8.27
17.6
8.27
17.6
TX
Cameron
9.90
25.7
9.90
25.7
TX
Dallas
8.24
19.7
8.24
19.7
TX
El Paso
8.84
23.5
8.84
23.5
TX
Harris
9.79
23.5
9.79
23.5
TX
Hidalgo
10.69
28.4
10.69
28.4
TX
Jefferson
8.46
21.4
8.46
21.4
TX
Nueces
8.65
24.5
8.65
24.5
TX
Orange
8.47
20.5
8.47
20.5
TX
Travis
9.22
21.9
9.22
21.9
140
-------
State
County
Annual
24-hour
Annual
24-hour
2032 DV
2032 DV
12/35 DV
12/35 DV
fag nr3}
fag nr3}
fugm-3}
fugm-3}
UT
Box Elder
6.65
31.7
6.65
31.7
UT
Cache
6.20
31.4
6.20
31.4
UT
Salt Lake
7.64
30.9
7.64
30.9
WA
King
8.13
25.6
8.13
25.6
WA
Okanogan
8.27
32.6
8.27
32.6
WA
Yakima
8.28
38.4
7.54
35.4
aThe Hidden Valley site in Pinal County (04-021-3015) was not compared with the annual NAAQS in this
analysis because it is a replacement site for the Cowtown Road site that was not comparable to the annual
NAAQS (PCAQCD, 2020).
2A.3 Developing Air Quality Ratios and Estimating Emission Reductions
As in the RIAs for the 2012 PM2.5 NAAQS review (USEPA, 2012a, 2012b), air quality
ratios are used here to estimate the emission reductions beyond the 2032 modeling case
that are needed to meet the existing, revised, and alternative standards. Air quality ratios
are developed from sensitivity modeling with CMAQ and relate a change in PM2.5 DV to a
change in emissions. Air quality ratios have units of |Lxg nr3 per kton of emissions. The
remainder of this section describes the development of air quality ratios and their
application to estimating emission reductions for meeting the existing, revised, and
alternative standards.
2A.3.1 Developing Air Quality Ratios for Primary PM2.5 Emissions
To develop air quality ratios that relate the change in DV in a county to the change in
primary PM2.5 emissions in that county, CMAQ sensitivity modeling was performed with
reductions in primary PM2.5 emissions in selected counties. The modeling was conducted
using CMAQ version 5.2.1 for a 2028 modeling case similar to that of recent regional haze
modeling (USEPA, 2019b) due to the availability of the 2028 (but not 2032) modeling
platform at the time of the work.
To develop air quality ratios for primary PM2.5 emissions, a 2028 CMAQ sensitivity
simulation was conducted with 50% reductions in primary PM2.5 emissions from
anthropogenic sources in counties with annual 2028 DVs greater than 8 |Lxg nr3 (Figure 2A-
28). The change in annual and 24-hour PM2.5 DVs in these counties was then divided by the
141
-------
change in emissions in the respective counties to determine the air quality ratio at
individual monitors as follows;
AQrutio^j^j^^x 1000 (2A-1)
where ADVis the change in design value (jj,g nr3) between the 2028 base case and the
simulation with 50% reduction in primary PM2.5 emissions at a monitor z in a county j,
AEmissCty is the change in primary PM2.5 emissions (tons) in county/between the 2028
base case and the simulation with 50% reduction in primary PM2.5 emissions, and the
factor of 1000 converts units from (|ig nr3 per ton) to (|„ig nr3 per kton).
50
45
(D
~o 40
3
03
_l
35
30
25
Figure 2A-28 Counties with 50% Reduction in Anthropogenic Primary PM2.5
Emissions in 2028 Sensitivity Modeling
Representative air quality ratios for regions of the US were developed from the
ratios at individual monitors as in the 2012 PM2.5 NAAQS review (USEPA, 2012b), Regional
ratios were calculated as the 75th percentile of air quality ratios at monitors within five
regions: Northeast, Southeast, Northern California, Southern California, and West (Figure
2A-29). The Northeast region was defined by combining the Upper Midwest, Ohio Valley,
and Northeast US climate regions (Figure 2A-2); the Southeast region was defined by
combining the Southeast and South climate regions (Figure 2A-2); and California was
separated into Southern and Northern regions as done previously (USEPA, 2012b). The air
1 I
1 1 »
* ¦
* „
>*
-120 -110 -100 -90 -80 -70
Longitude
142
-------
quality ratios for primary PM2.5 emissions used in estimating the emission reductions
needed to just meet standards are listed in Table 2A-9.
45°N
40°N
35°N
30°N
25°N
Figure 2A-29 Regional Groupings for Calculating Air Quality Ratios
Table 2A-9 Annual and 24-Hour Air Quality Ratios for Primary PM2.5 Emissions
Region Annual Air Quality Ratio 24-hour Air Quality Ratio
flig m ! per kton) (ug m 3 per kton)
Northeast 1.37 4.33
Southeast 1.22 3.51
West 2.14 8.70
Northern California 3.15 9.97
Southern California 1.18 2.56
The air quality ratios in Table 2A-9 relate the change in DV in a given county to a
change in emissions in that county. The ratios are developed for local spatial scales because
concentrations are most responsive to changes in local emissions. However, emission
controls may not always be identified in the local county, and emission reductions in
neighboring counties may sometimes be appropriate, such as in the eastern US where
counties are relatively small, and terrain is relatively flat. To apply emission reductions in
the neighboring counties in the eastern US, the responsiveness of annual PM2.5 DVs to
emission reductions within a county was compared with the responsiveness for
neighboring counties as estimated from the 2028 sensitivity modeling.
143
-------
First, county groups of most relevance were identified from the 2028 sensitivity
modeling. These groups were selected as eastern counties where emission reductions were
applied and whose neighbors were not also neighbors of another county where emission
reductions were applied. This set of county groups was then subset from the full list of
counties and filtered to ensure that at least one monitor was included in the neighbor
counties for the county group. The resulting county groups are shown in Figure 2A-30. The
average relative responsiveness of annual DVs in the east for emission reductions in a core
county to reductions in a neighboring county was then calculated as follows:
mean(ADVcore)
ImpactRatio =
= 4
(2A-2)
mean(ADVneighb
or)
where the numerator is the average impact on annual PM2.5 DVs in the core counties with
50% reduction in anthropogenic primary PM2.5 emissions, and the denominator is the
average impact on annual PM2.5 DVs in neighboring counties. The resulting impact ratio
suggests that primary PM2.5 emission reductions in neighboring counties would be 4x less
effective as in the core county.
50-
45-
0
40-1
cc
(
"*s f*l~r
M t*
- ft
A
-j
35-
30-
rf
a
Core
| Neighbor
X
—A
25-
-120
-110
-100
-90
Longitude
-80
-70
Figure 2A-30 Counties Used in Estimating the Relative Impact of Emissions in Core
and Neighboring Counties
144
-------
2A.3.2 Developing Air Quality Ratios for NOx in Southern California
As described above, PM2.5 DVs exceeded the existing standards at monitors in the
South Coast Air Basin in the 2032 modeling case. PM2.5 DVs were adjusted to meet the
existing standards in these counties in creating the 12/35 analytical baseline. Since
concentrations of ammonium nitrate are elevated in South Coast, NOx emission reductions
were applied in these counties in addition to primary PM2.5 emission reductions to meet
12/35. For this purpose, air quality ratios were developed that relate a change in PM2.5 DVs
to a change in NOx emissions at monitors in Southern California.
The air quality ratios were developed from a CMAQ sensitivity simulation with 50%
reductions in anthropogenic NOx emissions relative to the 2028 modeling case. The 50%
emission reductions were applied in counties with annual 2028 DVs greater than 8 |Lxg nr3
and their neighboring counties (Figure 2A-31). The change in annual and 24-hour DVs in
these counties was then divided by the change in emissions in the respective county groups
to determine the air quality ratio at individual monitors as follows:
AQratioPM2 5 / / = —-1 x 1000 (2A-3)
'AEmissCtyGroupj
where ADVis the change in design value (|j,g nr3) between the 2028 base case and the
simulation with 50% reduction in NOx emissions at monitor /, AEmissCtyGroup is the
change in NOx emissions (ton) in the county group associated with county j between the
2028 base case and the simulation with 50% reduction in NOx emissions, and the factor of
1000 converts units from (|Lxg nr3 per ton) to (|Lxg nr3 per kton).
145
-------
Figure 2A-31 Counties with 50% Reduction in Anthropogenic NOx Emissions in
2028 Sensitivity Modeling
-100 -90
Longitude
The county groups for determining the emission change to associate with the DV
change in Equation 2A-3 are defined in Table 2A-10. These county groups were identified
by first selecting the county of focus plus the neighboring counties to reflect the regional
nature of ammonium nitrate formation from NOx emissions. These county groups were
then refined to account for the influence of terrain, which limits air mixing between
different air basins, on meteorology and air pollution. For instance, although Kern County
neighbors Los Angeles County, Kern is part of the SJV air basin while Los Angeles is part of
the South Coast Air Basin. Kern is therefore not included in the county group associated
with Los Angeles County, because Kern is separated from Los Angeles County by mountain
ranges.
Table 2A-10 County Groups for Calculating Air Quality Ratios for NOx Emission
Changes in Southern California
FIPS
County
County Group
06025
Imperial
Imperial; San Diego
06037
Los Angeles
Los Angeles; Orange; San Bernardino; Ventura
06065
Riverside
Riverside; Orange; San Bernardino
06071
San Bernardino
San Bernardino; Los Angeles; Orange; Riverside
06073
San Diego
San Diego; Imperial; Orange
06111
Ventura
Ventura; Los Angeles; Santa Barbara
146
-------
To develop representative air quality ratios (USEPA, 2012b) for Southern California,
the 75th percentile of the air quality ratios for individual monitors in the six counties in
Table 2A-10 was calculated. The resulting air quality ratio for the annual standard is 0.004
Hg nr3 per kton and for the 24-hour standard ratio is 0.038 |Lxg nr3 per kton. These ratios
were applied to adjust 2032 PM2.5 DVs according to 75% reductions in anthropogenic NOx
emissions for counties in South Coast Air Basin (i.e., LA, San Bernardino, Riverside, and
Orange). The 75% reduction in emissions corresponded to 81,200 tons. The 2032 DVs and
the NOx-adjusted DVs are shown in Table 2A-11 for the highest annual and 24-hour DV
monitors in the county. Note that these emission reductions were applied in meeting the
existing standards (12/35) and are therefore not part of the incremental cost and benefits
of meeting revised and alternative standards relative to the existing standards.
Table 2A-11 2032 PM2.5 DVs and NOx-adjusted PM2.5 DVs for the Highest Annual
and 24-Hour DV Monitors in South Coast Counties
Site ID
County
AQ Ratio
Annual
Qig nr3 per kton)
AQ Ratio
24-hour
(|Jg [11 ; per kton)
2032 DV
Annual
(ugm')
2032 DV
2 4-hour
(M-gm3)
NOx-Adj DV
Annual
(ugm')
NOx-Adj
DV
2 4-hour
(Mfim3)
060371302
Los Angeles
0.004
0.038
12.87
37.7
12.54
34.6
060590007
Orange
0.004
0.038
11.07
32.7
10.74
29.6
060658005
Riverside
0.004
0.038
13.79
36.9
13.46
33.8
060710027
San Bernardino
0.004
0.038
14.33
35.0
14.00
31.9
2A.3.3 Developing Air Quality Ratios for NOx in SJV, CA
As in the South Coast Air Basin, PM2.5 DVs exceed existing standards in SJV in the
2032 modeling case, and concentrations of ammonium nitrate are elevated in SJV. To
develop PM2.5 DVs for SJV counties in the 12/35 analytical baseline, NOx emission
reductions were applied in addition to primary PM2.5 emission reductions. To develop air
quality ratios for NOx emission changes in SJV, information was used from Appendix K of
the 2018 SJV PM2.5 Plan (SJVAPCD, 2018). The Plan was based on fine-scale CMAQ modeling
and provides useful information for characterizing the responsiveness of PM2.5 DVs to NOx
emissions.
The California Air Resources Board (CARB) modeled PM2.5 concentrations in SJV
corresponding to 30% reductions in NOx emissions relative to a 2024 base case. The
147
-------
change in annual and 24-hour PM2.5 DVs at monitors in SJV was reported for the sensitivity
simulation. Using this information, along with PM2.5 DVs and emissions information from
the 2032 CMAQ modeling developed by EPA, air quality ratios were calculated at monitors
in SJV from the following equation:
The parenthesis labeled CARB indicates values from the SJV Plan, and the
parenthesis labeled 2032 indicates values from the 2032 CMAQ modeling. %Chg.DVi
indicates the percent change in the design value for a given monitor, i, from Table 49 and
50 of Appendix K of the SJV Plan (SJVAPCD, 2018). %Chg.DVt ranged from 2.3% to 7.5% for
annual DVs and from 7.0% to 16.3% for 24-hour DVs. %Chg.Emissionsjv indicates the
percent change in NOx emissions in SJV and equaled 30%. DVt corresponds to the design
value at monitor i, and Emissionsjv corresponds to the anthropogenic NOx emissions in SJV
in the 2032 CMAQ modeling. Equation 2A-4 normalizes the percent changes from CARB's
2024 modeling to the PM2.5 DVs and emissions from the 2032 case for application here.
Air quality ratios were calculated as above for all monitors in SJV, except for the
Tranquillity monitor. The Tranquillity monitor is in the Western part of Fresno County,
away from the urban exceedance monitors, and has a low PM2.5 concentration (e.g., 2024
annual DV for CARB modeling is 5.6 |Lxg nr3). To develop representative air quality ratios for
counties in SJV, the 75th percentile of air quality ratios over monitors in the SJV counties
was calculated. These ratios were applied to adjust 2032 PM2.5 DVs according to 75%
reductions in anthropogenic NOx emissions for counties in SJV. The 75% reduction in
emissions corresponded to 39,700 tons. The 2032 DVs and the NOx-adjusted DVs are
shown in Table 2A-12 for the highest annual and 24-hour DV monitors in the county. Note
that these emission reductions were applied in meeting the existing standards (12/35) and
are therefore not part of the incremental cost and benefits of meeting revised and
alternative standards relative to the existing standards.
(2A-4)
148
-------
Table 2A-12 2032 PM2.5 DVs and NOx-adjusted PM2.5 DVs for the Highest Annual
and 24-Hour DV Monitors in SJV Counties
Site ID County AQ Ratio AQ Ratio 2032 DV 2032 DV NOx-Adj DV NOx-Adj DV
Annual 24-hour Annual 24-hour Annual 24-hour
Qigm 3perkton) Qg m 'per kton) fugm3) (ug 111 ') (ug 111 ') (ug 111 ')
060195025
Fresno
0.033
0.337
13.66
48.2
12.36
34.9
060290016
Kern
0.041
0.418
15.34
53.0
13.70
36.4
060311004
Kings
0.072
0.467
14.51
44.0
11.64
25.5
060392010
Madera
0.038
0.216
11.47
37.3
9.96
28.7
060470003
Merced
0.027
0.178
11.85
36.7
10.77
29.6
060771002
San Joaquin
0.048
0.164
11.91
37.2
10.00
30.7
060990006
Stanislaus
0.034
0.222
12.32
38.5
10.99
29.7
061072002
Tulare
0.047
0.472
14.43
44.0
12.56
25.3
2A.3.4 Applying Air Quality Ratios to Estimate Emission Reductions
The emissions reductions needed to just meet standards were estimated using the
primary PM2.5 air quality ratios in combination with the required incremental change in
concentration. The emission reductions required to meet the DV target for a standard were
calculated as follows:
AEmissionstd = w^e;,std-wTarflet,std x 1QQQ
AQTdtios^d
where AEmissionstd is the emission reduction required to meet an annual or 24-hour
standard; DVTarget,std is the level of the annual or 24-hour standard to be met; DVModeistd is the
modeled PM2.5 DV for the annual or 24-hour standard at the county highest monitor;
Aqratiostd is the air quality ratio for that standard; and the factor of 1000 converts units
from kton to ton.
For example, the highest 2032 annual PM2.5 DV in Kern County is 13.70 |Lxg nr3 at site
06-029-0016 after applying the 75% NOx emission reduction to the 2032 DVs. The annual
air quality ratio for primary PM2.5 emissions in Northern California is 3.15 |Lxg nr3 per kton.
Therefore, to meet an annual standard of 12 |Lxg nr3, a total of 527 tons of primary PM2.5
emissions would be needed (i.e., (13.70-12.04)/3.15 x 1000). The highest 2032 24-hour
PM2.5 DV in Kern County is 36.4 |Lxg nr3 at site 06-029-0016 after applying the 75% NOx
emission reduction to the 2032 DVs. The 24-hour air quality ratio for primary PM2.5
emissions in Northern California is 9.97 |Lxg nr3 per kton. Therefore, to meet a 24-hour
149
-------
standard of 35 |Lxg nr3, a total of 100 tons of primary PM2.5 emissions would be needed (i.e.,
(36.4-35.4)/9.97 x 1000). To determine the emission reductions needed to meet an annual
and 24-hour standard combination, the maximum needed emission reductions across
standards is calculated as follows:
A Emissionstdcombo = max(AEmissionAnnuai, AEmission24fir) (2A-6)
For the Kern County example, a total 527 tons of primary PM2.5 emission reductions are
needed to meet the 12/35 standard combination (i.e., maximum of 527 and 100).
The PM2.5 DVs associated with meeting a standard combination at the highest
monitor in a county are calculated using the required emission reductions as follows:
^^Annual,std.combo ~ ^^Innuai.tntttai — AElTlissiOTlgtd combo X AQrCLtioAnnuai (2A-8)
DVoaily,std.combo ~ ^^Innuai.tntttai — AElTlissiOTlgtd combo X AQratiOjjany (2A-9)
In the Kern County example, the adjusted annual DV for the 12/35 case is 12.04 ngrrr3
(13.70-527*3.15/1000) and the adjusted 24-hour DV is 31.1 [ig nr3 (36.4-527*9.97/1000).
2A.3.4.1 Emission Reductions Needed to Meet 12/35
In the 2032 projections, PM2.5 DVs exceeded the existing standards for some
counties in the west (Figure 2A-26). To create the PM2.5 DVs for 12/35 analytical baseline,
the reductions in primary PM2.5 emissions needed to just meet 12/35 at the highest DV
monitor by county were calculated using the air quality ratios in Table 2A-9. PM2.5 DVs
were then adjusted according to those emission reductions. In Table 2A-13, the primary
PM2.5 emission reductions needed to meet 12/35 is shown by county for counties with
annual DVs greater than 8 |Lxg nr3 or 24-hour DVs greater than 30 |Lxg nr3 (note that required
emission reductions are zero for counties with DVs below 12/35). Table 2A-13 also
includes the corresponding air quality ratios, the 2032 PM2.5 DVs (or NOx-adjusted DVs for
South Coast and SJV counties), and the PM2.5 DVs that define the 12/35 analytical baseline.
150
-------
Table 2A-13 Summary of Primary PM2.5 Emissions Reductions by County Needed to
Meet the Existing Standards (12/35) for Counties with 2032a Annual
DVs greater than 8 (j,g m-3 or 24-Hour DVs Greater than 30 |ig m3
State
County
AEmission
2032 to
12/35
(ton)
AQ Ratio
Annual
Qigm3
perkton)
AQ Ratio
2 4-hour
Qjgnv3per
kton)
2032a DV
Annual
(ugm')
2032a
DV
2 4-hour
km3)
12/35 DV
Annual
(Hgm3)
12/35 DV
2 4-hour
(ligm3)
AL
Jefferson
0
1.22
3.51
9.00
19.4
9.00
19.4
AL
Russell
0
1.22
3.51
8.44
17.1
8.44
17.1
AZ
Maricopa
0
2.14
8.70
9.59
29.2
9.59
29.2
AZ
Pinal
0
2.14
8.70
7.95
34.6
7.95
34.6
AZ
Santa Cruz
0
2.14
8.70
8.71
24.9
8.71
24.9
AZ
Yuma
0
2.14
8.70
8.30
20.8
8.30
20.8
AR
Pulaski
0
1.22
3.51
8.92
21.3
8.92
21.3
CA
Alameda
0
3.15
9.97
10.37
27.6
10.37
27.6
CA
Butte
0
3.15
9.97
8.82
33.4
8.82
33.4
CA
Calaveras
0
3.15
9.97
9.60
30.0
9.60
30.0
CA
Colusa
201
3.15
9.97
8.24
37.4
7.61
35.4
CA
Contra Costa
0
3.15
9.97
9.44
25.2
9.44
25.2
CA
Fresno
102
3.15
9.97
12.36
34.9
12.04
33.8
CA
Imperial
272
1.18
2.56
12.36
33.5
12.04
32.8
CA
Kern
525
3.15
9.97
13.70
36.4
12.04
31.2
CA
Kings
0
3.15
9.97
11.64
25.5
11.64
25.5
CA
Los Angeles
423
1.18
2.56
12.54
34.6
12.04
33.5
CA
Madera
0
3.15
9.97
9.96
28.7
9.96
28.7
CA
Mendocino
0
3.15
9.97
8.07
32.7
8.07
32.7
CA
Merced
0
3.15
9.97
10.77
29.6
10.77
29.6
CA
Mono
331
3.15
9.97
9.63
38.7
8.59
35.4
CA
Napa
0
3.15
9.97
8.90
25.6
8.90
25.6
CA
Orange
0
1.18
2.56
10.74
29.6
10.74
29.6
CA
Placer
0
3.15
9.97
8.07
29.9
8.07
29.9
CA
Plumas
953
3.15
9.97
14.02
44.9
11.01
35.4
CA
Riverside
1206
1.18
2.56
13.46
33.8
12.04
30.7
CA
Sacramento
0
3.15
9.97
9.55
32.4
9.55
32.4
CA
San Bernardino
1665
1.18
2.56
14.00
31.9
12.04
27.6
CA
San Diego
0
1.18
2.56
9.32
24.8
9.32
24.8
CA
San Francisco
0
3.15
9.97
8.62
24.3
8.62
24.3
CA
San Joaquin
0
3.15
9.97
10.00
30.7
10.00
30.7
CA
San Luis Obispo
0
3.15
9.97
8.08
24.8
8.08
24.8
CA
San Mateo
0
3.15
9.97
8.40
24.1
8.40
24.1
CA
Santa Barbara
0
1.18
2.56
8.12
33.5
8.12
33.5
CA
Santa Clara
0
3.15
9.97
9.79
28.3
9.79
28.3
CA
Shasta
0
3.15
9.97
7.86
30.7
7.86
30.7
CA
Siskiyou
191
3.15
9.97
8.25
37.3
7.65
35.4
CA
Solano
0
3.15
9.97
9.55
23.7
9.55
23.7
151
-------
State
County
AEmission
2032 to
12/35
(ton)
AQ Ratio
Annual
Cngm3
perkton)
AQ Ratio
2 4-hour
(|Agnv3per
kton)
2032a DV
Annual
(Hgm')
2032a
DV
2 4-hour
km')
12/35 DV
Annual
(M-gm3)
12/35 DV
2 4-hour
(M-gm3)
CA
Stanislaus
0
3.15
9.97
10.99
29.7
10.99
29.7
CA
Sutter
0
3.15
9.97
9.41
32.3
9.41
32.3
CA
T ehama
0
3.15
9.97
7.51
33.3
7.51
33.3
CA
Tulare
164
3.15
9.97
12.56
25.3
12.04
23.7
CA
Ventura
0
1.18
2.56
9.09
34.8
9.09
34.8
CO
Adams
0
2.14
8.70
9.13
24.2
9.13
24.2
CO
Denver
0
2.14
8.70
9.06
25.6
9.06
25.6
CO
Weld
0
2.14
8.70
8.76
24.9
8.76
24.9
DC
District of Columbia
0
1.22
3.51
8.37
20.1
8.37
20.1
FL
Broward
0
1.22
3.51
9.06
19.1
9.06
19.1
GA
Bibb
0
1.22
3.51
8.18
16.5
8.18
16.5
GA
Chatham
0
1.22
3.51
8.14
18.2
8.14
18.2
GA
Dougherty
0
1.22
3.51
8.29
21.5
8.29
21.5
GA
Fulton
0
1.22
3.51
8.86
18.5
8.86
18.5
GA
Gwinnett
0
1.22
3.51
8.78
21.2
8.78
21.2
GA
Muscogee
0
1.22
3.51
8.29
27.6
8.29
27.6
GA
Richmond
0
1.22
3.51
9.10
20.4
9.10
20.4
ID
Benewah
0
NA
8.70
NA
34.1
NA
34.1
ID
Canyon
0
2.14
8.70
8.66
32.7
8.66
32.7
ID
Lemhi
126
2.14
8.70
9.85
36.5
9.58
35.4
ID
Shoshone
0
2.14
8.70
10.56
35.3
10.56
35.3
IL
Cook
0
1.37
4.33
9.77
22.5
9.77
22.5
IL
DuPage
0
1.37
4.33
8.05
18.7
8.05
18.7
IL
McLean
0
1.37
4.33
8.08
17.1
8.08
17.1
IL
Macon
0
1.37
4.33
8.34
17.8
8.34
17.8
IL
Madison
0
1.37
4.33
9.15
19.1
9.15
19.1
IL
Saint Clair
0
1.37
4.33
8.44
18.2
8.44
18.2
IN
Lake
0
1.37
4.33
9.18
21.9
9.18
21.9
IN
Marion
0
1.37
4.33
10.00
23.1
10.00
23.1
KS
Shawnee
0
1.22
3.51
8.38
20.5
8.38
20.5
KS
Wyandotte
0
1.22
3.51
8.41
22.7
8.41
22.7
KY
Jefferson
0
1.37
4.33
8.68
20.8
8.68
20.8
LA
Caddo
0
1.22
3.51
9.53
20.5
9.53
20.5
LA
East Baton Rouge
0
1.22
3.51
8.44
21.1
8.44
21.1
LA
West Baton Rouge
0
1.22
3.51
8.58
19.7
8.58
19.7
MI
Wayne
0
1.37
4.33
10.34
26.3
10.34
26.3
MS
Hinds
0
1.22
3.51
8.42
17.6
8.42
17.6
MT
Flathead
0
2.14
8.70
7.72
31.7
7.72
31.7
MT
Lewis and Clark
126
2.14
8.70
8.79
36.5
8.52
35.4
MT
Lincoln
0
2.14
8.70
11.63
31.6
11.63
31.6
MT
Missoula
0
2.14
8.70
9.09
28.2
9.09
28.2
NV
Clark
0
2.14
8.70
9.04
26.3
9.04
26.3
152
-------
State
County
AEmission
2032 to
12/35
(ton)
AQ Ratio
Annual
Cngm3
perkton)
AQ Ratio
2 4-hour
(|Agnv3per
kton)
2032a DV
Annual
(Hgm')
2032a
DV
2 4-hour
km')
12/35 DV
Annual
(M-gm3)
12/35 DV
2 4-hour
(M-gm3)
NJ
Bergen
0
1.37
4.33
9.78
21.6
9.78
21.6
NJ
Camden
0
1.37
4.33
9.27
22.2
9.27
22.2
NJ
Union
0
1.37
4.33
8.35
20.5
8.35
20.5
NY
New York
0
1.37
4.33
8.94
22.4
8.94
22.4
NC
Forsyth
0
1.22
3.51
8.07
22.1
8.07
22.1
NC
Mecklenburg
0
1.22
3.51
8.35
18.1
8.35
18.1
OH
Butler
0
1.37
4.33
9.42
21.1
9.42
21.1
OH
Cuyahoga
0
1.37
4.33
9.87
21.6
9.87
21.6
OH
Franklin
0
1.37
4.33
8.11
20.2
8.11
20.2
OH
Hamilton
0
1.37
4.33
10.44
21.8
10.44
21.8
OH
Jefferson
0
1.37
4.33
8.06
19.4
8.06
19.4
OH
Stark
0
1.37
4.33
8.29
18.8
8.29
18.8
OK
Cleveland
0
1.22
3.51
8.63
19.1
8.63
19.1
OK
Oklahoma
0
1.22
3.51
8.71
19.2
8.71
19.2
OK
Tulsa
0
1.22
3.51
8.66
21.9
8.66
21.9
OR
Crook
0
2.14
8.70
8.23
32.8
8.23
32.8
OR
Harney
0
2.14
8.70
9.47
31.4
9.47
31.4
OR
Jackson
0
2.14
8.70
9.55
30.5
9.55
30.5
OR
Josephine
0
2.14
8.70
8.62
26.0
8.62
26.0
OR
Klamath
586
2.14
8.70
10.69
40.5
9.44
35.4
OR
Lake
310
2.14
8.70
8.68
38.1
8.02
35.4
OR
Lane
0
2.14
8.70
8.48
32.1
8.48
32.1
PA
Allegheny
0
1.37
4.33
10.71
32.7
10.71
32.7
PA
Beaver
0
1.37
4.33
8.20
19.2
8.20
19.2
PA
Cambria
0
1.37
4.33
8.46
20.7
8.46
20.7
PA
Chester
0
1.37
4.33
8.69
21.8
8.69
21.8
PA
Delaware
0
1.37
4.33
10.08
24.9
10.08
24.9
PA
Lancaster
0
1.37
4.33
8.71
24.6
8.71
24.6
PA
Lebanon
0
1.37
4.33
8.27
24.9
8.27
24.9
PA
Philadelphia
0
1.37
4.33
8.74
21.4
8.74
21.4
PA
York
0
1.37
4.33
8.42
19.6
8.42
19.6
RI
Providence
0
1.37
4.33
8.12
17.6
8.12
17.6
TN
Davidson
0
1.37
4.33
8.46
17.2
8.46
17.2
TX
Bowie
0
1.22
3.51
8.27
17.6
8.27
17.6
TX
Cameron
0
1.22
3.51
9.90
25.7
9.90
25.7
TX
Dallas
0
1.22
3.51
8.24
19.7
8.24
19.7
TX
El Paso
0
1.22
3.51
8.84
23.5
8.84
23.5
TX
Harris
0
1.22
3.51
9.79
23.5
9.79
23.5
TX
Hidalgo
0
1.22
3.51
10.69
28.4
10.69
28.4
TX
Jefferson
0
1.22
3.51
8.46
21.4
8.46
21.4
TX
Nueces
0
1.22
3.51
8.65
24.5
8.65
24.5
TX
Orange
0
1.22
3.51
8.47
20.5
8.47
20.5
153
-------
State
County
AEmission
2032 to
12/35
(ton)
AQ Ratio
Annual
Cms 111:1
perkton)
AQ Ratio
2 4-hour
(Hgnv3per
kton)
2032a DV
Annual
(Hgm')
2032a
DV
2 4-hour
(Hgm3)
12/35 DV
Annual
(M-gm3)
12/35 DV
2 4-hour
(M-gm3)
TX
T ravis
0
1.22
3.51
9.22
21.9
9.22
21.9
UT
Box Elder
0
2.14
8.70
6.65
31.7
6.65
31.7
UT
Cache
0
2.14
8.70
6.20
31.4
6.20
31.4
UT
Salt Lake
0
2.14
8.70
7.64
30.9
7.64
30.9
WA
King
0
2.14
8.70
8.13
25.6
8.13
25.6
WA
Okanogan
0
2.14
8.70
8.27
32.6
8.27
32.6
WA
Yakima
345
2.14
8.70
8.28
38.4
7.54
35.4
a For South Coast and SJV counties, these are DVs that result from applying 75% NOx emission reduction to
the 2032 DVs.
2A.3.4.2 Emission Reductions Needed to Meet 10/35, 9/35,8/35, and 10/30
The primary PM2.5 emission reductions needed to meet the revised and alternative
standard levels of 10/35,10/30, 9/35, and 8/35 relative to the 12/35 analytical baseline
were calculated to inform identification of emission controls. These emission amounts
were calculated using Equations 2A-5 and 2A-6 and the air quality ratios in the Table 2A-9
and are shown in Table 2A-14. The total emission reductions needed in the eastern and
western US is also shown in Figure 2A-32 for the standard combinations.
40,000
i—
IT 30,000
o
-1—»
120,000
CO
CO
£ 10,000
LU
0
East
West
Figure 2A-32 Total Primary PM2.5 Emission Reductions Needed to Meet the Revised
and Alternative Standard Levels of 10/35,10/30,9/35, and 8/35
Relative to the 12/35 Analytical Baseline in the East and West
¦
12/35 10/35 10/30 9/35 8/35
Standard Level
154
-------
Table 2A-14 Primary PM2.5 Emission Reductions Needed to Meet the Revised and
Alternative Standard Levels of 10/35,10/30, 9/35, and 8/35 Relative
to the 12/35 Analytical Baseline
State
County
Emission
10/35
(ton)
Emission
9/35
(ton)
Emission
8/35
(ton)
Emission
10/30
(ton)
Alabama
Jefferson
0
0
785
0
Alabama
Russell
0
0
327
0
Arizona
Maricopa
0
257
725
0
Arizona
Pinal
0
0
0
483
Arizona
Santa Cruz
0
0
313
0
Arizona
Yuma
0
0
122
0
Arkansas
Pulaski
0
0
719
0
California
Alameda
105
422
739
105
California
Butte
0
0
247
301
California
Calaveras
0
178
495
0
California
Colusa
0
0
0
502
California
Contra Costa
0
127
444
0
California
Fresno
634
951
1268
634
California
Imperial
1701
2551
3402
1701
California
Kern
634
951
1268
634
California
Kings
508
825
1142
508
California
Los Angeles
1701
2551
3402
1701
California
Madera
0
292
609
0
California
Mendocino
0
0
10
231
California
Merced
232
549
866
232
California
Mono
0
0
173
502
California
Napa
0
0
273
0
California
Orange
593
1444
2294
593
California
Placer
0
0
10
0
California
Plumas
309
626
943
502
California
Riverside
1701
2551
3402
1701
California
Sacramento
0
162
479
201
California
San Bernardino
1701
2551
3402
1701
California
San Diego
0
238
1089
0
California
San Francisco
0
0
184
0
California
San Joaquin
0
304
621
31
California
San Luis Obispo
0
0
13
0
California
San Mateo
0
0
114
0
California
Santa Barbara
0
0
68
1213
California
Santa Clara
0
238
555
0
California
Shasta
0
0
0
30
California
Siskiyou
0
0
0
502
California
Solano
0
162
479
0
California
Stanislaus
301
618
935
301
155
-------
State County Emission Emission Emission Emission
10/35
(ton)
9/35
(ton)
8/35
(ton)
10/30
(ton)
California
Sutter
0
117
434
191
California
Tehama
0
0
0
291
California
Tulare
634
951
1268
634
California
Ventura
0
43
893
1721
Colorado
Adams
0
42
510
0
Colorado
Denver
0
9
477
0
Colorado
Weld
0
0
337
0
District Of Columbia
District of Columbia
0
0
270
0
Florida
Broward
0
16
834
0
Georgia
Bibb
0
0
114
0
Georgia
Chatham
0
0
82
0
Georgia
Dougherty
0
0
204
0
Georgia
Fulton
0
0
670
0
Georgia
Gwinnett
0
0
605
0
Georgia
Muscogee
0
0
204
0
Georgia
Richmond
0
49
867
0
Idaho
Benewah
0
0
0
425
Idaho
Canyon
0
0
290
264
Idaho
Lemhi
0
252
720
574
Idaho
Shoshone
243
711
1178
563
Illinois
Cook
0
534
1266
0
Illinois
DuPage
0
0
7
0
Illinois
McLean
0
0
29
0
Illinois
Macon
0
0
220
0
Illinois
Madison
0
80
812
0
Illinois
Saint Clair
0
0
293
0
Indiana
Lake
0
102
834
0
Indiana
Marion
0
702
1434
0
Kansas
Shawnee
0
0
278
0
Kansas
Wyandotte
0
0
303
0
Kentucky
Jefferson
0
0
468
0
Louisiana
Caddo
0
401
1218
0
Louisiana
East Baton Rouge
0
0
327
0
Louisiana
West Baton Rouge
0
0
442
0
Michigan
Wayne
220
951
1683
220
Mississippi
Hinds
0
0
311
0
Montana
Flathead
0
0
0
149
Montana
Lewis and Clark
0
0
224
574
Montana
Lincoln
744
1211
1679
744
Montana
Missoula
0
23
491
0
Nevada
Clark
0
0
468
0
New Jersey
Bergen
0
541
1273
0
156
-------
State
County
Emission
10/35
(ton)
Emission
9/35
(ton)
Emission
8/35
(ton)
Emission
10/30
(ton)
New Jersey
Camden
0
168
900
0
New Jersey
Union
0
0
227
0
New York
New York
0
0
659
0
North Carolina
Forsyth
0
0
25
0
North Carolina
Mecklenburg
0
0
253
0
Ohio
Butler
0
278
1010
0
Ohio
Cuyahoga
0
607
1339
0
Ohio
Franklin
0
0
51
0
Ohio
Hamilton
293
1024
1756
293
Ohio
Jefferson
0
0
15
0
Ohio
Stark
0
0
183
0
Oklahoma
Cleveland
0
0
482
0
Oklahoma
Oklahoma
0
0
548
0
Oklahoma
Tulsa
0
0
507
0
Oregon
Crook
0
0
89
276
Oregon
Harney
0
201
669
115
Oregon
Jackson
0
238
706
11
Oregon
Josephine
0
0
271
0
Oregon
Klamath
0
186
653
574
Oregon
Lake
0
0
0
574
Oregon
Lane
0
0
206
195
Pennsylvania
Allegheny
490
1222
1954
532
Pennsylvania
Beaver
0
0
117
0
Pennsylvania
Cambria
0
0
307
0
Pennsylvania
Chester
0
0
476
0
Pennsylvania
Delaware
29
761
1493
29
Pennsylvania
Lancaster
0
0
490
0
Pennsylvania
Lebanon
0
0
168
0
Pennsylvania
Philadelphia
0
0
512
0
Pennsylvania
York
0
0
278
0
Rhode Island
Providence
0
0
59
0
Tennessee
Davidson
0
0
307
0
Texas
Bowie
0
0
188
0
Texas
Cameron
0
703
1521
0
Texas
Dallas
0
0
164
0
Texas
El Paso
0
0
654
0
Texas
Harris
0
613
1431
0
Texas
Hidalgo
531
1349
2167
531
Texas
Jefferson
0
0
343
0
Texas
Nueces
0
0
499
0
Texas
Orange
0
0
352
0
Texas
Travis
0
147
965
0
157
-------
State
County
Emission
10/35
(ton)
Emission
9/35
(ton)
Emission
8/35
(ton)
Emission
10/30
(ton)
Utah
Box Elder
0
0
0
149
Utah
Cache
0
0
0
115
Utah
Salt Lake
0
0
0
57
Washington
King
0
0
42
0
Washington
Okanogan
0
0
108
253
Washington
Yakima
0
0
0
574
2A.4 Calculating PM2.5 Concentration Fields for Standard Combinations
National PM2.5 concentration fields corresponding to meeting the existing, revised,
and alternative standards were developed to inform health benefit calculations. First, a
gridded PM2.5 concentration field for the 2032 CMAQ modeling case was developed using
the enhanced Voronoi Neighbor Average (eVNA) method (Ding et al., 2016; Gold et al.,
1997; USEPA, 2007). Next, the incremental difference in annual PM2.5 DVs between the
2032 case and cases of meeting standard combinations was calculated at monitors and
interpolated to the spatial grid. The resulting field of incremental PM2.5 concentration was
then subtracted from the 2032 eVNA field to create the gridded field for the standard
combination. The steps in developing the PM2.5 concentration fields are described further
below.
2A.4.1 Creating the PM2.5 Concentration Field for 2032
The gridded field of annual average PM2.5 concentrations for 2032 was developed
using the eVNA method that combines information from the model and monitors to predict
PM2.5 concentrations. The eVNA approach was applied using SMAT-CE and has been
previously described in EPA's modeling guidance document (USEPA, 2018) and the user's
guide for the predecessor software to SMAT-CE (Abt, 2014). The method is briefly
described here, and more details are available in the primary references.
Quarterly average PM2.5 component concentrations measured during the 2017-2019
period were interpolated to the spatial grid using inverse distance-squared-weighting of
monitored concentrations that were further weighted by the ratio of the 2018 CMAQ value
in the prediction grid cell to CMAQ value in the monitor-containing grid cell. Weighting by
the ratio of CMAQ values adjusts the interpolation of monitor data to account for spatial
158
-------
gradients in the CMAQ fields. This step results in an interpolated field of gradient-adjusted
observed concentrations for each PM2.5 component and each quarter:
eVNAs q 2ois = 1 VKeightxMonitorxs q (2A-10)
where eVNAs,q,2018 is the gradient-adjusted quarterly-average concentration of PM2.5
component species, s, during quarter, q, at the prediction grid cell; Weightx is the inverse-
distance-squared weight for monitor, x, at the location of the prediction grid cell;
Monitorx,s,q is the average of the quarterly-average monitored concentrations for species, s,
at monitor, x, during quarter, q, in 2017-2019; Models,<7,2018 is the quarterly-average 2018
CMAQ concentration of species, s, during quarter, q, in the prediction grid cell; and
Modelx,s,q,2018 is the quarterly-average 2018 CMAQ concentration of species, s, during
quarter, q, in the grid cell of monitor, x.
The 2018 eVNA fields for quarterly-average PM2.5 component concentrations are the
starting point for developing the 2032 PM2.5 concentration field. To create eVNA fields for
PM2.5 components in 2032, the 2018 eVNA component concentration in each grid cell is
multiplied by the corresponding ratio of the quarterly-average CMAQ concentration
predictions in 2032 and 2018:
eVNAs,q,2032 = eVNAs,q,2018^^ (2A-11)
The PM2.5 concentration fields for quarters in 2032 are calculated by summing the
2032 PM2.5 component concentration by quarter. The 2032 PM2.5 concentration field is
then calculated by averaging the 2032 quarterly PM2.5 concentrations. The resulting 2032
PM2.5 concentration fields is shown in Figure 2A-33.
159
-------
ug/m3
> 15
12
-120
-110 -100 -90
Longitude
-80
-70
Figure 2A-33 PM2.5 Concentration for 2032 based on eVNA Method
2A.4.2 Creating Spatial Fields Corresponding to Meeting Standards
To create spatial fields corresponding to meeting standard levels, the 2032
concentration field was adjusted according to the change in PM2.5 concentrations
associated with the difference in annual PM2.5 DVs between the 2032 case and the cases
where standards are met. To implement this adjustment, the difference in annual PM2.5 DVs
was calculated at the county highest monitor between the 2032 case and cases of meeting
the 12/35,10/30,10/35, 9/35, and 8/35 standard combinations for counties with
monitors exceeding the standard level. The PM2.5 DVs for the cases where revised and
alternative standard levels are met were developed by applying the air quality ratios to the
emission reductions for the county (i.e., Eqn. 2-2). For the county non-highest monitors, the
difference in PM2.5 DVs was estimated by proportionally adjusting the DVs according to the
percent change in PM2.5 DV at the highest monitor.
Due to the relatively large size and complex terrain of counties in the western US,
the proportional adjustment of DVs within counties was limited in some cases.
Proportional adjustment was not applied to six sites that have 2032 annual PM2.5 DVs less
than 7 jj,g nr3 and are located within counties that exceed 12/35 standard combination: i.e.,
06-029-0011 and 06-029-0018 (Kern, CA), 06-037-9033 (Los Angeles, CA), 06-065-5001
(Riverside, CA), 06-071-8001 (San Bernardino, CA), 30-049-0004 (Lewis and Clark, MT).
The relatively low annual PM2.5 DVs for these sites compared with the highest-DV monitor
160
-------
suggests they are influenced by different air pollution processes than the highest-DV
monitor. Additionally, the annual PM2.5 DV at the 06-065-2002 site in Riverside County was
not adjusted due to its location outside of the portion of the county within the South Coast
Air Basin that contains the highest-DV monitor. The annual PM2.5 DV at the 06-051-0001
site in Mono County, CA was also not adjusted because it is outside of the San Joaquin
Valley.
After adjusting the annual PM2.5 DVs at county monitors and calculating the
difference in annual DVs between the 2032 case and cases of meeting the standard
combinations, the annual PM2.5 DV differences were interpolated to the spatial grid using
inverse-distance-squared VNA interpolation. The interpolated field was then clipped to
grid cells within 50 km of monitors whose design values changed in meeting the standard
level (USEPA, 2012a). National PM2.5 concentration fields were then developed for each
standard combination by subtracting the corresponding VNA field of incremental PM2.5
concentration from the 2032 eVNA concentration field. The resulting PM2.5 concentration
fields were then compared with regional estimates of background PM2.5 concentrations
based on a previous CMAQ modeling study with North American anthropogenic emissions
set to zero (see Table 3-23 of USEPA, 2009). For a small number of grid cells (three for the
9/35 case and seven for the 8/35 case) in the full attainment scenario, adjusted
concentrations were below the Southern California background level of 0.84 |Lxg nr3 and
were reset to that value. These grid cells are over the mountain ranges downwind of Los
Angeles and Bakersfield where concentrations are much lower than in the South Coast Air
Basin and SJV. In the partial attainment case, all concentrations were above the regional
background levels and no adjustments were applied.
In Figure 2A-34, the spatial field for the incremental change in PM2.5 concentration
between the 12/35 analytical baseline and the case of meeting the 9/35 standard
combination is shown.
161
-------
50-
CD
"§ 40-
n3
30-
¦ i i i
-120 -110 -100 -90 -80 -70
Longitude
jig m
¦
Figure 2A-34 PM2.5 Concentration Improvement Associated with Meeting 9/35
Relative to the 12/35 Analytical Baseline
162
-------
2A.5 References
Abt (2014). User's Guide: Modeled Attainment Test Software. Abt Associates, Prepared for
US EPA Office of Air Quality Planning and Standards
Appel, KW, Bash, JO, Fahey, KM, Foley, KM, Gilliam, RC, Hogrefe, C, Hutzell, WT, Kang, D,
Mathur, R, Murphy, BN, Napelenok, SL, Nolte, CG, Pleim, JE, Pouliot, GA, Pye, HOT, Ran, L,
Roselle, SJ, Sarwar, G, Schwede, DB, Sidi, FI, Spero, TL and Wong, DC (2021). The
Community Multiscale Air Quality (CMAQ) model versions 5.3 and 5.3.1: system
updates and evaluation. Geosci. Model Dev. 14(5): 2867-2897.
Appel, KW, Napelenok, S, Hogrefe, C, Pouliot, G, Foley, KM, Roselle, SJ, Pleim, JE, Bash, J, Pye,
HOT, Heath, N, Murphy, B and Mathur, R (2018). Overview and Evaluation of the
Community Multiscale Air Quality (CMAQ) Modeling System Version 5.2. Air Pollution
Modeling and its Application XXV, Cham, Springer International Publishing.
Appel, KW, Napelenok, SL, Foley, KM, Pye, HOT, Hogrefe, C, Luecken, DJ, Bash, JO, Roselle, SJ,
Pleim, JE, Foroutan, H, Hutzell, WT, Pouliot, GA, Sarwar, G, Fahey, KM, Gantt, B, Gilliam,
RC, Heath, NK, Kang, DW, Mathur, R, Schwede, DB, Spero, TL, Wong, DC and Young, JO
(2017). Description and evaluation of the Community Multiscale Air Quality (CMAQ)
modeling system version 5.1. Geoscientific Model Development 10(4): 1703-1732.
Baker, KR, Simon, H and Kelly, JT (2011). Challenges to Modeling "Cold Pool" Meteorology
Associated with High Pollution Episodes. Environmental Science & Technology 45(17):
7118-7119.
Bash, JO, Baker, KR and Beaver, MR (2016). Evaluation of improved land use and canopy
representation in BEIS v3.61 with biogenic VOC measurements in California. Geosci.
Model Dev. 9(6): 2191-2207.
Cohan, DS and Chen, R (2014). Modeled and observed fine particulate matter reductions
from state attainment demonstrations. Journal of the Air & Waste Management
Association 64(9): 995-1002.
Ding, D, Zhu, Y, Jang, C, Lin, C-J, Wang, S, Fu, J, Gao, J, Deng, S, Xie, J and Qiu, X (2016).
Evaluation of health benefit using BenMAP-CE with an integrated scheme of model and
monitor data during Guangzhou Asian Games. Journal of Environmental Sciences 42: 9-
18.
Emery, C, Jung, J, Koo, B and Yarwood, G (2015). Improvements to CAMx snow cover
treatments and Carbon Bond chemical mechanism for winter ozone. Final Report,
prepared for Utah Department of Environmental Quality, Salt Lake City, UT. Prepared
by Ramboll Environ, Novato, CA, August 2015.
Fountoukis, C and Nenes, A (2007). ISORROPIA II: a computationally efficient
thermodynamic equilibrium model for K+, Ca2+, Mg2+, NH4+, Na+, S042-, N03-, C1-,
H20 aerosols. Atmos. Chem. Phys. 7(17): 4639-4659.
163
-------
Frank, NH (2006). Retained Nitrate, Hydrated Sulfates, and Carbonaceous Mass in Federal
Reference Method Fine Particulate Matter for Six Eastern U.S. Cities. Journal of the Air &
Waste Management Association 56(4): 500-511.
Gantt, B, Kelly, JT and Bash, JO (2015). Updating sea spray aerosol emissions in the
Community Multiscale Air Quality (CMAQ) model version 5.0.2. Geosci. Model Dev.
8(11): 3733-3746.
Gold, C, Remmele, PR and Roos T. (1997). In: Algorithmic Foundation of Geographic
Information Systems. In: Lecture Notes in Computer Science, Vol. 1340 (van Kereveld
M, Nievergelt J, Roos T, Widmayer P, eds) Berlin, Germany: Springer-Verlag. Voronoi
methods in GIS. pp. 21-35.
Iacono, MJ, Delamere, JS, Mlawer, EJ, Shephard, MW, Clough, SA and Collins, WD (2008).
Radiative forcing by long-lived greenhouse gases: Calculations with the AER radiative
transfer models. Journal of Geophysical Research: Atmospheres 113(D13).
Kain, JS (2004). The Kain-Fritsch Convective Parameterization: An Update. Journal of
Applied Meteorology 43(1): 170-181.
Karl, TR and Koss, WJ (1984). Regional and National Monthly, Seasonal, and Annual
Temperature Weighted by Area, 1895-1983. Historical Climatology Series 4-3, National
Climatic Data Center, Asheville, NC, 38 pp. https://www.ncdc.noaa.gov/monitoring-
references/maps/us-climate-regions.php
Kelly, JT, Koplitz, SN, Baker, KR, Holder, AL, Pye, HOT, Murphy, BN, Bash, JO, Henderson, BH,
Possiel, NC, Simon, H, Eyth, AM, Jang, C, Phillips, S and Timin, B (2019). Assessing PM2.5
model performance for the conterminous U.S. with comparison to model performance
statistics from 2007-2015. Atmospheric Environment 214: 116872.
Ma, L-M and Tan, Z-M (2009). Improving the behavior of the cumulus parameterization for
tropical cyclone prediction: Convection trigger. Atmospheric Research 92(2): 190-211.
Morrison, H, Curry, JA and Khvorostyanov, VI (2005). A New Double-Moment Microphysics
Parameterization for Application in Cloud and Climate Models. Part I: Description.
Journal of the Atmospheric Sciences 62(6): 1665-1677.
Morrison, H and Gettelman, A (2008). A New Two-Moment Bulk Stratiform Cloud
Microphysics Scheme in the Community Atmosphere Model, Version 3 (CAM3). Part I:
Description and Numerical Tests. Journal of Climate 21(15): 3642-3659.
Murphy, BN, Woody, MC, Jimenez, JL, Carlton, AMG, Hayes, PL, Liu, S, Ng, NL, Russell, LM,
Setyan, A, Xu, L, Young, J, Zaveri, RA, Zhang, Q and Pye, HOT (2017). Semivolatile POA
and parameterized total combustion SOA in CMAQv5.2: impacts on source strength and
partitioning. Atmos. Chem. Phys. 17(18): 11107-11133.
164
-------
NRC (2004). Air Quality Management in the United States, National Research Council. The
National Academies Press, Washington, DC.
Pinal County Air Quality Control District (2020). 2020 Ambient Monitoring Network Plan
and 2019 Data Summary. https://www.pinal.gov/ArchiveCenter/ViewFile/Item/217
Pleim, JE (2007). A Combined Local and Nonlocal Closure Model for the Atmospheric
Boundary Layer. Part I: Model Description and Testing. Journal of Applied Meteorology
and Climatology 46(9): 1383-1395.
Pleim, JE, Xiu, A, Finkelstein, PL and Otte, TL (2001). A Coupled Land-Surface and Dry
Deposition Model and Comparison to Field Measurements of Surface Heat, Moisture,
and Ozone Fluxes. Water, Air and Soil Pollution: Focus 1(5): 243-252.
Simon, H, Baker, KR and Phillips, S (2012). Compilation and interpretation of
photochemical model performance statistics published between 2006 and 2012.
Atmospheric Environment 61: 124-139.
Simon, H and Bhave, PV (2012). Simulating the Degree of Oxidation in Atmospheric Organic
Particles. Environmental Science & Technology 46(1): 331-339.
SJVAPCD (2018). San Joaquin Valley Air Pollution Control District, 2018 Plan for the 1997,
2006, and 2012 PM2.5 Standards, Available:
http://valleyair.org/pmplans/documents/2018/pm-plan-adopted/2018-Plan-for-the-
1997-2006-and-2012-PM2.5-Standards.pdf.
Skamarock, WC, Klemp, JB, Dudhia, J, Gill, DO, Barker, DM, Duda, MG, Huang, X, Wang, W
and Powers, JG (2008). A description of the Advanced Reserch WRF version 3. NCAR
Technical Note NCAR/TN-475+STR.
USEPA (2007). Technical Report on Ozone Exposure, Risk, and Impact Assessments for
Vegetation. EPA 452/R-07-002. Prepared by Abt Associates Inc. for U.S. Environmental
Protection Agency, Office of Air Quality Planning and Standards, Health and
Environmental Impacts Division. Research Triangle Park, NC. Available:
https://www3.epa.gov/ttn/naaqs/standards/ozone/data/2007_01_environmental_tsd.
pdf.
USEPA (2009). Integrated Science Assessment (ISA) for Particulate Matter (Final Report,
Dec 2009). U.S. Environmental Protection Agency, Washington, DC, EPA/600/R-
08/139F. Available: https://cfpub.epa.gov/ncea/isa/recordisplay.cfm?deid=216546.
USEPA (2012a). Regulatory Impact Analysis for the Final Revisions to the National Ambient
Air Quality Standards for Particulate Matter, U.S. Environmental Protection Agency,
Office of Air Quality Planning and Standards, Research Triangle Park, NC. EPA-452/R-
12-005. Available: https://www.epa.gov/sites/default/files/2020-
07/documents/naaqs-pm_ria_final_2012-12.pdf.
165
-------
USEPA (2012b). Regulatory Impact Analysis for the Proposed Revisions to the National
Ambient Air Quality Standards for Particulate Matter. U.S. Environmental Protection
Agency, Office of Air Quality Planning and Standards, Research Triangle Park, NC. EPA-
452/R-12-003. Available: https://www.epa.gov/sites/default/files/2020-
07/documents/naaqs-pm_ria_proposed_2012-06.pdfo
USEPA (2018). Modeling Guidance for Demonstrating Attainment of Air Quality Goals for
Ozone, PM2.5, and Regional Haze. EPA -454/B-07-002 U.S. EPA, Office of Air Quality
Planning and Standards. Research Triangle Park, NC. EPA454/R-18-009. Available:
https://www3.epa.gov/ttn/scram/guidance/guide/03-PM-RH-Modeling_Guidance-
2018.pdf. http://www.epa.gov/ttn/scram/guidance/guide/final-03-pm-rh-
guidance.pdf.
USEPA (2019a). Additional Methods, Determinations, and Analyses to Modify Air Quality
Data Beyond Exceptional Events. U.S. Environmental Protection Agency, Office of Air
Quality Planning and Standards, Research Triangle Park, NC, EPA-457/B-19-002.
Available: https://www.epa.gov/sites/default/files/2019-
04/documents/clarification_memo_on_data_modification_methods.pdf.
USEPA (2019b). Technical Support Document for EPA's Updated 2028 Regional Haze
Modeling. U.S. Environmental Protection Agency, Office of Air Quality Planning and
Standards, Research Triangle Park, NC. Available:
https://www.epa.gov/sites/default/files/2019-
10/documents/updated_2028_regional_haze_modeling-tsd-2019_0.pdf
USEPA (2019c). Meteorological Model Performance for Annual 2016 Simulation WRF v3.8.
Available: https://www.epa.gov/sites/production/files/2020-
10/documents/met_model_performance-2016_wrf.pdf.
USEPA (2019d). Integrated Science Assessment (ISA) for Particulate Matter (Final Report,
2019). U.S. Environmental Protection Agency, Washington, DC, EPA/600/R-19/188.
USEPA (2021). Bayesian Space-time Downscaling Fusion Model (Downscaler) Derived
Estimates of Air Quality for 2018. U.S. Environmental Protection Agency, Office of Air
Quality Planning and Standards, Air Quality Assessment Division, Research Triangle
Park, NC. EPA-454/R-21-003. Available:
https://www.epa.gov/system/files/documents/2021-
ll/2018_ds_annual_report_final.pdf
USEPA (2023a). Technical Support Document (TSD): Preparation of Emissions Inventories
for the 2018v2 North American Emissions Modeling Platform. U.S. Environmental
Protection Agency, EPA-454/B-23-003 Available: https://www.epa.gov/air-emissions-
modeling/2018v2-platform.
USEPA (2023b). Meteorological Model Performance for Annual 2018 Simulation WRF v3.8.
US Environmental Protection Agency. Office of Air Quality Planning & Standards. Air
Quality Assessment Division. Research Triangle Park, NC. EPA-454/R-23-006.
166
-------
Wang, H, Zhu, Y, Jang, C, Lin, C-J, Wang, S, Fu, JS, Gao, J, Deng, S, Xie, J, Ding, D, Qiu, X and
Long, S (2015). Design and demonstration of a next-generation air quality attainment
assessment system for PM2.5 and 03. Journal of Environmental Sciences 29: 178-188.
Xiu, A and Pleim, JE (2001). Development of a Land Surface Model. Part I: Application in a
Mesoscale Meteorological Model. Journal of Applied Meteorology 40(2): 192-209.
167
-------
CHAPTER 3: CONTROL STRATEGIES AND PM2.5 EMISSIONS REDUCTIONS
Overview
The current primary annual PM2.5 standard is 12 ng/m3, and the current 24-hour
standard is 35 |~ig/m3. The EPA Administrator is revising the current annual PM2.5 standard
to a level of 9 |~ig/m3. The EPA Administrator is also retaining the current 24-hour standard
of 35 ng/m3. In this Regulatory Impact Analysis (RIA), we analyze the revised annual and
current 24-hour standard levels of 9/35 ng/m3, as well as the following less and more
stringent alternative standard levels: (1) a less stringent alternative annual standard level
of 10 ng/m3 in combination with the current 24-hour standard (i.e., 10/35 |j,g/m3), (2) a
more stringent alternative annual standard level of 8 |~ig/m3 in combination with the
current 24-hour standard (i.e., 8/35 |j,g/m3), and (3) a more stringent alternative 24-hour
standard level of 30 |~ig/m3 in combination with an alternative annual standard level of 10
Hg/m3 (i.e., 10/30 |j,g/m3). Because the EPA is not changing the current secondary PM
standards at this time, we did not evaluate alternative secondary standard levels in this
RIA.
As discussed in Chapter 1 in the Overview of the Regulatory Impact Analysis, the
analyses in this RIA rely on national-level data (emissions inventory and control measure
information) for use in national-level assessments (air quality modeling, control strategies,
environmental justice, and benefits estimation). However, the ambient air quality issues
being analyzed are highly complex and local in nature, and the results of these national-
level assessments therefore contain uncertainty. It is beyond the scope of this RIA to
develop detailed local information for the areas being analyzed, including populating the
local emissions inventory, obtaining local information to increase the resolution of the air
quality modeling, and obtaining local information on emissions controls, all of which would
reduce some of the uncertainty in these national-level assessments. For example, having
more refined data would be ideal for agricultural dust and burning, prescribed burning,
and non-point (area) sources due to their large contribution to primary PM2.5 emissions
168
-------
and the limited availability of emissions controls.1 In addition, for residential wood
combustion emissions, people will respond differently to the various regulations and
incentives offered for controlling PM2.5 emissions from wood burning, making it important
to identify the right balance of controls for each area.
We assume that areas will be designated such that they are required to reach
attainment by 2032, and we developed our projected baselines for emissions and air
quality for 2032. To estimate the costs and benefits of the revised and less and more
stringent annual and 24-hour PM2.5 alternative standard levels, we first prepared an
analytical baseline for 2032 that assumes full compliance with the current standards of
12/35 ng/m3. From that baseline, we then analyze illustrative control strategies that areas
might employ toward attaining the revised and less and more stringent annual and 24-hour
PM2.5 alternative standard levels.2 Because PM2.5 concentrations are most responsive to
primary PM emissions reductions, as discussed in Chapter 2, Section 2.1.3, we analyze
direct, local PM2.5 emissions reductions by individual counties. Section 2.1.3 also includes a
discussion of historical and projected emissions trends for direct PIVh.sand precursor
emissions (i.e., SO2, NOx, VOC, and ammonia), as well as a discussion of the "urban
increment" of consistently higher PM2.5 concentrations over urban areas. The projections of
additional, large reductions in SO2 and NOx emissions in the 2032 case further motivate the
need for control of local primary PM2.5 sources to address the highest PM2.5 concentrations
in urban areas.
For the eastern U.S. where counties are relatively small and terrain is relatively flat,
we identified potential PM2.5 emissions reductions within each county and in adjacent
counties within the same state, where needed. As discussed below in Section 3.2.2, when
we applied the emissions reductions from adjacent counties, we used a |~ig/m3 per ton PM2.5
air quality ratio that was four times less responsive than the ratio used when applying in-
county emissions reductions. Because the counties in the western U.S. are generally large
1 Examples of area source emissions include area fugitive dust, residential wood combustion, and commercial
cooking emissions.
2 We define control strategy as a group of end-of-pipe control technologies and area source controls. In this
analysis, we developed control strategies for the revised and alternative standard levels analyzed.
169
-------
and the terrain is more complex, we only identified potential PM2.5 emissions reductions
within each county.
Next, we prepare illustrative control strategies. We apply end-of-pipe control
technologies to non-electric generating unit (non-EGU) stationary sources (e.g., fabric
filters, electrostatic precipitators, venturi scrubbers) and area source controls to non-point
(area) sources (e.g., installing controls on charbroilers), to residential wood combustion
sources (e.g., converting woodstoves to gas logs), and for area fugitive dust emissions (e.g.,
paving unpaved roads) in analyzing PM2.5 emissions reductions needed toward attaining
the revised and alternative standard levels. We did not apply controls to EGUs or mobile
sources; Chapter 2, Section 2.1.3 includes a discussion of SO2 and NOx emissions decreases
reflected in the projections between 2018 and 2032, noting that over the period (1) NOx
emissions are projected to decrease by 3.6 million tons (41 percent), with the greatest
reductions from mobile source and EGU emissions inventory sectors, and (2) SO2 emissions
are projected to decrease by 1.1 million tons (48 percent), with the greatest reductions
from the EGU emissions inventory sector. In addition, Chapter 2, Section 2.2.1.2 includes a
discussion of the EGU and non-EGU rules reflected in the projections for this analysis.
The illustrative control strategy analyses indicate that counties in the northeast and
southeast U.S. do not need additional emissions reductions after the application of controls
to meet alternative standard levels of 10/35 |~ig/m3 and 10/30 |j,g/m3; however, counties in
the northeast and southeast U.S. would need additional PM2.5 emissions reductions to meet
revised and alternative standard levels of 9/35 |~ig/m3 and 8/35 ng/m3. Also, the analysis
indicates that counties in the west and California would need additional PM2.5 emissions
reductions after the application of controls to meet all of the standard levels analyzed.
The remainder of this chapter is organized into four sections. Section 3.1 provides a
summary of the steps that we took to create the analytical baseline. Section 3.2 presents
the illustrative control strategies identified to assess the revised and less and more
stringent annual and 24-hour alternative standard levels in the continental U.S., along with
the resulting estimated emissions reductions. Section 3.3 includes a summary of the key
limitations and uncertainties associated with the control strategy analyses. Finally, Section
170
-------
3.4 includes the references for the chapter. We present the costs associated with the
estimated PM2.5 emissions reductions in Chapter 4.
3.1 Preparing the 12/35 (ig/m3 Analytical Baseline
In the 2032 projections, PM2.5 DVs exceeded the current standards for some
counties in the west and California. As a result, we adjusted the PM2.5 DVs for 2032 to
correspond with just meeting the current standards to form the 12/35 p.g/m3 analytical
baseline used in estimating the incremental costs and benefits associated with control
strategies for the revised and less and more stringent alternative standard levels relative to
the current standards. Figure 3-1 includes a map of the U.S. with the areas identified as
northeast, southeast, west, and California; results are summarized for these areas. Table
3-1 presents a summary of the PM2.5 emissions reductions needed by area to meet the
current standards.
Figure 3-1 Geographic Areas Used in Analysis
171
-------
Table 3-1 Summary of PM2.5 Emissions Reductions Needed by Area in 2032 to
Meet Current Primary Annual and 24-hour Standards of 12/35 (ig/m3
(tons/year)
Area
12/35
Northeast
0
Southeast
0
West
1,494
CA
6,032
Total
7,526
Sixteen counties need PM2.5 emissions reductions to meet the current standards in
2032 - 11 counties in California and 5 counties in the west.3 The counties in California
include several counties in the San Joaquin Valley Air Pollution Control District and the
South Coast Air Quality Management District, as well as Plumas County, Colusa County, and
Siskiyou County in Northern California, Mono County in Eastern California, and Imperial
County in Southern California. No counties in the northeast or southeast U.S. need PM2.5
emissions reductions to meet the current annual and 24-hour standards.
3.2 Illustrative Control Strategies and PM2.5 Emissions Reductions from the
Analytical Baseline
To analyze counties projected to exceed the revised and less and more stringent
annual and 24-hour alternative standard levels in 2032, we estimate total PM2.5 emissions
reductions needed by county for the standard levels analyzed. To estimate the PM2.5
emissions reductions needed, we start with projected future DVs, DV targets for each area,
and the sensitivity of PM2.5 DVs to direct PM2.5 emissions reductions. For each of the
standard levels, we estimate PM2.5 emissions reductions needed by county and then
identify end-of-pipe and area source controls that can achieve PM2.5 emissions reductions.
In Section 3.2.1, we discuss the approach for estimating the direct PM2.5 emissions
reductions needed and present them by area for the revised and alternative standard levels
analyzed. In Sections 3.2.2 and 3.2.3, respectively, we present information on the controls
and the estimated emissions reductions, from the analytical baseline, associated with
applying controls by area for the revised and alternative standard levels analyzed. In
Sections 3.2.4 and 3.2.5, we discuss areas with other types of influences affecting PM2.5
3 The 16 counties require primary PM emissions reductions to meet the current standards of 12/35 |J.g/m3
following application of the NOx emission reductions in San Joaquin Valley and the South Coast to adjust the
2032 DVs. For additional discussion, see Appendix 2A, Section 2A.3.2 and Section 2A.3.3.
172
-------
concentrations. As noted in Chapter 2, Section 2.4, there are certain types of areas for
which our illustrative control strategies may not capture important local emissions and air
quality dynamics. For these areas, we note that local emissions inventory information and
information on potential additional controls for emissions inventory sectors that are
traditionally challenging to control may be needed. Section 3.2.4 presents the emissions
reductions still needed, and for each area Section 3.2.5 includes a qualitative discussion of
the remaining area-specific air quality challenges. Appendix 3A, Table 3A-2 through Table
3A-7 summarize estimated PM2.5 emissions reductions by county for the revised and
alternative standard levels for the northeast, the adjacent counties in the northeast, the
southeast, the adjacent counties in the southeast, the west, and California.
3.2.1 Estimating PM2.5 Emissions Reductions Needed for Annual and 24-hour
Revised and Alternative Standard Levels Analyzed
We apply regional PIVh.sair quality ratios to estimate the emissions reductions
needed to reach the revised and less and more stringent annual and 24-hour alternative
standard levels analyzed. To develop air quality ratios that relate the change in DV in a
county to the change in primary PM2.5 emissions in that county, we performed air quality
sensitivity modeling with reductions in primary PM2.5 emissions in selected counties. More
specifically, we conducted a 2028 Community Multiscale Air Quality Modeling System
(CMAQ) sensitivity modeling simulation with 50 percent reductions in primary PM2.5
emissions from anthropogenic sources in counties with annual 2028 DVs greater than 8
Hg/m3.4 We divided the change in annual and 24-hour PM2.5 DVs in these counties by the
change in emissions in the respective counties to determine the air quality ratio at
individual monitors.
We developed representative air quality ratios for regions of the U.S. from the ratios
at individual monitors as in the 2012 PM2.5 NAAQS review (U.S. EPA, 2012). We calculated
regional ratios as the 75th percentile of air quality ratios at monitors within five regions:
Northeast, Southeast, Northern California, Southern California, and West. The Northeast
4 The modeling sensitivity runs were based on 50 percent reductions in emissions to provide estimates of
PM2.5 sensitivity across the full range of potential emissions changes. Since the response of PM2.5
concentrations to changes in primary PM2.5 emissions is approximately linear (Kelly et al., 2015, 2019), the
air quality ratios are insensitive to the percent emissions change applied in the sensitivity simulations.
173
-------
region was defined by combining the Upper Midwest, Ohio Valley, and Northeast U.S.
climate regions; the Southeast region was defined by combining the Southeast and South
climate regions; and California was separated into Southern and Northern regions as done
previously. (These regions are shown in Figure 2-7 in Chapter 25, and the air quality ratios
for primary PM2.5 emissions used in estimating the emission reductions needed to just
meet the revised and alternative standard levels analyzed are listed in Table 2-1 in Chapter
2.) We estimated the emissions reductions needed to just meet the revised and alternative
standard levels analyzed using the primary PM2.5 air quality ratios in combination with the
required incremental change in concentration. (Chapter 2, Section 2.3.1 includes a brief
discussion of developing air quality ratios and estimated emissions reductions needed to
just meet the revised and alternative standard levels analyzed, and Appendix 2A, Section
2A.3 includes more detailed discussions.)
Table 3-2 presents a summary of the estimated emissions reductions needed by
area to reach the annual and 24-hour alternative standard levels. For the revised and each
alternative standard level, Table 3-2 also includes an area's percent of the total estimated
emissions reductions needed nationwide to reach that standard level in all locations. For
example, for the alternative standard levels of 10/35 |~ig/m3, California's 10,753 estimated
tons needed is 81 percent of the total estimated emissions reductions needed nationwide
to meet 10/35 |~ig/m3. (See Appendix 2A, Table 2A-14 for the estimated PM2.5 emissions
reductions, from the analytical baseline, needed by county for the revised and alternative
standard levels analyzed.) Figure 3-2 shows the counties projected to exceed the annual
and 24-hour revised and alternative standard levels of 10/35 |~ig/m3, 9/35 |~ig/m3, and
8/35 ng/m3 in the analytical baseline. Figure 3-3 shows the counties projected to exceed
the annual and 24-hour alternative standard levels of 10/30 |~ig/m3 in the analytical
baseline. Additional information on the air quality modeling, as well as information about
projected future DVs, DV targets, and air quality ratios is provided in Chapter 2 and
Appendix 2A.
5 To present results throughout this RIA, we combined the Northern California and Southern California
regions.
174
-------
Table 3-2 By Area, Summary of PM2.5 Emissions Reductions Needed, in
Tons/Year and as Percent of Total Reductions Needed Nationwide, for
Revised and Alternative Primary Standard Levels of 10/35 \xg/m3,
10/30 ng/m3, 9/35 jig/m3, and 8/35 |ug/m3 in 2032
Area
10/35
10/30
9/35
8/35
Northeast
1,032
1,073
6,974
20,620
Southeast
531
531
3,279
18,658
West
987
6,673
3,132
10,277
CA
10,753
16,660
19,402
31,518
Total
13,303
24,938
32,786
81,073
Area
10/35
10/30
9/35
8/35
Northeast
8%
4%
21%
25%
Southeast
4%
2%
10%
23%
West
7%
27%
10%
13%
CA
81%
67%
59%
39%
| Reductions required for 10/35, 9/35, and 8/35
Reductions required for 9/35 and 8/35
I Reductions required for 8/35
Figure 3-2 Counties Projected to Exceed in Analytical Baseline for Revised and
Alternative Standard Levels of 10/35 jig/m3, 9/35 pig/m3, and 8/35
Hg/m3
175
-------
Reductions required for 10/30
Figure 3-3 Counties Projected to Exceed in Analytical Baseline for Alternative
Standard Levels of 10/30 jig/in3
As presented previously, for the revised and alternative standard levels, Chapter 2,
Section 2.3.3 includes a discussion of the number of counties that are projected to exceed in
2032, and Figure 2-9 includes maps of counties projected to exceed along with the number
of counties. The following summarizes the number of counties, by revised and alternative
standard levels, in each geographic area that needs PM2.5 emissions reductions from the
analytical baseline.
• 10/35 j.ig/m3— 20 counties need PM2.5 emissions reductions. This includes 4
counties in the northeast, 1 county in the southeast, 2 counties in the west,
and 13 counties in California.
• 10/30 |j.g/m3" 49 counties need PM2.5 emissions reductions. This includes 4-
counties in the northeast, 1 county in the southeast, 19 counties in the west,
and 25 counties in California.
• 9/35 |_ig/m3 - 52 counties need PM2.5 emissions reductions. This includes 12
counties in the northeast, 7 counties in the southeast, 10 counties in the west,
and 23 counties in California.
176
-------
• 8/35 ng/m3 - 117 counties need PM2.5 emissions reductions. This includes 31
counties in the northeast, 33 counties in the southeast, 21 counties in the
west, and 32 counties in California.
3.2.2 Applying End-of-Pipe and Area Source Controls
To identify controls and estimate emissions reductions, we used information about
the emissions reductions needed, by county, in the northeast, southeast, west, and
California. Given the different county sizes between eastern and western states, as well as
different terrain or other topographical features, we estimated potential PM2.5 emissions
reductions for the eastern U.S. and western U.S. as detailed below. Note that we included a
total of 129 counties in the analyses. The total number of counties below (129 counties)
does not directly match the number of counties that would need emissions reductions for
the more stringent alternative standard levels of 8/35 |~ig/m3 (117 counties) in Section
3.2.1 above. This difference is because there are twelve counties that do not need PM2.5
emissions reductions to meet alternative standard levels of 8/35 |~ig/m3 but do need PM2.5
emissions reductions to meet alternative standard levels of 10/30 |j,g/m3.
1. Northeast (31 counties) and Southeast (33 counties) - In the eastern U.S. where
counties are relatively small, we were not always able to identify controls within a
given county. We identified controls and emissions reductions from neighboring
counties because the terrain is relatively flat, and the application of these controls is
appropriate in such cases. Because concentrations are most responsive to changes
in local emissions and for SIP planning purposes, any emissions reductions from
neighboring counties were identified in adjacent counties within the same state.
To apply emissions reductions in the neighboring counties in the eastern U.S., we
compared the responsiveness of annual PM2.5 DVs to emissions reductions within a
county to the responsiveness for neighboring counties. The resulting impact ratio
suggests that primary PM2.5 emissions reductions in neighboring counties would be
4 times less effective as in the core county. (Appendix 2A, Section 2A.3.1 includes a
more detailed discussion of the comparison.) As such, when we applied the
emissions reductions from adjacent counties, we used a |~ig/m3 per ton PIVh.sair
177
-------
quality ratio that was four times less responsive than the ratio used when applying
in-county emissions reductions (i.e., we applied four tons of PM2.5 emissions
reductions from an adjacent county for one ton of emissions reduction needed in a
given county).
2. West (29 counties) and California (36 counties) - Because these counties are
generally large and the terrain is complex, we only identified potential PM2.5
emissions reductions within each county.
We identified end-of-pipe and area source controls using the EPA's Control Strategy
Tool (CoST) (U.S. EPA, 2019a) and the control measures database.6'7 CoST estimates
emissions reductions and engineering costs associated with end-of-pipe control
technologies and area source controls applied to non-electric generating unit (non-EGU)
point, non-point (area), residential wood combustion, and area fugitive dust sources of air
pollutant emissions by matching controls to emissions sources by source classification
code (SCC). For these control strategy analyses, to maximize the number of emissions
sources included we applied controls to emissions sources with greater than 5 tons per
year of PM2.5 emissions at a marginal cost threshold of up to a $160,000/ton. Figure 3-4
presents estimated PM2.5 emissions reductions for 5 tons per year (tpy), 10 tpy, 15 tpy, 25
tpy, and 50 tpy emissions unit/source sizes up to the $160,000/ton marginal cost
threshold; the figure includes all emissions inventory and control technology data for the
counties in the analysis. We selected the $160,000/ton marginal cost threshold because it
is around that cost level that (i) road paving controls get selected and applied (as seen by
the slight uptick in the curves), and (ii) opportunities for additional emissions reductions
diminish (as seen by the flattening of the curve around that cost threshold). While the 2012
PM NAAQS RIA used a $20,000/ton marginal cost threshold and a 50 tpy emissions source
size threshold, this analysis uses a higher cost per ton threshold and a lower source size
threshold in recognition of the challenges that some areas will experience in identifying
6 More information about CoST and the control measures database can be found at the following link:
https://www.epa.gov/economic-and-cost-analysis-air-pollution-regulations/cost-analysis-modelstools-air-
pollution.
7 The 2032 emissions inventory data, the CoST run results, the CMDB, and the R code that processed these
data to prepare the summaries in Chapters 3 and 4 are in the docket and available upon request with the
Docket Office.
178
-------
controls to meet the current, revised, and alternative standard levels analyzed (U.S. EPA,
2012). The estimated costs of the control applications are presented in Chapter 4.
For counties in the northeast and southeast, because we identified controls and
reductions from adjacent or neighboring counties, we identified controls and emissions
reductions in 2 rounds. Note that a county can be both a core/home county and an
adjacent/neighboring county. In round 1 we identified controls and reductions in the home
counties for application in the home counties. In round 2 we identified controls and
reductions in neighboring counties for application in home counties that still needed
reductions. If a county was both a home county and a neighboring county in round 1,
before round 2 we adjusted any potential remaining reductions needed for a home county
to account for reductions that were applied in round 1 for application in its neighboring
county. In addition, in some cases more emissions reductions are selected by CoST than
may be needed for some areas to meet the revised and alternative standard levels. There
are two primary reasons this may occur. First, because CoST employs a least cost algorithm
to determine the bundle of controls that achieves the required emissions reductions at the
lowest possible cost, there are instances when a non-point or area fugitive dust source will
be selected for control due to its cost-effectiveness. Because the emissions from these
sources are summarized at the county level and the controls specify a percent reduction,
selection of these sources for control can sometimes lead to overshooting the emissions
reduction target.
Second, for counties in the northeast and southeast, we considered emissions
reductions from adjacent counties. There are some instances where a neighboring county
may be adjacent to multiple counties that need reductions. Furthermore, it is sometimes
the case that one of the multiple counties to which a neighboring county is adjacent needs
substantially more reductions than the other counties. In these cases, an adjacent
(neighboring) county may be called upon to provide reductions to help the county that
needs the most reductions, and in so doing it may cause the other counties to which it is
adjacent to overshoot their emissions reductions targets.
179
-------
100,000
80,000
0-P . . . . . . . . . . . .
$0 $20,000 $40,000 $60,000 $80,000 $100,000 $120,000 $140,000 $160,000 $180,000 $200,000 $220,000 $240,000
CPT Threshold
Figure 3-4 PM2.5 Emissions Reductions and Costs Per Ton (CPT) in 2032 (tons,
2017$)
We identified end-of-pipe and area source controls for non-electric generating unit
point sources (non-EGU point, oil & gas point), non-point (area) sources, residential wood
combustion sources, and area fugitive dust emissions. Controls applied for the analyses of
the current standards of 12/35 |~ig/m3 and the annual and 24-hour PM2.5 revised and
alternative standard levels of 10/35 |~ig/m3,10/30 |~ig/m3, 9/35 |~ig/m3, and 8/35 ng/m3 are
listed in Table 3-3 by emissions inventory sector, with an "X" indicating which controls
were applied in analyzing each standard level. See Appendix 3A, Table 3A.1 for a more
detailed presentation of controls applied for the revised and alternative standard levels
both by geographic area and by emissions inventory sector, as well as a discussion of some
of the controls. The non-EGU point source control Install new drift eliminator was applied at
different rule penetration (RP) rates depending on the reductions needed in particular
areas at different standard levels. In addition, with the exception of the Substitute chipping
for burning and Watering (Agriculture - Crops and Livestock Dust) controls, the non-point
180
-------
(area) source, residential wood combustion, and area source fugitive dust controls were
applied at different RP rates depending on the reductions needed in particular areas at
different standard levels. RP is the percent of the area source inventory emissions that the
control is applied to at a specified percent control efficiency. The controls were applied at
between 5 percent and 35 percent RP at 5 percent increments.
Non-EGU point source controls are applied to individual point sources. Non-point
(area), residential wood combustion, and area fugitive dust emissions data are generated at
the county level, and therefore controls for these emissions inventory sectors were applied
at the county level.8 End-of-pipe and area source controls were applied to non-EGU point,
non-point (area), residential wood combustion, and area fugitive dust sources of PM2.5
emissions including: industrial, commercial, and institutional boilers; industrial processes
located in the cement manufacturing, chemical manufacturing, pulp and paper, mining,
ferrous and non-ferrous metals, and refining industries; commercial cooking; residential
wood combustion; and fugitive construction, agriculture, and road dust.
8 The point source emissions inventories used may not accurately reflect existing controls on emissions units,
and we may inadvertently apply a control to a source that already has that, or a different, control. The non-
point emissions inventories may not accurately reflect the portion of existing emissions sources that have
already installed some non-point (area) source controls. As such, we may be applying controls that are
already accounted for in the underlying emissions inventory.
181
-------
Table 3-3 By Inventory Sector, Controls Applied in Analyses of the Current
Standards and the Revised and Alternative Primary Standard Levels
Inventory
Sector
Control Technology
12/35
10/35
10/30
9/35
8/35
Non-EGU Point
Electrostatic Precipitator-All Types
X
X
Fabric Filter-All Types
X
X
X
X
X
Install new drift eliminator
X
X
X
Venturi Scrubber
X
X
X
X
X
Oil & Gas Point
Fabric Filter-All Types
X
X
X
X
Non-Point
Add-on Scrubber
X
[Area]
Annual tune-up
X
X
X
X
X
Biennial tune-up
X
X
X
X
X
Catalytic oxidizers
X
X
X
X
Electrostatic Precipitator
X
X
X
X
X
HE PA filters
X
X
Smokeless Broiler
X
X
X
X
X
Substitute chipping for burning
X
X
X
X
X
Residential
Convert to Gas Logs
X
X
X
X
X
Wood
EPA Phase 2 Qualified Units
X
X
X
X
Combustion
EPA-certified wood stove
X
X
X
Install Cleaner Hydronic Heaters
X
X
X
X
X
Install Retrofit Devices
X
X
New gas stove or gas logs
X
X
X
X
X
Area Source
Chemical Stabilizer
X
X
X
X
Fugitive Dust
Pave Unpaved Roads
X
X
X
X
X
Pave existing shoulders
X
X
X
X
X
Watering
X
X
X
X
X
Note: For residential wood combustion emissions, we did not apply wood stove removal and burn ban
controls in the control strategy analyses because we did not want to potentially eliminate or ban any lone
home heating source without replacing a heating source.
3.2.3 Estimates of PM2.5 Emissions Reductions Resulting from Applying Control
Technologies
By area, Table 3-4 includes a summary of the estimated emissions reductions from
control applications for the revised and alternative standards analyzed. These emissions
reductions were used to create the PM2.5 spatial surfaces described in Appendix 2A, Section
2A.4.2 for the human health benefits assessments presented in Chapter 5.
182
-------
Table 3-4 Summary of PM2.5 Estimated Emissions Reductions from CoST by Area
for the Revised and Alternative Primary Standard Levels of 10/35 (j,g/-
m3,10/30 |ig/m3,9/35 |ig/m3, and 8/35 jig/m3 in 2032 (tons/year)
PM2.5 Emissions Reductions
Area
10/35
10/30
9/35
8/35
Northeast
1,032
1,074
7,226
14,036
Northeast (Adjacent Counties)
0
0
2,599
11,911
Southeast
521
521
1,959
13,995
Southeast (Adjacent Counties)
45
45
354
3,086
West
470
2,715
1,386
5,323
CA
3,010
4,652
5,069
7,181
Total
5,078
9,006
18,592
55,532
Note: Totals may not match related tables due to independent rounding. In the northeast and southeast
when we applied the emissions reductions from adjacent counties, we used a ppb/ton PM2.5 air quality
ratio that was four times less responsive than the ratio used when applying in-county emissions
reductions.
By emissions inventory sector, Table 3-5 includes a summary of PM2.5 emissions and
estimated emissions reductions from control applications for the revised and alternative
standard levels analyzed. The PM2.5 emissions in Table 3-5 are the total emissions
associated with the emissions units/sources that get controls applied within each of the
inventory sectors for each of the alternative standard levels (not the total emissions
associated with the entire inventory sector). Across the revised and alternative standard
levels analyzed, overall total emissions reductions are approximately 28 to 35 percent of
the PM2.5 emissions from the sources selected by CoST for control. In general, a large
percentage of the emissions are being controlled for the revised and alternative standard
levels analyzed, while different inventory sectors are selected for control in different areas
and additional reductions may be possible in some areas.
The emissions inventory sector with the highest percent of emissions reductions
relative to total potentially controllable emissions for that sector is the non-EGU point
sector - the estimated emissions reductions are between 88 and 99 percent of total PM2.5
emissions from the sources selected for control, with that percent decreasing as the
alternative standard level gets more stringent. The emissions inventory sector with the
lowest percent of emissions reductions relative to total potentially controllable emissions
for that sector is the area fugitive dust sector - the estimated emissions reductions are
between 18 and 27 percent of total PM2.5 emissions from the sources selected for control,
183
-------
with that percent decreasing as the alternative standard level gets more stringent. The
residential wood combustion sector's emissions reductions relative to total potentially
controllable emissions are between 31 and 34 percent across the revised and alternative
standard levels analyzed. It is worth noting that the control efficiencies associated with
area source controls for the non-point (area), area fugitive dust, and residential wood
combustion sectors are generally lower than control efficiencies associated with end-of-
pipe controls for the non-EGU point and oil and gas point inventory sectors, and many of
the controls for these sectors are only applied to a portion of the inventory. For the revised
standard levels of 9/35 |~ig/m3, the inventory sectors with the most potentially controllable
emissions are the non-point (area) and area fugitive dust sectors. The inventory sectors
with the most estimated emissions reductions are the non-point (area) and non-EGU point
sectors.
184
-------
Table 3-5 Summary of PM2.5 Emissions and Estimated Emissions Reductions
from CoST by Inventory Sector for Revised and Alternative Primary
Standard Levels of 10/35 |ig/m3,10/30 |ig/m3, 9/35 |ig/m3, and 8/35
(ig/m3 in 2032 (tons/year)
Emissions Inventory
Sector
10/35
10/30
9/35
8/35
Non-EGU Point
PM2.5 Emissions
940
1,288
4,887
14,136
PM2.5 Emissions
Reductions
928
1,269
4,277
12,534
Oil & Gas Point
PM2.5 Emissions
5
5
5
21
PM2.5 Emissions
Reductions
5
5
5
21
Non-Point (Area)
PM2.5 Emissions
7,379
10,794
27,286
70,847
PM2.5 Emissions
Reductions
2,234
3,264
8,766
25,947
Residential Wood
Combustion
PM2.5 Emissions
1,127
2,730
5,298
12,994
PM2.5 Emissions
Reductions
378
872
1,636
4,190
Area Source Fugitive
Dust
PM2.5 Emissions
5,676
16,866
15,608
71,086
PM2.5 Emissions
Reductions
1,533
3,596
3,909
12,840
Total
PM2.5 Emissions
15,127
31,683
53,085
169,086
PM2.5 Emissions
Reductions
5,078
9,006
18,592
55,532
Note: The PM2.5 emissions in the table are for the emissions sources that get controls applied within
each of the inventory sectors (not the total emissions associated with the entire inventory sector)
for each of the standard levels.
By emissions inventory sector and by control technology, Table 3-6 includes a
summary of estimated PM2.5 emissions reductions from control applications for the revised
and alternative standard levels analyzed. Across standard levels analyzed, estimated PM2.5
emissions reductions from control applications in the (i) non-EGU point and oil and gas
point inventory sectors account for between 14 and 23 percent of estimated reductions; (ii)
non-point (area) inventory sector account for between 36 and 47 percent of estimated
reductions; (hi) residential wood combustion inventory sector account for between 7 and
185
-------
10 percent; and (iv) area fugitive dust inventory sector account for between 21 and 40
percent.
Also, across standard levels analyzed, six end-of-pipe and area source controls
comprise between approximately 76 and 90 percent of the estimated emissions reductions.
Those controls include:
• Fabric Filter- All Types (non-EGU point inventory sector) - the control
technology is generally applied to industrial, commercial, and institutional
boilers and industrial processes located in the cement manufacturing,
chemical manufacturing, pulp and paper, mining, ferrous and non-ferrous
metals, and refining industries.
• Electrostatic Precipitator (non-point (area) inventory sector) - the control is
applied to area source commercial cooking emissions.
• Substitute Chipping for Burning (non-point (area) inventory sector) - the
control is applied to area source waste disposal emissions.
• Convert to Gas Logs (residential wood combustion inventory sector) - the
control is applied to area source residential wood combustion emissions.
• Pave Existing Shoulders (area fugitive dust inventory sector) - the control is
applied to road dust emissions.
• Pave Unpaved Roads (area fugitive dust inventory sector) - the control is
applied to road dust emissions.
The three controls that result in the most emissions reductions for revised and alternative
standard levels of 10/35 |~ig/m3, 9/35 |~ig/m3, and 8/35 |~ig/m3 are Fabric Filter- All Types,
Electrostatic Precipitator, and Substitute Chipping for Burning. The three controls that
result in the most emissions reductions for alternative standard levels of 10/30 |~ig/m3 are
Electrostatic Precipitator, Substitute Chipping for Burning, and Pave Unpaved Roads.
186
-------
Table 3-6 Summary of Estimated Emissions Reductions from CoST by Inventory
Sector and Control Technology for Revised and Alternative Primary
Standard Levels of 10/35 |ig/m3,10/30 |ig/m3, 9/35 |ig/m3, and 8/35
(ig/m3 in 2032 (tons/year)
Inventory
Sector
Control Technology
10/35
10/30
9/35
8/35
Non-EGU Point
Electrostatic Precipitator-All Types
0
0
24
0
Fabric Filter-All Types
923
1,222
3,444
10,748
Install new drift eliminator
0
0
171
512
Venturi Scrubber
5
47
638
1,274
Oil & Gas Point
Fabric Filter-All Types
5
5
5
21
Non-Point
Add-on Scrubber
0
0
0
4
(Area)
Annual tune-up
58
121
825
1,940
Biennial tune-up
136
160
24
198
Catalytic oxidizers
0
75
318
99
Electrostatic Precipitator
1,154
1,362
3,633
8,721
HE PA filters
0
0
1
0
Smokeless Broiler
74
79
437
909
Substitute chipping for burning
811
1,466
3,528
14,075
Residential
Convert to Gas Logs
357
674
1,212
2,572
Wood
EPA Phase 2 Qualified Units
0
78
13
197
Combustion
EPA-certified wood stove
1
1
0
0
Install Cleaner Hydronic Heaters
10
65
178
654
Install Retrofit Devices
0
0
52
33
New gas stove or gas logs
9
55
180
733
Area Source
Chemical Stabilizer
0
610
234
4,737
Fugitive Dust
Pave Unpaved Roads
838
1,493
1,364
2,185
Pave existing shoulders
468
734
1,655
4,056
Watering
226
759
656
1,862
Total
5,078
9,006
18,592
55,532
By emissions inventory sector and by inventory source classification code (SCC)
sector, Table 3-7 includes a summary of estimated PM2.5 emissions reductions from control
applications for the revised and alternative standard levels analyzed. As seen in Table 3-6,
across standard levels analyzed, estimated PM2.5 emissions reductions from control
applications in the (i) non-EGU point and oil and gas point inventory sectors account for
between 14 and 23 percent of estimated reductions; (ii) non-point (area) inventory sector
account for between 36 and 47 percent of estimated reductions; (iii) residential wood
combustion inventory sector account for between 7 and 10 percent; and (iv) area fugitive
dust inventory sector account for between 21 and 40 percent.
187
-------
Across standard levels analyzed, estimated PM2.5 emissions reductions from control
applications in the Industrial Processes - Cement Manufacturing, Industrial Processes -
Ferrous Metals, Industrial Processes - Not Elsewhere Classified, and Industrial Processes -
Petroleum Refineries inventory SCC sectors account for between 48 percent and 69 percent
of reductions from the non-EGU point and oil and gas point inventory sectors. Estimated
PM2.5 emissions reductions from control applications in the Commercial Cooking and Waste
Disposal - All Categories inventory SCC sectors account for between 82 percent and 90
percent of reductions from the non-point (area) inventory sector. Estimated PM2.5
emissions reductions from control applications in the Fuel Combustion - Residential - Wood
inventory SCC sector account for all of the reductions from the residential wood
combustion inventory sector, and estimated PM2.5 emissions reductions from control
applications in the Agriculture - Crops & Livestock Dust, Dust - Paved Road Dust, and Dust -
Unpaved Road Dust inventory SCC sectors account for all of the reductions from the area
source fugitive dust inventory sector.
Table 3-7 Summary of Estimated PM2.5 Emissions Reductions from CoST by
Inventory Source Classification Code Sectors for Revised and
Alternative Primary Standard Levels of 10/35 |ig/m3,10/30 |ig/m3,
9/35 (ig/m3, and 8/35 (ig/m3 in 2032 (tons/year)
Sector
SCC Sector
10/35
10/30
9/35
8/35
Fuel Combustion -
Commercial/Institutional Boilers - Natural
0
0
0
5.2
Gas
Fuel Combustion -
Commercial/Institutional Boilers - Other
47.6
47.6
47.6
47.6
Fuel Combustion - Industrial Boilers, ICEs
- Biomass
0
0
14.7
316.6
Fuel Combustion - Industrial Boilers, ICEs
- Natural Gas
0
0
23.8
109.0
Fuel Combustion - Industrial Boilers, ICEs
-Oil
0
0
0
11.5
Fuel Combustion - Industrial Boilers, ICEs
- Other
96.6
96.6
379.3
762.0
Industrial Processes - Cement
Manufacturing
149.1
149.1
469.1
581.9
Industrial Processes - Chemical
Manufacturing
74.6
80.0
222.7
576.1
Industrial Processes - Ferrous Metals
0
0
532.2
1,675.4
Industrial Processes - Mining
0
0
0
249.1
Industrial Processes - Non-ferrous Metals
0
13.0
195.4
724.3
188
-------
Sector
SCC Sector
10/35
10/30
9/35
8/35
Industrial Processes - Not Elsewhere
Classified
210.1
294.1
1,525.7
4,348.0
Industrial Processes - Petroleum
Refineries
164.1
164.1
409.2
1,615.5
Industrial Processes - Pulp & Paper
5.3
182.8
124.6
850.4
Industrial Processes - Storage and
Transfer
111.4
172.4
182.9
506.4
Solvent - Degreasing
0
0
17.2
17.2
Solvent - Industrial Surface Coating &
Solvent Use
62.9
62.9
125.8
125.8
Waste Disposal - General Processes
6.7
6.7
6.7
11.8
Oil & Gas Point
Fuel Combustion -
Commercial/Institutional Boilers - Natural
0
0
0
16.1
Gas
Industrial Processes - Oil & Gas
5.1
5.1
5.1
5.1
Production
Non-Point
Commercial Cooking
1,228.3
1,516.6
4,388.4
9,733.5
(Area)
Fuel Combustion -
Commercial/Institutional Boilers -
16.0
22.5
116.6
252.8
Biomass
Fuel Combustion -
Commercial/Institutional Boilers - Natural
17.0
24.1
73.6
109.3
Gas
Fuel Combustion -
Commercial/Institutional Boilers - Oil
6.7
9.7
14.6
18.5
Fuel Combustion -
Commercial/Institutional Boilers - Other
0.6
0.6
0.6
0.6
Fuel Combustion - Industrial Boilers, ICEs
- Biomass
120.4
184.7
567.6
1,629.9
Fuel Combustion - Industrial Boilers, ICEs
- Coal
0.1
2.6
2.5
14.0
Fuel Combustion - Industrial Boilers, ICEs
- Natural Gas
32.3
34.2
68.0
105.8
Fuel Combustion - Industrial Boilers, ICEs
-Oil
1.7
2.5
5.7
7.5
Waste Disposal - All Categories
773.1
1,156.5
3,009.4
11,995.2
Waste Disposal - Residential
38.0
309.8
519.0
2,080.2
Residential
Wood
377.9
871.9
1,635.6
4,189.5
Combustion
Fuel Combustion - Residential - Wood
Area Source
Agriculture - Crops & Livestock Dust
226.3
758.5
656.4
1,861.7
Fugitive Dust
Dust - Paved Road Dust
468.0
733.8
1,654.6
4,055.8
Dust - Unpaved Road Dust
838.2
2,103.7
1,597.9
6,922.3
Total
5,078.1
9,006.1
18,592.5
55,531.7
189
-------
3.2.4 Estimates of PM2.5 Emissions Reductions Still Needed after Applying End-of-
Pipe and Area Source Controls
The percent of total PM2.5 emissions reductions estimated from CoST (shown in
Table 3-4 above) relative to total PM2.5 emissions reductions needed (shown in Table 3-2
above) varies by standard level and by area. Note that in the northeast and southeast when
we applied the emissions reductions from adjacent counties, we used a |~ig/m3 per ton PM2.5
air quality ratio that was four times less responsive than the ratio used when applying in-
county emissions reductions (i.e., we applied four tons of PM2.5 emissions reductions from
an adjacent county for one ton of emissions reduction needed in a given county). For the
revised standards of 9/35 |~ig/m3, for the northeast we were able to identify approximately
98 percent of the reductions needed. For the southeast we were able to identify
approximately 68 percent of the reductions needed. For the west, we were able to identify
approximately 44 percent of the reductions needed, and for California the percentage is
approximately 26 percent. For the more stringent standard levels of 8/35 |~ig/m3, for the
northeast we were able to identify approximately 84 percent of the reductions needed. For
the southeast we were able to identify approximately 81 percent of the reductions needed.
For the west, we were able to identify approximately 52 percent of the reductions needed,
and for California the percentage is approximately 23 percent.9
The higher percent of estimated emissions reductions relative to needed reductions
in the northeast and southeast is likely because as the standard level becomes more
stringent, more controls from counties projected to exceed and their adjacent counties are
available and applied. See Appendix 3A, Table 3A-2 through Table 3A-7 for more detailed
summaries of PM2.5 emissions reductions by county for the revised and alternative
standard levels for the northeast, the adjacent counties in the northeast, the southeast, the
adjacent counties in the southeast, the west, and California. Table 3A-7 for California
presents the counties organized by air districts.
9 Compared to standard levels of 9/35 \xg/m:i, the more stringent alternative standard levels of 8/35 |_ig/m:i
include more counties that need reductions, a different set of emission reduction targets, and in the NE and
SE more neighbor/adjacent counties. When these additional counties are included and emissions reductions
in these counties are identified, the percentages may increase, which is happening in the SE and W for the
more stringent standard levels of 8/35 |J.g/m3 compared to the revised standard levels of 9/35 \.ig/m:i.
190
-------
As indicated, the estimated PM2.5 emissions reductions from control applications do
not fully account for all the emissions reductions needed to reach the revised and less and
more stringent alternative standard levels in some counties in the northeast, southeast,
west, and California. By area, Table 3-8 includes a summary of the estimated emissions
reductions still needed after control applications for the revised and alternative standard
levels analyzed. By area and by county, Table 3-9 includes a more detailed summary of the
estimated emissions reductions still needed after control applications for the revised and
alternative standard levels analyzed. As seen in Table 3-9, some counties need emissions
reductions to meet a standard level of 10/30 |~ig/m3 that did not need emissions reductions
to meet a standard level of 10/35 |~ig/m3. These counties are in the west and California,
where there are small valleys with mountainous terrain and wintertime inversions, along
with residential woodsmoke emissions and some wildfire influence, and need emissions
reductions to meet the more stringent 24-hour standard level of 30 |~ig/m3. Figure 3-5
through Figure 3-8 show the counties that still need emissions reductions after control
applications for the revised and alternative standards analyzed.
The analysis indicates that counties in the northeast and southeast U.S. do not need
additional emissions reductions to meet alternative standard levels of 10/35 |~ig/m3 and
10/30 ng/m3. For the northeast, 2 (out of 12) counties need additional emissions
reductions to reach attainment of the revised alternative standard levels of 9/35 |~ig/m3,
and 5 (out of 31) counties need additional emissions reductions to reach attainment of the
more stringent alternative standard levels of 8/35 |~ig/m3. For the southeast, 2 (out of 7)
counties need additional emissions reductions to reach attainment of the revised
alternative standard levels of 9/35 |~ig/m3, and 5 (out of 33) counties need additional
emissions reductions to reach attainment of the more stringent alternative standard levels
of 8/35 ng/m3.
The analysis also indicates that counties in the west and California need additional
emissions reductions after the application of controls to meet all of the standard levels. For
the west, 2 (out of 2) counties need additional emissions reductions to reach attainment of
the less stringent alternative standard levels of 10/35 ng/m3,10 (out of 19) counties need
additional emissions reductions to reach attainment of the more stringent alternative
191
-------
standard levels of 10/30 |~ig/m3, 4 (out of 10) counties need additional emissions
reductions to reach attainment of the revised standard levels of 9/35 |~ig/m3, and 10 (out of
21) counties need additional emissions reductions to reach attainment of the more
stringent alternative standard levels of 8/35 |~ig/m3. For California, 12 (out of 13) counties
need additional emissions reductions to reach attainment of the less stringent alternative
standard levels of 10/35 |~ig/m3, 20 (out of 25) counties need additional emissions
reductions to reach attainment of the more stringent alternative standard levels of 10/30
Hg/m3,17 (out of 23) counties need additional emissions reductions to reach attainment of
the revised alternative standard levels of 9/35 |~ig/m3, and 24 (out of 32) counties need
additional emissions reductions to reach attainment of the more stringent alternative
standard levels of 8/35 ng/m3.
Table 3-8 Summary of PM2.5 Emissions Reductions Still Needed by Area for the
Revised and Alternative Primary Standard Levels of 10/35 (j,g/m3,
10/30 ng/m3,9/35 (ig/m3, and 8/35 (ig/m3 in 2032 (tons/year)
Area
10/35
10/30
9/35
8/35
Northeast
0
0
130
3,285
Southeast
0
0
1,038
3,519
West
516
3,959
1,747
4,982
CA
7,739
11,986
14,411
24,366
Total
8,255
15,945
17,327
36,152
Table 3-9 Summary of PM2.5 Emissions Reductions Still Needed by Area and by
County for the Revised and Alternative Primary Standard Levels of
10/35 |ig/m3,10/30 |ig/m3, 9/35 (j,g/m3, and 8/35 |ig/m3 in 2032
(tons/year)
Area
Area Name
10/35
10/30
9/35
8/35
Northeast
Marion County, IN
0
0
0
496
Bergen County, NJ
0
0
75
807
Camden County, NJ
0
0
0
535
Hamilton County, OH
0
0
55
784
Delaware County, PA
0
0
0
664
Southeast
District Of Columbia
0
0
0
69
Caddo Parish, LA
0
0
0
374
West Baton Rouge Parish, LA
0
0
0
0
Cameron County, TX
0
0
351
1,168
El Paso County, TX
0
0
0
402
Hidalgo County, TX
0
0
687
1,505
West
Maricopa County, AZ
0
0
0
4
Pinal County, AZ
0
162
0
0
192
-------
Area
Area Name
10/35
10/30
9/35
8/35
Santa Cruz County, AZ
0
0
0
274
Adams County, CO
0
0
0
161
Denver County, CO
0
0
0
353
Benewah County, ID
0
238
0
0
Lemhi County, ID
0
557
235
703
Shoshone County, ID
90
410
558
1,026
Lewis and Clark County, MT
0
523
0
173
Lincoln County, MT
426
426
894
1,361
Harney County, OR
0
0
0
398
Klamath County, OR
0
449
60
528
Lake County, OR
0
575
0
0
Okanogan County, WA
0
44
0
0
Yakima County, WA
0
575
0
0
CA
Alameda County, CA
0
0
34
351
Calaveras County, CA
0
0
50
367
Colusa County, CA
0
502
0
0
Fresno County, CA
124
124
441
758
Imperial County, CA
1,665
1,665
2,516
3,366
Kern County, CA
409
409
726
1,043
Kings County, CA
413
413
730
1,047
Los Angeles County, CA
599
599
1,450
2,300
Madera County, CA
0
0
38
355
Mendocino County, CA
0
109
0
0
Merced County, CA
22
22
339
656
Mono County, CA
0
502
0
173
Napa County, CA
0
0
0
203
Orange County, CA
329
329
1,179
2,030
Plumas County, CA
309
502
626
943
Riverside County, CA
1,701
1,701
2,551
3,402
Sacramento County, CA
0
0
0
204
San Bernardino County, CA
1,625
1,625
2,475
3,326
San Diego County, CA
0
0
0
326
San Francisco County, CA
0
0
0
91
San Joaquin County, CA
0
0
48
365
Santa Barbara County, CA
0
1,082
0
0
Siskiyou County, CA
0
138
0
0
Solano County, CA
0
0
0
283
Stanislaus County, CA
97
97
414
731
Sutter County, CA
0
105
31
348
Tehama County, CA
0
170
0
0
Tulare County, CA
446
446
763
1,080
Ventura County, CA
0
1,447
0
619
Total
8,255
15,945
17,327
36,152
Note: The table includes only those counties that still need reductions (e.g., in the Northeast there were 31
counties that needed emissions reductions, and only the 5 counties still need emissions reductions for an
alternative standard level of 8/35 |ag/m3].
193
-------
I Counties with Sufficient Identified Reductions to Meet 10/35
I Counties Still Needing Reductions to Meet 10/35
Figure 3-5 Counties that Still Need PM2.5 Emissions Reductions for Less Stringent
Alternative Standard Levels of 10/35 j-ig/m3
| Counties with Sufficient Identified Reductions to Meet 9/35
B Counties Still Needing Reductions to Meet 9/35
Figure 3-6 Counties that Still Need PM2.5 Emissions Reductions for Revised
Standard Levels of 9/35 ng/rn3
194
-------
|( Counties with Sufficient Identified Reductions to Meet 8/35
I Counties Still Needing Reductions to Meet 8/35
Figure 3-7 Counties that Still Need PM2.5 Emissions Reductions for More Stringent
Alternative Standard Levels of 8/35 (ig/m3
| Counties with Sufficient Identified Reductions to Meet 10/30
I Counties Still Needing Reductions to Meet 10/30
Figure 3-8 Counties that Still Need PM2.5 Emissions Reductions for More Stringent
Alternative Standard Levels of 10/30 (ig/m3
195
-------
Table 3A-8 in Appendix 3A includes information on counties that still need
emissions reductions and remaining PM2.5 emissions by emissions inventory sector in those
and adjacent counties. For some counties still needing emissions reductions, additional
controls from the CMDB (listed below), rules listed in Table 3-10, or voluntary measures
identified in Table 3-11 may provide information on potential additional area source
controls or policies to consider that are beyond the scope of this analysis.
As noted, we did not apply some area source controls from the CMDB in the control
strategy analyses. By emissions inventory sector, the controls include those identified
below. Note that the X in the control measure abbreviation is the rule penetration rate.
• Residential Wood Combustion Emissions (Fireplaces, Hydronic Heaters,
Wood Stoves)
o Remove and destroy old wood stoves (control measure abbreviation:
PBBFPHHWDSX)
o Curtailment program, Burn bans for fireplaces, hydronic heaters,
wood stoves (control measure abbreviation: PROWSWDSTVX)
• Construction Dust, Road Dust Controls
o Apply gravel (control measure abbreviation: PGVUNPAVEDX)
o Apply water (PWATUNPAVEDX)
o Truck system for soil moisture (PTRCONSTX)
o Trackout control devices (PTCDPAVEDX)
o Gravel bed trackout (PGBTPAVEDX)
o Dust control plan (control measure abbreviation: PDCPCONSTX)
o Pave interior roads (control measure abbreviation: PPIPAVEDX)
o Pipe grid trackout (control measure abbreviation: PPGTPAVEDX)
o Sprinkler system for soil moisture (control measure abbreviation:
PSPCONSTX)
196
-------
Table 3-10 below lists current rules from several California air districts that could
potentially address some of the remaining PM2.5 emissions in some areas in some inventory
sectors. Table 3-10 does not necessarily reflect additional measures. As shown in Table 3-3,
in the control strategy analyses we apply some area source controls that may be required
by or used to comply with some of the rules in Table 3-10 (e.g., catalytic oxidizers on
charbroilers and residential wood combustion controls). However, because we do not
know the portion of existing emissions sources that have already installed some non-point
(area) source controls, the control strategy analyses limit the percent of the non-point
(area), residential wood combustion, or area fugitive dust inventory emissions that the
area source controls are applied to (e.g., EPA Phase 2 Qualified Units at 35% RP). Some of
the non-point (area) source controls applied in the control strategy analyses could apply to
a higher percentage of the applicable inventory.
In addition, Table 3-11 includes voluntary measures identified in the 2014 Houston-
Galveston Area Council (H-GAC) Advance Plan for PIVh.sthat could potentially address some
of the remaining PM2.5 emissions in some areas (HGAC, 2014).10
Table 3-10 Current Rules from Several California Air Districts and City of Portola
for Area Fugitive Dust Emissions, Non-point Source Emissions, and
Residential Wood Combustion Emissions
Air District or City
Rule
Area Fugitive Dust Emissions
(Construction Activities, Unpaved Roads, Paved Roads)
San Joaquin Valley Air
Pollution Control District
Regulation VIII: 8011 General Requirements; 8021 Construction, Demolition, Excavation,
Extraction, and Other Earthmoving Activities; 8031 Bulk Materials; 8041 Carryout and
Trackout; 8051 Open Areas; 8061 Paved and Unpaved Roads; 8071 Unpaved
Vehicle/Equipment Traffic Areas; and 8081 Agricultural Sources
South Coast Air Quality
Management District
Rule 403 Fugitive Dust
-Applies to all construction activity sources listed in Table 1 of Rule, including but not
limited to disturbed soil, road shoulder maintenance, truck loading, unpaved
roads/parking lots, and vacant land.
-Requires that no person conduct active operations without using the applicable best
available control measures included in Table 1 of Rule to minimize fugitive dust emissions
from each fugitive dust source type within the active operation.
10 The H-GAC developed the Path Forward in partnership with the Regional Air Quality Planning Advisory
Committee [RAQPAC], as part of H-GAC's participation in the voluntary EPA Particulate Matter [PM]
Advance Program. H-GAC partnered with local and regional government agencies, citizen and
environmental groups, business and industry-based organizations, and other stakeholders to proactively
pursue air quality improvements within the region.
197
-------
Air District or City
Rule
Ventura County Air
Pollution Control District
Rule 55 Fugitive Dust
-Applies to any operation, disturbed surface area, or man-made condition capable of
generating fugitive dust, including bulk material handling, earth-moving, construction,
demolition, storage piles, unpaved roads, track-out, or off-field agricultural operations.
-Rule has visible dust beyond property line, opacity, and track-out requirements for the
above operations.
Non-point (Area) Source Emissions
(Commercial Cooking, Waste Disposal, Cooling Towers, Small Boilers)
San Joaquin Valley Air
Pollution Control District
Rule 4692 Commercial Charbroiling
-Cannot operate a chain-driven charbroiler unless the chain-driven charbroiler is
equipped and operated with a catalytic oxidizer. Requires a control efficiency of at least
83% for PMio emissions and a control efficiency of at least 86% for VOC emissions. Chain-
driven charbroiler/catalytic oxidizers combinations certified by the South Coast Air
Quality Management District are considered compliant for the purposes of this section.
Rule 4103 Open Burning
-The Air Pollution Control Officer allocates burning based on the predicted meteorological
conditions and whether total tonnage to be emitted would allow the volume of smoke and
other contaminants to cause a public nuisance, impact smoke sensitive areas, or create or
contribute to an exceedance of an ambient air quality standard. Except as otherwise
provided, no person should set, permit, or use an open outdoor fire for the purpose of
disposal or burning of petroleum wastes; demolition or construction debris; residential
rubbish; garbage or vegetation; tires; tar; trees; wood waste; or other combustible or
flammable solid, liquid or gaseous waste; or for metal salvage or burning of motor vehicle
bodies.
South Coast Air Quality
Management District
Rule 1138 Control of Emissions from Restaurant Operations
-New and existing chain-driven charbroilers should be equipped and operated with a
catalytic oxidizer control device.
Rule 444 Open Burning
-Cannot conduct open burning unless all of the following are met: (i] it is a permissive
burn day or a marginal burn day on which burning is permitted in the applicable
source/receptor area; (ii] the Executive Officer or the applicable fire protection agency has
issued a written permit for the burn; (iii] the Executive Officer has authorized the burn by
issuing a Burn Authorization Number for each day for each open burning event; and (iv] all
site-specific permit conditions are met.
Ventura County Air
Pollution Control District
Rule 74.25 Restaurant Cooking Operations
-The owner or operator of a conveyorized charbroiler should reduce both reactive organic
compound emissions and particulate matter emissions by at least 83 percent using an
emission control device certified pursuant to rule.
Rule 56 Open Burning
-Requires a valid permit of specified substances for use on Burn Days.
Bay Area Air Quality
Management District
Regulation 6 Rule 2 Commercial Cooking Equipment
-For chain-driven charb roilers, no person should operate a chain-driven charbroiler
unless it is equipped and operated with a certified catalytic oxidizer certified for use in
combination with the specific model of chain-driven charbroiler by limiting the PMio and
organic compound emissions to no more than 1.3 pounds of PMio and 0.32 pounds of
organic compounds per 1,000 pounds of beef cooked.
-For under-fired charbroilers, no person should operate an under-fired charbroiler in any
restaurant that contains one or more underfired charbroilers with an aggregate grill
surface area of 10 square feet or more, unless emissions from each under-fired charbroiler
198
-------
Air District or City
Rule
are exhausted through a certified control device certified as limiting the PMio emissions of
the under-fired charbroiler to no more than 1.0 pounds of PMio per 1000 pounds of beef
cooked.
Regulation 5 Open Burning
-Forbids open burning with certain exceptions.
Residential Wood Combustion Emissions
(Fireplaces, Hydronic Heaters, Wood Stoves}
San Joaquin Valley Air
Pollution Control District
Rule 4901 Wood Burning Fireplaces, Wood Burning Heaters
-Requirements for new wood burning heaters and used wood burning heaters.
Requirements for property sales and certifications. Requirements for number of wood
burning devices in new single or multi-family housing units.
South Coast Air Quality
Management District
Rule 445 Wood-burning Devices
-No person can permanently install a wood-burning device into any new development.
-No person can sell, offer for sale, supply, or install, a new or used permanently installed
indoor or outdoor wood-burning device or gaseous-fueled device unless it is one of the
following: U.S. EPA Certified wood-burning heater; pellet-fueled wood-burning heater;
masonry heater; or dedicated gaseous-fueled fireplace.
-No person can burn any product not intended for use as fuel in a wood-burning device
including, but not limited to, garbage, treated wood, particle board, plastic products,
rubber products, waste petroleum products, paints, coatings or solvents, or coal.
-No person can operate an indoor or outdoor wood-burning device, portable outdoor
wood-burning device, or wood-fired cooking device on a calendar day during the
woodburning season for PM2.5 declared by the Executive Officer to be a mandatory wood-
burning curtailment (No-Burn] day based on the specified geographic area below 3,000
feet above mean sea level and applicable daily PIVh.sair quality forecast.
Bay Area Air Quality
Management District
Regulation 6 Rule 3 Wood-burning Devices
-Burning Prohibited During Mandatory Burn Ban
-Requirements for Wood Heater Manufacturers and Retailers: no manufacturer or retailer
should advertise, sell, offer for sale or resale, supply, install or transfer a new or used
wood-burning device intended for use within District boundaries unless the device meets
or exceeds the requirements of Title 40 Code of Federal Regulations, Part 60, Subpart AAA.
-Sale, Resale, Transfer, or Installation of Wood-Burning Devices: no person should
advertise, sell, offer for sale or resale, supply, install or transfer a new or used wood-
burning device intended for use within District boundaries unless the device meets or
exceeds the requirements of Title 40 Code of Federal Regulations, Part 60, Subpart AAA.
Does not apply if a wood-burning device is an installed fixture included in the sale or
transfer of any real property.
-Disclosure Requirements for Real Property: any person selling, renting, or leasing real
property should provide sale or rental disclosure documents that describe the health
hazards of PM2.5 from burning wood or any solid fuel as a source of heat. Disclosure
documents must disclose PM2.5 health hazards in accordance with guidance made available
on the District's website.
-Requirements for Rental Properties: all real property offered for lease or rent in areas
with natural gas service should have a permanently-installed form of heat that does not
burn solid fuel.
-Requirements for New Building Construction: no person or builder should install a
wood-burning device in a new building construction.
199
-------
Air District or City
Rule
-Requirements for Remodeling a Fireplace or Chimney: no person should remodel a
fireplace or chimney unless a gas-fueled, electric, or EPA certified device is installed that
meets requirements in Title 40 Code of Federal Regulations, Part 60, Subpart AAA.
City of Portola
Ordinance 359 - Regulation ofWood Stoves and Fireplaces and the Prohibition of the
Open Burning of Yard Waste
-Requirements for new devices (EPA-certified], existing devices (certificate or exemption
before completing escrow for residential/commercial properties], permitted fuels,
mandatory curtailments during stagnant conditions. Outdoor wood-fired boiler
installation prohibited.
-Prohibits burning of all yard waste and debris
Table 3-11 Voluntary Measures from the Houston-Galveston Area Advance Plan
for PM
Initiative
Program or Measure
Area Fugitive Dust Emissions
Dust Suppression
TCEQ, EPA Region 6, City of Houston, Harris County Precinct 2, Port of
Houston Authority, Port Terminal Rail Authority, and local industry
partnered to address PM2.5 sources and implement dust suppression
strategies to reduce PM2.5 emissions. Strategies included:
-Pave parking lot
-Barriers to prevent trucks from driving on unpaved shoulder
-Apply emulsified asphalt to reduce dust emissions at steel yards within the
Terminal
-Cease steel loading activities in dirt area
-Implement new dust control best management practices
Mobile Source Emissions
(including Airports and Railroads)
H-GAC, Clean Vehicles and Clean
School Bus Programs
Replace older diesel engines in public and private fleets; clean school bus
projects
H-GAC, Clean Vessels for Texas
Waters
Repower high-emitting tug vessels with new, cleaner engines
H-GAC, Commute Solutions: Vanpool
Program
Regional vanpool and rideshare program
H-GAC, Regional TCEQ Emission
Reduction Plan (TERP)
Established by the 77th Texas Legislature in 2001, through enactment of
Senate Bill (SB] 5. TCEQ provides TERP funding for emission reduction
projects to participants in Texas. Projects include a number of voluntary
financial incentive programs to help improve the air quality in Texas.
Between 2008 - 2013 TCEQ regional TERP has funded over 3,200 vehicle
replacements.
City of Houston, EV Charging
Stations
Participated in US Department of Energy's EV Project to secure additional
charging stations
City of Houston, Anti-Idling Policy
Adopted anti-idling policy for municipal vehicles
200
-------
Initiative
Program or Measure
City of Houston, Houston Airport
System
-Improvements to airfield runways, taxiways, and gates/ramp reduced
aircraft taxi and idle times
- Installed gate electrification equipment so parked aircraft forego the use
of auxiliary power units
Harris County, Enhanced
Enforcement Program Smoking
Vehicles
Law enforcement personnel target high emitting vehicles, smoking vehicles,
and suspicious vehicles to verify that the state inspection certificates
attached to these vehicles are legitimate
Metropolitan Transit Authority of
Harris County, Hybrid Bus Fleet
Converted over 1/3 of the METRO bus fleet to clean-running, diesel-electric
hybrid technology
Port Authority of Houston, Cleaner
Cranes
Replaced older cranes with new cranes. The increased efficiency associated
with the cleaner, faster cranes reduces the truck idling and associated
emissions at the Port.
Port Authority of Houston, Gate
Automation, Idling Program
-Gate optical character recognition installation enabled Port to process
trucks twice as fast and reduced idling time
--Idling program in place for all landside engines at the port, including heavy
duty diesel trucks and cargo handling equipment
H-GAC, Anti-Idling
Approximately 60 percent of Union Pacific switcher engines operating in the
H-GAC area have anti-idling controls
3.2.5 Qualitative Assessment of the Remaining Air Quality Challenges and Emissions
Reductions Potentially Still Needed
The sections below discuss the remaining air quality challenges for areas in the
northeast and southeast, as well as in the west and California for the revised standard
levels of 9/35 |j,g/m3; the areas include counties with near-road monitors, counties in
border areas, counties in small western mountain valleys, and counties in California's air
basins and districts. The characteristics of the air quality challenges for these areas include
features of certain near-road sites with challenging local conditions, cross-border
transport, effects of complex terrain in the west and California, and identifying wildfire
influence on projected PM2.5 DVs that could potentially qualify for exclusion as atypical,
extreme, or unrepresentative events (U.S. EPA, 2019b).
Consistent with Chapter 2, Section 2.4, to discuss the remaining air quality
challenges for the revised standard levels of 9/35 |~ig/m3, we group counties into the
following "bins": near-road monitors, border areas, small mountain valleys, and California
areas. By bin, Table 3-12 below summarizes the counties that need additional emissions
reductions for the revised standard levels of 9/35 |~ig/m3.
201
-------
Table 3-12 Summary of Counties by Bin that Still Need Emissions Reductions for
Revised Primary Standard Levels of 9/35 (ig/m3
Bin
Area
Counties3 for 9/35 mg/m3
Near-Road Monitors
Northeast
Bergen County, NJ
Hamilton County, OH
Border Areas
Southeast
California
Cameron County, TX
Hidalgo County, TX
Imperial County, CA
Small Mountain Valleys
West
Plumas County, CA
Lemhi County, ID
Shoshone County, ID
Lincoln County, MT
Klamath County, OR
California Areas
Fresno County, CA (SJVAPCD)
Kern County, CA (SJVAPCD)
Kings County, CA (SJVAPCD)
Los Angeles County, CA (SCAQMD)
Madera County, CA (SJVAPCD)
Merced County, CA (SJVAPCD)
Orange County, CA (SCAQMD)
Riverside County, CA (SCAQMD)
San Bernardino County, CA (SCAQMD)
Stanislaus County, CA (SJVAPCD)
Tulare County, CA (SJVAPCD)
San Joaquin County, CA (SJVAPCD)
Alameda County, CA (BAAQMD)
Calaveras County, CA
Sutter County, CA
Note: For California counties that are part of multi-county air districts, the relevant district is indicated in parentheses;
BAAQMD = Bay Area Air Quality Management District, SCAQMD = South Coast Air Quality Management District, and
SJVAPCD= San Joaquin Valley Air Pollution Control District.
a The following counties had no identified PM2.5 emissions reductions because available controls were applied for the
current standard of 12/35 |ig/m3 and additional controls were not available to apply for analyses of the revised and
alternative standards: Colusa, Mono, Plumas, and Riverside, CA, Lake, OR, and Yakima, WA.
3.2.5.1 Near-Road Monitors (Northeast)
As shown in Table 3-9 above, the analysis indicates that Bergen County, New Jersey
and Hamilton County, Ohio need additional emissions reductions for the revised standard
levels of 9/35 |~ig/m3.
In analyzing the revised standard levels of 9/35 |~ig/m3, we estimated Bergen County
would need 542 tons of PM2.5 emissions reductions.11 The control strategy analysis
identified 338 tons of reductions within Bergen County from the application of several
11 Appendix 2 A, Table 2A-14 provides a summary of emissions reductions needed by county for the revised
and less and more stringent alternative standard levels.
202
-------
controls.12 The control applications within Bergen County included: Electrostatic
Precipitator applied to commercial cooking emissions in the non-point (area) inventory
sector; Pave Existing Shoulders applied to road dust emissions in the area fugitive dust
inventory sector; and Convert to Gas Logs and New Gas Stove or Gas Logs applied to area
source residential wood combustion emissions in the residential wood combustion
inventory sector.
To analyze the 204 tons of PM2.5 emissions reductions still needed, we identified
514 tons of PM2.5 emissions reductions from adjacent counties13, which is the equivalent of
129 tons of in-county emissions reductions after adjusting for the 4:1 ratio of adjacent
county reductions identified to in-county reductions needed. This left 75 tons of PM2.5
emissions reductions still needed. As shown in Table 3A-8, Bergen County has area fugitive
dust (afdust), non-point (area) (nonpt), non-electric generating unit point source
(ptnonipm), and residential wood combustion (rwc) emissions remaining in the inventory if
additional controls beyond the scope of this analysis can be identified.
In Chapter 2, Section 2.4.1 we indicate that the Fort Lee, New Jersey (Bergen County)
near-road monitor is close to the roadway interchange that leads to the George Washington
Bridge, where six major highways converge in the area leading to the bridge and the
location has been reported to be the most congested freight-significant highway location in
the U.S. The monitor is also located near the urban activity of downtown Fort Lee.
Understanding the nature of the local contributions and developing a plan to meet the
revised standard levels under the complex conditions at this monitor would require a
detailed local study beyond the scope of this RIA.
In analyzing the revised standard level of 9/35 |~ig/m3, we estimated Hamilton
County would need 1,025 tons of PM2.5 emissions reductions.14 The control strategy
analysis identified 581 tons of reductions within Hamilton County from the application of
12 Appendix 3A, Table 3A-2 provides a summary of in-county emissions reductions from control applications
by county for the northeast.
13 Appendix 3A, Table 3A-3 provides a summary of adjacent county emissions reductions from control
applications in the northeast.
14 Appendix 2A, Table 2A-14 provides a summary of emissions reductions needed by county for the revised
and less and more stringent alternative standard levels.
203
-------
several controls.15 The control applications within Hamilton County included: Electrostatic
Precipitator applied to commercial cooking emissions in the non-point (area) inventory
sector; Pave Existing Shoulders and Pave Unpaved Roads applied to road dust emissions in
the area fugitive dust inventory sector; Fabric Filter - All Types applied to industrial,
commercial, and institutional boilers and industrial processes in the non-EGU point
inventory sector; and Convert to Gas Logs and New Gas Stove or Gas Logs applied to area
source residential wood combustion emissions in the residential wood combustion
inventory sector.
To analyze the 444 tons of PM2.5 emissions reductions still needed, we identified
1,275 tons of PM2.5 emissions reductions from adjacent counties16'17, which is the equivalent
of 319 tons of in-county emissions reductions after adjusting for the 4:1 ratio of adjacent
county reductions identified to in-county reductions needed. This left 55 tons of PM2.5
emissions reductions still needed. As shown in Table 3A-8, Hamilton County has area
fugitive dust (afdust), non-point (area) (nonpt), non-electric generating unit point source
(ptnonipm), and residential wood combustion (rwc) emissions remaining in the inventory if
additional controls beyond the scope of this analysis can be identified. In addition, Butler
County, which is adjacent to Hamilton County, has emissions remaining in those inventory
sectors if additional controls beyond the scope of this analysis can be identified.
In Chapter 2, Section 2.4.1 we indicate that the challenges in Hamilton County
appear to be related to construction of a storage facility near the monitor during the
monitoring period used in the air quality projections. However, a detailed local analysis
beyond the scope of the national RIA would be needed to determine the full contribution of
the construction and other local influences on the PM2.5 DV at this monitor.
15 Appendix 3A, Table 3A-2 provides a summary of in-county emissions reductions from control applications
by county for the northeast.
16 Hamilton County was both a core/home county and an adjacent/neighboring county. Before the second
round of identifying emissions reductions, we adjusted any potential remaining reductions needed to
account for reductions that were applied in the first round but for application in the neighboring county,
resulting in slightly fewer remaining tons needed.
17 Appendix 3A, Table 3A-3 provides a summary of adjacent county emissions reductions from control
applications in the northeast.
204
-------
3.2.5.2 Border Areas (Southeast, California)
As shown in Table 3-9 above, Cameron County and Hidalgo County, Texas need
additional emissions reductions for the revised standard levels of 9/35 |~ig/m3.
We estimated Cameron County would need 703 tons of PM2.5 emissions reductions.
The control strategy analysis identified 212 tons of reductions within Cameron County
from the application of several controls.18 The control applications within Cameron County
included: Electrostatic Precipitator applied to commercial cooking emissions in the non-
point (area) inventory sector; Substitute Chipping for Burning applied to waste disposal
emissions in the non-point (area) inventory sector; Pave Existing Shoulders and Pave
Unpaved Roads applied to road dust emissions and Watering applied to crops and livestock
dust in the area fugitive dust inventory sector; and Convert to Gas Logs applied to area
source residential wood combustion emissions in the residential wood combustion
inventory sector.
To analyze the 361 tons of PM2.5 emissions reductions still needed, we identified 43
tons of PM2.5 emissions reductions from adjacent counties19'20, which was the equivalent of
10.75 tons of in-county emissions reductions after adjusting for the 4:1 ratio of adjacent
county reductions identified to in-county reductions needed. This left 350 tons of PM2.5
emissions reductions still needed. As shown in Table 3A-8, Cameron County has area
fugitive dust (afdust), non-point (area) (nonpt), non-electric generating unit point source
(ptnonipm), and residential wood combustion (rwc) emissions remaining in the inventory if
additional controls beyond the scope of this analysis can be identified; the majority of the
emissions remaining are area fugitive dust emissions.
In addition, we estimated Hidalgo County would need 1,349 tons of PM2.5 emissions
reductions. The control strategy analysis identified 521 tons of reductions within Hidalgo
18 Appendix 3A, Table 3A-4 provides a summary of in-county emissions reductions from control applications
by county for the southeast.
19 Cameron County was both a core/home county and an adjacent/neighboring county. Before the second
round of identifying emissions reductions, we adjusted any potential remaining reductions needed to
account for reductions that were applied in the first round but for application in the neighboring county,
resulting in slightly fewer remaining tons needed.
20 Appendix 3A, Table 3A-5 provides a summary of adjacent county emissions reductions from control
applications in the southeast.
205
-------
County from the application of several controls.21 Some of the control applications within
Hidalgo County included: Electrostatic Precipitator applied to commercial cooking
emissions in the non-point (area) inventory sector; Substitute Chipping for Burning applied
to waste disposal emissions in the non-point (area) inventory sector; Fabric Filter - All
Types applied to industrial, commercial, and institutional boilers in the non-EGU point
inventory sector; Pave Existing Shoulders and Pave Unpaved Roads applied to road dust
emissions and Watering applied to crops and livestock dust in the area fugitive dust
inventory sector; and Convert to Gas Logs and New Gas Stove or Gas Logs RP applied to
area source residential wood combustion emissions in the residential wood combustion
inventory sector.
To analyze the 776 tons of PM2.5 emissions reductions still needed, we identified
354 tons of PM2.5 emissions reductions from adjacent counties22'23, which was the
equivalent of 88.5 tons of in-county emissions reductions after adjusting for the 4:1 ratio of
adjacent county reductions identified to in-county reductions needed. This left 687.5 tons
of PM2.5 emissions reductions still needed. As shown in Table 3A-8, Hidalgo County has
area fugitive dust (afdust), non-point (area) (nonpt), non-point source oil and gas
(np_oilgas), non-electric generating unit point source (ptnonipm), and residential wood
combustion (rwc) emissions remaining in the inventory if additional controls beyond the
scope of this analysis can be identified; the majority of the emissions remaining are area
fugitive dust emissions.
In Chapter 2, Section 2.4.2.1 we note that the monitors in Cameron County and
Hidalgo County are in the Lower Rio Grande Valley, which includes the northern portion of
the state of Tamaulipas, Mexico. Addressing emissions reductions needed for the revised
21 Appendix 3A, Table 3A-4 provides a summary of in-county emissions reductions from control applications
by county for the southeast.
22 Hidalgo County was both a core/home county and an adjacent/neighboring county. Before the second
round of identifying emissions reductions, we adjusted any potential remaining reductions needed to
account for reductions that were applied in the first round but for application in the neighboring county,
resulting in slightly fewer remaining tons needed.
23 Appendix 3A, Table 3A-5 provides a summary of adjacent county emissions reductions from control
applications in the southeast.
206
-------
standard levels of 9/35 |~ig/m3 at the monitors is challenging because of the location of
these counties along the U.S.-Mexico border.
Area fugitive dust emissions make up the largest fraction of primary PM2.5 emissions
in Hidalgo County and Cameron County in the 2018 and 2032 air quality modeling cases
(Chapter 2, Figure 2-16). Paved-road dust emissions (in the area fugitive dust inventory
sector) are projected to increase in these counties between 2018 and 2032 as a result of
projected increases in the vehicle miles travelled; non-point (area) sources emissions are
also projected to increase as a result of population-based emissions projection factors.
Increases in area fugitive dust and non-point (area) emissions from 2018 to 2032 offset the
decreases in primary PM2.5 emissions projected for EGUs and mobile sources in the
counties. More detailed local analyses for these counties would be needed to better
understand the potential growth in area fugitive dust and non-point (area) source
emissions, as well as the potential contributions of international transport.
Further, for Imperial County, California the control strategy analysis identified 33
tons of PM2.5 reductions from the application of controls for the revised standard levels of
9/35 ng/m3. As shown in Table 3A-8, Imperial County has area fugitive dust (afdust), non-
point (area) (nonpt), non-electric generating unit point source (ptnonipm), and residential
wood combustion (rwc) emissions remaining in the inventory if controls beyond the scope
of this analysis can be identified; the majority of the emissions remaining are area fugitive
dust emissions.
As discussed in Chapter 2, Section 2.4.2, Imperial County is located in the southeast
corner of California and shares a southern border with Mexicali, Mexico. Imperial County
includes three PM2.5 monitoring sites, located in the cities of Calexico, El Centro, and
Brawley (Chapter 2, Figure 2-12). While these three cities are of similar size and have
similar emissions sources, the annual 2032 PM2.5 DV at the Calexico monitor, which is the
southern-most monitor and is less than a mile from the U.S.-Mexico border, is much greater
than the other two monitors (12.36 |~ig/m3, 9.70 |~ig/m3, and 8.69 |~ig/m3, respectively). In
addition, substantially greater NOx, SO2 and sulfate, and primary PM2.5 emissions have been
estimated for Mexicali, Mexico than for Calexico, California. For the revised standard levels,
Imperial County may not need the additional emissions reductions estimated because of
207
-------
the potential influence of Mexicali emissions on PM2.5 concentrations at the Calexico
monitor and Section 179B of the Clean Air Act; however, a detailed local analysis would be
needed.24
3.2.5.3 Small Mountain Valleys (West)
As shown in Table 3-9 above, the analysis also indicates that counties in the west
need additional emissions reductions after the application of controls for all of the
alternative standard levels analyzed. For the small mountain valleys bin, Table 3-13 below
summarizes the estimated PM2.5 emissions reductions needed and emissions reductions
identified by CoST for each of these counties for the revised standard levels of 9/35 |~ig/m3.
Table 3-13 Summary of Estimated PM2.5 Emissions Reductions Needed and
Emissions Reductions Identified by CoST for the West for the Revised
Primary Standard Levels of 9/35 ng/m3 in 2032 (tons/year)
County/State
PM2.5 Emissions Reductions
Needed
In-County PM2.5 Emissions
Reductions Identified by CoST
Plumas, CA
625.6
0
Lemhi, ID
242.8
7.6
Shoshone, ID
710.8
152.9
Lincoln, MT
1,211.2
317.4
Klamath, OR
185.5
125.1
Note: As shown in Table 3A-8, for Plumas, CA CoST identified controls to apply toward the current standard
of 12/35 |-ig/m:i. Additional controls were not available for the revised and less and more stringent
alternative standard levels.
As shown in Table 3-13, the control strategy analysis identified emissions
reductions for four of the counties. The controls applied included Electrostatic Precipitator
applied to commercial cooking emissions in the non-point (area) inventory sector;
Substitute Chipping for Burning applied to waste disposal emissions in the non-point
(area) inventory sector; Fabric Filter - All Types applied to industrial, commercial, and
institutional boilers in the non-EGU point inventory sector; Pave Existing Shoulders and
Pave Unpaved Roads applied to road dust emissions and Watering applied to crops and
livestock dust in the area fugitive dust inventory sector; and Convert to Gas Logs and New
24 Section 179B of the Clean Air Act [CAA] provides that a nonattainment area would be able to attain, or
would have attained, the relevant National Ambient Air Quality Standard but for emissions emanating from
outside the U.S.
208
-------
Gas Stove or Gas Logs applied to area source residential wood combustion emissions in the
residential wood combustion inventory sector.
As shown in Table 3A-8, these counties have area fugitive dust (afdust), non-point
(area) [nonpt], non-electric generating unit point source (ptnonipm), point oil and gas
(pt_oilgas), and residential wood combustion (rwc) emissions remaining in the inventory if
additional controls beyond the scope of this analysis can be identified; for each of the
counties the majority of the emissions remaining are area fugitive dust emissions.
Meteorological temperature inversions often occur in small northwestern mountain
valleys in winter and trap pollution emissions in a shallow atmospheric layer at the surface
(Chapter 2, Section 2.1.2). As discussed in Chapter 2, Section 2.4.3, primary PM2.5 emissions
can build up in the surface layer and produce high PM2.5 concentrations in winter (Chapter
2, Figure 2-18). These mountain valleys are often very small in size relative to the area of
the surrounding county and far smaller than the resolution of photochemical air quality
models (e.g., 12km grid cells). See Chapter 2, Figures 2-19 and 2-20 for maps of the Portola
nonattainment area (2012 PM2.5 NAAQS) relative to the city of Portola, California and the
Libby nonattainment area (1997 PM2.5 NAAQS) relative to the city of Libby, Montana. PM2.5
concentrations in these small mountain valleys can be influenced by the temperature
inversions, as well as by residential wood combustion and wildfire smoke.
Also as discussed in Chapter 2, Section 2.4.3, because of the small size of the urban
areas within the northwestern mountain valleys, air quality planning is commonly based
on linear rollback methods. To estimate emissions reductions needed for a standard level,
the linear rollback method relates wood-smoke contribution estimates at an exceeding
monitor to the local, or sub-county, wood combustion emissions totals. The PM2.5 response
factors from linear rollback methods estimate that relatively fewer residential wood
combustion emissions reductions can greatly influence PM2.5 concentrations in many of
these small mountain valleys. We did not apply linear rollback-based response factors for
the mountain valleys in this RIA because emissions inventory and area source control
information are available at the county level, preventing us from targeting residential wood
combustion controls in the local communities identified in the analyses. To better assess
the emissions reductions needed for the revised and alternative standard levels, more
209
-------
detailed analyses that include local PM2.5 response factors, emissions estimates, and
controls for each local area would be needed.
In addition to air quality challenges related to meteorological temperature
inversions and residential wood combustion, PM2.5 concentrations in these small mountain
valleys may also be influenced by wildfire emissions that could potentially qualify for
exclusion as atypical, extreme, or unrepresentative events.25 We performed sensitivity
projections to assess the potential for wildfire impacts. These projections indicate these
sites may include an important contribution from wildfire. Detailed local analyses would be
needed to fully characterize the wildfire influence in these areas. For more detailed
discussions of the residential wood combustion and wildfire smoke air quality challenges,
see Chapter 2, Section 2.4.3.
3.2.5.4 Califo rnia Areas
As shown in Table 3-9 above, the analysis also indicates that counties in California
need additional emissions reductions after the application of controls for the revised and
alternative standard levels analyzed. The sections below discuss the air quality challenges
by each air basin and/or district.
In the SJVAPCD, in analyzing the revised standard levels of 9/35 |~ig/m3, the District
needed 5,440 tons of PM2.5 emissions reductions. The control strategy analysis identified
1,940 tons of reductions from the application of several controls.26 Some of the control
applications included: Electrostatic Precipitator applied to commercial cooking emissions
in the non-point (area) inventory sector; Substitute Chipping for Burning applied to waste
disposal emissions in the non-point (area) inventory sector; Fabric Filter - All Types
applied to industrial, commercial, and institutional boilers and industrial processes in the
non-EGU point inventory sector; Pave Existing Shoulders and Pave Unpaved Roads applied
to road dust emissions and Watering applied to crop and livestock dust in the area fugitive
dust inventory sector; and Convert to Gas Logs applied to area source residential wood
25 Some wildfire influence likely persists in the projected 2032 PM2.5 DVs despite the exclusion of EPA-
concurred exceptional events and the wildfire screening (Appendix 2A, Section 2A.2.1],
26 Appendix 3A, Table 3A-7 provides a summary of in-county emissions reductions from control applications
by county for California.
210
-------
combustion emissions in the residential wood combustion inventory sector. As discussed
above, we did not attempt to identify additional PM2.5 emissions reductions in adjacent
counties or air districts.
As discussed in more detail in Chapter 2, Section 2.4.4, the air quality in SJVAPCD is
influenced by complex terrain and meteorological conditions that are best characterized
with a high-resolution air quality modeling platform developed for the specific conditions
of the valley. Air quality in the valley is influenced by emissions from large cities such as
Bakersfield and Fresno, a productive agricultural region, dust exacerbated by drought,
major goods transport corridors, and wildfires. The largest share of 2032 PM2.5 emissions
are from agricultural dust, the production of crops and livestock, agricultural burning,
paved and unpaved road dust, and prescribed burning (Chapter 2, Figure 2-24); wildfire
emissions also influence PM2.5 concentrations.
Specific, local information on area source controls to reduce emissions from
agricultural dust and burning and prescribed burning would be needed given the
magnitude of emissions from these sources. In addition, more detailed analyses would be
needed to characterize the influence of wildfires on PM2.5 concentrations and the potential
for some of these wildfires to be considered as atypical, extreme, or unrepresentative
events. Note that wildfire screening is particularly complex in California because different
parts of the state have different wildfire seasons.
In the SCAQMD, in analyzing the revised standard levels of 9/35 |j,g/m3, the District
needed 9,098 tons of PM2.5 emissions reductions. The control strategy analysis identified
1,442 tons of reductions from the application of several controls.27 Some of the control
applications included: Electrostatic Precipitator applied to commercial cooking emissions
in the non-point (area) inventory sector; Substitute Chipping for Burning applied to waste
disposal emissions in the non-point (area) inventory sector; Fabric Filter - All Types
applied to industrial, commercial, and institutional boilers and industrial processes in the
non-EGU point inventory sector; and Watering applied to crop and livestock dust in the
27 Appendix 3A, Table 3A-7 provides a summary of in-county emissions reductions from control applications
by county for California.
211
-------
area fugitive dust inventory sector. We did not attempt to identify additional PM2.5
emissions reductions in adjacent counties or air districts.
As discussed in more detail in Chapter 2, Section 2.4.4, the air quality in the SCAQMD
is influenced by complex terrain and meteorological conditions that are best characterized
with a high-resolution air quality modeling platform developed for the specific conditions
of the air basin. Air quality is influenced by diverse emissions sources associated with the
large population, the ports of Los Angeles and Long Beach, wildfires, and transportation of
goods. The largest share of 2032 PM2.5 emissions are from commercial and residential
cooking, on-road mobile sources, and paved and unpaved road dust (Chapter 2, Figure 2-
27).
Specific, local information on area source controls to reduce emissions from many of
the non-point (area) emissions sources (e.g., commercial and residential cooking) would be
needed given the magnitude of emissions from these sources. In addition, more detailed
analyses would be needed to characterize the influence of wildfires on PM2.5 concentrations
and the potential for some of these wildfires to be considered as atypical, extreme, or
unrepresentative events.
In the BAAQMD, in analyzing the revised standard levels of 9/35 |j,g/m3, the District
needed 948 tons of PM2.5 emissions reductions. The control strategy analysis identified
sufficient tons of reductions from the application of several controls for all the counties in
the District except Alameda County.28 Some of the control applications included:
Electrostatic Precipitator and Catalytic Oxidizer applied to commercial cooking emissions
in the non-point (area) inventory sector; and Substitute Chipping for Burning applied to
waste disposal emissions in the non-point (area) inventory sector; Fabric Filter - All Types
applied to industrial, commercial, and institutional boilers and industrial processes in the
non-EGU point inventory sector; Pave Existing Shoulders and Pave Unpaved Roads applied
to road dust emissions and Watering applied to crop and livestock dust in the area fugitive
dust inventory sector; and Convert to Gas Logs applied to area source residential wood
combustion emissions in the residential wood combustion inventory sector. We did not
28 Appendix 3A, Table 3A-7 provides a summary of in-county emissions reductions from control applications
by county for California.
212
-------
attempt to identify additional PM2.5 emissions reductions in adjacent counties or air
districts.
For the monitor in Alameda County, the ambient PM2.5 DV for the 2020-2022 period
at this monitor is 8.6 |~ig/m3 and the ambient PM2.5 DV for the 2019-2021 period is 8.5
Hg/m3. For the DV periods that overlap the 2016-2020 monitoring period reflected in the
air quality modeling projections, higher ambient DVs of 12.0 |~ig/m3, H-7 Hg/m3, and 10.8
Hg/m3 were measured. Wildfire influence during this period may explain the much higher
projected 2032 DV than the most recent ambient DV at the monitor. For example, the
California Department of Forestry and Fire Protection reports
(https://www.fire.ca.gov/incidents/) that 2017 was the most destructive wildfire year on
record in California at the time in terms of property damage; the 2018 wildfire year
included a total of over 7,500 fires burning an area of over 1.7 million acres; and the 2020
wildfire year had nearly 10,000 fires burning over 4.2 million acres, making 2020 the
largest wildfire season recorded in California's modern history. Although detailed analysis
of wildfire influence would be needed to determine the full extent of the fire impacts at the
Alameda monitor, the existing evidence suggests that wildfire has an important
contribution to the projected exceedances at this monitor.
In analyzing the revised standard levels of 9/35 |~ig/m3, we identified remaining air
quality challenges in two additional counties in California: Calaveras County and Sutter
County. These air quality challenges in these counties may also have been influenced by
wildfire during the 2016-2020 monitoring period used for air quality modeling projections.
Table 2-3 in Chapter 2, Section 2.4.4 includes DVs for Alameda, Calaveras, and Sutter
Counties, along with DVs using a more stringent wildfire screening threshold. At the Sutter
and Calaveras monitors, the 2032 DVs decrease from above to below the revised annual
standard of 9 |~ig/m3 when the more stringent wildfire screening threshold is applied. This
suggests the important influence of fires on the projected DVs, especially considering that
the more stringent wildfire screening threshold is close to 4x the revised annual standard
level of 9 |~ig/m3. Although detailed analysis of wildfire influence would be needed to
determine the full extent of the fire impacts at the Alameda, Calaveras, and Sutter monitors,
213
-------
the existing evidence suggests that wildfire has an important contribution to the projected
exceedances at these monitors.
3.3 Limitations and Uncertainties
The EPA's analysis is based on the best available information from engineering
studies of air pollution controls and a reasonable modeling framework for analyzing the
cost, emissions changes, and other impacts of emissions controls. However, the control
strategies above are subject to important limitations and uncertainties. Below we
summarize the most significant limitations and uncertainties.
• Control Strategies are Illustrative: A control strategy is a set of end-of-pipe
control technologies, area source controls, or policy actions that States may
take to comply with a revised standard. The illustrative control strategy
analyses in this RIA present only one potential pathway for controlling
emissions, and we do not presume that these control strategies represent an
exhaustive list of technologies, controls, or policy actions. Lastly, the
illustrative control strategies are not recommendations for how a revised
PM2.5 NAAQS should be implemented, and States will make the final decisions
regarding implementation of a revised NAAQS.
• Limitations of Emissions Inventories and Air Quality Modeling: Emissions
inventories and air quality modeling serve as a foundation for the projected
PM2.5 DVs, control strategies, and estimated emissions reductions and costs in
this analysis. The point source emissions inventories used may not accurately
reflect existing controls on emissions units. The non-point emissions
inventories - including inventories for non-point (area) sources, residential
wood combustion, and area fugitive dust emissions - are emissions estimates
at a county-level that reflect inventory-specific activity factors. These
inventories do not reflect county-specific tabulations of emissions sources
(e.g., the number of residential woodstoves or commercial cooking
establishments in a county). As such, the non-point emissions inventories used
may not accurately reflect the portion of existing emissions sources that have
already installed some non-point (area) source controls. For example, Ventura
214
-------
County Air Pollution Control District Rule 74.25 requires conveyorized
charbroilers to reduce both reactive organic compound emissions and
particulate matter emissions by at least 83 percent using an emissions control
device; we do not know what portion of the commercial cooking emissions
estimates for the county reflect those controls. The uncertainties introduced
through the point and non-point (area) source inventories could result in the
application of controls that are already on point sources or are reflected in
non-point (area) source inventories (resulting in overestimating emissions
reductions); these uncertainties could also result in not applying controls to a
large enough portion of a non-point (area) source inventory (resulting in
underestimating emissions reductions). The limitations and uncertainties
associated with the inventories also impact the future year emissions
projections and resulting estimated emissions reductions needed. Lastly, there
are other factors not reflected that affect emissions estimates and introduce
additional uncertainty, such as the economic base in a given area and
economic growth.
• Limits on Sizes of Sources Included: We included emissions sources with
greater than 5 tpy of PM2.5 emissions because emissions sources with fewer
tons are likely already controlled. We also limited the rule penetration rate
when applying controls to non-point (area), residential wood combustion, and
area fugitive emissions because the inventories used may not accurately
reflect the portion of existing emissions sources that have already installed
some non-point (area) source controls.
• Assumptions About the Baseline: There is significant uncertainty in the
illustration of the impact of rules on the baseline.
• Projecting the Level and Geographic Scope of Future Year Exceedances:
Estimates of the geographic areas that may exceed revised and alternative
standard levels in a future year, and the level to which those areas may exceed,
are approximations based on several factors. Any nonattainment
215
-------
determinations that would result from a revised NAAQS will likely depend on
the consideration of local air quality issues, changes in source operations
between the time of this analysis and implementation of a new standard, and
changes in control technologies over time.
Targeted Pollutants: Local knowledge of atmospheric chemistry in each
geographic area may result in a different prioritization of pollutants for
potential control.
Applicability of Control Technologies: The applicability of a control
technology to a specific source varies depending on a number of process
equipment factors such as age, design, capacity, fuel, and operating
parameters. These can vary considerably from source to source and over time.
While combinations of controls might be possible, we only select one control
per source/SCC combination, and we have no way of assessing their
appropriateness. This is especially limiting for area sources because we don't
have county-specific tabulations of specific emissions sources and emissions
are aggregated to the county level.
Limitations Related to Control Technologies Included: Given the decades
of progress in improving air quality, for many areas projected to exceed the
alternative standards analyzed, analyzing only traditional end-of-pipe control
technologies is limited. Future analyses should reflect changes in broader
energy supply, demand, and use patterns, as well as other innovative
technologies, policies, and strategies.
Advances in Control Technologies Over Time: The control technologies
applied do not reflect potential effects of technological change that may be
available in future years. All estimates of impacts associated with control
technologies applied reflect current knowledge, and not projections, of the
technology's effectiveness or costs.
216
-------
3.4 References
HGAC. 2014. Houston-Galveston-Brazoria (HGB) PM2.5 Advance Path Forward. Houston-
Galveston Area Council Regional Air Quality Planning Advisory Committee.
Kelly, J. T., CJ Jang, B Timin, B Gantt, A Reff, Y Zhu, S Long, A Hanna. 2019. A System for
Developing and Projecting PM2.5 Spatial Fields to Correspond to Just Meeting National
Ambient Air Quality Standards. Atmospheric Environment: X 100019.
https://doi.Org/10.1016/j.aeaoa.2019.100019
Kelly, JT, Koplitz, SN, Baker, KR, Holder, AL, Pye, HOT, Murphy, BN, Bash, JO, Henderson, BH,
Possiel, NC, Simon, H, Eyth, AM, Jang, C, Phillips, S and Timin, B (2019). Assessing PM2.5
model performance for the conterminous U.S. with comparison to model performance
statistics from 2007-2015. Atmospheric Environment 214: 116872
Kelly, J. T., K. R. Baker, S. N. Napelenok, and S. R. Roselle. 2015. Examining single-source
secondary impacts estimated from brute-force, decoupled direct method, and advanced
plume treatment approaches. Atmospheric Environment 111:10-19.
https://doi.Org/10.1016/j.atmosenv.2015.04.004
U.S. Environmental Protection Agency (U.S. EPA). 2019a. CoST v3.7 User's Guide. Office of
Air Quality Planning and Standards, Research Triangle Park, NC. November 2019.
Available at <
https://www.cmascenter.org/help/documentation. cfm?model=cost&version=3.7>.
United States Environmental Protection Agency (U.S. EPA). 2019b. Additional Methods,
Determinations, and Analyses to Modify Air Quality Data Beyond Exceptional Events.
U.S. Environmental Protection Agency, Office of Air Quality Planning and Standards,
Research Triangle Park, NC, EPA-457/B-19-002. Available:
https://www.epa.gov/sites/default/files/2019-
04/documents/clarification_memo_on_data_modification_methods.pdf.
U.S. Environmental Protection Agency (U.S. EPA). 2012. Regulatory Impact Analysis for the
Final Revisions to the National Ambient Air Quality Standards for Particulate Matter,
U.S. Environmental Protection Agency, Office of Air Quality Planning and Standards,
Research Triangle Park, NC. EPA-452/R-12-005. Available at:
https://www.epa.gov/sites/default/files/2020-07/documents/naaqs-
pm_ria_final_2 012-12.pdf.
217
-------
APPENDIX 3 A: CONTROL STRATEGIES AND PM2.5 EMISSIONS REDUCTIONS
Overview
Chapter 3 describes the approach that EPA used in applying the illustrative control
strategies for analyzing the following revised and less and more stringent alternative
annual and 24-hour standard levels - 10/35 |j,g/m3,10/30 ng/m3, 9/35 ng/m3, and 8/35
Hg/m3. This Appendix contains additional information about the end-of-pipe and area
source controls that were applied, as well as additional details on the estimated PM2.5
emissions reductions.
3A.1 Types of Control Technologies
Several types of controls were applied in the analyses for the analytical baseline and
revised and alternative standard levels. We identified controls using the EPA's Control
Strategy Tool (CoST) (U.S. EPA, 2019) and the control measures database.1 A brief
description of several of the controls is below.
3A.1.1 PM Controls for Non-EGU Point Sources
Non-EGU point source categories covered in this analysis include industrial boilers,
as well as industrial processes in the cement manufacturing, chemical manufacturing, pulp
and paper, mining, ferrous and non-ferrous metals, and refining industries. Several types of
end-of-pipe PM2.5 controls were applied for these sources, including venturi scrubbers,
fabric filters, and electrostatic precipitators, which are the primary controls analyzed for
non-EGU point sources.
• Venturi scrubbers - Venturi scrubbers are one of several types of wet
scrubbers that remove both acid gas and PM from waste gas streams of
stationary point sources. The pollutants are removed primarily through the
impaction, diffusion, interception and/or absorption of the pollutant onto
droplets of liquid. The liquid containing the pollutant is then collected for
disposal.
1 More information about CoST and the control measures database can be found at the following link:
https://www.epa.gov/economic-and-cost-analysis-air-pollution-regulations/cost-analysis-modelstools-air-
pollution.
218
-------
• Fabric filters - A fabric filter unit consists of one or more isolated
compartments containing rows of fabric bags in the form of round, flat, or
shaped tubes, or pleated cartridges. Particle-laden gas usually passes up
along the surface of the bags then radially through the fabric. Particles are
retained on the upstream face of the bags, and the cleaned gas stream is
vented to the atmosphere. Fabric filters collect particles with sizes ranging
from submicron to several hundred microns in diameter at efficiencies
generally in excess of 99 or 99.9 percent.
• Electrostatic precipitators - An ESP is a particle control device that uses
electrical forces to move the particles out of the flowing gas stream and onto
collector plates. The particles are given an electrical charge by forcing them
to pass through a corona, a region in which gaseous ions flow. The electrical
field that forces the charged particles to the walls comes from electrodes
maintained at high voltage in the center of the flow lane. Once the particles
are collected on the plates, they must be removed from the plates without re-
entraining them into the gas stream. This is usually accomplished by
knocking them loose from the plates, allowing the collected layer of particles
to slide down into a hopper from which they are evacuated.
3A.1.2 PM Controls for Non-point (Area) Sources
The non-point sector of the emissions inventory includes emissions sources that are
generally too small and/or numerous to estimate emissions for individual sources (e.g.,
commercial cooking, residential woodstoves, commercial or backyard waste burning). We
estimate the emissions from these sources for each county overall, typically using an
emissions factor that is applied to a surrogate of activity such as population or number of
houses. Controls for non-point sources are applied to the county level emissions. Several
area source controls were applied to PM2.5 emissions from non-point sources, including
catalytic oxidizers applied to charbroilers in commercial cooking, electrostatic precipitator
applied to under-fire charbroilers in commercial cooking, substitute chipping for open
burning in general and for households, converting to gas logs for residential wood
combustion, chemical stabilizers to suppress unpaved road dust, paving existing shoulders
219
-------
to suppress paved road dust, and managing crop and livestock dust to suppress fugitive
dust.
3A.2 EGU Trends Reflected in EPA's Integrated Planning Model (IPM) v6 Platform,
Post-IRA 2022 Reference Case Projections
The EPA's Integrated Planning Model (IPM) v6 Platform Post-IRA 2022 Reference
Case projections were used in the air quality modeling done for this RIA.2 A high level
summary of the input assumptions in the Post-IRA 2022 Reference Case is below. This
version features several input data updates, in addition to reflecting Inflation Reduction Act
(IRA) of 2022 provisions of tax incentives impacting electricity supply, other bottom-up
input data, and assumption updates3, including the following:
• Demand - Annual Energy Outlook (AEO) 2021 + Office of Transportation and
Air Quality 2027 GHG Rule
• Gas and Coal Market Assumptions - Updated as of December 2021
• Cost and Performance of Fossil Generation Technologies - AEO 2021
• Cost and Performance of Renewable Energy Generation Technologies -
National Renewable Energy Lab Annual Technology Baseline 2021 mid-case
• Environmental Rules and Regulations (On-the-Books) - 2023 Federal Good
Neighbor Plan Addressing Regional Ozone Transport for the 2015 Ozone
National Ambient Air Quality Standards (Final GNP), Revised Cross-State Air
Pollution Rule, Mercury and Air Toxics Standards, BART, California Assembly
Bill 32, Regional Greenhouse Gas Initiative, various renewable portfolio
standards and clean energy standards, non-air rules (Cooling Water Intake,
Steam Electric Power Generating Effluent Guidelines, Coal Combustion
Residuals), and State Rules
2 Documentation of the baseline run building on Post-IRA 2022 Reference Case, incorporating the Final GNP,
and the corresponding results are available at https://www.epa.gov/power-sector-modeling/final-pm-
naaqs.
3 For a complete reference summary, see Chapter 1, Table 1-1 available at
https://www.epa.gov/system/files/documents/2021-09/chapter-l-introduction.pdf
220
-------
• Financial Assumptions - Based on 2016-2020 data, reflects tax credit
extensions from Consolidated Appropriations Act of 2021
• Transmission - Updated data with build options
• Retrofits - Carbon capture and storage option for combined cycles
• Post-combustion control operation - Operate according to historical rates
and/or to performance levels established in the IPM documentation
• Operating Reserves (in select runs) - Greater detail in representing
interaction of load, wind, and solar, ensuring availability of quick response of
resources at higher levels of renewable energy penetration
• Fleet - NEEDS for Post-IRA 022 Reference Case (rev:10-14-22)
The Post-IRA 2022 Reference Case projections show a significant decline in
national-level annual SO2, NOx, and primary PM emissions because of displacement of
retired coal units with new natural gas generation and renewable energy. Greater near-
term renewable energy penetration is due to an increase in actual projects reflected in
NEEDS prior to the IPM projections; the long-term increase is largely driven by improved
renewable energy technology costs, especially enhanced with the IRA tax incentives.
California sees a significant decrease in projected emissions for all pollutants by
2030 due to the state's Clean Energy Standards (CES). California's Senate Bill No. 100
requires expansion of the Renewable Portfolio Standard through 2030 where generation
from qualifying renewables must achieve a 50 percent share of retail sales by 2026 and 60
percent by 2030.4 California's legislation requires a transition from the RPS to CES where
generation from qualifying "zero carbon resources" must equal 100 percent of retail sales
by 2045. Our projections show a significant shift from fossil to renewable energy
generation in California between 2025 and 2030 with the trend continuing thereafter.
4 https://leginfo.legislature.ca.gov/faces/billNavClient.xhtml?bill_id=201720180SB100
221
-------
3A.3 Applying End-of-Pipe and Area Source Controls
As mentioned in Chapter 3, Section 3.2.2, controls applied for the analyses of the
existing standards of 12/35 |~ig/m3 and the revised and less and more stringent annual and
24-hour PM2.5 alternative standard levels of 10/35 ng/m3,10/30 ng/m3, 9/35 ng/m3, and
8/35 ng/m3 are listed in Table 3A-1 by geographic area and by emissions inventory sector,
with an "X" indicating which controls were applied for each standard level.
Table 3A-2 through Table 3A-7 include detailed summaries of PM2.5 emissions
reductions by county for the revised and alternative standard levels for the northeast, the
adjacent counties in the northeast, the southeast, the adjacent counties in the southeast, the
west, and California. Table 3A-7 for California presents counties organized by air districts.
As shown in Table 3A-2 and Table 3A-3 for the northeast counties (31 counties) and
the adjacent counties (51 counties), for the alternative standard levels of 10/35 ng/m3 and
10/30 ng/m3, controls were applied in four counties and no additional emissions
reductions were needed in adjacent counties. For the revised standard levels of 9/35 |Lxg/-
m3, we estimated a total of 9,825 tons of PM2.5 emission reductions available from the
application of controls - approximately 74 percent of that total is available from within a
core county and 26 percent is from an adjacent county. For the alternative standard levels
of 8/35 ng/m3, we estimated a total of 25,947 tons of PM2.5 emission reductions -
approximately 54 percent of that total is available from within a core county and 46
percent is from an adjacent county.
As shown in Table 3A-4 and Table 3A-5 for the southeast counties (33 counties) and
the adjacent counties (34 counties), for the alternative standard levels of 10/35 ng/m3 and
10/30 ng/m3, controls were applied in one county and additional emissions reductions
were identified in three adjacent counties. For the revised standard levels of 9/35 ng/m3,
we estimated a total of 2,313 tons of PM2.5 emission reductions - approximately 85 percent
of that total is available from the application of controls from within a core county and 15
percent is from an adjacent county. For the alternative standard levels of 8/35 ng/m3, we
estimated a total of 17,080 tons of PM2.5 emission reductions - approximately 82 percent of
that total is available from within a core county and 18 percent is from an adjacent county.
222
-------
As shown in Table 3A-6 for the west (29 counties), for the alternative standard level
of 10/35 ng/m3 controls were applied in two counties. For the alternative standard levels
of 10/30 ng/m3 controls were applied in 17 counties; for the revised standard levels of
9/35 ng/m3 controls were applied in 10 counties; and for the alternative standard levels of
8/35 ng/m3 controls were applied in 21 counties.
As shown in Table 3A-7 for California (36 counties) of the eight counties in the San
Joaquin Valley Air Pollution Control District, we estimated that all eight need PM2.5
emissions reductions. For six counties, we identified some emissions reductions available
for alternative standard levels of 10/35 |~ig/m3 and no additional emissions reductions for
the revised and lower alternative standard levels analyzed. For one county, we identified
some emissions reductions available for alternative standard levels of 10/30 |~ig/m3 and
additional reductions available for revised standard levels of 9/35 |j,g/m3; for the
remaining county we identified some emissions reductions available for revised standard
levels of 9/35 ng/m3. We estimated that the four counties in the South Coast Air Quality
Management District need emissions reductions. For one county we did not identify any
emissions reductions from the application of controls for any of the revised and alternative
standard levels. For three counties, we identified some emissions reductions available for
alternative standard levels of 10/35 ng/m3.
Table 3A-8 includes information on PM2.5 emissions by emissions inventory sector,
on counties needing emissions reductions, and on estimated emissions reductions by the
revised and alternative standard levels analyzed. The column labeled Sector uses
abbreviations for emissions inventory sectors from the National Emissions Inventory. The
abbreviations and related sectors include: afdust or area fugitive dust emissions; nonpt or
non-point (area) source emissions; np_oilgas or non-point (area) source oil and gas
emissions; ptagflre or point source agriculture fire emissions; ptnonipm or non-electric
generating unit, point source emissions; pt_oilgas or point source oil and gas emissions;
and rwc or residential wood combustions emissions.
The first column includes names of adjacent counties and counties still needing
emissions reductions. The second column lists any counties that need emissions
223
-------
reductions. The columns with annual PM2.5 emissions and the PM2.5 emissions reductions
are related to the county in the first column. If the second column is blank, then the annual
PM2.5 emissions serves as an indicator of the county's own PM2.5 emissions that might be
controllable if a state or local jurisdiction knew how to control those emissions; in these
cases the maximum PM2.5 emissions reductions should be equal to the selected PM2.5
emissions reductions for one of the alternative standards being analyzed (e.g., Santa Cruz
County, AZ).
The table is intended to present information about potential nearby emissions
reductions that might be available to help counties attain a revised or alternative standard
level. The list of PM2.5 emissions is not exhaustive, as inventory sectors with reported
emissions less than 5 tons per year are excluded in general, and emissions from rail,
airports, and wildfires of all types are excluded regardless of their emissions because either
we do not have information on potential controls for these sectors or the emissions from
these sectors are not necessarily controllable (i.e., wildfires). While we considered
emissions from adjacent counties in the east, we did not do so in the west and California
due to uncertainty about the air quality impacts of emissions reductions from adjacent
counties. For the west and California, in addition to finding ways of controlling remaining
emissions within a county or adjacent counties (or within the same air district in
California), it will be necessary to determine how much emissions reductions in adjacent
counties may impact the DV at a monitor of interest.
224
-------
Table 3A-1 By Area and Emissions Inventory Sector, Controls Applied in Analyses
of the Current Standards, Revised, and Alternative Primary Standard
Levels
Inventory
Area Sector Control Technology 12/35 10/35 10/30 9/35 8/35
Northeast Non-EGU Point Electrostatic Precipitator-All Types x
Fabric Filter-All Types x x x x
Install new drift eliminator x x
Venturi Scrubber x x x x_
Non-Point Annual tune-up x x x x
(Area) Biennial tune-up x x x x
Electrostatic Precipitator x x
HE PA filters x x
Smokeless Broiler x x x x
Substitute chipping for burning x x x x_
Residential Convert to Gas Logs x x
Wood EPA Phase 2 Qualified Units x x
Combustion EPA-certified wood stove x x
Install Cleaner Hydronic Heaters x x x x
Install Retrofit Devices x
New gas stove or gas logs x x x x_
Area Source Chemical Stabilizer x x
Fugitive Dust Pave Unpaved Roads x x
Pave existing shoulders x x
Watering x x_
Northeast Non-EGU Point Fabric Filter-All Types x x
(Adjacent Install new drift eliminator x
Counties) Venturi Scrubber x x
Non-Point Annual tune-up x x
(Area) Biennial tune-up x x
Electrostatic Precipitator x x
Smokeless Broiler x x
Substitute chipping for burning x x_
Residential Convert to Gas Logs x x
Wood Install Cleaner Hydronic Heaters x x
Combustion New gas stove or gas logs x x_
Area Source Chemical Stabilizer x
Fugitive Dust Pave Unpaved Roads x x
Pave existing shoulders x x
Watering x x_
Southeast Non-EGU Point Fabric Filter-All Types x x x x
Install new drift eliminator x x
Venturi Scrubber x_
Non-Point Add-on Scrubber x
(Area) Annual tune-up x x
Biennial tune-up x x
Electrostatic Precipitator x x x x
Smokeless Broiler x x
225
-------
Inventory
Area Sector Control Technology 12/35 10/35 10/30 9/35 8/35
Substitute chipping for burning x x x x_
Residential Convert to Gas Logs x x x x
Wood EPA Phase 2 Qualified Units x x
Combustion EPA-certified wood stove x
Install Cleaner Hydronic Heaters x x
Install Retrofit Devices x
New gas stove or gas logs x x x x_
Area Source Chemical Stabilizer x x
Fugitive Dust Pave Unpaved Roads x x x x
Pave existing shoulders x x x x
Watering x x x x_
Southeast Non-EGU Point Fabric Filter-All Types x
(Adjacent Install new drift eliminator x
Counties) Venturi Scrubber x
Non-Point Annual tune-up x
(Area) Biennial tune-up x
Electrostatic Precipitator x x
Smokeless Broiler x
Substitute chipping for burning x x_
Residential Convert to Gas Logs x
Wood Install Cleaner Hydronic Heaters x
Combustion New gas stove or gas logs x_
Area Source Chemical Stabilizer x
Fugitive Dust Pave Unpaved Roads x x
Pave existing shoulders x x
Watering x x x x_
West Non-EGU Point Fabric Filter-All Types x x x x
Install new drift eliminator x x
Venturi Scrubber x x x
Oil & Gas Point Fabric Filter-All Types x_
Non-Point Annual tune-up x x x x x
(Area) Biennial tune-up x x x x
Electrostatic Precipitator x x x x
Smokeless Broiler x x x x
Substitute chipping for burning x x x x x_
Residential Convert to Gas Logs x x x
Wood EPA Phase 2 Qualified Units x x x
Combustion install Cleaner Hydronic Heaters x x x x x
Install Retrofit Devices x x
New gas stove or gas logs x x x x x_
Area Source Chemical Stabilizer x x x x
Fugitive Dust Pave Unpaved Roads x x x x x
Pave existing shoulders x x x x
Watering x x x x x_
226
-------
Area
Inventory
Sector
Control Technology
12/35
10/35
10/30
9/35
8/35
CA
Non-EGU Point
Electrostatic Precipitator-All Types
X
Fabric Filter-All Types
X
X
X
X
X
Install new drift eliminator
X
X
X
Venturi Scrubber
X
X
X
X
Oil & Gas Point
Fabric Filter-All Types
X
X
X
X
Non-Point
Annual tune-up
X
X
X
X
X
(Area)
Biennial tune-up
X
X
X
X
X
Catalytic oxidizers
X
X
X
X
Electrostatic Precipitator
X
X
X
X
X
Substitute chipping for burning
X
X
X
X
X
Residential
Convert to Gas Logs
X
X
X
X
X
Wood
Combustion
EPA Phase 2 Qualified Units
Install Retrofit Devices
X
X
X
X
Area Source
Chemical Stabilizer
X
X
X
X
Fugitive Dust
Pave Unpaved Roads
X
X
X
X
X
Pave existing shoulders
X
X
X
X
X
Watering
X
X
X
X
X
Table 3A-2 Summary of PM2.5 Estimated Emissions Reductions from CoST for the
Northeast (31 counties) for Revised and Alternative Primary Standard
Levels of 10/35 |jg/m3,10/30 ng/m3,9/35 (ig/m3, and 8/35 ng/m3 in
2032 (tons/year)
County
10/35
10/30
9/35
8/35
Cook County, IL
0
0
534
785
DuPage County, IL
0
0
0
394
Macon County, IL
0
0
0
220
Madison County, IL
0
0
81
812
McLean County, IL
0
0
0
29
St. Clair County, IL
0
0
0
293
Lake County, IN
0
0
102
834
Marion County, IN
0
0
510
510
Jefferson County, KY
0
0
0
468
Wayne County, MI
220
220
951
1,048
Bergen County, NJ
0
0
338
338
Camden County, NJ
0
0
168
183
Union County, NJ
0
0
0
217
New York County, NY
0
0
0
593
Butler County, OH
0
0
784
794
Cuyahoga County, OH
0
0
607
840
Franklin County, OH
0
0
0
51
Hamilton County, OH
293
293
581
581
Jefferson County, OH
0
0
0
15
Stark County, OH
0
0
0
183
Allegheny County, PA
490
532
1,222
1,628
Beaver County, PA
0
0
0
117
Cambria County, PA
0
0
0
308
227
-------
County
10/35
10/30
9/35
8/35
Chester County, PA
0
0
521
586
Delaware County, PA
29
29
384
384
Lancaster County, PA
0
0
0
491
Lebanon County, PA
0
0
0
168
Philadelphia County, PA
0
0
442
521
York County, PA
0
0
0
279
Providence County, RI
0
0
0
59
Davidson County, TN
0
0
0
307
Total
1,032
1,074
7,226
14,036
228
-------
Table 3A-3 Summary of PM2.5 Estimated Emissions Reductions from CoST for the
Adjacent Counties in the Northeast (51 counties) for Revised and
Alternative Primary Standard Levels of 10/35 |ig/m3,10/30 |ig/m3,
9/35 (ig/m3, and 8/35 (ig/m3 in 2032 (tons/year)
County Adjacent Counties 10/35 10/30 9/35 8/35
Kane County, IL
Cook County, IL
DuPage County, IL
0
0
0
450
Lake County, IL
Cook County, IL
0
0
0
326
McHenry County, IL
Cook County, IL
0
0
0
583
Will County, IL
Cook County, IL
DuPage County, IL
0
0
0
169
Boone County, IN
Marion County, IN
0
0
48
89
Hamilton County, IN
Marion County, IN
0
0
179
344
Hancock County, IN
Marion County, IN
0
0
52
94
Hendricks County, IN
Marion County, IN
0
0
127
255
Johnson County, IN
Marion County, IN
0
0
117
203
Morgan County, IN
Marion County, IN
0
0
78
173
Shelby County, IN
Marion County, IN
0
0
168
554
Macomb County, MI
Wayne County, MI
0
0
0
571
Monroe County, MI
Wayne County, MI
0
0
0
397
Oakland County, MI
Wayne County, MI
0
0
0
1,171
Washtenaw County, MI
Wayne County, MI
0
0
0
399
Atlantic County, NJ
Camden County, NJ
0
0
0
155
Burlington County, NJ
Camden County, NJ
0
0
0
257
Essex County, NJ
Bergen County, NJ
Union County, NJ
0
0
223
223
Gloucester County, NJ
Camden County, NJ
0
0
0
317
Hudson County, NJ
Bergen County, NJ
Union County, NJ
0
0
148
148
Passaic County, NJ
Bergen County, NJ
0
0
144
144
Bronx County, NY
New York County, NY
0
0
0
72
Kings County, NY
New York County, NY
0
0
0
64
Queens County, NY
New York County, NY
0
0
0
126
Clermont County, OH
Hamilton County, OH
0
0
333
333
Geauga County, OH
Cuyahoga County, OH
0
0
0
297
Lake County, OH
Cuyahoga County, OH
0
0
0
208
Lorain County, OH
Cuyahoga County, OH
0
0
0
394
Medina County, OH
Cuyahoga County, OH
0
0
0
401
Portage County, OH
Cuyahoga County, OH
Stark County, OH
0
0
0
308
Summit County, OH
Cuyahoga County, OH
Stark County, OH
0
0
0
388
Warren County, OH
Butler County, OH
Hamilton County, OH
0
0
437
437
Armstrong County, PA
Allegheny County, PA
0
0
0
88
Butler County, PA
Allegheny County, PA
Beaver County, PA
0
0
0
364
229
-------
County Adjacent Counties 10/35 10/30 9/35 8/35
Montgomery County, PA
Chester County, PA
Delaware County, PA
Philadelphia County, PA
0
0
546
672
Washington County, PA
Allegheny County, PA
Beaver County, PA
0
0
0
387
Westmoreland County, PA
Allegheny County, PA
Cambria County, PA
0
0
0
348
Total
0
0
2,599
11,911
230
-------
Table 3A-4 Summary of PM2.5 Estimated Emissions Reductions from CoST for the
Southeast (33 counties) for Revised and Alternative Primary Standard
Levels of 10/35 (j,g/m3,10/30 (j,g/m3, 9/35 (j,g/m3, and 8/35 jig/m3 in
2032 (tons/year)
County
10/35
10/30
9/35
8/35
Jefferson County, AL
0
0
0
785
Russell County, AL
0
0
0
327
Pulaski County, AR
0
0
0
720
District of Columbia
0
0
0
200
Broward County, FL
0
0
16
834
Bibb County, GA
0
0
0
115
Chatham County, GA
0
0
0
82
Dougherty County, GA
0
0
0
240
Fulton County, GA
0
0
0
648
Gwinnett County, GA
0
0
0
442
Muscogee County, GA
0
0
0
204
Richmond County, GA
0
0
49
620
Shawnee County, KS
0
0
0
278
Wyandotte County, KS
0
0
0
303
Caddo Parish, LA
0
0
401
519
East Baton Rouge Parish, LA
0
0
0
381
West Baton Rouge Parish, LA
0
0
0
292
Hinds County, MS
0
0
0
311
Forsyth County, NC
0
0
0
25
Mecklenburg County, NC
0
0
0
254
Cleveland County, OK
0
0
0
482
Oklahoma County, OK
0
0
0
548
Tulsa County, OK
0
0
0
507
Bowie County, TX
0
0
0
188
Cameron County, TX
0
0
212
212
Dallas County, TX
0
0
0
172
El Paso County, TX
0
0
0
238
Harris County, TX
0
0
613
1,431
Hidalgo County, TX
521
521
521
521
Jefferson County, TX
0
0
0
343
Nueces County, TX
0
0
0
499
Orange County, TX
0
0
0
352
Travis County, TX
0
0
147
923
Total
521
521
1,959
13,995
231
-------
Table 3A-5 Summary of PM2.5 Estimated Emissions Reductions from CoST for the
Adjacent Counties in the Southeast (34 counties) for Revised
Alternative Primary Standard Levels of 10/35 |ig/m3,10/30 |ig/m3,
9/35 (ig/m3, and 8/35 (ig/m3 in 2032 (tons/year)
County Adjacent Counties 10/35 10/30 9/35 8/35
Burke County, GA
Richmond County, GA
0
0
0
487
Columbia County, GA
Richmond County, GA
0
0
0
91
Forsyth County, GA
Fulton County, GA
0
0
0
3
Gwinnett County, GA
Jefferson County, GA
Richmond County, GA
0
0
0
308
McDuffie County, GA
Richmond County, GA
0
0
0
100
Bossier Parish, LA
Caddo Parish, LA
0
0
0
286
De Soto Parish, LA
Caddo Parish, LA
0
0
0
202
East Feliciana Parish, LA
East Baton Rouge Parish, LA
0
0
0
38
West Baton Rouge Parish, LA
Iberville Parish, LA
East Baton Rouge Parish, LA
0
0
0
109
West Baton Rouge Parish, LA
Pointe Coupee Parish, LA
West Baton Rouge Parish, LA
0
0
0
33
Red River Parish, LA
Caddo Parish, LA
0
0
0
813
West Feliciana Parish, LA
West Baton Rouge Parish, LA
0
0
0
38
Bastrop County, TX
Travis County, TX
0
0
0
70
Blanco County, TX
Travis County, TX
0
0
0
9
Brooks County, TX
Hidalgo County, TX
16
16
108
108
Burnet County, TX
Travis County, TX
0
0
0
15
Caldwell County, TX
Travis County, TX
0
0
0
28
Hays County, TX
Travis County, TX
0
0
0
14
Hudspeth County, TX
El Paso County, TX
0
0
0
56
Kenedy County, TX
Hidalgo County, TX
18
18
78
78
Starr County, TX
Hidalgo County, TX
0
0
125
125
Willacy County, TX
Cameron County, TX
11
11
43
43
Hidalgo County, TX
Williamson County, TX
Travis County, TX
0
0
0
30
Total
45
45
354
3,086
Table 3A-6 Summary of PM2.5 Estimated Emissions Reductions from CoST for the
West (29 counties) for Revised and Alternative Primary Standard
Levels of 10/35 (ig/m3,10/30 ng/m3,9/35 (ig/m3, and 8/35 ng/m3 in
2032 (tons/year)
County
10/35
10/30
9/35
8/35
Maricopa County, AZ
0
0
257
721
Pinal County, AZ
0
321
0
0
Santa Cruz County, AZ
0
0
0
39
Yuma County, AZ
0
0
0
122
Adams County, CO
0
0
43
348
Denver County, CO
0
0
9
124
Weld County, CO
0
0
0
373
Benewah County, ID
0
187
0
0
Canyon County, ID
0
264
0
290
232
-------
County 10/35 10/30 9/35 8/35
Lemhi County, ID
0
8
8
8
Shoshone County, ID
153
153
153
153
Flathead County, MT
0
149
0
0
Lewis and Clark County, MT
0
51
0
51
Lincoln County, MT
317
317
317
317
Missoula County, MT
0
0
24
491
Clark County, NV
0
0
0
468
Crook County, OR
0
276
0
89
Harney County, OR
0
115
212
271
Jackson County, OR
0
12
239
706
Josephine County, OR
0
0
0
271
Klamath County, OR
0
125
125
125
Lake County, OR
0
0
0
0
Lane County, OR
0
196
0
206
Box Elder County, UT
0
159
0
0
Cache County, UT
0
115
0
0
Salt Lake County, UT
0
57
0
0
King County, WA
0
0
0
42
Okanogan County, WA
0
209
0
108
Yakima County, WA
0
0
0
0
Total
470
2,715
1,386
5,323
233
-------
Table 3A-7
Summary of PM2.5 Estimated Emissions Reductions from CoST for
California (36 counties) for Revised and Alternative Primary Standard
Levels of 10/35 (j,g/m3,10/30 |ig/m3,9/35 (j,g/m3, and 8/35 |ig/m3 in
2032 (tons/year)
County
Air District
10/35
10/30
9/35
8/35
Alameda County, CA
Contra Costa County, CA
Napa County, CA
San Francisco County, CA
San Mateo County, CA
Santa Clara County, CA
Solano County, CA
Butte County, CA
Calaveras County, CA
Colusa County, CA
Sutter County, CA
Mono County, CA
Imperial County, CA
Mendocino County, CA
Plumas County, CA
Placer County, CA
Sacramento County, CA
San Diego County, CA
Fresno County, CA
Kern County, CA
Kings County, CA
Madera County, CA
Merced County, CA
San Joaquin County, CA
Stanislaus County, CA
Tulare County, CA
San Luis Obispo County, CA
Santa Barbara County, CA
Shasta County, CA
Siskiyou County, CA
Los Angeles County, CA
Orange County, CA
Riverside County, CA
San Bernardino County, CA
Tehama County, CA
Ventura County, CA
Bay Area AQMD 105 105 388 388
Bay Area AQMD 0 0 127 444
Bay Area AQMD 0 0 0 69
Bay Area AQMD 0 0 0 93
Bay Area AQMD 0 0 0 114
Bay Area AQMD 0 0 320 555
Bay Area AQMD 0 0 162 195
Butte County AQMD 0 301 0 248
Calaveras County APCD 0 0 128 128
Colusa County APCD 0 0 0 0
Feather River AQMD 0 86 86 86
Great Basin Unified APCD 0 0 0 0
Imperial County APCD 33 33 33 33
Mendocino County AQMD 0 121 0 13
Northern Sierra AQMD 0 0 0 0
Placer County APCD 0 0 0 24
Sacramento Metro AQMD 0 201 162 275
San Diego County APCD 0 0 238 763
San Joaquin Valley APCD 508 508 508 508
San Joaquin Valley APCD 225 225 225 225
San Joaquin Valley APCD 95 95 95 95
San Joaquin Valley APCD 0 0 255 255
San Joaquin Valley APCD 210 210 210 210
San Joaquin Valley APCD 0 38 256 256
San Joaquin Valley APCD 204 204 204 204
San Joaquin Valley APCD 188 188 188 188
San Luis Obispo County APCD 0 0 0 28
Santa Barbara County APCD 0 131 0 68
Shasta County AQMD 0 30 0 0
Siskiyou County APCD 0 339 0 0
South Coast AQMD 1,102 1,102 1,102 1,102
South Coast AQMD 264 264 264 264
South Coast AQMD 0 0 0 0
South Coast AQMD 76 76 76 76
Tehama County APCD 0 121 0 0
Ventura County APCD 0 274 43 274
Total
3,010
4,652
5,069
7,181
234
-------
Table 3A-8 Remaining PM2.5 Emissions and Potential Additional Reduction Opportunities
Adjacent Counties Maximum
(NE,SE,W) or Counties in Annual PM2.5 Selected PM2.5 Emissions Reductions
Same Air District (CA)
PM2.5
Emissions
County
Still Needing Reductions
Sector
Emissions
Reduction
12/35
10/35
10/30
9/35
8/35
Maricopa County, AZ
Pinal County, AZ
afdust
5,257
68
-
-
-
68
68
nonpt
2,299
457
-
-
-
166
457
ptnonipm
163
18
-
-
-
18
18
rwc
1,146
177
-
-
-
5
177
Pinal County, AZ
Maricopa County, AZ
afdust
3,337
118
-
-
118
-
-
nonpt
377
191
-
-
191
-
-
pt_oilgas
8
-
-
-
-
-
-
ptnonipm
88
-
-
-
-
-
-
rwc
99
11
-
-
11
-
-
Santa Cruz County, AZ
-
afdust
169
15
-
-
-
-
15
nonpt
68
23
-
-
-
-
23
rwc
13
-
-
-
-
-
-
Yuma County, AZ
Maricopa County, AZ
afdust
1,284
113
-
-
-
-
62
nonpt
187
79
-
-
-
-
56
ptnonipm
7
-
-
-
-
-
-
rwc
42
4
-
-
-
-
4
Alameda County, CA
Napa County, CA
afdust
531
89
-
8
8
89
89
San Francisco County, CA
nonpt
589
63
-
4
4
63
63
Solano County, CA
ptnonipm
369
114
-
93
93
114
114
rwc
356
122
-
-
-
122
122
Calaveras County, CA
-
afdust
189
49
-
-
-
49
49
nonpt
144
67
-
-
-
67
67
rwc
173
12
-
-
-
12
12
Colusa County, CA
-
afdust
346
56
56
-
-
-
-
nonpt
102
15
15
-
-
-
-
ptnonipm
19
-
-
-
-
-
-
rwc
45
4
4
-
-
-
-
235
-------
Adjacent Counties
Maximum
(NE,SE,W) or Counties in
Annual
PM2.5
Selected PM2.5 Emissions Reductions
Same Air District (CA)
PM2.5
Emissions
County
Still Needing Reductions
Sector
Emissions
Reduction
12/35
10/35
10/30
9/35
8/35
Contra Costa County, CA
Alameda County, CA
afdust
392
71
-
-
-
8
8
Napa County, CA
nonpt
502
55
-
-
-
-
3
San Francisco County, CA
ptnonipm
1,611
897
-
-
-
118
433
Solano County, CA
rwc
786
230
-
-
-
-
-
Fresno County, CA
Kern County, CA
afdust
2,240
373
67
306
306
306
306
Kings County, CA
nonpt
643
135
27
109
109
109
109
Madera County, CA
pt_oilgas
28
_
_
_
_
_
Merced County, CA
ptnonipm
219
65
10
55
55
55
55
San Joaquin County, CA
rwc
280
39
39
39
39
39
Stanislaus County, CA
Tulare County, CA
Imperial County, CA
-
afdust
3,610
275
275
-
-
-
-
nonpt
190
14
-
14
14
14
14
ptnonipm
47
15
-
15
15
15
15
rwc
18
4
-
4
4
4
4
Kern County, CA
Fresno County, CA
afdust
1,377
53
53
-
-
-
-
Kings County, CA
nonpt
936
325
303
7
7
7
7
Madera County, CA
np_oilgas
5
-
-
-
-
-
-
Merced County, CA
pt_oilgas
220
5
.
5
5
5
5
San Joaquin County, CA
Stanislaus County, CA
ptnonipm
684
351
169
177
177
177
177
Tulare County, CA
rwc
217
36
36
36
36
36
Kings County, CA
Fresno County, CA
afdust
845
74
-
74
74
74
74
Kern County, CA
nonpt
74
16
-
16
16
16
16
Madera County, CA
ptnonipm
50
-
-
-
-
-
-
Merced County, CA
rwc
30
5
_
5
5
5
5
San Joaquin County, CA
Stanislaus County, CA
Tulare County, CA
236
-------
Adjacent Counties
Maximum
(NE,SE,W) or Counties in
Annual
PM2.5
Selected PM2.5 Emissions Reductions
Same Air District (CA)
PM2.5
Emissions
County
Still Needing Reductions
Sector
Emissions
Reduction
12/35
10/35
10/30
9/35
8/35
Los Angeles County, CA
Orange County, CA
afdust
2,193
6
6
-
-
-
-
Riverside County, CA
nonpt
4,599
785
83
702
702
702
702
San Bernardino County,
np_solvents
247
-
-
-
-
-
-
CA
pt_oilgas
11
-
-
-
-
-
-
ptnonipm
1,720
583
334
249
249
249
249
rwc
917
151
-
151
151
151
151
Madera County, CA
Fresno County, CA
afdust
660
118
-
-
-
118
118
Kern County, CA
nonpt
213
34
-
-
-
34
34
Kings County, CA
ptnonipm
139
98
-
-
-
98
98
Merced County, CA
rwc
50
6
_
_
_
6
6
San Joaquin County, CA
Stanislaus County, CA
Tulare County, CA
Mendocino County, CA
-
afdust
244
55
-
-
55
-
13
nonpt
158
36
-
-
36
-
-
rwc
322
30
-
-
30
-
-
Merced County, CA
Fresno County, CA
afdust
1,287
156
-
156
156
156
156
Kern County, CA
nonpt
137
30
-
30
30
30
30
Kings County, CA
ptnonipm
80
11
-
11
11
11
11
Madera County, CA
rwc
110
13
_
13
13
13
13
San Joaquin County, CA
Stanislaus County, CA
Tulare County, CA
Mono County, CA
-
afdust
174
34
34
-
-
-
-
nonpt
12
-
-
-
-
-
-
ptnonipm
6
-
-
-
-
-
-
rwc
48
7
7
-
-
-
-
Napa County, CA
Alameda County, CA
afdust
111
20
-
-
-
-
20
San Francisco County, CA
nonpt
45
4
-
-
-
-
4
Solano County, CA
ptnonipm
46
24
-
-
-
-
24
rwc
119
22
-
-
-
-
22
237
-------
Adjacent Counties Maximum
(NE,SE,W) or Counties in Annual PM2.5 Selected PM2.5 Emissions Reductions
V*" J ...
Same Air District (CA)
PM2.5
Emissions
County
Still Needing Reductions
Sector
Emissions
Reduction
12/35
10/35
10/30
9/35
8/35
Orange County, CA
Los Angeles County, CA
afdust
654
-
-
-
-
-
-
Riverside County, CA
nonpt
1,217
184
-
184
184
184
184
San Bernardino County,
np_solvents
95
_
_
_
_
_
_
CA
ptnonipm
180
8
-
8
8
8
8
rwc
295
73
-
73
73
73
73
Plumas County, CA
-
afdust
489
149
149
-
-
-
-
nonpt
70
3
3
-
-
-
-
ptnonipm
5
-
-
-
-
-
-
rwc
316
12
12
-
-
-
-
Riverside County, CA
Los Angeles County, CA
afdust
2,525
23
23
-
-
-
-
Orange County, CA
nonpt
765
119
119
-
-
-
-
San Bernardino County,
np_solvents
98
_
_
_
_
_
_
CA
pt_oilgas
18
-
-
-
-
-
-
ptnonipm
177
33
33
-
-
-
-
rwc
454
46
46
-
-
-
-
Sacramento County, CA
-
afdust
1,002
25
-
-
15
15
25
nonpt
634
107
-
-
93
90
107
ptnonipm
79
21
-
-
21
21
21
rwc
1,734
122
-
-
73
37
122
San Bernardino County,
Los Angeles County, CA
afdust
2,357
73
64
-
-
-
-
CA
Orange County, CA
nonpt
1,733
403
403
-
-
-
-
Riverside County, CA
np_solvents
43
-
-
-
-
-
-
pt_oilgas
34
-
-
-
-
-
-
ptnonipm
2,606
1,252
1,156
76
76
76
76
rwc
456
42
42
-
-
-
-
San Diego County, CA
-
afdust
2,452
270
-
-
-
5
270
nonpt
1,772
435
-
-
-
222
435
ptnonipm
355
6
-
-
-
6
6
rwc
657
52
-
-
-
5
52
238
-------
Adjacent Counties
Maximum
(NE,SE,W) or Counties in
Annual
PM2.5
Selected PM2.5 Emissions Reductions
Same Air District (CA)
PM2.5
Emissions
County
Still Needing Reductions
Sector
Emissions
Reduction
12/35
10/35 10/30
9/35
8/35
San Francisco County, CA
Alameda County, CA
afdust
105
17
-
-
-
17
Napa County, CA
nonpt
352
55
-
-
-
55
Solano County, CA
ptnonipm
55
7
-
-
-
7
rwc
47
14
-
-
-
14
San Joaquin County, CA
Fresno County, CA
afdust
1,090
138
-
29
138
138
Kern County, CA
nonpt
350
70
-
-
70
70
Kings County, CA
ptnonipm
185
10
-
9
10
10
Madera County, CA
rwc
210
39
.
.
39
39
Merced County, CA
Stanislaus County, CA
Tulare County, CA
San Mateo County, CA
Alameda County, CA
afdust
242
42
-
-
-
6
Napa County, CA
nonpt
299
32
-
-
-
25
San Francisco County, CA
ptnonipm
145
64
-
-
-
58
Solano County, CA
rwc
162
36
-
-
-
25
Santa Barbara County, CA
-
afdust
478
57
-
57
-
13
nonpt
152
31
-
31
-
26
pt_oilgas
12
-
-
-
-
-
ptnonipm
40
-
-
-
-
-
rwc
294
42
-
42
-
29
Santa Clara County, CA
Alameda County, CA
afdust
709
124
-
-
-
7
Napa County, CA
nonpt
637
100
-
-
-
90
San Francisco County, CA
ptnonipm
611
493
-
-
320
425
Solano County, CA
rwc
594
166
-
-
-
33
Siskiyou County, CA
-
afdust
898
258
-
258
-
-
nonpt
488
219
215
4
-
-
ptnonipm
77
56
-
56
-
-
rwc
210
20
-
20
-
-
239
-------
Adjacent Counties Maximum
(NE,SE,W) or Counties in Annual PM2.5 Selected PM2.5 Emissions Reductions
V*" J ...
Same Air District (CA)
PM2.5
Emissions
County
Still Needing Reductions
Sector
Emissions
Reduction
12/35
10/35
10/30
9/35
8/35
Solano County, CA
Alameda County, CA
afdust
408
68
-
-
-
38
68
Napa County, CA
nonpt
209
31
-
-
-
29
31
San Francisco County, CA
ptnonipm
169
40
-
-
-
38
40
rwc
318
56
-
-
-
56
56
Stanislaus County, CA
Fresno County, CA
afdust
1,125
59
-
59
59
59
59
Kern County, CA
nonpt
289
59
-
59
59
59
59
Kings County, CA
ptnonipm
192
56
_
56
56
56
56
Madera County, CA
rwc
183
30
.
30
30
30
30
Merced County, CA
San Joaquin County, CA
Tulare County, CA
Sutter County, CA
-
afdust
279
39
-
-
39
39
39
nonpt
279
32
-
-
32
32
32
ptnonipm
34
-
-
-
-
-
-
rwc
193
15
-
-
15
15
15
Tehama County, CA
-
afdust
380
100
-
-
100
-
-
nonpt
137
10
-
-
10
-
-
ptnonipm
26
-
-
-
-
-
-
rwc
154
12
-
-
12
-
-
Tulare County, CA
Fresno County, CA
afdust
2,078
288
101
188
188
188
188
Kern County, CA
nonpt
282
50
49
-
-
-
-
Kings County, CA
ptnonipm
95
-
-
-
-
-
-
Madera County, CA
rwc
134
17
15
.
.
.
.
Merced County, CA
San Joaquin County, CA
Stanislaus County, CA
Ventura County, CA
-
afdust
519
73
-
-
73
3
73
nonpt
256
47
-
-
47
39
47
pt_oilgas
7
-
-
-
-
-
-
ptnonipm
99
7
-
-
7
-
7
rwc
655
147
-
-
147
-
147
240
-------
Adjacent Counties Maximum
(NE,SE,W) or Counties in Annual PM2.5 Selected PM2.5 Emissions Reductions
Same Air District (CA)
PM2.5
Emissions
County
Still Needing Reductions
Sector
Emissions
Reduction
12/35
10/35
10/30
9/35
8/35
Adams County, CO
Denver County, CO
afdust
1,884
118
-
-
-
11
118
nonpt
201
63
-
-
-
-
63
np_oilgas
12
-
-
-
-
-
-
pt_oilgas
36
16
-
-
-
-
16
ptnonipm
357
103
-
-
-
17
103
rwc
345
49
-
-
-
15
49
Denver County, CO
Adams County, CO
afdust
1,454
8
-
-
-
8
8
nonpt
248
75
-
-
-
-
75
ptnonipm
151
23
-
-
-
1
23
rwc
169
18
-
-
-
-
18
Weld County, CO
Adams County, CO
afdust
4,283
748
-
-
-
-
373
nonpt
181
67
-
-
-
-
-
np_oilgas
283
-
-
-
-
-
-
pt_oilgas
239
-
-
-
-
-
-
ptnonipm
612
193
-
-
-
-
-
rwc
270
42
-
-
-
-
-
District of Columbia, DC
-
afdust
447
49
-
-
-
-
49
nonpt
476
151
-
-
-
-
151
ptnonipm
36
-
-
-
-
-
-
rwc
17
-
-
-
-
-
-
Benewah County, ID
Shoshone County, ID
afdust
860
184
-
-
184
-
-
nonpt
33
3
-
-
3
-
-
ptnonipm
9
-
-
-
-
-
-
rwc
20
-
-
-
-
-
-
Lemhi County, ID
-
afdust
728
181
136
-
8
8
8
nonpt
12
-
-
-
-
-
-
rwc
19
-
-
-
-
-
-
Shoshone County, ID
Benewah County, ID
afdust
584
140
-
140
140
140
140
nonpt
24
11
-
11
11
11
11
rwc
27
2
-
2
2
2
2
241
-------
Adjacent Counties
(NE,SE,W) or Counties in
Same Air District (CA)
Annual
Maximum
PM2.5
Selected PM2.5 Emissions Reductions
PM2.5
Emissions
County
Still Needing Reductions
Sector
Emissions
Reduction
12/35 10/35 10/30 9/35
8/35
Boone County, IN
Marion County, IN
afdust
444
33
2
33
nonpt
98
49
42
49
rwc
70
7
3
7
Hamilton County, IN
Marion County, IN
afdust
789
90
2
90
nonpt
398
223
167
223
rwc
264
32
10
32
Hancock County, IN
Marion County, IN
afdust
315
32
2
32
nonpt
103
51
46
51
rwc
88
11
4
11
Hendricks County, IN
Marion County, IN
afdust
416
52
2
52
nonpt
242
135
105
135
ptnonipm
139
48
13
48
rwc
162
20
8
20
Johnson County, IN
Marion County, IN
afdust
395
48
3
48
nonpt
236
138
109
138
rwc
133
17
6
17
Marion County, IN
-
afdust
1,535
204
204
204
nonpt
697
182
182
182
ptnonipm
176
80
80
80
rwc
316
44
44
44
Morgan County, IN
Marion County, IN
afdust
369
41
3
41
nonpt
133
75
70
75
ptnonipm
49
45
-
45
rwc
97
13
5
13
Shelby County, IN
Marion County, IN
afdust
271
20
2
20
nonpt
76
33
30
33
ptnonipm
565
495
134
495
rwc
61
6
3
6
242
-------
Adjacent Counties Maximum
(NE,SE,W) or Counties in Annual PM2.5 Selected PM2.5 Emissions Reductions
Same Air District (CA)
PM2.5
Emissions
County
Still Needing Reductions
Sector
Emissions
Reduction 12/35 10/35 10/30
9/35
8/35
Bossier Parish, LA
Caddo Parish, LA
afdust
430
90
-
90
nonpt
445
189 ...
-
189
np_oilgas
34
-
-
-
pt_oilgas
10
-
-
-
ptnonipm
28
-
-
-
rwc
50
7 ...
-
7
Caddo Parish, LA
-
afdust
969
163 ...
74
163
nonpt
884
237
236
237
np_oilgas
76
-
-
-
pt_oilgas
15
-
-
-
ptnonipm
219
106
80
106
rwc
84
13
10
13
De Soto Parish, LA
Caddo Parish, LA
afdust
439
96
-
96
nonpt
128
41
-
41
np_oilgas
89
-
-
-
pt_oilgas
44
-
-
-
ptnonipm
348
65
-
65
rwc
14
-
-
-
East Baton Rouge Parish,
West Baton Rouge Parish,
afdust
1,735
178
-
7
LA
LA
nonpt
pt_oilgas
3,541
8
1,057
-
349
ptnonipm
1,652
1,049
-
24
rwc
104
12
-
-
East Feliciana Parish, LA
West Baton Rouge Parish,
afdust
274
60
-
9
LA
nonpt
pt_oilgas
rwc
69
24
10
30
-
29
243
-------
Adjacent Counties
Maximum
(NE,SE,W) or Counties in
Annual
PM2.5
Selected PM2.5 Emissions Reductions
Same Air District (CA)
PM2.5
Emissions
County
Still Needing Reductions
Sector
Emissions
Reduction
12/35
10/35 10/30 9/35
8/35
Iberville Parish, LA
West Baton Rouge Parish,
afdust
326
39
-
-
4
LA
nonpt
375
113
-
-
86
pt_oilgas
46
-
-
-
-
ptnonipm
541
83
-
-
19
rwc
13
-
-
-
-
Pointe Coupee Parish, LA
West Baton Rouge Parish,
afdust
533
80
-
-
11
LA
nonpt
66
21
-
-
19
ptnonipm
256
3
-
-
3
rwc
10
-
-
-
-
Red River Parish, LA
Caddo Parish, LA
afdust
201
45
-
-
45
nonpt
55
11
-
-
11
np_oilgas
21
-
-
-
-
pt_oilgas
7
-
-
-
-
ptnonipm
777
757
-
-
757
West Baton Rouge Parish,
-
afdust
249
48
-
-
48
LA
nonpt
282
77
-
-
77
ptnonipm
398
166
-
-
166
rwc
9
-
-
-
-
West Feliciana Parish, LA
West Baton Rouge Parish,
afdust
189
44
-
-
8
LA
nonpt
58
24
-
-
23
ptnonipm
133
129
-
-
7
rwc
5
-
-
-
-
Flathead County, MT
Lewis and Clark County,
afdust
4,052
1,105
-
19
-
MT
nonpt
289
123
-
101
-
Lincoln County, MT
ptnonipm
172
95
-
14
-
rwc
173
28
-
14
-
Lewis and Clark County,
-
afdust
1,686
454
79
22
22
MT
nonpt
135
65
42
22
22
rwc
82
13
6
7
7
244
-------
Adjacent Counties Maximum
(NE,SE,W) or Counties in Annual PM2.5 Selected PM2.5 Emissions Reductions
Same Air District (CA)
PM2.5
Emissions
County
Still Needing Reductions
Sector
Emissions
Reduction 12/35
10/35
10/30
9/35
8/35
Lincoln County, MT
-
afdust
1,026
296
296
296
296
296
nonpt
40
12
12
12
12
12
rwc
64
9
9
9
9
9
Atlantic County, NJ
Camden County, NJ
afdust
278
70
-
-
-
70
nonpt
175
44
-
-
-
44
ptnonipm
18
-
-
-
-
-
rwc
252
41
-
-
-
41
Bergen County, NJ
-
afdust
451
93
-
-
93
93
nonpt
770
204
-
-
204
204
ptnonipm
23
-
-
-
-
-
rwc
333
41
-
-
41
41
Burlington County, NJ
Camden County, NJ
afdust
479
108
-
-
-
108
nonpt
261
59
-
-
-
59
ptnonipm
40
-
-
-
-
-
rwc
539
90
-
-
-
90
Camden County, NJ
-
afdust
277
58
-
-
48
58
nonpt
319
78
-
-
73
78
ptnonipm
22
-
-
-
-
-
rwc
231
47
-
-
47
47
Essex County, NJ
Bergen County, NJ
afdust
345
71
-
-
71
71
nonpt
501
124
-
-
124
124
ptnonipm
46
15
-
-
15
15
rwc
148
14
-
-
14
14
Gloucester County, NJ
Camden County, NJ
afdust
269
52
-
-
-
52
nonpt
150
33
-
-
-
33
ptnonipm
265
189
-
-
-
189
rwc
284
44
-
-
-
44
Hudson County, NJ
Bergen County, NJ
afdust
196
36
-
-
36
36
nonpt
424
111
-
-
111
111
ptnonipm
21
-
-
-
-
-
rwc
11
-
-
-
-
-
245
-------
Adjacent Counties Maximum
(NE,SE,W) or Counties in Annual PM2.5 Selected PM2.5 Emissions Reductions
Same Air District (CA)
PM2.5
Emissions
County
Still Needing Reductions
Sector
Emissions
Reduction 12/35
10/35
10/30
9/35
8/35
Passaic County, NJ
Bergen County, NJ
afdust
220
37
-
-
37
37
nonpt
331
84
-
-
84
84
rwc
164
23
-
-
23
23
Butler County, OH
Hamilton County, OH
afdust
645
102
-
-
102
102
nonpt
460
199
-
-
193
199
ptnonipm
683
451
-
-
451
451
rwc
336
41
-
-
37
41
Clermont County, OH
Hamilton County, OH
afdust
513
95
-
-
95
95
nonpt
351
207
-
-
207
207
ptnonipm
13
-
-
-
-
-
rwc
251
31
-
-
31
31
Hamilton County, OH
-
afdust
1,193
129
-
-
129
129
nonpt
977
381
282
282
381
381
ptnonipm
171
16
10
10
16
16
rwc
357
55
-
-
55
55
Warren County, OH
Hamilton County, OH
afdust
533
87
-
-
87
87
nonpt
495
308
-
-
308
308
pt_oilgas
20
-
-
-
-
-
ptnonipm
28
10
-
-
10
10
rwc
242
31
-
-
31
31
Crook County, OR
Harney County, OR
afdust
1,134
326
-
246
-
65
nonpt
37
18
-
18
-
17
ptnonipm
6
-
-
-
-
-
rwc
89
12
-
12
-
7
Harney County, OR
Lake County, OR
afdust
nonpt
1,341
11
268
-
113
212
268
rwc
31
3
-
2
-
3
Jackson County, OR
Klamath County, OR
afdust
1,916
568
-
12
12
336
nonpt
379
233
-
-
206
230
ptnonipm
167
134
-
-
-
63
rwc
566
81
-
-
20
77
246
-------
Adjacent Counties Maximum
(NE,SE,W) or Counties in Annual PM2.5 Selected PM2.5 Emissions Reductions
Same Air District (CA)
PM2.5
Emissions
County
Still Needing Reductions
Sector
Emissions
Reduction
12/35
10/35
10/30
9/35
8/35
Klamath County, OR
Lake County, OR
afdust
2,945
836
425
-
38
38
38
nonpt
103
48
39
-
2
2
2
pt_oilgas
9
-
-
-
-
-
-
ptnonipm
217
184
99
-
85
85
85
rwc
223
32
23
-
-
-
-
Lake County, OR
Harney County, OR
afdust
1,118
248
248
-
-
-
-
Klamath County, OR
nonpt
13
4
4
-
-
-
-
rwc
35
5
5
-
-
-
-
Lane County, OR
Klamath County, OR
afdust
4,643
1,304
-
-
12
-
12
nonpt
622
377
-
-
127
-
131
ptnonipm
436
377
-
-
30
-
33
rwc
852
120
-
-
26
-
30
Chester County, PA
Delaware County, PA
afdust
877
97
-
-
-
5
97
nonpt
961
459
-
-
-
459
431
pt_oilgas
12
-
-
-
-
-
-
ptnonipm
84
9
-
-
-
-
9
rwc
417
56
-
-
-
56
49
Delaware County, PA
-
afdust
368
51
-
-
-
51
51
nonpt
554
110
-
28
28
110
110
ptnonipm
314
199
-
-
-
199
199
rwc
130
23
-
1
1
23
23
Montgomery County, PA
Delaware County, PA
afdust
985
103
-
-
-
2
103
nonpt
1,376
389
-
-
-
389
389
pt_oilgas
9
-
-
-
-
-
-
ptnonipm
288
129
-
-
-
104
129
rwc
415
52
-
-
-
52
52
Philadelphia County, PA
Delaware County, PA
afdust
596
78
-
-
-
-
11
nonpt
1,349
330
-
-
-
330
330
ptnonipm
439
176
-
-
-
109
176
rwc
40
5
-
-
-
3
4
247
-------
Adjacent Counties Maximum
(NE,SE,W) or Counties in Annual PM2.5 Selected PM2.5 Emissions Reductions
Same Air District (CA)
PM2.5
Emissions
County
Still Needing Reductions
Sector
Emissions
Reduction 12/35
10/35
10/30
9/35
8/35
Brooks County, TX
Hidalgo County, TX
afdust
466
108
16
16
108
108
np_oilgas
7
-
-
-
-
-
Cameron County, TX
Hidalgo County, TX
afdust
915
128
-
-
128
128
nonpt
217
80
-
-
80
80
ptnonipm
43
-
-
-
-
-
rwc
35
3
-
-
3
3
El Paso County, TX
-
afdust
1,605
78
-
-
-
78
nonpt
258
124
-
-
-
124
pt_oilgas
7
-
-
-
-
-
ptnonipm
120
28
-
-
-
28
rwc
58
8
-
-
-
8
Hidalgo County, TX
Cameron County, TX
afdust
1,755
261
261
261
261
261
nonpt
462
192
192
192
192
192
np_oilgas
19
-
-
-
-
-
ptnonipm
85
60
60
60
60
60
rwc
58
8
8
8
8
8
Hudspeth County, TX
El Paso County, TX
afdust
248
56
-
-
-
56
pt_oilgas
18
-
-
-
-
-
Kenedy County, TX
Hidalgo County, TX
afdust
268
78
18
18
78
78
Starr County, TX
Hidalgo County, TX
afdust
489
107
-
-
107
107
nonpt
50
18
-
-
18
18
np_oilgas
16
-
-
-
-
-
pt_oilgas
11
-
-
-
-
-
rwc
8
-
-
-
-
-
Willacy County, TX
Cameron County, TX
afdust
357
43
11
11
43
43
Hidalgo County, TX
nonpt
11
-
-
-
-
-
King County, WA
Yakima County, WA
afdust
3,808
154
-
-
-
9
nonpt
2,503
632
-
-
-
28
ptnonipm
90
51
-
-
-
5
rwc
1,864
204
-
-
-
-
248
-------
Adjacent Counties
(NE,SE,W) or Counties in
Same Air District (CA)
Annual
Maximum
PM2.5
Selected PM2.5 Emissions Reductions
PM2.5
Emissions
County
Still Needing Reductions
Sector
Emissions
Reduction
12/35 10/35 10/30 9/35 8/35
Okanogan County, WA
-
afdust
769
176
176 - 78
nonpt
41
10
I
o
1
1
rwc
149
23
23 - 23
Yakima County, WA
-
afdust
1,839
203
203 -
nonpt
220
60
60
ptnonipm
6
-
-
rwc
335
53
53
249
-------
CHAPTER 4: ENGINEERING COST ANALYSIS AND QUALITATIVE DISCUSSION OF
SOCIAL COSTS
Overview
This chapter provides estimates of the engineering costs of the illustrative control
strategies identified in Chapter 3 for the revised annual and current 24-hour alternative
standard levels of 9/35 ng/m3, as well as the following less and more stringent alternative
standard levels 10/35 ng/m3,10/30 ng/m3, and 8/35 |~ig/m3. Because the EPA is retaining
the current secondary PM standards, we did not evaluate alternative secondary standard
levels in this RIA. The chapter summarizes the methods, tools, and data sources used to
estimate the engineering costs presented. As discussed in Chapter 3, for the revised and
alternative standards analyzed we applied end-of-pipe control technologies and area
source controls to sources in the following emissions inventory sectors: non-electric
generating unit (non-EGU) point, oil and gas point, non-point (area), residential wood
combustion, and area fugitive dust.
The estimated costs for the revised and alternative standard levels are a function of
(i) assumptions used in the analysis, including assumptions about which areas will require
emissions controls and the sources and controls available in those areas; (ii) the level of
sufficient, detailed information on emissions sources and end-of-pipe and area source
controls needed to estimate engineering costs; and (iii) the future year baseline emissions
from which the emissions reductions are measured.
For the less stringent alternative standard level of 10/35 |~ig/m3, because 13 of the
20 counties that need emissions reductions are counties in California, the majority of the
estimated costs are incurred in California. As the alternative standard levels become more
stringent, more counties in the northeast and southeast need emissions reductions. For
revised and more stringent standard levels of 9/35 |~ig/m3 and 8/35 ng/m3, more controls
are available to apply in the northeast and their adjacent counties and the southeast and
their adjacent counties. As additional controls are applied, those areas account for a
relatively higher proportion of estimated costs compared to the west and California
because availability of additional controls is limited for those areas. Note that in the
northeast and southeast we identified controls and associated emissions reductions from
250
-------
adjacent counties and used a ppb/ton PIVh.sair quality ratio that was four times less
responsive than the ratio used when applying in-county emissions reductions (i.e., applied
four tons of PM2.5 emissions reductions from an adjacent county for one ton of emissions
reduction needed in a given county); the cost of the additional reductions from adjacent
counties contributes to the higher proportion of the estimated costs. Lastly, for the more
stringent alternative standard levels of 8/35 |~ig/m3, the largest share of estimated costs is
from controls for area fugitive dust emissions across all areas.
The remainder of the chapter is organized as follows. Section 4.1 presents the
engineering costs associated with the application of controls identified in EPA's national-
scale analysis. Section 4.2 provides a discussion of the uncertainties and limitations
associated with the engineering cost estimates. Section 4.3 includes a qualitative discussion
on social costs. Section 4.4 includes references.
4.1 Estimating Engineering Costs
The engineering costs described in this chapter generally include the costs of
purchasing, installing, operating, and maintaining the controls applied. The costs associated
with monitoring, testing, reporting, and recordkeeping for potentially affected sources are
not included in the annualized cost estimates. These cost estimates are presented for 2032
but reflect the annual cost that is expected to be incurred each year over a longer time
horizon. We calculate the present value of these annual costs over 20 years in Chapter 8
using 3 and 7 percent discount rates.
This analysis focuses on emissions reductions needed for the revised and alternative
standard levels. As discussed in this analysis, the end-of-pipe and area source controls and
strategies selected for analysis were from information available in EPA's control measures
database; these control strategies illustrate one way in which nonattainment areas could
work toward meeting a revised standard. There are many ways to construct and evaluate
potential control programs for a revised standard, and the EPA anticipates that state and
local governments will consider programs best suited for local conditions.
The EPA understands that some states will incur costs both designing State
Implementation Plans (SIPs) and implementing new control strategies to meet a revised
251
-------
standard. However, the EPA does not know what specific actions states will take to design
their SIPs to meet a revised standard. Therefore, we do not present estimated costs that
government agencies may incur for managing the requirement or implementing these (or
other) control strategies.
4.1.1 Methods, Tools, and Data
The EPA uses the Control Strategy Tool (CoST) (U.S. EPA, 2019a) to estimate
engineering control costs. CoST models emissions reductions and control costs associated
with the application of end-of-pipe and area source controls by matching the controls in the
control measures database (CMDB) to emissions sources in the future year projected
emissions inventory by source classification code (SCC).1*2 CoST was used in two ways in
the analysis. First, CoST was used to identify controls and related potential PM2.5 emissions
reductions in counties projected to exceed the revised and alternative annual and 24-hour
standard levels of 10/35 |~ig/m3,10/30 |~ig/m3, 9/35 |~ig/m3, and 8/35 |~ig/m3 in the
analytical baseline (see Chapter 3, Section 3.2.1 for a discussion of the counties and areas).
Second, CoST was used to estimate the control costs for the technologies identified. As
indicated in Chapter 3, Section 3.2.2, for the control strategy analyses in this RIA, to
maximize the number of emissions sources included we applied controls to emissions
sources with greater than 5 tons per year of PM2.5 emissions at a marginal cost threshold of
up to a $160,000/ton.
CoST calculates engineering costs using one of two different methods: (1) an
equation that incorporates key operating unit information, such as unit design capacity or
stack flow rate, or (2) an average annualized cost-per-ton factor multiplied by the total tons
of reduction of a pollutant. Most control cost information within CoST was developed based
on the cost-per-ton approach because (1) parameters used in the engineering equations
are not readily available or broadly representative across emissions sources within the
emissions inventory and (2) estimating engineering costs using an equation requires data
from the emissions inventory, which may not be available. The cost equations used in CoST
1 More information about CoST and the control measures database can be found at the following link:
https://www.epa.gov/economic-and-cost-analysis-air-pollution-regulations/cost-analysis-modelstools-air-
pollution.
2 We used a 2016-based modeling platform to project future-year emissions and air quality for 2032.
252
-------
estimate annual, capital and/or operating and maintenance (O&M) costs and are used
primarily for some larger emissions sources such as industrial, commercial, and
institutional (ICI) boilers, glass manufacturing furnaces, and cement kilns.
CoST gets key operating unit information from the emissions inventory data
submitted by state, local, and tribal air agencies (S/L/T), including detailed information by
source on emissions, installed control devices, and control device efficiency. Much of this
underlying emissions inventory data serves as key inputs into CoST and the control
strategy analyses. The information on whether a source is currently controlled, by what
control device, and control device efficiency is required under the Air Emissions Reporting
Rule (AERR) used to collect the emissions inventory data. However, control information
may not be fully reported by S/L/T agencies and would not be available for purposes of the
control strategy analyses, introducing the possibility that CoST applies controls to already
controlled emissions sources.
When sufficient information is available to estimate control costs using equations,
the capital costs of the control equipment must be annualized. Capital costs are converted
to annual costs using the capital recovery factor (CRF).3 The engineering cost analysis uses
the equivalent uniform annual costs (EUAC) method, in which annualized costs are
calculated based on the equipment life for the control and the interest rate incorporated
into the CRF. Annualized costs represent an equal stream of yearly costs over the period
the control is expected to operate. For more information on the EUAC method, refer to the
EPA Air Pollution Control Cost Manual (U.S. EPA, 2017a).
4.1.2 Cost Estimates for the Control Strategies
In this section, we provide engineering cost estimates for the end-of-pipe control
technologies and area source controls presented in Chapter 3 that include end-of-pipe
control technologies for non-EGU point sources, oil and gas point, non-point (area) sources,
residential wood combustion sources, and area fugitive dust emissions. The cost estimates
presented in Table 4-1 through Table 4-5 reflect the engineering costs annualized at 7
3 The capital recovery factor incorporates the interest rate and equipment life (in years) of the control
equipment. The capital recovery factor formula is expressed as r*(l+r]n/[(l+rOn -1], where r is the real rate
of interest and n is the number of time periods. The annualized costs assumed a 7 percent interest rate.
253
-------
percent, to the extent possible.4 When calculating the annualized costs we prefer to use the
interest rates faced by firms; however, we do not know what those rates will be.
By area, Table 4-1 includes a summary of estimated control costs from control
applications for the revised and alternative standard levels analyzed. Table 4A-1 through
Table 4A-6 in Appendix 4A include detailed information on estimated costs by area and by
county.
Table 4-1 By Area, Summary of Annualized Control Costs for Revised and
Alternative Primary Standard Levels of 10/35 |ig/m3,10/30 |ig/m3,
9/35 (ig/m3, and 8/35 (ig/m3 for 2032 (millions of 2017$)
Area
10/35
10/30
9/35
8/35
Northeast
$5.3
$5.5
$203.6
$371.1
Northeast (Adjacent Counties)
$0
$0
$62.1
$364.2
Southeast
$35.8
$35.8
$60.4
$299.7
Southeast (Adjacent Counties)
$0.02
$0.02
$25.5
$69.2
West
$39.7
$112.4
$57.7
$140.6
CA
$121.8
$186.1
$184.4
$256.7
Total
$202.5
$339.8
$593.8
$1,501.5
Note: Costs associated with monitoring, testing, reporting, and recordkeeping for potentially affected
sources are not included in these estimates.
For the less stringent alternative standard levels of 10/35 |~ig/m3, the majority of the
estimated costs are incurred in California because 13 of the 20 counties that need
emissions reductions are located in California. Looking at the alternative standard levels of
10/30 ng/m3 in the west, an additional 17 counties need emissions reductions, and the
estimated costs increase; estimated costs for the revised alternative standard levels of 9/35
Hg/m3 are higher than for 10/35 |~ig/m3 but lower than for 10/30 |~ig/m3 in this area. For
the revised and more stringent alternative standard levels of 9/35 |~ig/m3 and 8/35 ng/m3,
more controls are available to apply in the northeast and the southeast as compared to in
California and the west. Therefore, the estimated costs for the northeast and the southeast
4 Because we obtain control cost data from many sources, we are not always able to obtain consistent data
across original data sources. As a result, we do not know the interest rates used to calculate costs for some
of the controls included in this analysis. If disaggregated control cost data is available (i.e., where capital,
equipment life value, and O&M costs are separated out) we can calculate costs using a specified percent
interest rate. EPA may not know the interest rates used to calculate costs when disaggregated control cost
data is unavailable (i.e., where we only have a $/ton value and where capital, equipment life value, and O&M
costs are not separated out).
254
-------
are significantly higher for 9/35 |~ig/m3 and 8/35 |~ig/m3. See Table 3A-2 through Table
3A-7 for more details on emissions reductions available by area and county.
As discussed in Chapter 3, when we applied the emissions reductions from adjacent
counties in the northeast and southeast, we applied a ratio of 4:1 in which four tons of
PM2.5 emissions reductions from an adjacent county are needed to produce the equivalent
air quality change of one ton of emissions reduction if it had occurred within the county
needing the reduction (see Appendix 2A, Section 2A.3.1 for a discussion of how the ratio
was developed). Application of this ratio contributes to the higher cost estimates for
revised and alternative standard levels of 9/35 |~ig/m3 and 8/35 ng/m3. It is anticipated
that states will first attempt to find emissions reductions within the counties that need the
reductions. To the extent that states are able to identify control opportunities within those
counties beyond the reductions identified by CoST, the need for reductions from adjacent
counties will be reduced. Also, depending on local air quality factors, the resulting air
quality impact may be greater than a 4:1 ratio suggests. As a result, the estimate of costs for
adjacent counties maybe an overestimate.
By emissions inventory sector, Table 4-2 includes a summary of the estimated costs
from control applications for the revised and alternative standard levels analyzed. For all of
the standard levels analyzed, area source controls for area fugitive dust emissions
comprise the largest share of the estimated costs, ranging from 65 to 79 percent of the cost
estimates. Non-EGU point and non-point (area) controls represent the next largest shares
of the cost estimates.
By area and by emissions inventory sector, Table 4-3 includes a summary of the
estimated costs from control applications for the revised and alternative standard levels
analyzed. For the more stringent alternative standard levels of 8/35 ng/m3 across all areas
the largest share of estimated costs is from controls for area fugitive dust emissions. In
addition, as the standard levels become more stringent, more counties in the northeast and
southeast need emissions reductions and controls are applied in those areas (and less so in
the west and California because availability of additional controls is limited), resulting in a
relatively higher proportion of estimated costs for those areas.
255
-------
Table 4-2 By Emissions Inventory Sector, Summary of Annualized Control Costs
for Revised and Alternative Primary Standard Levels of 10/35 ng/m3,
10/30 |ig/m3,9/35 (j,g/m3, and 8/35 jig/m3 for 2032 (millions of
2017$)
Sector
10/35
10/30
9/35
8/35
Non-EGU Point
$29.0
$35.6
$114.1
$275.5
Oil & Gas Point
$0.4
$0.4
$0.4
$0.8
Non-Point (Area)
$20.4
$27.4
$75.3
$203.0
Residential Wood Combustion
$4.3
$9.3
$16.6
$39.3
Area Source Fugitive Dust
$148.5
$267.1
$387.4
$983.0
Total
$202.5
$339.8
$593.8
$1,501.5
Table 4-3 By Area and by Emissions Inventory Sector, Summary of Annualized
Control Costs for Revised and Alternative Primary Standard Levels of
10/35 |ig/m3,10/30 |ig/m3, 9/35 (j,g/m3, and 8/35 |ig/m3 for 2032
(millions of 2017$)
Area
Sector
10/35
10/30
9/35
8/35
Northeast
Non-EGU Point
$0.5
$0.5
$69.4
$144.2
Non-Point (Area)
$4.7
$5.0
$33.0
$57.0
Residential Wood Combustion
$0.03
$0.03
$7.1
$8.8
Area Source Fugitive Dust
$0
$0
$94.1
$161.1
Northeast
Non-EGU Point
$0
$0
$2.4
$26.8
(Adjacent
Non-Point (Area)
$0
$0
$12.9
$45.7
Counties)
Residential Wood Combustion
$0
$0
$1.4
$8.9
Area Source Fugitive Dust
$0
$0
$45.3
$282.7
Southeast
Non-EGU Point
$2.9
$2.9
$4.1
$52.0
Non-Point (Area)
$1.8
$1.8
$8.0
$53.9
Residential Wood Combustion
$0.08
$0.08
$0.5
$5.3
Area Source Fugitive Dust
$31.0
$31.0
$47.9
$188.6
Southeast
Non-EGU Point
$0
$0
$0
$6.0
(Adjacent
Non-Point (Area)
$0
$0
$0.1
$6.5
Counties)
Residential Wood Combustion
$0
$0
$0
$0.2
Area Source Fugitive Dust
$0.02
$0.02
$25.4
$56.6
West
Non-EGU Point
$0
$3.9
$3.9
$8.1
Oil & Gas Point
$0
$0
$0
$0.4
Non-Point (Area)
$0.1
$3.9
$2.4
$13.9
Residential Wood Combustion
$0.05
$0.6
$0.1
$3.5
Area Source Fugitive Dust
$39.5
$104.0
$51.4
$114.7
CA
Non-EGU Point
$25.6
$28.3
$34.2
$38.3
Oil & Gas Point
$0.4
$0.4
$0.4
$0.4
Non-Point (Area)
$13.7
$16.8
$18.9
$25.9
Residential Wood Combustion
$4.2
$8.6
$7.6
$12.7
Area Source Fugitive Dust
$77.9
$132.1
$123.4
$179.3
Total
$202.5
$339.8
$593.8
$1,501.5
256
-------
By control technology, Table 4-4 includes a summary of the estimated costs from
control applications for the revised and alternative standard levels analyzed. Across all of
the standard levels analyzed, the end-of-pipe and area source controls that comprise more
than 80 percent of the cost estimates include Pave Existing Shoulders and Pave Unpaved
Roads (area fugitive dust inventory sector), Fabric Filter-All Types (non-EGU point
inventory sector), and Electrostatic Precipitator (non-point (area) inventory sector).
By emissions inventory sector and by control technology, Table 4-5 includes a
summary of the cost estimates. Across all of the standard levels analyzed, for the non-EGU
point sector, the application of Fabric Filter-All Types results in the highest portion of
estimated costs for that inventory sector; for the non-point (area) sector, the application of
Electrostatic Precipitator and Substitute Chipping for Burning result in the highest portion
of estimated costs for that inventory sector; for the residential wood combustion sector,
the application of Convert to Gas Logs results in the highest portion of estimated costs for
that inventory sector; and for the area fugitive dust sector, the application of Pave Existing
Shoulders and Pave Unpaved Roads result in the highest portion of estimated costs for that
inventory sector.
257
-------
Table 4-4 By Control Technology, Summary of Annualized Control Costs for
Revised and Alternative Primary Standard Levels of 10/35 |jg/m3,
10/30 |ig/m3,9/35 (j,g/m3, and 8/35 jig/m3 for 2032 (millions of
2017$)
Control Technology 10/35 10/30 9/35 8/35
Add-on Scrubber
$0
$0
$0
$0.05
Annual tune-up
$0.4
$0.9
$6.2
$14.7
Biennial tune-up
$0.8
$0.9
$0.1
$1.2
Catalytic oxidizers
$0
$0.6
$2.4
$0.7
Chemical Stabilizer
$0
$18.7
$7.5
$152.5
Convert to Gas Logs
$4.2
$8.0
$14.4
$30.5
EPA Phase 2 Qualified Units
$0
$0.8
$0.1
$2.0
EPA-certified wood stove
$0.01
$0.01
$0.0
$0
Electrostatic Precipitator
$14.5
$17.2
$45.8
$109.9
Electrostatic Precipitator-All Types
$0
$0
$0.4
$0
Fabric Filter-All Types
$29.4
$35.9
$107.9
$266.0
HE PA filters
$0
$0
$0.01
$0.01
Install Cleaner Hydronic Heaters
$0.01
$0.05
$0.1
$0.5
Install Retrofit Devices
$0
$0
$0.5
$0.3
Install new drift eliminator
$0
$0
$0.7
$2.2
New gas stove or gas logs
$0.07
$0.4
$1.5
$5.9
Pave Unpaved Roads
$75.6
$133.9
$122.4
$199.2
Pave existing shoulders
$72.7
$114.1
$257.2
$630.4
Smokeless Broiler
$0.6
$0.6
$3.3
$6.9
Substitute chipping for burning
$4.0
$7.2
$17.4
$69.6
Venturi Scrubber
$0.01
$0.05
$5.5
$8.0
Watering
$0.1
$0.4
$0.3
$1.0
Total
$202.5
$339.8
$593.8
$1,501.5
258
-------
Table 4-5 By Emissions Inventory Sector and Control Technology, Summary of
Annualized Control Costs for Revised and Alternative Primary
Standard Levels of 10/35 |ig/m3,10/30 |ig/m3, 9/35 |ig/m3, and 8/35
(ig/m3 for 2032 (millions of 2017$)
Inventory Sector
Control Technology
10/35
10/30
9/35
8/35
Non-EGU Point
Electrostatic Precipitator-All Types
$0
$0
$0.4
$0
Fabric Filter-All Types
$29.0
$35.5
$107.5
$265.2
Install new drift eliminator
$0
$0
$0.7
$2.2
Venturi Scrubber
$0.01
$0.05
$5.5
$8.0
Oil & Gas Point
Fabric Filter-All Types
$0.4
$0.4
$0.4
$0.8
Non-Point (Area)
Add-on Scrubber
$0
$0
$0
$0.05
Annual tune-up
$0.4
$0.9
$6.2
$14.7
Biennial tune-up
$0.8
$0.9
$0.1
$1.2
Catalytic oxidizers
$0
$0.6
$2.4
$0.7
Electrostatic Precipitator
$14.5
$17.2
$45.8
$109.9
HE PA filters
$0
$0
$0.01
$0.01
Smokeless Broiler
$0.6
$0.6
$3.3
$6.9
Substitute chipping for burning
$4.0
$7.2
$17.4
$69.6
Residential Wood
Convert to Gas Logs
$4.2
$8.0
$14.4
$30.5
Combustion
EPA Phase 2 Qualified Units
$0
$0.8
$0.1
$2.0
EPA-certified wood stove
$0.01
$0.01
$0.0
$0
Install Cleaner Hydronic Heaters
$0.01
$0.05
$0.1
$0.5
Install Retrofit Devices
$0
$0
$0.5
$0.3
New gas stove or gas logs
$0.07
$0.4
$1.5
$5.9
Area Source
Chemical Stabilizer
$0
$18.7
$7.5
$152.5
Fugitive Dust
Pave Unpaved Roads
$75.6
$133.9
$122.4
$199.2
Pave existing shoulders
$72.7
$114.1
$257.2
$630.4
Watering
$0.1
$0.4
$0.3
$1.0
Total
$202.5
$339.8
$593.8
$1,501.5
As discussed in Chapter 2, Section 2.4 and Chapter 3, Section 3.2.5 for the revised
standard levels of 9/35 ng/m3, there are remaining air quality challenges for areas in the
northeast and southeast, as well as in the west and California; the areas include 2 counties
with near-road monitors, 3 counties in border areas, 5 counties in small western mountain
valleys, and 15 additional counties in California's air districts and basins. The
characteristics of the air quality challenges for these areas include features of certain near-
road sites with challenging local conditions, cross-border transport, effects of complex
terrain in the west and California, and identifying wildfire influence on projected PM2.5 DVs
that could qualify for exclusion as atypical, extreme, or unrepresentative events (U.S. EPA,
2019b). To the extent that state and local areas are able to find alternative lower-cost
approaches to reducing emissions, the annualized control costs above may be
259
-------
overestimated. To the extent that additional PM2.5 emissions reductions are required that
were not identified in our analysis of these areas, the annualized control costs above may
be underestimated.
4.2 Limitations and Uncertainties in Engineering Cost Estimates
The EPA acknowledges several important limitations of this analysis, including:
• Exclusions from the Cost Analysis: As indicated, recordkeeping, reporting,
testing, and monitoring costs are not included. In addition, the costs some
states will incur both designing SIPs and implementing new control
strategies to meet a revised standard are not included.
• Cost and Effectiveness of Controls: We are not able to account for regional
or local variation in capital and annual cost items such as energy, labor, or
materials. The estimates of control efficiencies assume that the control
devices are properly installed and maintained. The estimates of control
efficiencies do not account for differences in individual applications because
we use a single value for each control that does not account for differences in
individual applications; a control may operate more or less effectively than
the specified efficiency. In addition, variability in scale of control application
is difficult to reflect for small area sources of emissions.
• Interest Rate: Because we obtain control cost data from many sources, we
are not always able to obtain consistent data across original data sources. If
disaggregated control cost data is available (i.e., where capital, equipment life
value, and O&M costs are separated out) we can calculate costs using a
specified percent interest rate. The EPA may not know the interest rates used
to calculate costs if disaggregated control cost data is unavailable (i.e., where
we only have a $/ton value and where capital, equipment life value, and O&M
costs are not separated out). In general, we have some disaggregated data
available for non-EGU point source controls, but we do not have any
disaggregated control cost data for non-point (area) source controls.
260
-------
• Differences Between ex ante and ex post Compliance Cost Estimates: In
comparing regulatory cost estimates before and after regulation, ex ante cost
estimate predictions may differ from actual costs. Harrington etal. (2000)
surveyed the predicted and actual costs of 28 federal and state rules,
including 21 issued by the U.S. Environmental Protection Agency and the
Occupational Safety and Health Administration (OSHA). In 14 of the 28 rules,
predicted total costs were overestimated, while analysts underestimated
costs in three of the remaining rules. In EPA rules where per-unit costs were
specifically evaluated, costs of regulations were overestimated in five cases,
underestimated in four cases, and accurately estimated in four cases
(Harrington et al., 2000). The collection of literature regarding the accuracy
of cost estimates seems to reflect these splits. A recent EPA report, the
"Retrospective Study of the Costs of EPA Regulations" that examined the
compliance costs of five EPA regulations in four case studies,5 found that
several of the case studies suggested that cost estimates were over-estimated
ex ante and did not find the evidence to be conclusive. The EPA stated in the
report that the small number of regulatory actions covered, as well as
significant data and analytical challenges associated with the case studies
limited the certainty of this conclusion (U.S. EPA, 2014).
4.3 Social Costs
As discussed in EPA's Guidelines for Preparing Economic Analyses, social costs are
the total economic burden of a regulatory action (U.S. EPA, 2016). This burden is the sum of
all opportunity costs incurred due to the regulatory action, where an opportunity cost is
the value lost to society of any goods and services that will not be produced and consumed
as a result of reallocating some resources toward pollution mitigation. Estimates of social
costs may be compared to the social benefits expected as a result of a regulation to assess
5 The four case studies in the 2014 Retrospective Study of the Costs of EPA Regulations examine five EPA
regulations: the 2001/2004 National Emission Standards for Hazardous Air Pollutants and Effluent
Limitations Guidelines, Pretreatment Standards, and New Source Performance Standards on the Pulp and
Paper Industry; Critical Use Exemptions for Use of Methyl Bromide for Growing Open Field Fresh
Strawberries in California for the 2004-2008 Seasons; the 2001 National Primary Drinking Water
Regulations for Arsenic; and the 1998 Locomotive Emission Standards.
261
-------
its net impact on society. Computable General Equilibrium (CGE) models are a class of
economy-wide models that can be used to evaluate the broader impacts of a regulatory
action and can therefore be used to estimate social costs. While CGE modeling was not
conducted for this analysis, we include a qualitative discussion of its use in evaluating
social costs and economic impact modeling.
Economic impacts focus on the behavioral response to the costs imposed by a policy
being analyzed. The responses typically analyzed are market changes in prices, quantities
produced and purchased, changes in international trade, changes in profitability, facility
closures, and employment. Sometimes these behavioral changes can be used to estimate
social costs if there is indication that the social costs differ from the estimate of control
costs because behavioral change results in other ways of meeting the requirements (e.g.,
facilities choosing to reduce emissions by producing less rather than adding pollution
control devices). Changes in production in a directly regulated sector may have indirect
effects on a myriad of other markets when output from that sector is used as an input in the
production of many other goods. It may also affect upstream industries that supply goods
and services to the sector, along with labor and capital markets, as these suppliers alter
production processes in response to changes in factor prices. In addition, households may
change their demand for particular goods and services due to changes in the price of those
goods.
When new regulatory requirements are expected to result in effects outside of
regulated and closely related sectors, a key challenge is determining whether they are of
sufficient magnitude to warrant explicit evaluation (Hahn and Hird 1990). It is not possible
to estimate the magnitude and direction of all of these potential effects outside of the
regulated sector(s) without an economy-wide modeling approach. For example, studies of
air pollution regulations for the power sector have found that the social costs and benefits
may be greater or lower than when secondary market impacts are taken into account, and
that the direction of the estimates may depend on the form of the regulation (e.g., Goulder
et al. 1999, Williams 2002, Goulder et al. 2016).
The revised and alternative standard levels analyzed are anticipated to impact
multiple markets in many places over time. CGE models are a class of economy-wide
262
-------
models that could be used to evaluate the impacts of a regulation on the broader economy
because they explicitly capture interactions between markets across the entire economy.
While a CGE model captures the effects of behavioral responses on the part of consumers
or other producers to changes in price that are missed by an engineering estimate of
compliance costs, most CGE models do not model the environmental externality or the
benefits that accrue to society from mitigating the externality. When benefits from a
regulation are expected to be substantial, social cost cannot be interpreted as a complete
characterization of economic welfare.6 To the extent that the benefits affect behavioral
responses in markets, the social cost measure may also be potentially biased.
A CGE-based approach to cost estimation concurrently considers the effect of a
regulation across all sectors in the economy. It is structured around the assumption that,
for some discrete period of time, an economy can be characterized by a set of equilibrium
conditions in which supply equals demand in all markets. When the imposition of a
regulation alters conditions in one market, a general equilibrium approach will determine a
new set of prices for all markets that will return the economy to equilibrium. These prices
in turn determine the outputs and consumption of goods and services in the new
equilibrium. In addition, a new set of prices and demands for the factors of production
(labor, capital, and land), the returns to which compose the income of businesses and
households, will be determined in general equilibrium. The social cost of the regulation can
then be estimated by comparing the value of variables in the pre-regulation "baseline"
equilibrium with those in the post-regulation, simulated equilibrium.
In 2015, the EPA established a Science Advisory Board (SAB) panel to consider the
technical merits and challenges of using economy-wide models to evaluate costs, benefits,
and economic impacts in regulatory development. In its final report (U.S. EPA, 2017b), the
SAB recommended that the EPA begin to integrate CGE modeling into regulatory analysis
6 The EPA included specific types of health benefits in a CGE model for the prospective analysis — The Benefits
and Costs of the Clean Air Act from 1990 to 2020 (EPA 2011) — and demonstrated the importance of their
inclusion when evaluating the economic welfare effects of policy. However, while the external Council on
Clean Air Compliance Analysis (Council) peer review of the EPA report (Hammitt, 2010) stated that
inclusion of benefits in an economy-wide model, specifically adapted for use in that study, "represent[ed] a
significant step forward in benefit-cost analysis", serious technical challenges remain when attempting to
evaluate the benefits and costs of potential regulatory actions using economy-wide models.
263
-------
to offer a more comprehensive assessment of the effects of air regulations. The SAB noted
that CGE models can provide insight into the likely social costs of a regulation even when
they do not typically include a characterization of the likely social benefits of the regulation.
CGE models may also offer insights into the ways costs are distributed across regions,
industry sectors, or households.
In response to the SAB's recommendations, the EPA built a new CGE model called
SAGE. A second SAB panel performed a peer review of SAGE, and the review concluded in
2020. While the theory and inherent model structure, including key underlying
assumptions and model parameterization, have been through rigorous peer review,
applying SAGE to specific regulatory contexts requires the EPA to determine when and how
the tool can best be leveraged to gain insights. U.S. EPA (2017b) noted that there is "no
hard and fast rule" for deciding when an economy-wide modeling approach will add value
beyond other tools typically utilized by the EPA to quantify costs, though they suggest
several relevant factors, including strong cross-price effects between markets, pre-existing
distortions present in those markets, and impacts that are not small relative to the
precision of the model.
There are additional considerations that are equally important when considering
whether a CGE model will add value, on net, beyond the set of tools already being leveraged
to estimate costs. For example, care must be given when preparing engineering costs to be
used as an input in an economy-wide model to avoid double counting taxes and transfers,
translate capital costs to a consistent measure within the economy-wide model, and
attribute engineering costs to specific inputs. Using SAGE or any other CGE model in a
rulemaking requires significant time and resources to adapt engineering cost estimates for
use in the model and to modify the model, as needed, to capture important sector-specific
nuances in modeling the behavioral response to a regulation. Therefore, in deciding
whether and how to utilize CGE models to analyze the social costs of regulations requires a
weighing of the value added of additional insights that can be gained from an economy-
wide analysis against the time and resource costs of developing a careful approach to
accurately capture key compliance pathways and sector-specific behavioral responses
264
-------
within the CGE model. The EPA will continue to evaluate the appropriateness of conducting
an economy-wide analysis using SAGE or another CGE model for rulemakings.
In May 2023, the EPA used SAGE for the first time to analyze the social costs of a
proposed regulation, the Greenhouse Gas Standards and Guidelines for Fossil Fuel-fired
Power Plants. The analysis appears in an appendix of the Regulatory Impact Analysis and
outlines the approach taken to ensure careful calibration of compliance cost estimates from
the EPA's electricity sector model, the Integrated Planning Model (IPM), for use in SAGE.
Connecting the outputs from a sectoral partial equilibrium model to a CGE model required
significant attention and resources. The EPA needed to develop an approach to linking the
SAGE model and the results from IPM that could adequately represent the regulatory
requirements and detailed compliance response information from the technologically rich
partial equilibrium model of the power sector in the CGE model. The EPA requested public
comment on the use of the SAGE model and presentation of results, which the Agency will
review in developing the analysis for the final Greenhouse Gas Standards and Guidelines
for Fossil Fuel-fired Power Plants rule.
The EPA does not currently have the capacity to estimate the social costs of this rule
using the SAGE model and anticipates that significant modeling and data development
would be needed to adequately characterize the economy-wide impacts of the rule. This
rulemaking differentially affects sectors and regions across the economy in ways that make
capturing the economy-wide impacts difficult. The SAB's 2017 report noted that, "The more
spatially, sectorally, and/or temporally detailed the regulation, the more challenging it is to
represent in a modeling framework. For example, the National Ambient Air Quality
Standards (NAAQS) are determined at the national level, with implementation occurring at
the state level in accordance with air basin-specific considerations. As a result, the
implementation of the standard can vary widely across air basins, making it difficult to
capture in an economy- wide model, which usually are too spatially and sectorally
aggregated to capture air basin specific regulations. It is also difficult to predict what each
state will do to comply with the NAAQS, particularly for those compliance actions states
must take that are not attributable to specific control measures and which may cost more
than EPA's upper-bound action. However, this difficulty is not unique to CGE analysis: other
265
-------
methodologies must confront it as well." (U.S. EPA, 2017b). We therefore did not conduct
an economy-wide analysis using SAGE for this rulemaking.
266
-------
4.4 References
Goulder, L., I. Parry, R. Williams, and D. Burtraw (1999). The cost-effectiveness of
alternative instruments for environmental protection in a second-best setting. Journal
of Public Economics, 72(3): 329-360.144
Goulder, L., M. Hafstead, and R. Williams III (2016). General equilibrium impacts of a
federal clean energy standard. American Economic Journal - Economic Policy 8(2):
186-218.
Hahn, R., and J. Hird (1990). The costs and benefits of regulation: review and synthesis. Yale
Journal of Regulation 8: 233-278.
Hammitt, J.K. (2010). Review of the final integrated report for the second section 812
prospective study of the benefits and costs of the clean air act. Available at:
https://council.epa.gov/ords/sab/f?p=104:12:11754097140249::: 12::.
Harrington, W., R.D. Morgenstern, and P. Nelson (2000). On the accuracy of regulatory cost
estimates. Journal of Policy Analysis and Management 19, 297-322.
U.S. EPA (2011). The Benefits and Costs of the Clean Air Act from 1990 to 2020. Final
Report. Office of Air and Radiation, Washington, DC. Available at:
https://www.epa.gov/sites/default/files/2015-07/documents/fullreport_rev_a.pdf.
U.S. EPA (2014). Retrospective Study of the Costs of EPA Regulations: A Report of Four Case
Studies, https://www.epa.gov/sites/default/files/2017-09/documents/ee-0575_0.pdf.
U.S. EPA (2016). EPA Guidelines for Preparing Economic Analyses. Available at:
https://www.epa.gov/sites/default/files/2017-08/documents/ee-0568-50.pdf.
U.S. EPA (2017a). EPA Air Pollution Control Cost Manual, Section 1, Chapter 2. Office of Air
Quality Planning and Standards, Research Triangle Park, NC. Available at
https://www.epa.gov/sites/default/files/2017-
12/documents/epaccmcostestimationmethodchapter_7thedition_2017.pdf.
U.S. EPA (2017b). SAB Advice on the Use of Economy-Wide Models in Evaluating the Social
Costs, Benefits, and Economic Impacts of Air Regulations. EPA-SAB-17-012.
U.S. EPA (2019a). CoST v3.7 User's Guide. Office of Air Quality Planning and Standards,
Research Triangle Park, NC. November 2019. Available at:
https://www.cmascenter.org/help/documentation. cfm?model=cost&version=3.7.
U.S. EPA (2019b). Additional Methods, Determinations, and Analyses to Modify Air Quality
Data Beyond Exceptional Events. U.S. Environmental Protection Agency, Office of Air
Quality Planning and Standards, Research Triangle Park, NC, EPA-457/B-19-002.
Available: https://www.epa.gov/sites/default/files/2019-
04/documents/clarification_memo_on_data_modification_methods.pdf.
267
-------
Williams III, R. (2002). Environmental tax interactions when pollution affects health or
productivity. Journal of Environmental Economics and Management 44(2): 261-270.
268
-------
APPENDIX 4A: ENGINEERING COST ANALYSIS
Overview
Chapter 4 describes the engineering cost analysis approach that EPA used to analyze
the following revised and alternative annual and 24-hour standard levels in this regulatory
impact analysis (RIA) --10/35 ng/m3,10/30 ng/m3, 9/35 ng/m3, and 8/35 |~ig/m3. This
Appendix contains more detailed information about the estimated costs from application of
controls by area and by county for the northeast and their adjacent counties, the southeast
and their adjacent counties, the west, and California.
4A.1 Estimated Costs by County for Revised and Alternative Standard Levels
The cost estimates presented in Table 4A-1 through Table 4A-6 reflect the
engineering costs annualized at 7 percent, to the extent possible.1 When calculating the
annualized costs we prefer to use the interest rates faced by firms; however, we do not
know what those rates are.
Table 4A-1 and Table 4A-2 present the cost estimates for the northeast counties and
their adjacent counties. Table 4A-3 and Table 4A-4 present the cost estimates for the
southeast counties and their adjacent counties. Table 4A-5 presents the cost estimates for
the counties in the west, and Table 4A-6 presents the cost estimates for the counties in
California, organized by air district.
Table 4A-1 Summary of Estimated Annual Control Costs for the Northeast (31
counties) for Revised and Alternative Primary Standard Levels of
10/35 ng/m3,10/30 ng/m3, 9/35 (ig/m3, and 8/35 ng/m3 for 2032
(millions of 2017$)
County
10/35
10/30
9/35
8/35
Cook County, IL
$0
$0
$5.4
$12.8
DuPage County, IL
$0
$0
$0
$9.5
Macon County, IL
$0
$0
$0
$4.9
Madison County, IL
$0
$0
$0.4
$22.1
McLean County, IL
$0
$0
$0
$0.08
1 Because we obtain control cost data from many sources, we are not always able to obtain consistent data
across original data sources. As a result, we do not know the interest rates used to calculate costs for some
of the controls included in this analysis. If disaggregated control cost data is available (i.e., where capital,
equipment life value, and O&M costs are separated out) we can calculate costs using a specified percent
interest rate. EPA may not know the interest rates used to calculate costs when disaggregated control cost
data is unavailable (i.e., where we only have a $/ton value and where capital, equipment life value, and O&M
costs are not separated out).
269
-------
County
10/35
10/30
9/35
8/35
St. Clair County, IL
$0
$0
$0
$7.7
Lake County, IN
$0
$0
$0.1
$1.7
Marion County, IN
$0
$0
$39.3
$39.3
Jefferson County, KY
$0
$0
$0
$3.4
Wayne County, MI
$1.2
$1.2
$21.4
$31.1
Bergen County, NJ
$0
$0
$17.3
$17.3
Camden County, NJ
$0
$0
$8.9
$10.4
Union County, NJ
$0
$0
$0
$13.5
New York County, NY
$0
$0
$0
$7.2
Butler County, OH
$0
$0
$33.4
$33.5
Cuyahoga County, OH
$0
$0
$6.5
$15.6
Franklin County, OH
$0
$0
$0
$0.2
Hamilton County, OH
$1.6
$1.6
$23.6
$23.6
Jefferson County, OH
$0
$0
$0
$0.04
Stark County, OH
$0
$0
$0
$0.7
Allegheny County, PA
$2.3
$2.6
$18.4
$60.3
Beaver County, PA
$0
$0
$0
$0.5
Cambria County, PA
$0
$0
$0
$0.5
Chester County, PA
$0
$0
$3.3
$17.5
Delaware County, PA
$0.2
$0.2
$19.3
$19.3
Lancaster County, PA
$0
$0
$0
$1.9
Lebanon County, PA
$0
$0
$0
$0.9
Philadelphia County, PA
$0
$0
$6.2
$12.6
York County, PA
$0
$0
$0
$1.2
Providence County, RI
$0
$0
$0
$0.3
Davidson County, TN
$0
$0
$0
$1.5
Total
$5.3
$5.5
$203.6
$371.1
Table 4A-2 Summary of Estimated Annual Control Costs for Adjacent Counties in
the Northeast (51 counties) for Revised and Alternative Primary
Standard Levels of 10/35 ng/m3,10/30 ng/m3, 9/35 ng/m3, and 8/35
(j,g/m3 for 2032 (millions of 2017$)
County
Adjacent Counties
10/35
10/30
9/35
8/35
Kane County, IL
Cook County, IL
$0
$0
$0
$11.4
DuPage County, IL
Lake County, IL
Cook County, IL
$0
$0
$0
$7.0
McHenry County, IL
Cook County, IL
$0
$0
$0
$15.5
Will County, IL
Cook County, IL
$0
$0
$0
$1.5
DuPage County, IL
Boone County, IN
Marion County, IN
$0
$0
$0.2
$5.2
Hamilton County, IN
Marion County, IN
$0
$0
$0.8
$15.4
Hancock County, IN
Marion County, IN
$0
$0
$0.2
$5.1
Hendricks County, IN
Marion County, IN
$0
$0
$0.5
$13.3
Johnson County, IN
Marion County, IN
$0
$0
$0.5
$8.0
Morgan County, IN
Marion County, IN
$0
$0
$0.4
$7.5
Shelby County, IN
Marion County, IN
$0
$0
$0.9
$14.3
Macomb County, MI
Wayne County, MI
$0
$0
$0
$18.1
Monroe County, MI
Wayne County, MI
$0
$0
$0
$11.0
270
-------
County
Adjacent Counties
10/35
10/30
9/35
8/35
Oakland County, MI
Wayne County, MI
$0
$0
$0
$36.3
Washtenaw County, MI
Wayne County, MI
$0
$0
$0
$7.3
Atlantic County, NJ
Camden County, NJ
$0
$0
$0
$9.5
Burlington County, NJ
Camden County, NJ
$0
$0
$0
$16.0
Essex County, NJ
Bergen County, NJ
$0
$0
$13.0
$13.0
Union County, NJ
Gloucester County, NJ
Camden County, NJ
$0
$0
$0
$12.1
Hudson County, NJ
Bergen County, NJ
$0
$0
$7.0
$7.0
Union County, NJ
Passaic County, NJ
Bergen County, NJ
$0
$0
$7.0
$7.0
Bronx County, NY
New York County, NY
$0
$0
$0
$0.5
Kings County, NY
New York County, NY
$0
$0
$0
$0.4
Queens County, NY
New York County, NY
$0
$0
$0
$0.9
Clermont County, OH
Hamilton County, OH
$0
$0
$13.1
$13.1
Geauga County, OH
Cuyahoga County, OH
$0
$0
$0
$10.8
Lake County, OH
Cuyahoga County, OH
$0
$0
$0
$4.7
Lorain County, OH
Cuyahoga County, OH
$0
$0
$0
$13.3
Medina County, OH
Cuyahoga County, OH
$0
$0
$0
$14.7
Portage County, OH
Cuyahoga County, OH
$0
$0
$0
$7.6
Stark County, OH
Summit County, OH
Cuyahoga County, OH
$0
$0
$0
$16.0
Stark County, OH
Warren County, OH
Butler County, OH
$0
$0
$13.9
$13.9
Hamilton County, OH
Armstrong County, PA
Allegheny County, PA
$0
$0
$0
$0.2
Butler County, PA
Allegheny County, PA
$0
$0
$0
$1.7
Beaver County, PA
Montgomery County, PA
Chester County, PA
$0
$0
$4.4
$21.5
Delaware County, PA
Philadelphia County, PA
Washington County, PA
Allegheny County, PA
$0
$0
$0
$1.8
Beaver County, PA
Westmoreland County, PA
Allegheny County, PA
$0
$0
$0
$1.6
Cambria County, PA
Total
$0
$0
$62.1
$364.2
Table 4A-3 Summary of Estimated Annual Control Costs for the Southeast (33
counties) for Revised and Alternative Primary Standard Levels of
10/35 ng/m3,10/30 ng/m3, 9/35 (ig/m3, and 8/35 ng/m3 for 2032
(millions of 2017$)
County
10/35
10/30
9/35
8/35
Jefferson County, AL
$0
$0
$0
$0.8
Russell County, AL
$0
$0
$0
$3.1
Pulaski County, AR
$0
$0
$0
$5.5
District of Columbia
$0
$0
$0
$9.6
Broward County, FL
$0
$0
$0.1
$12.1
Bibb County, GA
$0
$0
$0
$1.4
271
-------
County
10/35
10/30
9/35
8/35
Chatham County, GA
$0
$0
$0
$0.3
Dougherty County, GA
$0
$0
$0
$1.2
Fulton County, GA
$0
$0
$0
$42.9
Gwinnett County, GA
$0
$0
$0
$32.0
Muscogee County, GA
$0
$0
$0
$13.5
Richmond County, GA
$0
$0
$0.1
$23.7
Shawnee County, KS
$0
$0
$0
$6.1
Wyandotte County, KS
$0
$0
$0
$4.3
Caddo Parish, LA
$0
$0
$4.2
$21.7
East Baton Rouge Parish, LA
$0
$0
$0
$1.8
West Baton Rouge Parish, LA
$0
$0
$0
$10.4
Hinds County, MS
$0
$0
$0
$3.0
Forsyth County, NC
$0
$0
$0
$0.1
Mecklenburg County, NC
$0
$0
$0
$2.5
Cleveland County, OK
$0
$0
$0
$9.0
Oklahoma County, OK
$0
$0
$0
$3.2
Tulsa County, OK
$0
$0
$0
$3.2
Bowie County, TX
$0
$0
$0
$2.7
Cameron County, TX
$0
$0
$15.8
$15.8
Dallas County, TX
$0
$0
$0
$0.9
El Paso County, TX
$0
$0
$0
$7.7
Harris County, TX
$0
$0
$2.9
$9.2
Hidalgo County, TX
$35.8
$35.8
$35.8
$35.8
Jefferson County, TX
$0
$0
$0
$1.6
Nueces County, TX
$0
$0
$0
$3.3
Orange County, TX
$0
$0
$0
$4.6
Travis County, TX
$0
$0
$1.6
$6.9
Total
$35.8
$35.8
$60.4
$299.7
Table 4A-4 Summary of Estimated Annual Control Costs for Adjacent Counties in
the Southeast (34 counties) for Revised and Alternative Primary
Standard Levels of 10/35 |ig/m3,10/30 |ig/m3, 9/35 |ig/m3, and 8/35
(ig/m3 for 2032 (millions of 2017$)
County
Adjacent Counties
10/35
10/30
9/35
8/35
Burke County, GA
Richmond County, GA
$0
$0
$0
$4.0
Columbia County, GA
Richmond County, GA
$0
$0
$0
$1.3
Forsyth County, GA
Fulton County, GA
$0
$0
$0
$0.0
Gwinnett County, GA
Jefferson County, GA
Richmond County, GA
$0
$0
$0
$4.1
McDuffie County, GA
Richmond County, GA
$0
$0
$0
$1.1
Bossier Parish, LA
Caddo Parish, LA
$0
$0
$0
$10.8
De Soto Parish, LA
Caddo Parish, LA
$0
$0
$0
$10.5
East Feliciana Parish, LA
East Baton Rouge Parish, LA
$0
$0
$0
$0.1
West Baton Rouge Parish, LA
Iberville Parish, LA
East Baton Rouge Parish, LA
$0
$0
$0
$0.5
West Baton Rouge Parish, LA
Pointe Coupee Parish, LA
West Baton Rouge Parish, LA
$0
$0
$0
$0.1
Red River Parish, LA
Caddo Parish, LA
$0
$0
$0
$6.2
West Feliciana Parish, LA
West Baton Rouge Parish, LA
$0
$0
$0
$0.1
272
-------
County Adjacent Counties 10/35 10/30 9/35 8/35
Bastrop County, TX
Travis County, TX
$0
$0
$0
$0.2
Blanco County, TX
Travis County, TX
$0
$0
$0
$0.01
Brooks County, TX
Hidalgo County, TX
$0.01
$0.01
$9.1
$9.1
Burnet County, TX
Travis County, TX
$0
$0
$0
$0.01
Caldwell County, TX
Travis County, TX
$0
$0
$0
$0.06
Hays County, TX
Travis County, TX
$0
$0
$0
$0.02
Hudspeth County, TX
El Paso County, TX
$0
$0
$0
$4.6
Kenedy County, TX
Hidalgo County, TX
$0.01
$0.01
$5.9
$5.9
Starr County, TX
Hidalgo County, TX
$0
$0
$7.4
$7.4
Willacy County, TX
Cameron County, TX
Hidalgo County, TX
$0.01
$0.01
$3.2
$3.2
Williamson County, TX
Travis County, TX
$0
$0
$0
$0.02
Total $0.02 $0.02 $25.5 $69.2
Table 4A-5 Summary of Estimated Annual Control Costs for the West (29
counties) for Revised and Alternative Primary Standard Levels of
10/35 |ig/m3,10/30 |ig/m3, 9/35 (j,g/m3, and 8/35 |ig/m3 for 2032
(millions of 2017$)
County
10/35
10/30
9/35
8/35
Maricopa County, AZ
$0
$0
$1.2
$7.2
Pinal County, AZ
$0
$1.5
$0
$0
Santa Cruz County, AZ
$0
$0
$0
$0.8
Yuma County, AZ
$0
$0
$0
$0.4
Adams County, CO
$0
$0
$0.04
$16.0
Denver County, CO
$0
$0
$0.01
$1.8
Weld County, CO
$0
$0
$0
$0.2
Benewah County, ID
$0
$17.0
$0
$0
Canyon County, ID
$0
$4.2
$0
$6.8
Lemhi County, ID
$0
$1.2
$1.2
$1.2
Shoshone County, ID
$12.8
$12.8
$12.8
$12.8
Flathead County, MT
$0
$0.6
$0
$0
Lewis and Clark County, MT
$0
$3.7
$0
$3.7
Lincoln County, MT
$26.9
$26.9
$26.9
$26.9
Missoula County, MT
$0
$0
$0.02
$9.6
Clark County, NV
$0
$0
$0
$2.3
Crook County, OR
$0
$7.1
$0
$1.2
Harney County, OR
$0
$1.9
$4.7
$19.0
Jackson County, OR
$0
$0.01
$1.1
$12.8
Josephine County, OR
$0
$0
$0
$5.1
Klamath County, OR
$0
$9.8
$9.8
$9.8
Lake County, OR
$0
$0
$0
$0
Lane County, OR
$0
$0.7
$0
$0.7
Box Elder County, UT
$0
$3.3
$0
$0
Cache County, UT
$0
$3.2
$0
$0
Salt Lake County, UT
$0
$0.06
$0
$0
King County, WA
$0
$0
$0
$0.2
Okanogan County, WA
$0
$18.4
$0
$2.0
273
-------
Yakima County, WA $0 $0 $0 $0
Total $39.7 $112.4 $57.7 $140.6
Table 4A-6 Summary of Estimated Annual Control Costs for California (36
counties) for Revised and Alternative Primary Standard Levels of
10/35 ng/m3,10/30 ng/m3, 9/35 (ig/m3, and 8/35 ng/m3 for 2032
(millions of 2017$)
County
Air District
10/35
10/30
9/35
8/35
Alameda County, CA
Bay Area AQMD
$0.1
$0.1
$16.4
$16.4
Contra Costa County, CA
Bay Area AQMD
$0
$0
$0.2
$1.4
Napa County, CA
Bay Area AQMD
$0
$0
$0
$3.7
San Francisco County, CA
Bay Area AQMD
$0
$0
$0
$3.3
San Mateo County, CA
Bay Area AQMD
$0
$0
$0
$0.5
Santa Clara County, CA
Bay Area AQMD
$0
$0
$0.3
$1.7
Solano County, CA
Bay Area AQMD
$0
$0
$2.4
$8.1
Butte County, CA
Butte County AQMD
$0
$3.6
$0
$1.7
Calaveras County, CA
Calaveras County APCD
$0
$0
$4.0
$4.0
Colusa County, CA
Colusa County APCD
$0
$0
$0
$0
Sutter County, CA
Feather River AQMD
$0
$4.2
$4.2
$4.2
Mono County, CA
Great Basin Unified APCD
$0
$0
$0
$0
Imperial County, CA
Imperial County APCD
$1.8
$1.8
$1.8
$1.8
Mendocino County, CA
Mendocino County AQMD
$0
$5.1
$0
$0.01
Plumas County, CA
Northern Sierra AQMD
$0
$0
$0
$0
Placer County, CA
Placer County APCD
$0
$0
$0
$0.02
Sacramento County, CA
Sacramento Metro AQMD
$0
$1.4
$1.0
$4.2
San Diego County, CA
San Diego County APCD
$0
$0
$1.7
$40.2
Fresno County, CA
San Joaquin Valley APCD
$44.0
$44.0
$44.0
$44.0
Kern County, CA
San Joaquin Valley APCD
$10.1
$10.1
$10.1
$10.1
Kings County, CA
San Joaquin Valley APCD
$5.0
$5.0
$5.0
$5.0
Madera County, CA
San Joaquin Valley APCD
$0
$0
$16.8
$16.8
Merced County, CA
San Joaquin Valley APCD
$12.4
$12.4
$12.4
$12.4
San Joaquin County, CA
San Joaquin Valley APCD
$0
$0.02
$15.6
$15.6
Stanislaus County, CA
San Joaquin Valley APCD
$2.6
$2.6
$2.6
$2.6
Tulare County, CA
San Joaquin Valley APCD
$21.7
$21.7
$21.7
$21.7
San Luis Obispo County, CA
San Luis Obispo County APCD
$0
$0
$0
$0.01
Santa Barbara County, CA
Santa Barbara County APCD
$0
$6.6
$0
$0.6
Shasta County, CA
Shasta County AQMD
$0
$0.03
$0
$0
Siskiyou County, CA
Siskiyou County APCD
$0
$23.6
$0
$0
Los Angeles County, CA
South Coast AQMD
$12.5
$12.5
$12.5
$12.5
Orange County, CA
South Coast AQMD
$3.2
$3.2
$3.2
$3.2
Riverside County, CA
South Coast AQMD
$0
$0
$0
$0
San Bernardino County, CA
South Coast AQMD
$8.4
$8.4
$8.4
$8.4
Tehama County, CA
Tehama County APCD
$0
$7.3
$0
$0
Ventura County, CA
Ventura County APCD
$0
$12.5
$0.3
$12.5
Total
$121.8
$186.1
$184.4
$256.7
274
-------
CHAPTER 5: BENEFITS ANALYSIS APPROACH AND RESULTS
Overview
This chapter presents the estimated human health-related and welfare benefits of
the illustrative control strategies discussed in Chapter 3 for the revised annual and current
24-hour alternative standard levels of 9/35 |j,g/m3, as well as the following less and more
stringent alternative standard levels 10/35 ng/m3,10/30 ng/m3, and 8/35 |~ig/m3. Because
the EPA is retaining the current secondary PM standards, we did not evaluate alternative
secondary standard levels in this chapter. We quantify the number and economic value of
the estimated avoided premature deaths and illnesses attributable to applying hypothetical
national control strategies for the revised, less stringent, and more stringent annual PM2.5
NAAQS standards with a sensitivity analysis for a more stringent 24-hour standard that
reduces fine particulate matter (PM2.5) concentrations in 2032. Reducing directly emitted
PM2.5 and PM2.5 precursor emissions would also improve environmental quality (U.S. EPA,
2019c, 2022a) and reduce the ecological effects of nitrogen and sulfur deposition. Because
the EPA is retaining the current secondary PM NAAQS standards, we did not evaluate
alternative secondary standard levels in this RIA, or any visibility-, climate change-, or
materials-damage-related benefits of the final rule (Cox, 2019; U.S. EPA, 2019c).
The analysis in this chapter aims to characterize the benefits of the air quality
changes resulting from the implementation of revised PM standard levels by answering
two key questions:
1. What is the estimated number and geographic distribution of avoided PM2.5-
attributable premature deaths and illnesses expected to result from applying
hypothetical national control strategies for the revised and alternative PM2.5
NAAQS standards? This chapter presents these results. As discussed in Chapter
3, Section 3.2.4, the estimated PM2.5 emissions reductions from control
applications do not fully account for all the emissions reductions needed to reach
the revised, less, and more stringent alternative standard levels in some counties
in the northeast, southeast, west, and California. In Chapter 2, Section 2.4 and
Chapter 3, Section 3.2.5, we discuss the remaining air quality challenges for
275
-------
areas in the northeast and southeast, as well as in the west and California for the
revised standard levels of 9/35 |~ig/m3.
2. What is the estimated number and geographic distribution of avoided PM2.5-
attributable premature deaths and illnesses expected to result if we assume that
areas identify all of the controls needed for compliance with the revised and
alternative PM2.5 NAAQS standard levels? Appendix 5A presents these results.
3. What is the estimated economic value of these avoided impacts?
To answer these questions we perform a human health benefits analysis (NRC,
2002). Starting first with the Integrated Science Assessment (ISA) for Particulate Matter
(U.S. EPA, 2019b) and the Supplement to the ISA for Particulate Matter (U.S. EPA, 2022a),
we identify the human health effects associated with ambient particles, which include
premature death and a variety of morbidity effects associated with acute (hours-long) and
chronic (months- or years- long) exposures. Table 5-2 summarizes human health
categories monetized and reflected in the total value of the benefits reported and those
categories not monetized due to limited data or resources. The list of benefits categories is
neither exhaustive nor completely quantified. We excluded effects not identified as having
a causal or likely to be causal relationship with the affected pollutants in the most recent
PM ISA (U.S. EPA, 2019b),(U.S. EPA, 2022a). In the Estimating PM2.5- and Ozone-
Attributable Health Benefits Technical Support Document (TSD) (U.S. EPA, 2023b), we
specify in detail our approach for identifying, selecting, and parametrizing concentration-
response relationships and economic unit values to support this benefits analysis.
This chapter contains a subset of the estimated health benefits of the revised and
alternative PM2.5 standard levels in 2032 that EPA was able to quantify, given available
resources and methods. This benefits analysis relies on an array of data inputs—including
air quality modeling, health impact functions and valuation estimates—which are
themselves subject to uncertainty and may also in turn contribute to the overall
uncertainty in this analysis. We employ several techniques to characterize this uncertainty,
which are described in detail in Section 5.3.
276
-------
As described in Chapter 1, the analytical objectives of the NAAQS RIA are unique as
compared to other RIAs. The NAAQS RIAs illustrate the potential costs and benefits of
attaining one or more revised air quality standard(s) nationwide; these estimated costs and
benefits are estimated after we first assume the current standards have been attained. In
this RIA, we illustrate the potential costs and benefits for the revised, less, and more
stringent alternative standard levels nationwide. The NAAQS RIAs hypothesize the control
strategies that States may choose to enact when implementing a revised NAAQS, but they
cannot do so with perfect foresight; individual states will formulate air quality
management plans whose mix of emissions controls may differ substantially from those we
simulate here. Hence, NAAQS RIAs are illustrative. The benefits and costs estimated in a
NAAQS RIA are not intended to be added to the costs and benefits of other regulations that
result in specific costs of control and emissions reductions. By contrast, EPA is generally
confident in the emissions projected to be reduced from rules affecting specific and well-
characterized sources—such as mobile and Electric Generating Units (U.S. EPA, 2019a).
Hence, the emissions reduced by final rules affecting such sources are accounted for when
simulating attainment with the revised and alternative NAAQS.
In the following sections of this chapter, we estimate health benefits occurring as an
increment to a 2032 baseline in which the nation fully attains the current primary PM2.5
standards (i.e., an annual standard of 12 |ig/m3 and a 24-hour standard of 35 |ig/m3,
hereafter referred to as "12/35"). This baseline accounts for: (1) promulgated regulations
(Chapter 1, Section 1.3.); and (2) any additional illustrative emissions reductions needed to
simulate attainment with 12/35 (Chapter 3, Section 3.1). We project PM2.5 levels in 2032 in
certain areas that would exceed 10/35,10/30, 9/35 and 8/35, even after illustrative
controls applied to simulate attainment with 12/35 and estimate emissions reductions
needed to attain the revised and alternative standard levels (Chapter 3, Table 3-2). Table
5-1 summarizes the total national monetized benefits resulting from applying the control
strategies in 2032. Because the analyses in the RIA are national-level assessments and the
ambient air quality issues are complex and local in nature, we do not currently have
sufficiently detailed local information for the areas being analyzed, including local
inventory information on emissions sources, higher resolution air quality modeling, and
277
-------
local information on emissions controls to estimate the control strategies that might result
in meeting the range of revised annual and 24-hour alternative standard levels.
Whereas the main analysis in this chapter presents the benefits of the applied
control strategies for the revised and alternative standards levels (Table 5-5 through Table
5-9), in Appendix 5A, we present the potential health and monetized benefits of full
compliance with the revised and alternative standard levels; the tables in Appendix 5A
present potential health benefits regardless of whether the control technologies or
strategies to achieve them are currently available or whether an agency submits
information on cross-border transport or wildfire influence on projected PM2.5 DVs that
could potentially qualify for exclusion as atypical, extreme, or unrepresentative events,
potentially affecting the amount of any additional control needed. The estimates reflect the
value of the avoided PIVh.s-attributable deaths and the value of morbidity impacts,
including, for example, hospital admissions and emergency department visits for
cardiovascular and respiratory health issues.
Table 5-1 Estimated Monetized Benefits of the Applied Control Strategies for the
Revised and Alternative Primary PM2.5 Standard Levels in 2032,
Incremental to Attainment of 12/35 (billions of 2017$)
ft r l*g/m3 annual & 10 jig/m3 annual & 9 jig/m3 annual & 8 ng/m3 annual &
Benefits Estimate 3 5 ng/m3 24-hour 30 iig/m3 24-hour 35 ng/m3 24-hour 35 ng/m3 24-hour
Economic value of avoided PIVh.s-related morbidities and premature deaths using PM2.5 mortality
estimate from Pope III et al., 2019
3% discount $17+ B $21 + B $46 + B $99 + B
rate
7% discount $16+ B $19 + B $42 + B $89 + B
rate
Economic value of avoided PIVh.s-related morbidities and premature deaths using PM2.5 mortality
estimate from Wu et al., 2020
3% discount 5 + g $10 + B $22 + B $48 + B
rate
7% discount $7.6+ B $9.2 + B $20+ B $43 + B
rate
Note: Rounded to two significant figures. Avoided premature deaths account for over 98% of monetized
benefits here, which are discounted over the SAB-recommended 20-year segmented lag. It was not possible to
quantify all benefits due to data limitations in this analysis. "B" is the sum of all unquantified health and
welfare benefits.
Because the method used in this analysis to simulate the control strategies does not
also simulate changes in ambient concentrations of other pollutants, we were not able to
quantify the additional benefits associated with reduced exposure to other pollutants. We
278
-------
also did not estimate the additional benefits from improvements in welfare effects, such as
climate effects, ecosystem effects, and visibility (Cox, 2019; U.S. EPA, 2019c). With regard to
potential climate benefits, we note that because the available evidence suggests direct PM
controls will be most effective in reducing ambient PM2.5 concentrations, and because we
lack information on the C02-related emissions changes that may result from such controls,
we do not quantitatively estimate CCh-related climate benefits in this RIA.
5.1 Human Health Benefits Analysis Methods
We estimate the quantity and economic value of air pollution-related effects using a
"damage-function." This approach quantifies counts of air pollution-attributable cases of
adverse health outcomes and assigns dollar values to those counts, while assuming that
each outcome is independent of one another. We construct this damage function by
adapting primary research—specifically, air pollution epidemiology studies and economic
value studies—from similar contexts. This approach is sometimes referred to as "benefits
transfer." Below we describe the procedure we follow for: (1) selecting air pollution health
endpoints to quantify; (2) calculating counts of air pollution effects using a health impact
function; (3) calculating the economic value of the health impacts.
5.1.1 Selecting Air Pollution Health Endpoints to Quantify
As a first step in quantifying PIVh.s-related human health impacts, the Agency
consults the most recent PM ISA and the Supplement to the ISA for Particulate Matter (U.S.
EPA, 2019b, 2022a). This document synthesizes the toxicological, clinical and
epidemiological evidence to determine whether PM is causally related to an array of
adverse human health outcomes associated with either acute (i.e., hours or days-long) or
chronic (i.e., years-long) exposure; for each outcome, the ISA reports this relationship to be
causal, likely to be causal, suggestive of a causal relationship, inadequate to infer a causal
relationship or not likely to be a causal relationship. Historically, the Agency estimates the
incidence of air pollution effects for those health endpoints that the ISA classified as either
causal or likely-to-be-causal.
Consistent with economic theory, the willingness-to-pay (WTP) for reductions in
exposure to environmental hazard will depend on the expected impact of those reductions
on human health and other outcomes. All else equal, WTP is expected to be higher when
279
-------
there is stronger evidence of a causal relationship between exposure to the contaminant
and changes in a health outcome (McGartland et al., 2017). For example, in the case where
there is no evidence of a potential relationship the WTP would be expected to be zero and
the effect should be excluded from the analysis. Alternatively, when there is some evidence
of a relationship between exposure and the health outcome, but that evidence is
insufficient to definitively conclude that there is a causal relationship, individuals may have
a positive WTP for a reduction in exposure to that hazard (U.S. EPA-SAB, 2020; Kivi and
Shogren, 2010). Lastly, the WTP for reductions in exposure to pollutants with strong
evidence of a relationship between exposure and effect are likely positive and larger than
for endpoints where evidence is weak, all else equal. Unfortunately, the economic literature
currently lacks a settled approach for accounting for how WTP may vary with uncertainty
about causal relationships.
Given this challenge, the Agency draws its assessment of the strength of evidence on
the relationship between exposure to PM2.5 and potential health endpoints from the ISAs
that are developed for the NAAQS process as discussed above. The focus on categories
identified as having a "causal" or "likely to be causal" relationship with the pollutant of
interest is to estimate the pollutant-attributable human health benefits in which we are
most confident. All else equal, this approach may underestimate the benefits of PM2.5
exposure reductions as individuals may be willing to pay to avoid specific risks where the
evidence is insufficient to conclude they are "likely to be caus[ed]" by exposure to these
pollutants.6 At the same time, WTP may be lower for those health outcomes for which
causality has not been definitively established. This approach treats relationships with ISA
causality determinations of "likely to be causal" as if they were known to be causal, and
therefore benefits could be overestimated. Table 5-2 reports the effects we quantified and
those we did not quantify in this RIA. The list of benefit categories not quantified is not
exhaustive. The table below omits welfare effects such as acidification and nutrient
enrichment.
280
-------
Table 5-2
Pollutant
Human Health Effects of Pollutants Potentially Affected by Attainment
of the Primary PM2.5 NAAQS
Effect (age)
Effect Effect More
Quantified Monetized Information
Adult premature mortality based on cohort study
estimates [>17 or >64]
~
~
PM ISA
Infant mortality (<1]
~
~
PM ISA
Non-fatal heart attacks (>18]
~
V
PM ISA
Hospital admissions - cardiovascular (all]
~
V
PM ISA
Hospital admissions - respiratory (<19 and >64]
~
V
PM ISA
Hospital admissions - Alzheimer's disease (>64]
~
V
PM ISA
Hospital admissions - Parkinson's disease (>64]
~
V
PM ISA
Emergency department visits - cardiovascular (all]
~
V
PM ISA
Emergency department visits - respiratory (all]
~
~
PM ISA
Emergency hospital admissions (>65]
~
~
PM ISA
Non-fatal lung cancer (>29]
~
~
PM ISA
Out-of-hospital cardiac arrest (all]
~
—
PM ISA
Stroke incidence (50-79]
~
V
PM ISA
PM2.5
New onset asthma (<12]
~
~
PM ISA
Exacerbated asthma - albuterol inhaler use (asthmatics,
6-13]
~
~
I'M ISA
Lost work days (18-64]
~
~
PM ISA
Minor restricted-activity days (18-64]
~
—
PM ISA
Other cardiovascular effects (e.g., doctor's visits,
prescription medication]
—
—
PM ISA1
Other respiratory effects (e.g., pulmonary function, other
ages]
—
—
PM ISA1
Other cancer effects (e.g., mutagenicity, genotoxicity]
—
—
PM ISA1
Other nervous system effects (e.g., dementia]
—
—
PM ISA1
Metabolic effects (e.g., diabetes, metabolic syndrome]
—
—
PM ISA1
Reproductive and developmental effects (e.g., low birth
weight, pre-term births]
—
PM ISA1
1 We assess these benefits qualitatively due to epidemiological or economic data limitations.
5.1.2 Calculating Counts of Air Pollution Effects Using the Health Impact Function
We use the environmental Benefits Mapping and Analysis Program—Community
Edition (BenMAP-CE) software program to quantify counts of premature deaths and
illnesses attributable to photochemical modeled changes in annual mean PM2.5 for the year
2032 using a health impact function (Sacks et al., 2018).1 A health impact function
combines information regarding: the concentration-response relationship between air
1 The 2032 air quality modeling surface input files, BenMAP configuration files and script to produce the
health benefits analyses in Chapters 5 and Appendix 5A are in the docket and available upon request with
the Docket Office.
281
-------
quality changes and the risk of a given adverse outcome; the population exposed to the air
quality change; the baseline rate of death or disease in that population; and, the air
pollution concentration to which the population is exposed.
The following provides an example of a PM2.5 mortality risk health impact function.
We estimate counts of PIVh.s-related total deaths (yij) during each year i (i=2032) among
adults aged 18 and older (a) in each county in the contiguous U.S. j (j=l,...,J where J is the
total number of counties) as
yij= Eayija
yija = moija x(eP ACij-l) x Pija, Eq[l]
where moija is the baseline total mortality rate for adults aged a= 18-99 in county j in
year i stratified in 10-year age groups, (3 is the risk coefficient for total mortality for adults
associated with annual average PM2.5 exposure, Cij is the annual mean PM2.5 concentration
in county j in year i, and Pija is the number of county adult residents aged a= 18-99 in county
j in year i stratified into 5-year age groups.2
To assess economic value in a damage-function framework, the changes in
environmental quality must be translated into effects on people or on the things that
people value. In some cases, the changes in environmental quality can be directly valued. In
other cases, such as for changes in ozone and PM, a health and welfare impact analysis
must first be conducted to convert air quality changes into effects that can be assigned
dollar values. For the purposes of this RIA, the health impacts analysis is limited to those
health effects that are directly and specifically linked to PM2.5.
We note at the outset that EPA rarely has the time or resources to perform extensive
new research to measure directly either the health outcomes or their values for regulatory
2 In this illustrative example, the air quality is resolved at the county level. For this RIA, we simulate air
quality concentrations at 12 x 12km grids. The BenMAP-CE tool assigns the rates of baseline death and
disease stored at the county level to the 12 x 12km grid cells using an area-weighted algorithm. This
approach is described in greater detail in the appendices to the BenMAP-CE user manual appendices (U.S.
EPA, 2023a).
282
-------
analyses. Thus, similar to (Kiinzli et al., 2000) and other, more recent health impact
analyses, our estimates are based on the best available methods of benefits transfer.
5.1.3 Calculating the Economic Valuation of Health Impacts
After quantifying the change in adverse health impacts, the final step is to estimate
the economic value of these avoided impacts. The appropriate economic value for a change
in a health effect depends on whether the health effect is viewed ex ante (before the effect
has occurred) or ex post (after the effect has occurred). Reductions in ambient
concentrations of air pollution generally lower the risk of future adverse health effects by a
small amount for a large population. The appropriate economic measure is therefore ex
ante WTP for changes in risk. However, epidemiological studies generally provide
estimates of the relative risks of a particular health effect avoided due to a reduction in air
pollution. A convenient way to use these data in a consistent framework is to convert
probabilities to units of avoided statistical incidences. This measure is calculated by
dividing individual WTP for a risk reduction by the related observed change in risk. For
example, suppose a regulation reduces the risk of premature mortality from 2 in 10,000 to
1 in 10,000 (a reduction of 1 in 10,000). If individual WTP for this risk reduction is $100,
then the WTP for an avoided statistical premature mortality amounts to $1 million
($100/0.0001 change in risk). The same type of calculation can produce values for
statistical incidences of other health endpoints.
For some health effects, such as hospital admissions, WTP estimates are generally
not available. In these cases, we instead use the cost of treating or mitigating the effect to
economically value the health impact. For example, for the valuation of hospital admissions
we use the avoided medical costs as an estimate of the value of avoiding the health effects
causing the admission. These cost-of-illness (COI) estimates generally (although not in
every case) understate the true value of reductions in risk of a health effect. They tend to
reflect the direct expenditures related to treatment but not the value of avoided pain and
suffering from the health effect.
5.2 Benefits Analysis Data Inputs
In Figure 5-1, we summarize the key data inputs to the health impact and economic
valuation estimates, which were calculated using BenMAP-CE model version 1.5.1 (Sacks et
283
-------
al., 2018). In the sections below we summarize the data sources for each of these inputs,
including demographic projections, incidence and prevalence rates, effect coefficients, and
economic valuation.
Census
Population
Data
Modeled
Baseline and
Post-Control
Ambient PM2.5
2032
Population
Projections
PM2.5 Incremental
Air Quality Change
Woods & Poole
Population
Projections
PM -.r. Health
Functions
PM2.5-Related
Health Impacts
Background
Incidence and
Prevalence Rates
Economic
Valuation
Functions
Monetized PM2.5-
related Benefits
Blue identifies a user-selected input within the BenMAP model
Green identifies a data input generated outside of the BenMAP
Figure 5-1 Data Inputs and Outputs for the BenMAP-CE Model
5.2.1 Demographic Data
Quantified and monetized human health impacts depend on the demographic
characteristics of the population, including age, location, and income. We use projections
based on economic forecasting models developed by Woods & Poole, Inc. (Woods & Poole,
2015). The Woods & Poole database contains county-level projections of population by age,
sex, and race out to 2060, relative to a baseline using the 2010 Census data. Projections in
each county are determined simultaneously with every other county in the U.S. to consider
patterns of economic growth and migration. The sum of growth in county-level populations
is constrained to equal a previously determined national population growth, based on
Bureau of Census estimates (Hollmann et al., 2000). According to Woods & Poole, linking
county-level growth projections together and constraining the projected population to a
284
-------
national-level total growth avoids potential errors introduced by forecasting each county
independently (for example, the projected sum of county-level populations cannot exceed
the national total). County projections are developed in a four-stage process:
• First, national-level variables such as income, employment, and populations are
forecasted.
• Second, employment projections are made for 179 economic areas defined by the
Bureau of Economic Analysis (U.S. BEA, 2004)3, using an "export-base" approach,
which relies on linking industrial-sector production of non-locally consumed
production items, such as outputs from mining, agriculture, and manufacturing with
the national economy. The export-based approach requires estimation of demand
equations or calculation of historical growth rates for output and employment by
sector.
• Third, population is projected for each economic area based on net migration rates
derived from employment opportunities and following a cohort-component method
based on fertility and mortality in each area.
• Fourth, employment and population projections are repeated for counties, using the
economic region totals as bounds. The age, sex, and race distributions for each
region or county are determined by aging the population by single year by sex and
race for each year through 2060 based on historical rates of mortality, fertility, and
migration.
5.2.2 Baseline Incidence and Prevalence Estimates
Epidemiological studies of the association between pollution levels and adverse
health effects generally provide a direct estimate of the relationship of air quality changes
to the relative risk of a health effect, rather than estimating the absolute number of avoided
cases. For example, a typical result might be that a 5 |ig/m3 decrease in daily PM2.5 levels is
associated with a decrease in hospital admissions of 3%. A baseline incidence rate,
necessary to convert this relative change into a number of cases, is the estimate of the
3 According to the Bureau of Economic Analysis (BEA) website, due to the impact of sequestration and
reduced FY 2013 funding levels, statistics will not be updated or made available after November 21, 2013.
285
-------
number of cases of the health effect per year in the assessment location, as it corresponds
to baseline pollutant levels in that location. To derive the total baseline incidence per year,
this rate must be multiplied by the corresponding population number. For example, if the
baseline incidence rate is the number of cases per year per million people, that number
must be multiplied by the millions of people in the total population.
Table 12 (reproduced below as Table 5-3) from the Estimating PM2.5- and Ozone-
Attributable Health Benefits TSD (referred to as the TSD in rest of this chapter)
summarizes the sources of baseline incidence rates and reports average incidence rates for
the endpoints included in the analysis. For both baseline incidence and prevalence data, we
used age-specific rates where available. We applied concentration-response functions to
individual age groups and then summed over the relevant age range to provide an estimate
of total population benefits. National-level incidence rates were used for most morbidity
endpoints4, whereas county-level data are available for premature mortality. Whenever
possible, the national rates used are national averages, because these data are most
applicable to a national assessment of benefits. For some studies, however, the only
available incidence information comes from the studies themselves; in these cases,
incidence in the study population is assumed to represent typical incidence at the national
level.
Table 5-3 Baseline Incidence Rates for Use in Impact Functions
Endpoint
Parameter
Rates
Value
Source
Mortality1
Daily or annual projected
incidence to 2060 in 5-year
increments fO—99")
Age-, cause-, race-, and
county-stratified rates
CDC WONDER (2012-2014)
(U.S. Census Bureau, 2012)
Hospitalizations2
Daily incidence rates for all
ages
Age-,
region/state/county-, and
cause- stratified rates
2011-2014 HCUP data files
and data requested from and
supplied by individual states
Emergency
Department Visits2
Daily emergency
department visit incidence
rates for all ages
Age-, region-, state-,
county-, and cause-
stratified rates
2011-2014 HCUP data files
and data requested from and
supplied by individual states
4 Data availability from HCUP has changed since the 2012 PM NAAQS RIA, with state-level incidence data
replacing regional-level data. As some states have low populations, many data points are unavailable, either
because they are missing or have been censored to protect health record privacy. To avoid interpolating the
missing values, we used national-level incidence data, which corresponds appropriately with the national-
level epidemiology effect coefficients used in these analyses.
286
-------
Rates
Endpoint
Parameter
Value
Source
Nonfatal Acute
Myocardial
Infarction
Daily nonfatal AMI
incidence rate per person
aged 18-99
Age-, region-, state-, and
county- stratified rates
(AHRQ, 2016)
Asthma Symptoms
Daily incidence among
asthmatic children
Age- and race- stratified
rates
(B. Ostro et al., 2001)
Wheeze (ages 5-12)
Cough (ages 5-12)
Shortness of breath (ages 5-
12)
Albuterol use (ages 6-13)
2.2 puffs per day
(Rabinovitch et al., 2006)
Asthma Onset
Annual incidence
0-4
5-11
12-17
0.0234
0.0111
0.0044
(Winer et al., 2012)
Alzheimer's Disease
Daily incidence rates for all
ages
Age-, region-, state-, and
county- stratified rates
2011-2014 HCUP data files
Parkinson's Disease
Annual incidence
18-44
45-64
65-84
85-99
0.0000011
0.0000366
0.0002001
0.0002483
HCUPnet
Allergic Rhinitis
Respondents aged 3-17
experiencing allergic
rhinitis/hay fever
symptoms within the year
prior to the survey
0.192
(Parker et al., 2009)
Cardiac Arrest
Daily nonfatal incidence
rates
0-17
18-39
40-64
65-99
0.00000002
0.00000009
0.00000056
0.00000133
(Ensor et al., 2013; Rosenthal
et al., 2008; Silverman et al.,
2010)
Lung Cancer
Annual nonfatal incidence
25-34
35-44
45-54
55-64
65-74
75-84
95-99
0.000001746
0.000014919
0.000067463
0.000208053
0.000052370
0.000576950
0.000557130
(SEER, 2015) and
(Gharibvand et al., 2017)
Stroke
Annual nonfatal incidence
in ages 65-99
0.00446
(Kloog et al., 2012)
Work Loss Days
Daily incidence rate per
person (18-64)
Aged 18-24
Aged 25-44
Aged 45-64
0.00540
0.00678
0.00492
(Adams et al., 1999), Table
41; (U.S. Census Bureau,
2000)
School Loss Days
Rate per person per year,
assuming 180 school days
per year
9.9
(Adams et al., 1999), Table 47
287
-------
Rates
Endpoint
Parameter
Value
Source
Minor Restricted-
Daily MRAD incidence rate
0.02137
(B. D. Ostro and Rothschild,
Activity Days
per person (18-64")
1989), p. 243
CDC-Centers for Disease Control; NHS-National Health Interview Survey. Detailed references associated with
this table are located in the TSD.
'Mortality rates are only available in 5-year increments. The Healthcare Cost and Utilization Program (HCUP)
database contains individual level, state and regional-level hospital and emergency department discharges
for a variety of International Classification of Diseases (ICD) codes (AHRQ, 2016).
2Baseline incidence rates now include corrections from the states of Indiana and Montana.
We projected mortality rates such that future mortality rates are consistent with our
projections of population growth (U.S. EPA, 2018). To perform this calculation, we began
first with an average of 2007-2016 cause-specific mortality rates. Using Census Bureau
projected national-level annual mortality rates stratified by age range, we projected these
mortality rates to 2060 in 5-year increments (U.S. Census Bureau, 2009; U.S. EPA, 2023b).
Further information regarding this procedure may be found in the TSD and the appendices
to the BenMAP user manual (U.S. EPA, 2023a).
5.2.3 Effect Coefficients
Our approach for selecting and parametrizing effect coefficients for the benefits
analysis is described fully in the TSD. Because of the substantial economic value associated
with estimated counts of PIVh.s-attributable deaths, we describe our rationale for selecting
among long-term exposure epidemiologic studies below; a detailed description of all
remaining endpoints may be found in the TSD.
5.2.3.1 PM2.5 Premature Mortality Effect Coefficients for Adults
A substantial body of published scientific literature documents the association
between PM2.5 concentrations and the risk of premature death (U.S. EPA, 2019b) (U.S. EPA,
2022a). This body of literature reflects thousands of epidemiology, toxicology, and clinical
studies. The PM ISA, completed as part of this review of the PM standards and reviewed by
the Clean Air Scientific Advisory Committee (CASAC) (Sheppard, 2022), concluded that
there is a causal relationship between mortality and both long-term and short-term
exposure to PM2.5 based on the full body of scientific evidence (U.S. EPA, 2019b) (U.S. EPA,
2022a). The size of the mortality effect estimates from epidemiologic studies, the serious
288
-------
nature of the effect itself, and the high monetary value ascribed to reducing its risk, make
premature mortality the most significant health endpoint quantified in this analysis. EPA
selects Hazard Ratios from cohort studies to estimate counts of PM-related premature
death, following a systematic approach detailed in the TSD that is generally consistent with
previous (e.g., U.S. EPA, 2011a, U.S. EPA, 2011b, U.S. EPA, 2011c, U.S. EPA, 2012a, U.S. EPA,
2012b, U.S. EPA, 2015a, U.S. EPA, 2019a).
As premature mortality typically constitutes the vast majority of monetized benefits
in a PM2.5 benefits assessment, quantifying effects using risk estimates reported from
multiple long-term exposure studies using different cohorts helps account for uncertainty
in the estimated number of PM-related premature deaths. Below we summarize the two
identified studies and hazard ratios used in this analysis and then describe our rationale for
quantifying premature PM-attributable deaths using two of these studies.
Wu et al., 2020 evaluated the relationship between long-term PM2.5 exposure and
all-cause mortality in more than 68.5 million Medicare enrollees (over the age of 64), using
Medicare claims data from 2000-2016 representing over 573 million person-years of
follow up and over 27 million deaths. This cohort included over 20% of the U.S. population
and was, at the time of publishing, the largest air pollution study cohort to date. The
authors modeled PM2.5 exposure at a 1-km2 grid resolution using a hybrid ensemble-based
prediction model that combined three machine learning models and relied on satellite data,
land-use information, weather variables, chemical transport model simulation outputs, and
monitor data. Wu et al., 2020 fit five different statistical models: a Cox proportional hazards
model, a Poisson regression model, and three causal inference approaches (GPS estimation,
GPS matching, and GPS weighting). All five statistical approaches provided consistent
results; we report the results of the Cox proportional hazards model here. The authors
adjusted for numerous individual-level and community-level confounders, and sensitivity
analyses suggest that the results are robust to unmeasured confounding bias. In a single-
pollutant model, the coefficient and standard error for PM2.5 are estimated from the hazard
ratio (1.066) and 95% confidence interval (1.058-1.074) associated with a change in
annual mean PM2.5 exposure of 10.0 ug/m3 (Wu et al., 2020, Table S3, Main analysis, 2000-
2016 Cohort, Cox PH). We use a risk estimate from this study in place of the risk estimate
289
-------
from Di et al., 2017. These two epidemiologic studies share many attributes, including the
Medicare cohort and statistical model used to characterize population exposure to PM2.5. As
compared to Di et al., 2017, Wu et al., 2020 includes a longer follow-up period and reflects
more recent PM2.5 concentrations.
Pope III et al., 2019 examined the relationship between long-term PM2.5 exposure
and all-cause mortality in a cohort of 1,599,329 U.S. adults (aged 18-84 years) who were
interviewed in the National Health Interview Surveys (NHIS) between 1986 and 2014 and
linked to the National Death Index (NDI) through 2015. The authors also constructed a sub-
cohort of 635,539 adults from the full cohort for whom body mass index (BMI) and
smoking status data were available. The authors employed a hybrid modeling technique to
estimate annual-average PM2.5 concentrations derived from regulatory monitoring data
and constructed in a universal kriging framework using geographic variables including
land use, population, and satellite estimates. Pope III et al., 2019 assigned annual-average
PM2.5 exposure from 1999-2015 to each individual by census tract and used complex
(accounting for NHIS's sample design) and simple Cox proportional hazards models for the
full cohort and the sub-cohort. We select the Hazard Ratio calculated using the complex
model for the sub-cohort, which controls for individual-level covariates including age, sex,
race-ethnicity, inflation-adjusted income, education level, marital status, rural versus
urban, region, survey year, BMI, and smoking status. In a single-pollutant model, the
coefficient and standard error for PM2.5 are estimated from the hazard ratio (1.12) and
95% confidence interval (1.08-1.15) associated with a change in annual mean PM2.5
exposure of 10.0 ug/m3 (Pope III et al., 2019, Table 2, Subcohort). This study exhibits two
key strengths that makes it particularly well suited for a benefits analysis: (1) it includes a
long follow-up period with recent (and thus relatively low) PM2.5 concentrations; (2) the
NHIS cohort is representative of the U.S. population, especially with respect to the
distribution of individuals by race, ethnicity, income, and education.
5.2.4 Unquantified Human Health Benefits
Although we have quantified many of the health benefits associated with reducing
exposure to PM2.5, as shown in Table 5-2, we are unable to quantify the health benefits of
implementing the illustrative control strategies described in Chapter 3 associated with
290
-------
reducing ozone exposure, SO2 exposure, or NO2 exposure. This is because we focused on
reducing direct PM emissions and do not have air quality modeling data for these
pollutants. Although we used air quality surfaces that reflect applying the control strategies
for the impact of each combination of standard levels on ambient levels of PM2.5, this
method does not simulate how the illustrative emissions reductions would affect ambient
levels of ozone, SO2, or NO2. Below we provide a qualitative description of these health
benefits. In general, previous analyses have shown that the monetized value of these
additional health benefits is much smaller than PIVh.s-related benefits (U.S. EPA, 2010,
2015a). The extent to which ozone, SO2, and/or NOx would be reduced would depend on
the specific control strategies used to reduce PM2.5 in a given area.
Exposure to ambient ozone is associated with human health effects, including
respiratory and metabolic morbidity (U.S. EPA, 2020a). Epidemiological researchers have
associated ozone exposure with adverse health effects in numerous toxicological, clinical
and epidemiological studies (U.S. EPA, 2020a). When adequate data and resources are
available, EPA generally quantifies several health effects associated with exposure to ozone
(e.g., U.S. EPA, 2014b, 2015a). These health effects include respiratory morbidity such as
asthma attacks, hospital admissions, emergency department visits, and school loss days.
The scientific literature suggests that exposure to ozone is also associated with chronic
respiratory damage and premature aging of the lungs, but EPA has not quantified these
effects in benefits analyses previously.
Following an extensive evaluation of health evidence from epidemiologic and
laboratory studies, the Integrated Science Assessment for Sulfur Dioxide—Health Criteria
(SO2 ISA) concluded that there is a causal relationship between respiratory health effects
and short-term exposure to SO2 (U.S. EPA, 2017). The immediate effect of SO2 on the
respiratory system in humans is bronchoconstriction. Asthmatics are more sensitive to the
effects of SO2 likely resulting from preexisting inflammation associated with this disease. A
clear concentration-response relationship has been demonstrated in laboratory studies
following exposures to SO2, both in terms of increasing severity of effect and percentage of
asthmatics adversely affected. Based on our review of this information, we identified three
short-term morbidity endpoints that the SO2 ISA identified as a "causal relationship":
291
-------
asthma exacerbation, respiratory-related emergency department visits, and respiratory -
related hospitalizations. The differing evidence and associated strength of the evidence for
these different effects is described in detail in the SO2 ISA (U.S. EPA, 2017). The SO2 ISA also
concluded that the relationship between short-term SO2 exposure and premature mortality
was "suggestive of a causal relationship" because it is difficult to attribute the mortality risk
effects to SO2 alone. Although the SO2 ISA stated that studies are generally consistent in
reporting a relationship between SO2 exposure and premature mortality, the number of
studies was limited. Because we focused on reducing primary PM emissions, we did not
quantify these benefits.
Epidemiological researchers have associated NO2 exposure with adverse health
effects in numerous toxicological, clinical and epidemiological studies, as described in the
Integrated Science Assessment for Oxides of Nitrogen—Health Criteria (NO2 ISA) (U.S. EPA,
2016). The NO2 ISA provides a comprehensive review of the current evidence of health and
environmental effects of NO2. The NO2 ISA concluded that "evidence for asthma attacks
supports a causal relationship between short-term NO2 exposure and respiratory effects,"
and "evidence for development of asthma supports a likely to be causal relationship
between long-term NO2 exposure and respiratory effects." These are stronger conclusions
than those determined in the 2008 NO2 ISA (U.S. EPA, 2008).These epidemiologic and
experimental studies encompass a number of endpoints including emergency department
visits and hospitalizations, respiratory symptoms, airway hyperresponsiveness, airway
inflammation, and lung function. Effect estimates from epidemiologic studies conducted in
the United States and Canada generally indicate a 2-20% increase in risks for ED visits and
hospital admissions and higher risks for respiratory symptoms. The NO2 ISA concluded that
the relationship between short-term NO2 exposure and premature mortality was
"suggestive but not sufficient to infer a causal relationship" because it is difficult to
attribute the mortality risk effects to NO2 alone. Although the NO2 ISA stated that studies
consistently reported a relationship between NO2 exposure and mortality, the effect was
generally smaller than that for other pollutants such as PM. Because we focused on
reducing primary PM emissions, we did not quantify these benefits.
292
-------
Illustrative controls to meet the revised and alternative standard levels are expected
to reduce PM2.5 emissions from fossil fuel and wood combustion, as well as industrial
processes, and consequentially is expected to lead to reduced Hazardous Air Pollutant
(HAP) emissions. HAP emissions from EGUs and other industrial sources may contribute to
increased cancer risks and other serious health effects, including damage to the immune
system, as well as neurological, reproductive (e.g., reduced fertility), developmental,
respiratory and other health problems. These public health implications of exposure to
HAPs can be particularly pronounced for segments of the population that are especially
vulnerable to some of these effects [e.g., children are especially vulnerable to neurological
effects because their brains are still developing). Some HAPs can also detrimentally affect
ecosystems used for recreational and commercial purposes.
5.2.5 Unquantified Welfare Benefits
The Clean Air Act definition of welfare effects includes, but is not limited to, effects
on soils, water, wildlife, vegetation, visibility, weather, and climate, as well as effects on
man-made materials, economic values, and personal comfort and well-being. Detailed
information regarding the ecological effects of nitrogen and sulfur deposition is available in
the Integrated Science Assessment for Oxides of Nitrogen, Oxides of Sulfur, and Particulate
Matter— Ecological Criteria (ISA) (U.S. EPA, 2020b).
Particulate matter (PM) is composed of some or all of the following components:
nitrate (NO3-), sulfate (SO42-), ammonium (NH4+), metals, minerals (dust), and organic and
elemental carbon. Nitrate, sulfate, and ammonium contribute to nitrogen (N) and sulfur (S)
deposition, which causes substantial ecological effects. The ecological effects of deposition
are grouped into three main categories: acidification, N enrichment/N driven
eutrophication, and S enrichment. Ecological effects are further subdivided into terrestrial,
wetland, freshwater, and estuarine/near-coastal ecosystems. These ecosystems and effects
are linked by the connectivity of terrestrial and aquatic habitats through biogeochemical
pathways of N and S.
In the ISA, information on ecological effects from controlled exposure, field addition,
ambient deposition, and toxicological studies, among others, are integrated to form
conclusions about the causal relationships between NOy, SOx, and PM and ecological
293
-------
effects. A consistent and transparent framework (U.S. EPA, 2015b, Table II) is applied to
classify the ecological effect evidence according to a five-level hierarchy:
1. Causal relationship
2. Likely to be a causal relationship
3. Suggestive of, but not sufficient to infer, a causal relationship
4. Inadequate to infer a causal relationship
5. Not likely to be a causal relationship
Table 5-4 summarizes the causal determinations for relationships between N and S
deposition and ecological effects. Though not quantified in this RIA, it is reasonable to infer
that reducing fine particle levels by controlling emissions of NOx and SOx will yield the
ecological benefits detailed below.
Table 5-4 Causal Determinations Identified in Integrated Science Assessment for
Oxides of Nitrogen, Oxides of Sulfur, and Particulate Matter —
Ecological Criteria 2020b
Effect Category
Causal Determination
N and acidifying deposition to terrestrial
ecosystems
N and S deposition and alteration of soil
Causal relationship
biogeochemistry in terrestrial ecosystems
Section IS.5.1 and Appendix 4.1
N deposition and the alteration of the physiology and
Causal relationship
growth of terrestrial organisms and the productivity
of terrestrial ecosystems
Section IS.5.2 and Appendix 6.6.1
N deposition and the alteration of species richness,
Causal relationship
community composition, and biodiversity in
terrestrial ecosystems
Section IS.5.2 and Appendix 6.6.2
Acidifying N and S deposition and the alteration of the
Causal relationship
physiology and growth of terrestrial organisms and
the productivity of terrestrial ecosystems
Section IS.5.3 and Appendix 5.7.1
Acidifying N and S deposition and the alteration of
Causal relationship
species richness, community composition, and
biodiversity in terrestrial ecosystems
Section IS.5.3 and Appendix 5.7.2
N and acidifying deposition to freshwater
ecosystems
N and S deposition and alteration of freshwater
Causal relationship
biogeochemistry
Section IS.6.1 and Appendix 7.1.7
294
-------
Effect Category
Causal Determination
Acidifying N and S deposition and changes in biota,
including physiological impairment and alteration of
species richness, community composition, and
biodiversity in freshwater ecosystems
Section IS.6.3 and Appendix 8.6
Causal relationship
N deposition and changes in biota, including altered
growth and productivity, species richness, community
composition, and biodiversity due to N enrichment in
freshwater ecosystems
Section IS.6.2 and Appendix 9.6
Causal relationship
N deposition to estuarine ecosystems
N deposition and alteration of biogeochemistry in
estuarine and near-coastal marine systems
Section IS.7.1 and Appendix 7.2.10
Causal relationship
N deposition and changes in biota, including altered
growth, total primary production, total algal
community biomass, species richness, community
composition, and biodiversity due to N enrichment in
estuarine environments
Section IS.7.2 and Appendix 10.7
Causal relationship
N deposition to wetland ecosystems
N deposition and the alteration of biogeochemical
cycling in wetlands
Section IS.8.1 and Appendix 11.10
Causal relationship
N deposition and the alteration of growth and
productivity, species physiology, species richness,
community composition, and biodiversity in wetlands
Section IS.8.2 and Appendix 11.10
Causal relationship
S deposition to wetland and freshwater
ecosystems
S deposition and the alteration of mercury
methylation in surface water, sediment, and soils in
wetland and freshwater ecosystems
Section IS.9.1 and Appendix 12.7
Causal relationship
S deposition and changes in biota due to sulfide
phytotoxicity, including alteration of growth and
productivity, species physiology, species richness,
community composition, and biodiversity in wetland
and freshwater ecosystems
Section IS.9.2 and Appendix 12.7
Causal relationship
5.2.5.1 Visibility Impairment Benefits
Reducing PM2.5 would improve levels of visibility in the U.S. because suspended
particles and gases degrade visibility by scattering and absorbing light (U.S. EPA, 2009).
Fine particles with significant light-extinction efficiencies include sulfates, nitrates, organic
carbon, elemental carbon, and soil (Sisler, 1996). Visibility has direct significance to
people's enjoyment of daily activities and their overall sense of wellbeing. Good visibility
295
-------
increases the quality of life where individuals live and work, and where they engage in
recreational activities. Particulate sulfate is the dominant source of regional haze in the
eastern U.S. and particulate nitrate is an important contributor to light extinction in
California and the upper Midwestern U.S., particularly during winter (U.S. EPA, 2009).
Previous analyses (U.S. EPA, 2011d) show that visibility benefits can be a significant
welfare benefit category. Without air quality modeling, we are unable to estimate visibility-
related benefits, and we are also unable to determine whether the emission reductions
would be likely to have a significant impact on visibility in urban areas or Class I areas.
5.2.6 Climate Effects of PM2.5
Atmospheric particles influence climate in multiple ways: directly absorbing light,
scattering light, changing the reflectivity ("albedo") of snow and ice through deposition,
and interacting with clouds. The particle's composition, the timing of emissions, and where
the particle is in the atmosphere determine if it contributes to cooling or warming.
Pollutants that affect the energy balance of the earth are referred to as climate forcers. A
pollutant that increases the amount of energy in the Earth's climate system is said to exert
"positive radiative forcing," which leads to warming and climate change. In contrast, a
pollutant that exerts negative radiative forcing reduces the amount of energy in the Earth's
system and leads to cooling. The short atmospheric lifetime of particles, lasting from days
to weeks, and the mechanisms by which particles affect climate, distinguish it from long-
lived greenhouse gases like CO2. This means that actions taken to reduce PM2.5 will have
near term effects on climate change. The Intergovernmental Panel on Climate Change Sixth
Assessment Report concludes that for forcers with short lifetimes, "the response in surface
temperature occurs strongly, as soon as a sustained change in emissions is implemented"
(Naik et al., 2021). The potential to affect near-term climate change and the rate of climate
change with policies to address these emissions is gaining attention nationally and
internationally (e.g., Black Carbon Report to Congress, Arctic Council, Climate and Clean Air
Coalition, and Convention on Long-Range Transboundary Air Pollution of the United
Nations Economic Commission for Europe). Recent reports have concluded that short-lived
compounds play a prominent role in keeping global warming below 1.5° C (IPCC, 2018),
and are especially important in the rapidly warming Arctic (AMAP, 2021). While reducing
296
-------
long-lived GHGs such as CO2 is necessary to protect against long-term climate change,
reducing short-lived forcers and would slow the rate of climate change within the first half
of this century (UNEP, 2011).
5.2.6.1 Climate Effects of Carbonaceous Particles
The illustrative control strategies are focused on emissions sources that are
significant sources of carbonaceous particles, including black carbon and organic carbon.
Black Carbon (BC), also called soot, is the most strongly light-absorbing component of
PM2.5, and is formed by incomplete combustion of fossil fuels, biofuels, and biomass.
Another contributor to carbonaceous particles is organic carbon (OC), which in addition to
carbon are also composed of oxygen and hydrogen. Organic carbon particles can be directly
emitted from the same sources as black carbon or formed in the atmosphere from chemical
reactions. They can be light-absorbing, but most have a larger light-scattering component.
Both BC and organic carbon in the atmosphere influence climate in multiple ways:
directly absorbing or reflecting light, modifying the rate of vertical mixing, and interacting
with clouds. Light-absorbing particles also have an additional climate effect when
deposited on snow and ice. These particles darken the surface and decrease albedo,
thereby increasing absorption and accelerating melting (Hock et al., 2019; Meredith et al.,
2019). Regional climate impacts of BC are highly variable, and sensitive regions such as the
Arctic are particularly vulnerable to the warming and melting effects of BC. Snow and ice
cover in the western U.S. has also been affected by BC. Specifically, deposition of BC on
mountain glaciers and snowpacks produces a positive snow and ice albedo effect,
contributing to the melting of snowpack earlier in the spring and reducing the amount of
snowmelt that normally would occur later in the spring and summer (Hadley et al. 2010).
This has implications for freshwater resources in regions of the U.S. dependent on snow-
fed or glacier-fed water systems. In the Sierra Nevada mountain range, Hadley et al. (2010)
found BC at different depths in the snowpack, deposited over the winter months by
snowfall. In the spring, the continuous uncovering of the BC contributed to the early melt. A
model capturing the effects of soot on snow in the western U.S. shows significant decreases
in snowpack between December and May (Qian et al., 2009). Snow water equivalent (the
amount of water that would be produced by melting all the snow) is reduced 2-50
297
-------
millimeters (mm) in mountainous areas, particularly over the Central Rockies, Sierra
Nevadas, and western Canada. A study found that biomass burning emissions in Alaska and
the Rocky Mountain region during the summer can enhance snowmelt (McKenzie, Skiles et
al. 2018). Light-absorbing particles and especially BC can have an additional warming
effect when deposited on snow and ice, and this effect is highly seasonal and regional.
Relative to greenhouse gases, the net effect of carbonaceous particles is both more
regionally variable and more uncertain (Naik et al., 2021). Particles have a relatively short
lifetime in the atmosphere, leading to spatial concentration differences, while greenhouse
gases are more well mixed and have less global variability. The amount of light absorption
by particles depends on the season, with different effects in the summer and winter. Lastly,
even light-absorbing particles can also contribute to cooling (e.g., by shading the surface).
5.2.6.2 Climate Effects: Summary and Conclusions
The net climate change effect of carbonaceous aerosols in the illustrative control
strategies depends on the location, timing, and type of the emissions controls. As described
above, the black carbon emissions are more likely to contribute to warming and organic
aerosols more likely to contribute to cooling. Emissions sources with larger amounts of
light-absorbing aerosols, like diesel vehicles, or with emissions near snow or the Arctic, like
residential wood combustion, are more likely to contribute to warming (Bond et al., 2013).
However, assessing the net effect is beyond the scope of this RIA and requires
climate atmospheric modeling that has not been undertaken. Furthermore, there are
uncertainties relevant to the assessment of the net climate change effects of PM2.5,
especially at a regional scale (U.S. EPA, 2019b). Strategies that could be implemented by
State and Local governments that would likely provide climate change mitigation benefits
include prioritizing (/) emissions control actions that also achieve emissions reductions for
warming agents like carbon dioxide, methane, and ozone precursors (carbon monoxide and
volatile organic compounds), and (/'/') sources of light-absorbing carbonaceous aerosols,
especially diesel engines and residential wood combustion.
5.2.7 Economic Valuation Estimates
To directly compare benefits estimates associated with a rulemaking to cost
estimates, the number of instances of each air pollution-attributable health impact must be
298
-------
converted to a monetary value. This requires a valuation estimate for each unique health
endpoint, and potentially also discounting if the benefits are expected to accrue over more
than a single year, as recommended by the Guidelines for Preparing Economic Analyses (U.S.
EPA, 2016a).
5.3 Characterizing Uncertainty
In any complex analysis using estimated parameters and inputs from numerous
models, there are likely to be many sources of uncertainty. This analysis is no exception.
The TSD details our approach to characterizing uncertainty in both quantitative and
qualitative terms. The TSD describes the sources of uncertainty associated with key input
parameters including emissions inventories, air quality data from models (with their
associated parameters and inputs), population data, population estimates, health effect
estimates from epidemiology studies, economic data for monetizing benefits, and
assumptions regarding the future state of the country (i.e., regulations, technology, and
human behavior). Each of these inputs is uncertain and affects the size and distribution of
the estimated benefits. When the uncertainties from each stage of the analysis are
compounded, even small uncertainties can have large effects on the total quantified
benefits.
To characterize uncertainty and variability into this assessment, we incorporate
three quantitative analyses described below and in greater detail within the TSD (Section
7.1):
1. A Monte Carlo assessment that accounts for random sampling error and
between study variability in the epidemiological and economic valuation studies;
2. The quantification of PM-related mortality using alternative PM2.5 mortality
effect estimates drawn from two long-term cohort studies; and
3. Presentation of 95th percentile confidence interval around each risk estimate.
Quantitative characterization of other sources of PM2.5 uncertainties are discussed
only in Section 7.1 of the TSD:
1. For adult all-cause mortality:
299
-------
a. The distributions of air quality concentrations experienced by the
original cohort population (TSD Section 7.1.2.1);
b. Methods of estimating and assigning exposures in epidemiologic studies
(TSD Section 7.1.2.2);
c. Confounding by ozone (TSD Section 7.1.2.3); and
d. The statistical technique used to generate hazard ratios in the
epidemiologic study (TSD Section 7.1.2.4).
2. Plausible alternative risk estimates for asthma onset in children (TSD Section
7.1.3), cardiovascular hospital admissions (TSD Section 7.1.4,), and respiratory
hospital admissions (TSD Section 7.1.5);
3. Effect modification of PIVh.s-attributable health effects in at-risk populations
(TSD Section 7.1.6).
Quantitative consideration of baseline incidence rates and economic valuation
estimates are provided in Section 7.3 and 7.4 of the TSD, respectively. Qualitative
discussions of various sources of uncertainty can be found in Section 7.5 of the TSD.
5.3.1 Monte Carlo Assessment
Similar to other recent RIAs, we used Monte Carlo methods for characterizing
random sampling error associated with the concentration response functions from
epidemiological studies and random effects modeling to characterize both sampling error
and variability across the economic valuation functions. The Monte Carlo simulation in the
BenMAP-CE software randomly samples from a distribution of incidence and valuation
estimates to characterize the effects of uncertainty on output variables. Specifically, we
used Monte Carlo methods to generate confidence intervals around the estimated health
impact and monetized benefits. The reported standard errors in the epidemiological
studies determined the distributions for individual effect estimates for endpoints derived
from a single study. For endpoints estimated using a pooled estimate of multiple studies,
the confidence intervals reflect both the standard errors and the variance across studies.
The confidence intervals around the monetized benefits incorporate the epidemiology
standard errors as well as the distribution of the valuation function. These confidence
300
-------
intervals do not reflect other sources of uncertainty inherent within the estimates, such as
baseline incidence rates, populations exposed, and transferability of the effect estimate to
diverse locations. As a result, the reported confidence intervals and range of estimates give
an incomplete picture about the overall uncertainty in the benefits estimates.
5.3.2 Sources of Uncertainty Treated Qualitatively
Although we strive to incorporate as many quantitative assessments of uncertainty
as possible, there are several aspects we are only able to address qualitatively. These
attributes are summarized below and described more fully in the TSD.
Key assumptions underlying the estimates for premature mortality, which account
for over 98% of the total monetized benefits in this analysis, include the following:
1. We assume that all fine particles, regardless of their chemical composition,
are equally potent in causing premature mortality. This is an important
assumption, because PM2.5 varies considerably in composition across
sources, but the scientific evidence is not yet sufficient to allow
differentiation of effect estimates by particle type. The PM ISA, which was
reviewed by CASAC, concluded that "across exposure durations and health
effects categories ... the evidence does not indicate that any one source or
component is consistently more strongly related with health effects than
PM2.5 mass" (U.S. EPA, 2019b).
2. We assume that the health impact function for fine particles is log-linear
down to the lowest air quality levels modeled in this analysis. Thus, the
estimates include health benefits from reducing fine particles in areas with
varied concentrations of PM2.5, including both regions that are in attainment
with the fine particle standard and those that do not meet the standard down
to the lowest modeled concentrations. The PM ISA concluded that "the
majority of evidence continues to indicate a linear, no-threshold
concentration-response relationship for long-term exposure to PM2.5 and
total (nonaccidental) mortality" (U.S. EPA, 2019b).
301
-------
3. We assume that there is a "cessation" lag between the change in PM
exposures and the total realization of changes in mortality effects.
Specifically, we assume that some of the incidences of premature mortality
related to PM2.5 exposures occur in a distributed fashion over the 20 years
following exposure based on the advice of the SAB-HES (Cameron, 2004),
which affects the valuation of mortality benefits at different discount rates.
Similarly, we assume there is a cessation lag between the change in PM
exposures and both the development and diagnosis of lung cancer.
5.4 Benefits Results
5.4.1 Benefits of the Applied Control Strategies for the Revised and Alternative
Combinations of Primary PM2.5 Standard Levels
Applying the impact and valuation functions described previously in this chapter to
the estimated changes in PM2.5 yields estimates of the changes in physical damages (e.g.,
premature mortalities, cases of hospital admissions and emergency department visits) and
the associated monetary values for those changes. Not all known PM health effects could be
quantified or monetized.
We present two sets of tables - one set in this chapter and one set in Appendix 5A.
First, Table 5-5 through Table 5-9 present benefits associated with the illustrative control
strategies identified in Chapter 3. More specifically, for the revised standard levels of 9/35
Hg/m3, for the northeast we were able to identify approximately 98 percent of the
reductions needed. For the southeast we were able to identify approximately 68 percent of
the reductions needed. For the west, we were able to identify approximately 44 percent of
the reductions needed, and for California the percentage is approximately 26 percent. As
such, these tables present the benefits associated with the illustrative control strategies
and reflect the remaining air quality challenges (discussed in Chapter 2, Section 2.4 and
Chapter 3, Section 3.2.5). For a more detailed description of the geographic distribution of
the emissions reductions needed for the revised and alternative standard levels, see the
discussion in Chapter 3, Section 3.2.4. Second, Table 5A-1 through 5A-5 in Appendix 5A
present the potential benefits associated with fully meeting the revised and alternative
standards.
302
-------
Table 5-5 through Table 5-9 present the benefits results of applying the control
strategies for the revised annual and current 24-hour alternative standard levels of 9/35
Hg/m3, as well as the less stringent alternative standard levels of 10/35 |~ig/m3 and the
following two more stringent alternative standard levels: (1) an alternative annual
standard level of 8 |~ig/m3 in combination with the current 24-hour standard (i.e., 8/35 |Lxg/-
m3), and (2) an alternative 24-hour standard level of 30 ng/m3 in combination with an
annual standard level of 10 ng/m3 (i.e., 10/30 |j,g/m3).
Table 5-5 presents the estimated avoided incidences of PM-related illnesses and
premature mortality resulting from the control strategies applied to the revised and
alternative standard levels in 2032. Table 5-6 and Table 5-7 present the monetized
valuation benefits (discounted at a 3% and 7% discount rate, respectively) of the avoided
health outcomes presented in Table 5-5.
Table 5-8 and Table 5-9 present a summary of the monetized benefits associated
with the revised and the alternative standard levels, both nationally and by area. The
monetized benefits in Table 5-8 are presented in four areas: California (CA), the
Northeastern (NE) states, the Southeastern (SE) states, and the Western (W) states. For
Table 5-8 and Table 5-9, the monetized value of unquantified effects is represented by
adding an unknown "B" to the aggregate total. This B represents both uncertainty and a
bias in this analysis, as it reflects health and welfare benefits that we are unable to
quantify.5
5 The health and monetized benefits of fully attaining the revised and alternative standard levels in all areas
can be found in Appendix 5A.
303
-------
Table 5-5 Estimated Avoided PM-Related Premature Mortalities and Illnesses of
the Applied Control Strategies for the Revised and Alternative Primary
Standard Levels of 10/35 |ig/m3,10/30 |ig/m3, 9/35 |ig/m3, and 8/35
(ig/m3 for 2032 (95% Confidence Interval)
Avoided Mortality3
10/35
10/30
9/35
8/35
(Pope Illetal., 2019)
(adult mortality ages 18-
99 years}
1,700
(1,200 to 2,100)
2,000
(1,400 to 2,600)
4,500
(3,200 to 5,700)
9,500
(6,800 to 12,000)
(Wu et al., 2020) (adult
mortality ages 65-99
years)
810
(710 to 900)
970
(850 to 1,100)
2,100
(1,900 to 2,400)
4,500
(4,000 to 5,100)
(Woodruff et al., 2008)
(infant mortality)
1.7
(-1.0 to 4.3)
2.0
(-1.2 to 5.1)
5.0
(-3.1 to 13)
11
(-7.2 to 29)
Avoided Morbidity
10/35
10/30
9/35
8/35
Hospital admissions—
cardiovascular (age > 18)
140
(100 to 180)
160
(120 to 200)
330
(240 to 420)
690
(500 to 870)
Hospital admissions—
respiratory
90
(31 to 150)
100
(35 to 170)
230
(79 to 370)
480
(170 to 780)
ED visits-cardiovascular
260
(-98 to 600)
300
(-110 to 690)
660
(-250 to 1,500)
1,400
(-550 to 3,300)
ED visits—respiratory
470
(93 to 990)
560
(110 to 1,200)
1,300
(250 to 2,700)
2,900
(570 to 6,000)
Acute Myocardial
Infarction
30
(17 to 42)
35
(20 to 49)
72
(42 to 100)
150
(86 to 210)
Cardiac arrest
14
(-5.9 to 33)
17
(-6.9 to 38)
36
(-15 to 81)
75
(-31 to 170)
Hospital admissions-
Alzheimer's Disease
360
(270 to 440)
400
(300 to 500)
910
(690 to 1,100)
2,000
(1,500 to 2,400)
Hospital admissions-
Parkinson's Disease
47
(24 to 69)
56
(29 to 81)
130
(67 to 190)
280
(140 to 400)
Stroke
54
(14 to 93)
65
(17 to 110)
140
(35 to 230)
290
(74 to 490)
Lung cancer
65
(20 to 110)
77
(23 to 130)
160
(49 to 270)
340
(100 to 560)
Hay Fever/Rhinitis
15,000
(3,600 to 26,000)
17,000
(4,200 to 30,000)
38,000
(9,100 to 65,000)
79,000
(19,000 to 140,000)
Asthma Onset
2,300
(2,200 to 2,300)
2,600
(2,500 to 2,700)
5,700
(5,500 to 6,000)
12,000
(12,000 to 13,000)
Asthma symptoms -
Albuterol use
310,000
(-150,000 to
760,000)
370,000
(-180,000 to
900,000)
800,000
(-390,000 to
1,900,000)
1,700,000
(-820,000 to
4,100,000)
Lost work days
110,000
(96,000 to
130,000)
130,000
(110,000 to
150,000)
290,000
(240,000 to
330,000)
610,000
(510,000 to
700,000)
Minor restricted-activity
days
670,000
(540,000 to
790,000)
780,000
(630,000 to
920,000)
1,700,000
(1,400,000 to
2,000,000)
3,500,000
(2,900,000 to
4,200,000)
Note: Values rounded to two significant figures.
a Reported here are two alternative estimates of the number of premature deaths among adults due to long-
term exposure to PM2.5. These values should not be added to one another.
304
-------
Table 5-6 Monetized PM-Related Premature Mortalities and Illnesses of the
Applied Control Strategies for the Revised and Alternative Primary
Standard Levels of 10/35 ng/m3,10/30 ng/m3, 9/35 ng/m3, and 8/35
(j,g/m3 for 2032 (Millions of 2017$, 3% discount rate; 95% Confidence
Interval)
Avoided Mortality3
10/35
10/30
9/35
8/35
(Pope III et al., 2019) (adult
mortality ages 18-99 years)
17,000
(1,600 to 47,000)
21,000
(1,900 and 56,000)
46,000
(4,100 to 120,000)
98,000
(8,900 to 260,000)
(Wu et al., 2020) (adult
mortality ages 65-99 years)
8,300
(760 to 22,000)
9,900
(920 to 26,000)
22,000
(2,000 to 58,000)
47,000
(4,300 to 120,000)
(Woodruff et al., 2008)
(infant mortality)
19
(-11 and 75)
22
(-12 and 89)
56
(-31 and 220)
130
(-72 and 510)
Avoided Morbidity
10/35
10/30
9/35
8/35
Hospital admissions—
cardiovascular (age > 18)
2.3
(1.7 to 2.9)
2.7
(1.9 and 3.4)
5.5
(4 and 7)
11
(8.2 and 14)
Hospital admissions—
respiratory
1.5
(0.34 and 2.7)
1.8
(0.39 and 3.1)
3.8
(0.87 and 6.6)
7.9
(1.8 and 14)
ED visits--cardiovascular
0.32
(-0.12 and 0.74)
0.37
(-0.14 and 0.86)
0.82
(-0.31 and 1.9)
1.8
(-0.68 and 4.1)
ED visits—respiratory
0.44
(0.087 and 0.92)
0.52
[0.1 and 1.1]
1.2
(0.24 and 2.5)
2.7
(0.53 and 5.6)
Acute Myocardial Infarction
1.5
(0.9 and 2.2)
1.8
(1 and 2.5]
3.7
(2.2 and 5.2)
7.7
[4.5 and 11]
Cardiac arrest
0.55
(-0.22 and 1.2)
0.64
(-0.26 and 1.5)
1.4
(-0.56 and 3.1)
2.9
(-1.2 and 6.5)
Hospital admissions-
Alzheimer's Disease
4.6
(3.5 and 5.6)
5.2
(3.9 and 6.3)
12
(8.8 and 14)
25
(19 and 31)
Hospital admissions-
Parkinson's Disease
0.65
(0.33 and 0.94)
0.77
(0.39 and 1.1)
1.8
(0.92 and 2.6)
3.8
(2 and 5.5)
Stroke
2
(0.51 and 3.4)
2.3
(0.6 and 4)
4.9
(1.3 and 8.4)
10
(2.7 and 18)
Lung cancer
1.8
(0.55 and 3)
2.1
(0.65 and 3.5)
4.5
(1.4 and 7.4)
9.3
(2.9 and 15)
Hay Fever/Rhinitis
9.5
(2.3 and 16)
11
(2.7 and 19)
24
(5.8 and 41)
50
(12 and 87)
Asthma Onset
110
(99 and 110)
120
(120 and 130)
270
(250 and 290)
570
(530 and 610)
Asthma symptoms -
Albuterol use
0.12
(-0.056 and 0.28)
0.14
(-0.066 and 0.33)
0.29
(-0.14 and 0.71)
0.62
(-0.3 and 1.5)
Lost work days
20
(17 and 24)
24
(20 and 27)
51
(43 and 59)
110
(92 and 130)
Minor restricted-activity
days
52
(27 and 79)
61
(32 and 92)
130
(69 and 200)
280
(150 and 420)
Note: Values rounded to two significant figures.
a Reported here are two alternative estimates of the number of premature deaths among adults due to long-
term exposure to PM2.5. These values should not be added to one another.
305
-------
Table 5-7 Monetized PM-Related Premature Mortalities and Illnesses of the
Applied Control Strategies for the Revised and Alternative Primary
Standard Levels of 10/35 |ig/m3,10/30 |ig/m3, 9/35 |ig/m3, and 8/35
(ig/m3 for 2032 (Millions of 2017$, 7% discount rate; 95% Confidence
Interval)
Avoided Mortality3
10/35
10/30
9/35
8/35
(Pope Illetal., 2019)
(adult mortality ages 18-
99 years}
16,000
(1,400 and 42,000)
19,000
(1,700 and 50,000)
41,000
(3,700 and
110,000)
88,000
(8,000 and 240,000)
(Wu et al., 2020) (adult
mortality ages 65-99
years)
7400(690 and
20000)
8900 (830 and
24000)
20000 (1800 and
52000)
42000 (3900 and
110000)
(Woodruff et al., 2008)
(infant mortality)
19
(-11 and 75)
22
(-12 and 89)
56
(-31 and 220)
130
(-72 and 510)
Avoided Morbidity
10/35
10/30
9/35
8/35
Hospital admissions—
cardiovascular (age > 18)
2.3
(1.7 to 2.9)
2.7
(1.9 and 3.4)
5.5
(4 and 7)
11
(8.2 and 14)
Hospital admissions—
respiratory
1.5
(0.34 and 2.7)
1.8
(0.39 and 3.1)
3.8
(0.87 and 6.6)
7.9
(1.8 and 14)
ED visits-cardiovascular
0.32
(-0.12 and 0.74)
0.37
(-0.14 and 0.86)
0.82
(-0.31 and 1.9)
1.8
(-0.68 and 4.1)
ED visits—respiratory
0.44
(0.087 and 0.92)
0.52
(0.1 and 1.1)
1.2
(0.24 and 2.5)
2.7
(0.53 and 5.6)
Acute Myocardial
Infarction
1.5 (0.87 and 2.1)
1.8 (land 2.5)
3.6 (2.1 and 5.1)
7.5 (4.4 and 11)
Cardiac arrest
0.54
(-0.22 and 1.2)
0.64
(-0.26 and 1.4)
1.3
(-0.55 and 3)
2.8
(-1.2 and 6.4)
Hospital admissions-
Alzheimer's Disease
4.6
(3.5 and 5.6)
5.2
(3.9 and 6.3)
12
(8.8 and 14)
25
(19 and 31)
Hospital admissions-
Parkinson's Disease
0.65
(0.33 and 0.94)
0.77
(0.39 and 1.1)
1.8
(0.92 and 2.6)
3.8
(2 and 5.5)
Stroke
2
(0.51 and 3.4)
2.3
(0.6 and 4)
4.9
(1.3 and 8.4)
10
(2.7 and 18)
Lung cancer
1.3
(0.41 and 2.2)
1.6
(0.49 and 2.6)
3.3
(1 and 5.5)
7
(2.1 and 12)
Hay Fever/Rhinitis
9.5
(2.3 and 16)
11
(2.7 and 19)
24
(5.8 and 41)
50
(12 and 87)
Asthma Onset
66 (61 and 70)
77 (72 and 82)
170(160 and 180)
350 (330 and 380)
Asthma symptoms -
Albuterol use
0.12
(-0.056 and 0.28)
0.14
(-0.066 and 0.33)
0.29
(-0.14 and 0.71)
0.62
(-0.3 and 1.5)
Lost work days
20
(17 and 24)
24
(20 and 27)
51
(43 and 59)
110
(92 and 130)
Minor restricted-activity
days
52
(27 and 79)
61
(32 and 92)
130
(69 and 200)
280
(150 and 420)
Note: Values rounded to two significant figures.
a Reported here are two alternative estimates of the number of premature deaths among adults due to long-
term exposure to PM2.5. These values should not be added to one another.
306
-------
Table 5-8 Estimated Monetized Benefits of the Applied Control Strategies for the
Revised and Alternative Primary Standard Levels of 10/35 |jg/m3,
10/30 |ig/m3,9/35 (j,g/m3, and 8/35 jig/m3 for 2032, Incremental to
Attainment of 12/35 (billions of 2017$)
ft f I® l*g/m3 annual & 10 jig/m3 annual & 9 jig/m3 annual & 8 ng/m3 annual &
Benents Estimate 35 ng/m3 24-hour 30 iig/m3 24-hour 35 ng/m3 24-hour 35 ng/m3 24-hour
Economic value of avoided PIVh.s-related morbidities and premature deaths using PM2.5 mortality
estimate from (Pope III et al., 2019)
3% discount $17+ B $21 + B $46 + B $99 + B
rate
7% discount $16+ B $19 + B $42 + B $89 + B
rate
Economic value of avoided PIVh.s-related morbidities and premature deaths using PM2.5 mortality
estimate from (Wu etal., 2020)
3% discount 5 + g $10 + B $22 + B $48 + B
rate
7% discount $7.6+ B $9.2 + B $20+ B $43 + B
rate
Note: Rounded to two significant figures. Avoided premature deaths account for over 98% of monetized
benefits here, which are discounted over the SAB-recommended 20-year segmented lag. It was not all
possible to quantify all benefits due to data limitations in this analysis. "B" is the sum of all unquantified
health and welfare benefits.
Table 5-9 is a summary of the monetized benefits associated with applying the
control strategies for the revised and alternative standard levels by four areas: the
Northeast, the Southeast, the West, and California. The monetized benefits differ by area
and for each standard level. For the revised standard levels of 9/35 |~ig/m3, because 23 of
the 52 counties that need emissions reductions are counties in California, a large share of
the benefits is incurred in California (Table 5-9). For California, we were able to identify
approximately 26 percent of the reductions needed. In addition, as the alternative standard
levels become more stringent, more counties in the northeast and southeast need
emissions reductions. As additional controls are applied in those areas, those areas account
for a relatively higher proportion of the benefits. For example, for more alternative
standard levels of 8/35 |~ig/m3, more controls are available to apply in the northeast and
their adjacent counties and the southeast and their adjacent counties.6 The benefits for
those areas are higher than for the west and California.
6 Note that in the northeast and southeast we identified controls and associated emissions reductions from
adjacent counties and used a ppb/ton PM2.5 air quality ratio that was four times less responsive than the
307
-------
Table 5-9 Estimated Monetized Benefits by Area of the Applied Control
Strategies for the Revised and Alternative Primary Standard Levels of
10/35 |ig/m3,10/30 |ig/m3, 9/35 (j,g/m3, and 8/35 |ig/m3 for 2032,
Incremental to Attainment of 12/35 (billions of 2017$)
Benefits
Estimate
Area
10 jig/m3
annual &
35 Jig/m3 24-
hour
10 jig/m3
annual &
30 Jig/m3 24-
hour
9 jig/m3
annual &
35 Jig/m3 24-
hour
8 jig/m3
annual &
35 Jig/m3 24-
hour
Economic value of avoided PIVh.s-related morbidities and premature deaths using PM2.5 mortality estimate
from (Pope III etal., 2019)
3%
discount
rate
7%
discount
rate
Northeast
Southeast
West
California
$2.4+ B
$0.51+ B
$0,059 + B
$15+ B
$2.5+ B
$0.51+ B
$1.3+ B
$17+ B
$18+ B
$5.3+ B
$2.3+ B
$21+ B
$37+ B
$25+ B
$10+ B
$27+ B
Northeast
Southeast
West
California
$2.1+ B
$0.46 + B
$0,053 + B
$13+ B
$2.2 + B
$0.46 + B
$1.2 + B
$15+ B
$16+ B
$4.8 + B
$2.1+ B
$19+ B
$34+ B
$22+ B
$9.3 + B
$24+ B
Economic value of avoided PIVh.s-related morbidities and premature deaths using PM2.5 mortality estimate
from (Wu et al., 2020)
3%
discount
rate
7%
discount
rate
Northeast
Southeast
West
California
$1.2 + B
$0.24+ B
$0.03+ B
$7.0+ B
$1.2+ B
$0.24+ B
$0.64+ B
$8.1 + B
$8.7 + B
$2.5+ B
$1.1+ B
$10+ B
$18+ B
$12+ B
$5.0+ B
$13+ B
Northeast
Southeast
West
California
$1.0+ B
$0.21+ B
$0.027+ B
$6.3 + B
$1.1+ B
$0.21+ B
$0.58 + B
$7.3+ B
$7.8 + B
$2.2 + B
$1.0+ B
$9.1 + B
$16+ B
$10+ B
$4.5 + B
$12+ B
Note: Rounded to two significant figures. Avoided premature deaths account for over 98% of monetized
benefits here, which are discounted over the SAB-recommended 20-year segmented lag. It was not possible to
quantify all benefits due to data limitations in this analysis. "B" is the sum of all unquantified health and
welfare benefits.
5.5 Discussion
The estimated benefits to human health and the environment of the revised and
alternative PM2.5 daily and annual standard levels are substantial. We estimate that by
2032 the emissions reduced by the applied control strategies for the revised and
alternative annual primary standards would decrease the number of PIVh.s-related
premature deaths and illnesses. The emissions reduction strategies will also yield
ratio used when applying in-county emissions reductions (i.e., applied four tons of PM2.5 emissions
reductions from an adjacent county for one ton of emissions reduction needed in a given county); the
benefits of the additional reductions from adjacent counties also contributes to the higher proportion of the
benefits.
308
-------
significant welfare benefits (see Section 5.2.5), though this RIA does not quantify those
endpoints.
Inherent to any complex analysis quantifying the benefits of improved air quality,
such as this one, are multiple sources of uncertainty. Some of these we characterized
through our use of Monte Carlo techniques to sample the statistical error reported in the
epidemiologic and economic studies supplying concentration-response parameters and
economic unit values. Other key sources of uncertainty that affect the size and distribution
of the estimated benefits—including projected atmospheric conditions and source-level
emissions, projected baseline rates of illness and disease, incomes and expected advances
in healthcare—remain unquantified. When evaluated within the context of these
uncertainties, the estimated health impacts and monetized benefits in this RIA provide
important information regarding the public health benefits associated with a revised PM
NAAQS.
There are two important differences worth noting in the design and analytical
objectives of NAAQS RIAs compared to RIAs for implementation rules. First, the NAAQS
RIAs illustrate the potential costs and benefits of a revised air quality standard nationwide
based on an array of emission reduction strategies for different sources. Second, those
costs and benefits are calculated incremental to implementation of existing regulations as
well as additional controls applied to reach the current standards and create the analytical
baseline for the analysis. In short, NAAQS RIAs hypothesize, but do not predict, the
strategies that States may follow to reduce emissions when implementing previous and
revised NAAQS options. Setting a NAAQS does not directly result in costs or benefits, and as
such, NAAQS RIAs illustrate potential benefits and costs; these estimated values cannot be
added, or directly compared, to the costs and benefits of regulations that require specific
emissions control strategies to be implemented.
This latter type of regulatory action—often referred to as an implementation rule—
reduces emissions for specific, well-characterized sources. In general, the EPA is more
confident in the magnitude and location of the emissions reductions for these
implementation rules. As such, emissions reductions achieved under promulgated
implementation rules have been reflected in the baseline of this NAAQS analysis. For this
309
-------
reason, the benefits estimated in this RIA and all other NAAQS RIAs should not be added to
the benefits estimated for implementation rules.
In setting the NAAQS, the EPA accounts for the variability in PM2.5 concentrations
over space and time. While the standard is designed to limit concentrations at the highest
monitor in an area, EPA acknowledges that emissions controls implemented to meet the
standard at the highest monitor will simultaneously result in lower PM2.5 concentrations in
neighboring areas. In fact, the Policy Assessment for the Review of the National Ambient
Air Quality Standards for Particulate Matter (U.S. EPA, 2022c) shows how different
standard levels would affect the distribution of PM2.5 concentrations, as well as people's
risk, across urban areas. For this reason, it is inappropriate to use the NAAQS level as a
bright line for health effects.
The NAAQS are not set at levels that eliminate the risk of air pollution completely.
Instead, the Administrator sets the NAAQS at a level requisite to protect public health with
an adequate margin of safety, taking into consideration effects on susceptible populations
based on the scientific literature. The risk analysis prepared in support of this PM NAAQS
reported risks below these levels, while acknowledging that the confidence in those effect
estimates is higher at levels closer to the standard (U.S. EPA, 2022c). While benefits
occurring below the standard may be somewhat more uncertain than those occurring
above the standard, the EPA considers these to be legitimate components of the total
benefits estimate. Though there are greater uncertainties at lower PM2.5 concentrations,
there is no evidence of a threshold in PIVh.s-related health effects in the epidemiology
literature. Given that the epidemiological literature in most cases has not provided
estimates based on threshold models, there would be additional uncertainties imposed by
assuming thresholds or other non-linear concentration response functions for the purposes
of benefits analysis.
310
-------
5.6 References
AHRQ (2016). HCUPnet, Healthcare Cost and Utilization Project. Agency for Healthcare
Research and Quality. Agency for Healthcare Research and Quality.
AMAP (2021). Impacts of Short-lived Climate Forcers on Arctic Climate, Air Quality, and
Human Health. Summary for Policy-makers. Arctic Monitoring and Assessment
Programme (AMAP), Troms0, Norway. Available at:
https://www.amap.no/documents/download/6760/inline.
Bond, TC, Doherty, SJ, Fahey, DW, Forster, PM, Berntsen, T, DeAngelo, BJ, Flanner, MG,
Ghan, S, Karcher, B, Koch, D, Kinne, S, Kondo, Y, Quinn, PK, Sarofim, MC, Schultz, MG,
Schulz, M, Venkataraman, C, Zhang, H, Zhang, S, Bellouin, N, Guttikunda, SK, Hopke, PK,
Jacobson, MZ, Kaiser, JW, Klimont, Z, Lohmann, U, Schwarz, JP, Shindell, D, Storelvmo, T,
Warren, SG, Zender, CS (2013). Bounding the role of black carbon in the climate system:
A scientific assessment. Journal of Geophysical Research - Atmospheres 118 (11):
5380-5552.
Cameron, TA. (2004). Letter from Trudy Cameron, Chair, Advisory Council on Clean Air
Compliance Analysis, to Administrator Michael Leavitt. Re: Review of the Draft
Analytical Plan for EPA's Second Prospective Analysis - Benefits and Costs of the Clean
Air Act, 1990-2020: An Advisory by the Advisory Council for Clean Air Compliance
Analysis. May 20, 2004. EPA-SAB-COUNCIL-ADV-04-004. Office of the Administrator,
Science Advisory Board U.S. EPA HQ, Washington DC.
Cox, LA. (2019). Letter from Louis Anthony Cox, Jr., Chair, Clean Air Scientific Advisory
Committee, to Administrator Andrew R. Wheeler. Re: CASAC Review of the EPA's Policy
Assessment for the Review of the National Ambient Air Quality Standards for Particulate
Matter (External Review Draft - September 2019). December 16, 2019. EPA-CASAC-20-
001. Office of the Administrator, Science Advisory Board U.S. EPA HQ, Washington DC.
Di, Q, Wang, Y, Zanobetti, A, Wang, Y, Koutrakis, P, Choirat, C, Dominici, F and Schwartz, JD
(2017). Air pollution and mortality in the Medicare population. New England Journal of
Medicine 376(26): 2513-2522.
Hadley, OL, Corrigan, CE, Kirchstetter, TW, Cliff, SS and Ramanathan V (2010). Measured
black carbon deposition on the Sierra Nevada snow pack and implication for snow pack
retreat. Atmos. Chem. Phys. 10(15): 7505-7513.
Hock, R., Rasul, G, Adler, C, Caceres, B, Gruber, S, Hirabayashi, Y, Jackson, M, Kaab, A, Kang,
S, Kutuzov, S, Milner, A, Molau, U, Morin, S, Orlove, B and H. Steltzer, H (2019). High
Mountain Areas. Cambridge University Press, Cambridge, UK and New York, NY, USA,
pp. 131-202. Available at: https://doi.org/10.1017/9781009157964.004.
Hollmann, F, Mulder, T and Kalian, JJW, DC: US Bureau of the Census (2000). Methodology
and assumptions for the population projections of the United States: 1999 to 2100
(Population Division Working Paper No. 38). 338.
311
-------
IPCC (2018). Summary for Policy Makers In: Global Warming of 1.5° C. An IPCC Special
Report on the impacts of global warming of 1.5° C above pre-industrial levels and
related global greenhouse gas emission pathways, in the context of strengthening the
global response to the threat of climate change, sustainable development, and efforts to
eradicate poverty Cambridge University Press, Cambridge, UK and New York, NY, USA,
pp. 3-24, Available at: https://doi.org/10.1017/9781009157940.001.
Kivi, PA and Shogren, JF (2010). Second-order ambiguity in very low probability risks: food
safety valuation. J Agric Resour Econ: 443-456.
Krewski D, Jerrett M, Burnett RT, Ma R, Hughes E, Shi Y, Turner MC, Pope CA III, Thurston G,
Calle EE, Thun MJ (2009). Extended Follow-Up and Spatial Analysis of the American
Cancer Society Study Linking Particulate Air Pollution and Mortality. HEI Research
Report 140. Health Effects Institute, Boston, MA
Kiinzli, N, Kaiser, R, Medina, S, Studnicka, M, Chanel, 0, Filliger, P, Heriy, M, Horakjr, F,
Puybonnieux-Texier, V and Quenel, P (2000). Public-health impact of outdoor and
traffic-related air pollution: a European assessment. The Lancet 356(9232): 795-801.
McGartland, A., Revesz R, Axelrad DA, Dockins C, Sutton P, Woodruff TJ (2017). Estimating
the health benefits of environmental regulations. Science 357(6350): 457-458.
McKenzie Skiles, S., Flanner, M., Cook, JM, Dumont, M, Painter, TH (2018). Radiative forcing
by light-absorbing particles in snow. Nature Climate Change 8, 964-971.
Meredith, M, Sommerkorn, M, Cassotta, S, Derksen, C, Ekaykin, A, Hollowed, A, Kofinas, G,
Mackintosh, A, Melbourne-Thomas, J, Muelbert, MMC, Ottersen, G, Pritchard, H and
Schuur, EAG (2019). Polar Regions. In: IPCC Special Report on the Ocean and
Cryosphere in a Changing Climate. Cambridge University Press, Cambridge, UK and New
York, NY, USA, pp. 203-320. Available at:
https://doi.org/10.1017/9781009157964.005.
Naik, V, Szopa, S, Adhikary, B, Artaxo, P, Berntsen, T, Collins, WD, Fuzzi, S, Gallardo, L,
Kiendler Scharr, A, Klimont, Z, Liao, H, Unger, N and Zanis, P (2021). Short-Lived
Climate Forcers. In Climate Change 2021: The Physical Science Basis. Contribution of
Working Group I to the Sixth Assessment Report of the Intergovernmental Panel on
Climate Change. Cambridge University Press, Cambridge, United Kingdom and New
York, NY, USA, pp. 817-922, Available at:
https://doi.org/10.1017/9781009157896.008
NRC (2002). Estimating the public health benefits of proposed air pollution regulations.
0309086094. National Academies Press.
Pope, C.A., III, M.J. Thun, M.M. Namboodiri, D.W. Dockeiy, J.S. Evans, F.E. Speizer, and C.W.
Heath, Jr. (1995). Particulate air pollution as a predictor of mortality in a prospective
study of US adults. American Journal of Respiratory Critical Care Medicine 151(3): 669-
674.
312
-------
Pope, C.A., III, R.T. Burnett, M.J. Thun, E.E. Calle, D. Krewski, K. Ito, and G.D. Thurston
(2002). Lung cancer, cardiopulmonary mortality, and long-term exposure to fine
particulate air pollution. Journal of the American Medical Association 287:1132-1141.
Pope, CA, Turner, MC, Burnett, R, Jerrett, M, Gapstur, SM, Diver, WR, Krewski, D and Brook,
RD (2015). Relationships between fine particulate air pollution, cardiometabolic
disorders and cardiovascular mortality. Circulation Research 116(1): 108-U258.
Pope III, CA, Lefler, JS, Ezzati, M, Higbee, JD, Marshall, JD, Kim, S-Y, Bechle, M, Gilliat, KS,
Vernon, SE and Robinson, AL (2019). Mortality risk and fine particulate air pollution in
a large, representative cohort of US adults. Environmental health perspectives 127(7):
077007.
Qian Y, Gustafson WI, Leung LR, and Ghan SJ (2009). Effects of soot-induced snow albedo
change on snowpack and hydrological cycle in western United States based on Weather
Research Forecasting chemistry and regional climate simulations. Journal of
Geophysical Research 114 (D02108).
Sacks, JD, Lloyd, JM, Zhu, Y, Anderton, J, Jang, CJ, Hubbell, B and Fann, N (2018). The
Environmental Benefits Mapping and Analysis Program-Community Edition (BenMAP-
CE): A tool to estimate the health and economic benefits of reducing air pollution.
Environmental Modelling Software 104: 118-129.
Sheppard, EL (2022). Letter from Elizabeth A. (Lianne) Sheppard, Chair, Clean Air Scientific
Advisory Committee, to Administrator Regan. Re:CASAC Review of the EPA's
Supplement to the 2019 Integrated Science Assessment for Particulate Matter (External
Review Draft - October 2021). March 18, 2022. EPA-CASAC-22-001. Office of the
Adminstrator, Science Advisory Board Washington, DC.
Sisler, JF (1996). Spatial and Seasonal Patterns and Long-Term Variability of the
Composition of the Haze in the United States: An analysis of data from the IMPROVE
network. CIRA Report, ISSN 0737-5352-32, Colorado State University.
Turner, MC, Jerrett, M, Pope, A, III, Krewski, D, Gapstur, SM, Diver, WR, Beckerman, BS,
Marshall, JD, Su, J, Crouse, DL and Burnett, RT (2016). Long-term ozone exposure and
mortality in a large prospective study. American Journal of Respiratory and Critical Care
Medicine 193(10): 1134-1142.
UNEP (2011). Integrated assessment of black carbon and tropospheric ozone: summary for
decision makers. Available at: https://wedocs.unep.org/20.500.11822/8028.
U.S. BEA (2004). New BEA Economic Areas For 2004. U.S., Bureau of Economic Analysis,
Washington, DC, BEA 04-52. Available at: https://www.bea.gov/news/2004/new-bea-
economic-areas-2004
313
-------
U.S. Census Bureau (2009). Statistical Abstract of the United States 2010. Government
Printing Office. Available at: https://www.census.gov/programs-
surveys/acs/about.html.
U.S. EPA (2008). Integrated Science Assessment For Oxides Of Nitrogen - Health Criteria
(Final Report). EPA/600/R-08/071. Office of Air Quality Planning and Standards.
Research Triangle Park, NC. Available at:
https://cfpub.epa.gov/ncea/isa/recordisplay.cfm?deid=194645.
U.S. EPA (2009). Integrated Science Assessment for Particulate Matter (Final Report). EPA-
600-R-08-139F. National Center for Environmental Assessment - RTP Division,
Research Triangle Park, NC. December. Available at:
http://cfpub.epa.gov/ncea/cfm/recordisplay.cfm?deid=216546.
U.S. EPA (2010). Regulatory Impact Analysis for the Proposed Federal Transport Rule.
Office of Air Quality Planning and Standards, Health and Environmental Impacts
Division. Research Triangle Park, NC. U.S. EPA. June 2010. Available at:
https://www3.epa.gov/ttnecasl/regdata/RIAs/proposaltrria_final.pdf.
U.S. EPA (2011a). Regulatory Impact Analysis: National Emission Standards for Hazardous
Air Pollutants for Industral, Commercial, and Institutional Boilers and Process Heaters.
U.S. Environmental Protection Agency, Office of Air Quality Planning and Standards,
Health and Environmental Impact Division. Research Triangle Park, NC. U.S EPA.
Available at:
https://www3.epa.gov/ttn/ecas/regdata/RIAs/boilersriafinalll0221_psg.pdf.
U.S. EPA (2011b). Regulatory Impact Analysis for the Final Mercury and Air Toxics
Standards. U.S. Environmental Protection Agency, Office of Air Quality Planning and
Standards, Health and Environmental Impact Division. Research Triangle Park, NC. U.S.
EPA. EPA-452/R-11-011. December 2011. Available at:
http://www.epa.gov/ttn/ecas/regdata/RIAs/matsriafinal.pdf.
U.S. EPA (2011c). Regulatory Impact Analysis for the Federal Implementation Plans to
Reduce Interstate Transport of Fine Particulate Matter and Ozone in 27 States;
Correction of SIP Approvals for 22 States. U.S. Environmental Protection Agency, Office
of Air Quality Planning and Standards, Health and Environmental Impact Division.
Research Triangle Park, NC. U.S. EPA. June 2011. Available at:
https://www3.epa.gov/ttn/ecas/docs/ria/transport_ria_final-csapr_2011-06.pdf.
U.S. EPA (2011d). Policy Assessment for the Review of the Particulate Matter National
Ambient Air Quality Standards. EPA-452/D-11-003. Office of Air Quality Planning and
Standards, Health and Environmental Impacts Division. April. Available at:
https://www3.epa.gov/ttn/naaqs/standards/pm/data/20110419pmpafinal.pdf.
314
-------
U.S. EPA (2012a). Regulatory Impact Analysis for the Proposed Revisions to the National
Ambient Air Quality Standards for Particulate Matter. U.S. Environmental Protection
Agency, Office of Air Quality Planning and Standards, Health and Environmental Impact
Division. Research Triangle Park, NC. U.S. EPA. EPA-452/R-12-003. June 2012. Available
at:
https://www3.epa.gov/ttnecasl/regdata/RIAs/PMRIACombinedFile_Bookmarked.pdf.
U.S. EPA (2012b). Regulatory Impact Analysis for the Final Revisions to the National
Ambient Air Quality Standards for Particulate Matter. U.S. Environmental Protection
Agency, Office of Air Quality Planning and Standards, Health and Environmental Impact
Division. Research Triangle Park, NC. U.S. EPA. EPA-452/R-12-005. December 2012.
Available at: https://www3.epa.gov/ttnecasl/regdata/RIAs/finalria.pdf.
U.S. EPA (2014b). Health Risk and Exposure Assessment for Ozone. (Final Report). Office of
Air Quality Planning and Standards. Research Triangle Park, NC. U.S. EPA. EPA-452/R-
14-004a. August 2014. Available at:
https://nepis.epa.gov/Exe/ZyPURL.cgi?Dockey=P 100KBUF.txt.
U.S. EPA (2015a). Regulatory Impact Analysis of the Final Revisions to the National
Ambient Air Quality Standards for Ground-Level Ozone. Office of Air Quality Planning
and Standards, Health and Environmental Impacts Division. Research Triangle Park, NC.
U.S. EPA. EPA-452/R-15-007. September 2015. Available at:
https://www3.epa.gov/ttnecasl/docs/20151001ria.pdf.
U.S. EPA (2015b). Preamble to the Integrated Science Assessments. U.S. Environmental
Protection Agency, Office of Research and Development, National Center for
Environmental Assessment, RTP Division. Research Triangle Park, NC. U.S. EPA.
EPA/600/R-15/067. November 2015. Available at:
https://cfpub.epa.gov/ncea/isa/recordisplay.cfm?deid=310244.
U.S. EPA (2016). Integrated Science Assessment for Oxides of Nitrogen - Health Criteria.
Office of Research and Development, National Center for Environmental Assessment.
Research Triangle Park, NC. U.S. EPA. EPA/600/R-15/068. January 2016. Available at:
https://cfpub.epa.gov/ncea/isa/recordisplay.cfm?deid=310879.
U.S. EPA (2016a). Guidelines for Preparing Economic Analyses National Center for
Environmental Economics. Available at:
https://www.epa.gov/sites/default/files/2017-08/documents/ee-0568-50.pdf
U.S. EPA (2017). Integrated Science Assessment for Sulfur Oxides - Health Criteria. Office of
Research and Development, National Center for Environmental Assessment. Research
Triangle Park, NC. U.S. EPA. EPA/600/R-17/451. December 2017. Available at:
https://cfpub.epa.gov/ncea/isa/recordisplay.cfm?deid=338596.
315
-------
U.S. EPA (2019a). Regulatory Impact Analysis for the Repeal of the Clean Power Plan, and
the Emission Guidelines for Greenhouse Gas Emissions from Existing Electric Utility
Generating Units. U.S. Environmental Protection Agency, Office of Air Quality Planning
and Standards, Health and Environmental Impact Division. Research Triangle Park, NC.
U.S. EPA. EPA-452/R-19-003. June 2019. Available at:
https://www.epa.gov/sites/production/files/2019-
06/documents/utilities_ria_final_cpp_repeal_and_ace_2019-06.pdf.
U.S. EPA (2019b). Integrated Science Assessment (ISA) for Particulate Matter (Final
Report). U.S. Environmental Protection Agency, Office of Research and Development,
Center for Public Health and Environmental Assessment. Research Triangle Park, NC.
U.S. EPA. EPA/600/R-19/188. December 2019. Available at:
https://www.epa.gov/naaqs/particulate-matter-pm-standards-integrated-science-
assessments-current-review.
U.S. EPA (2019c). Policy Assessment for the Review of the National Ambient Air Quality
Standards for Particulate Matter, External Review Draft. U.S. Environmental Protection
Agency, Office of Air Quality Planning and Standards, Health and Environmental
Impacts Division. Research Triangle Park, NC. U.S. EPA. EPA-452/P-19-001. September
2019. Available at: https://www.epa.gov/naaqs/particulate-matter-pm-standards-
policy-assessments-current-review-O.
U.S. EPA (2020a). Integrated Science Assessment for Ozone and Related Photochemical
Oxidants (Final Report). U.S. Environmental Protection Agency. Washington, DC. Office
of Research and Development. EPA/600/R-20/012. April 2020. Available at:
https://www.epa.gov/isa/integrated-science-assessment-isa-ozone-and-related-
photochemical-oxidants.
U.S. EPA (2020b). Integrated Science Assessment (ISA) for Oxides of Nitrogen, Oxides of
Sulfur and Particulate Matter Ecological Criteria. U.S. Environmental Protection Agency.
Washington, DC. Office of Research and Development. EPA/600/R-20/278. Available at:
https://cfpub.epa.gov/ncea/isa/recordisplay.cfm?deid=349473.
U.S. EPA (2020c). Regulatory Impact Analysis for the for the Proposed Revised Cross-State
Air Pollution Rule (CSAPR) Update for the 2008 Ozone NAAQS. U.S. Environmental
Protection Agency, Office of Air Quality Planning and Standards, Health and
Environmental Impact Division. Research Triangle Park, NC. U.S. EPA. EPA-452/P-20-
003. October 2020. Available at: https://www.epa.gov/sites/production/files/2020-
10/documents/revised_csapr_update_ria_proposal.pdf.
U.S. EPA (2020d). Policy Assessment for the Review of the National Ambient Air Quality
Standards for Particulate Matter. U.S. Environmental Protection Agency, Office of Air
Quality Planning and Standards, Health and Environmental Impacts Division. Research
Triangle Park, NC. U.S. EPA. EPA-452/R-20-002. January 2020. Available at:
https://www.epa.gov/naaqs/particulate-matter-pm-standards-policy-assessments-
current-review-O.
316
-------
U.S. EPA (2022a). Supplement to the 2019 Integrated Science Assessment for Particulate
Matter (Final Report). U.S. Environmental Protection Agency, Office of Research and
Development, Center for Public Health and Environmental Assessment. Research
Triangle Park, NC. U.S. EPA. EPA/600/R-22/028. May 2022. Available at:
https://www.epa.gov/isa/integrated-science-assessment-isa-particulate-matter.
U.S. EPA (2022c). Policy Assessment for the Reconsideration of the National Ambient Air
Quality Standards for Particulate Matter (Final Report). U.S. Environmental Protection
Agency, Office of Air Quality Planning and Standards, Heath and Environmental Impacts
Division. Research Triangle Park, NC. U.S. EPA. EPA-452/R-22-004. May 2022.
U.S. EPA (2023a). BenMAP-CE User Manual and Appendices. Office of Air Quality Planning
and Standards. Research Triangle Park, NC. U.S. EPA. Available at: Available at:
https://www.epa.gov/sites/production/files/2015-04/documents/benmap-
ce_user_manual_march_2015.pdf
U.S. EPA (2023b). Technical Support Document (TSD) for the 2022 PM NAAQS
Reconsideration Proposal RIA: Estimating PM2.5- and Ozone-Attributable Health
Benefits U.S. Environmental Protection Agency. Durham, NC. Office of Air Quality
Planning and Standards. January 2023. Available at:
https://www.epa.gov/system/files/documents/2023-01/Estimating%20PM2.5-
%20and%200zone-Attributable%20Health%20Benefits%20TSD_0.pdf.
U.S. EPA-SAB (2020). Letter from Michael Honeycutt Chair, Scientific Advisory, to
Administrator Lisa Jackson. Re: Science Advisory Board (SAB) Consideration of the
Scientific and Technical Basis of EPA's Proposed Rule titled "Increasing Consistency and
Transparency in Considering Benefits and Costs in the Clean Air Act Rulemaking
Process.". September 30, 2020. EPA-SAB-20-012. Office of the Administrator, Science
Advisory Board U.S. EPA HQ, Washington DC.
Wei, Y, Wang, Y, Di, Q, Choirat, C, Wang, Y, Koutrakis, P, Zanobetti, A, Dominici, F and
Schwartz, JD (2019). Short term exposure to fine particulate matter and hospital
admission risks and costs in the Medicare population: time stratified, case crossover
study, bmj 367.
Woodruff, TJ, Darrow, LA and Parker, JD (2008). Air pollution and postneonatal infant
mortality in the United States, 1999-2002. Environmental Health Perspectives 116(1):
110-115.
Woods & Poole (2015). Complete Demographic Database.
https: //www. woodsandpoole.com/our-databases/united-states / cedds /
Wu, X, Braun, D, Schwartz, J, Kioumourtzoglou, M and Dominici, F (2020). Evaluating the
impact of long-term exposure to fine particulate matter on mortality among the elderly.
Science advances 6(29): eaba5692.
317
-------
APPENDIX 5A: BENEFITS OF THE REVISED AND ALTERNATIVE STANDARD LEVELS
Overview
In this Appendix, we estimate the potential health benefits resulting from
identifying controls and emissions reductions to comply fully with the revised and
alternative standard levels, incremental to a 2032 baseline in which the nation fully attains
the current primary PM2.5 standards (i.e., an annual standard of 12 |ig/m3 and a 24-hour
standard of 35 [ig/m3). In contrast, in the main analysis in Chapter 5, we present the
national health impacts and monetized benefits resulting only from the applied control
strategies identified in Chapter 3 for the revised and alternative PM2.5 standard levels in
2032. After applying the control strategies for the main analysis, we estimated that PM2.5
emissions reductions would still be needed in certain areas to meet the revised standard of
9/35 and the 10/35,10/30, and 8/35 alternative standard levels. Additional information
on estimating the emission reductions needed to meet the revised and alternative
standards is available in section 2A.3.4.2 of Appendix 2A. Also, additional information on
the emissions reductions still needed is available in Chapter 3, Section 3.2.4. Lastly, Chapter
2, Section 2.4 and Chapter 3, Section 3.2.5 discuss the remaining air quality challenges for
areas in the northeast and southeast, as well as in the west and California that may still
need emissions reductions. These challenges limit our ability to characterize how standard
levels might be met given highly local influences that require more specific information
beyond what is available for this type of national analysis. In this Appendix, we assume the
remaining emissions reductions are identified to meet the revised, less, and more stringent
alternative standard levels, and we present the resulting health and monetized benefits
below. To the extent that the additional PM2.5 emissions reductions are not achieved, the
health benefits reported below may be overestimated.
For this appendix, the annual-mean PM2.5 concentration fields where existing,
revised, and alternative NAAQS standard levels are just met were developed to estimate the
emission changes resulting from fully meeting the revised, less, and more stringent
alternative standard levels. Using the methods described in Chapter 5 of this RIA and the
Estimating PM2.5- and Ozone-Attributable Health Benefits TSD (U.S. EPA, 2023b), we
estimate health benefits from achieving the revised, less, and more stringent alternative
318
-------
standard levels occurring as an increment to a 12/35 baseline. These benefits reflect the
value of the avoided PIVh.s-attributable deaths and the value of avoided morbidity impacts,
including, for example, hospital admissions and emergency department visits for
cardiovascular and respiratory health issues.
5A.1 Benefits of the Revised, Less, and More Stringent Alternative Standard Levels of
Primary PM2.5 Standards
Applying the impact and valuation functions described in Chapter 5 and the TSD to
the projected changes in PM2.5 yields estimates of the changes in physical damages (e.g.,
premature mortalities, cases of hospital admissions and emergency department visits) and
the associated monetary values for those changes. Not all known PM health effects could be
quantified or monetized. Tables 5A-1 through 5A-5 present the benefits results for the
revised, less, and more stringent alternative annual primary PM2.5 standard levels. Table
5A-1 presents the estimated avoided incidences of PM-related illnesses and premature
mortality for achieving the revised and alternative standard levels in 2032. Tables 5A-2 and
5A-3 present the monetized valuation benefits of the avoided morbidity and premature
mortality (at a 3% and 7% discount rate respectively) of the health outcomes in Table 5A-1
for the revised and alternative standard levels in 2032.
Tables 5A-4 and 5A-5 present a summary of the monetized benefits nationally and
by area of achieving the revised and alternative standard levels. The monetized benefits in
Table 5A-5 are presented in four areas: the Northeast, the Southeast, the West, and
California. For Tables 5A-4 and 5A-5, the monetized value of unquantified effects is
represented by adding an unknown "B" to the aggregate total. The estimate of total
monetized health benefits is thus equal to the subset of monetized PM-related health
benefits plus B, the sum of the non-monetized health and welfare benefits; this B
represents both uncertainty and a bias in this analysis, as it reflects those benefits
categories that we are unable to quantify in this analysis.
319
-------
Table 5A-1 Estimated Avoided PM-Related Premature Mortalities and Illnesses of
Meeting the Revised and Alternative Primary PM2.5 Standard Levels for
2032 (95% Confidence Interval)
Avoided Mortality
10/35
10/30
9/35
8/35
Pope et al. (adult mortality
ages 18-99 years)
3,100
(2,200 to 4,000)
3,800
(2,700 to 4,900)
7,400
(5,300 to 9,400)
15,000
(11,000 to 20,000)
Wu et al. (adult mortality
ages 65-99 years)
1,500
(1,300 to 1,700)
1,800
(1,600 to 2,000)
3,500
(3,100 to 3,900)
7,400
(6,500 to 8,200)
Woodruff et al. (infant
mortality)
3.3
(-2.1 to 8.4)
3.9
(-2.5 to 9.9)
8.2
(-5.2 to 21)
18
(-11 to 45)
Avoided Morbidity
10/35
10/30
9/35
8/35
Hospital admissions—
cardiovascular (age > 18)
260
(190 to 320)
310
(220 to 390)
580
(420 to 730)
1,200
(840 to 1,500)
Hospital admissions—
respiratory
170
(60 to 280)
200
(69 to 330)
400
(140 to 650)
820
(290 to 1,300)
ED visits-cardiovascular
490
(-190 to 1,100)
580
(-220 to 1,300)
1,100
(-440 to 2,600)
2,400
(-910 to 5,500)
ED visits—respiratory
950
(190 to 2,000)
1,100
(220 to 2,300)
2,200
(440 to 4,700)
4,700
(920 to 9,800)
Acute Myocardial
Infarction
56
(33 to 79)
67
(39 to 94)
130
(73 to 180)
250
(150 to 350)
Cardiac arrest
27
(-11 to 61)
33
(-13 to 74)
62
(-25 to 140)
130
(-52 to 290)
Hospital admissions-
Alzheimer's Disease
600
(460 to 730)
700
(530 to 850)
1,400
(1,100 to 1,700)
2,900
(2,300 to 3,600)
Hospital admissions-
Parkinson's Disease
84
(44 to 120)
100
(53 to 150)
200
(110 to 290)
420
(220 to 600)
Stroke
100
(26 to 170)
120
(32 to 210)
230
(61 to 400)
480
(120 to 820)
Lung cancer
120
(37 to 200)
140
(45 to 240)
270
(85 to 450)
560
(170 to 900)
Hay Fever/Rhinitis
29,000
(7,100 to 50,000)
35,000
(8,500 to 60,000)
67,000
(16,000 to
110,000)
140,000
(33,000 to 230,000)
Asthma Onset
4,400
(4,300 to 4,600)
5,300
(5,100 to 5,500)
10,000
(9,700 to 11,000)
20,000
(20,000 to 21,000)
Asthma symptoms -
Albuterol use
620,000
(-300,000 to
1,500,000)
740,000
(-360,000 to
1,800,000)
1,400,000
(-700,000 to
3,500,000)
2,900,000
(-1,400,000 to
7,000,000)
Lost work days
220,000
(190,000 to
250,000)
260,000
(220,000 to
300,000)
500,000
(430,000 to
580,000)
1,000,000
(870,000 to
1,200,000)
Minor restricted-activity
days
1,300,000
(1,000,000 to
1,500,000)
1,500,000
(1,200,000 to
1,800,000)
3,000,000
(2,400,000 to
3,500,000)
6,000,000
(4,900,000 to
7,100,000)
Note: Values rounded to two significant figures.
320
-------
Table 5A-2 Monetized Avoided PM-Related Premature Mortalities and Illnesses of
Meeting the Revised and Alternative Primary PM2.5 Standard Levels for
2032 (Millions of 2017$, 3% discount rate; 95% Confidence Interval)
Avoided Mortality
10/35
10/30
9/35
8/35
Pope et al. (adult mortality
ages 18-99 years)
32,000
(2,900 and 86,000)
39,000
(3,600 and
110,000)
76,000
(6,900 and
200,000)
160,000
(14,000 and
430,000)
Wu et al. (adult mortality
ages 65-99 years)
15,000
(1,400 and 40,000)
19,000
(1,700 and 50,000)
36,000
(3,300 and
96,000)
75,000
(7,000 and 200,000)
Woodruff et al. (infant
mortality)
37
(-21 and 150)
44
(-25 and 170)
92
(-52 and 370)
200
(-110 and 790)
Avoided Morbidity
10/35
10/30
9/35
8/35
Hospital admissions—
cardiovascular (age > 18)
4.2
(3.1 and 5.3)
5
(3.6 and 6.4)
9.5
(6.9 and 12)
19
(14 and 24}
Hospital admissions—
respiratory
2.9
(0.66 and 5.1)
3.4
(0.76 and 5.9)
6.6
(1.5 and 12)
13
(3.1 and 2 3)
ED visits-cardiovascular
0.6
(-0.23 and 1.4)
0.71
(-0.27 and 1.7)
1.4
(-0.54 and 3.3)
2.9
(-1.1 and 6.8)
ED visits—respiratory
0.88
(0.17 and 1.8)
1
(0.2 and 2.2)
2.1
(0.41 and 4.3)
4.4
(0.86 and 9.1)
Acute Myocardial
Infarction
2.9
(1.7 and 4.1)
3.5
(2 and 4.9_)
6.6
(3.8 and 9.2)
13
(7.6 and 18)
Cardiac arrest
1
(-0.42 and 2.3)
1.2
(-0.51 and 2.8)
2.4
(-0.97 and 5.3)
4.8
(-2 and 11)
Hospital admissions—
Alzheimer's Disease
7.7
(5.9 and 9.4)
8.9
(6.8 and 11)
18
(14 and 22)
38
(29 and 46)
Hospital admissions-
Parkinson's Disease
1.1
(0.6 and 1.7)
1.4
(0.73 and 2)
2.8
(1.4 and 4)
5.8
(3 and 8.2)
Stroke
3.7
(0.95 and 6.2)
4.5
(1.2 and 7.6)
8.4
(2.2 and 14)
17
(4.5 and 30)
Lung cancer
3.3
(1 and 5.4}
4
(1.2 and 6.6)
7.5
(2.3 and 12)
15
(4.8 and 25)
Hay Fever/Rhinitis
19
(4.6 and 32)
22
(5.4 and 38)
43
(10 and 73)
86
(21 and 150)
Asthma Onset
210
(200 and 220)
250
(230 and 260)
480
(450 and 510)
960
(900 and 1000)
Asthma symptoms -
Albuterol use
0.23
(-0.11 and 0.56)
0.27
(-0.13 and 0.66)
0.53
(-0.26 and 1.3)
1.1
(-0.52 and 2.6)
Lost work days
40
(33 and 46)
47
(40 and 54)
91
(77 and 100)
190
(160 and 210)
Minor restricted-activity
days
100
(53 and 150)
120
(63 and 180)
230
(120 and 350)
470
(250 and 720)
Note: Values rounded to two significant figures.
321
-------
Table 5A-3 Monetized Avoided PM-Related Premature Mortalities and Illnesses of
Meeting the Revised and Alternative Primary PM2.5 Standard Levels
for 2032 (Millions of 2017$, 7% discount rate; 95% Confidence
Interval)
Avoided Mortality
10/35
10/30
9/35
8/35
Pope et al. (adult mortality
ages 18-99 years)
29,000
(2,600 and 78,000)
35,000
(3,200 and 95,000)
68,000
(6,200 and
180,000)
140,000
(13,000 and
380,000)
Wu et al. (adult mortality
ages 65-99 years)
14,000
(1,300 and 36,000)
17,000
(1,600 and 45,000)
33,000
(3,000 and
86,000)
68,000
(6,300 and 180,000)
Woodruff et al. (infant
mortality)
37
(-21 and 150)
44
(-25 and 170)
92
(-52 and 370)
200
(-110 and 790)
Avoided Morbidity
10/35
10/30
9/35
8/35
Hospital admissions—
cardiovascular (age > 18)
4.2
(3.1 and 5.3)
5
(3.6 and 6.4)
9.5
(6.9 and 12)
19
(14 and 24}
Hospital admissions—
respiratory
2.9
(0.66 and 5.1)
3.4
(0.76 and 5.9)
6.6
(1.5 and 12)
13
(3.1 and 2 3)
ED visits-cardiovascular
0.6
(-0.23 and 1.4)
0.71
(-0.27 and 1.7)
1.4
(-0.54 and 3.3)
2.9
(-1.1 and 6.8)
ED visits—respiratory
0.88
(0.17 and 1.8)
1
(0.2 and 2.2)
2.1
(0.41 and 4.3)
4.4
(0.86 and 9.1)
Acute Myocardial
Infarction
2.9
(1.7 and 4)
3.4
(2 and 4.8_)
6.4
(3.7 and 9)
13
(7.4 and 18)
Cardiac arrest
1
(-0.42 and 2.3)
1.2
(-0.5 and 2.8)
2.3
(-0.96 and 5.3)
4.8
(-2 and 11)
Hospital admissions-
Alzheimer's Disease
7.7
(5.9 and 9.4)
8.9
(6.8 and 11)
18
(14 and 22)
38
(29 and 46)
Hospital admissions-
Parkinson's Disease
1.1
(0.6 and 1.7)
1.4
(0.73 and 2)
2.8
(1.4 and 4)
5.8
(3 and 8.2)
Stroke
3.7
(0.95 and 6.2)
4.5
(1.2 and 7.6)
8.4
(2.2 and 14)
17
(4.5 and 30)
Lung cancer
2.5
(0.76 and 4)
3
(0.93 and 4.9)
5.6
(1.7 and 9.2)
11
(3.6 and 19)
Hay Fever/Rhinitis
19
(4.6 and 32)
22
(5.4 and 38)
43
(10 and 73)
86
(21 and 150)
Asthma Onset
130
(120 and 140)
150
(140 and 160)
300
(280 and 310)
600
(560 and 630)
Asthma symptoms -
Albuterol use
0.23
(-0.11 and 0.56)
0.27
(-0.13 and 0.66)
0.53
(-0.26 and 1.3)
1.1
(-0.52 and 2.6)
Lost work days
40
(33 and 46)
47
(40 and 54)
91
(77 and 100)
190
(160 and 210)
Minor restricted-activity
days
100
(53 and 150)
120
(63 and 180)
230
(120 and 350)
470
(250 and 720)
Note: Values rounded to two significant figures.
322
-------
Table 5A-4 Total Estimated Monetized Benefits of Meeting the Revised and
Alternative Primary Standard Levels in 2032, Incremental to
Attainment of 12/35 (billions of 2017$)
ft r l*g/m3 annual & 10 jig/m3 annual & 9 jig/m3 annual & 8 ng/m3 annual &
Benefits Estimate 35 ng/m3 24-hour 30 iig/m3 24-hour 35 ng/m3 24-hour 35 ng/m3 24-hour
Economic value of avoided PIVh.s-related morbidities and premature deaths using PM2.5 mortality
estimate from Pope (2019)
3% discount ^ + g $40 + B $77 + B $160 + B
rate
7% discount $29+ B $36 + B $69 + B $140+B
rate
Economic value of avoided PIVh.s-related morbidities and premature deaths using PM2.5 mortality
estimate from Wu et al. (2020)
3% discount $16 +B $19 + B $37 + B $77 + B
rate
7% discount $14+ B $17 + B $33 + B $70 + B
rate
Note: Rounded to two significant figures. Avoided premature deaths account for over 98% of monetized
benefits here, which are discounted over the SAB-recommended 20-year segmented lag. It was not possible to
quantify all benefits due to data limitations in this analysis. "B" is the sum of all unquantified health and
welfare benefits.
323
-------
Table 5A-5
Total Estimated Monetized Benefits by Area of Meeting the Revised
and Alternative Primary Standard Levels in 2032, Incremental to
Attainment of 12/35 (billions of 2017$)
Benefits
Estimate
Area
10 jig/m3
annual &
35 Jig/m3 24-
hour
10 jig/m3
annual &
30 Jig/m3 24-
hour
9 jig/m3
annual &
35 Jig/m3 24-
hour
8 jig/m3
annual &
35 Jig/m3 24-
hour
Economic value of avoided PIVh.s-related morbidities and premature deaths using PM2.5 mortality estimate
from Pope (2019)
3%
discount
rate
Northeast
Southeast
West
California
$2.4+ B
$0.51+ B
$0.11 + B
$29+ B
$2.5+ B
$0.51+B
$2.1+ B
$35+ B
$15+ B
$6.2 + B
$2.6+ B
$53+ B
$39+ B
$27+ B
$12+ B
$82+ B
Northeast $2.1 + B $2.2 + B $14 + B $35 + B
Southeast $0.46 + B $0.46 + B $5.6 + B $24 + B
West $0.1 + B $1.9 + B $2.3 + B $11 + B
California $26 + B $31 + B $47 + B $74+ B
Economic value of avoided PIVh.s-related morbidities and premature deaths using PM2.5 mortality estimate
from Wu et al. (2020)
Northeast $1.2 + B $1.2 + B $7.5 + B $19 + B
Southeast $0.24 + B $0.24 + B $2.9 + B $13 + B
West $0.059+ B $1.1 + B $1.3 + B $5.8 + B
California $14 + B $17 + B $25+ B $40 + B
7%
discount
rate
30/0
discount
rate
7%
discount
rate
Northeast
Southeast
West
California
$1.0+ B
$0.21+ B
$0.053+ B
$13+ B
$1.1+ B
$0.21+ B
$0.95+ B
$15+ B
$6.8 + B
$2.6+ B
$1.1+ B
$23+ B
$17+ B
$11+ B
$5.2 + B
$36+ B
Note: Rounded to two significant figures. Avoided premature deaths account for over 98% of monetized
benefits here, which are discounted over the SAB-recommended 20-year segmented lag. It was not all
possible to quantify all benefits due to data limitations in this analysis. "B" is the sum of all unquantified
health and welfare benefits.
324
-------
5A.2 References
Pope III, CA, Lefler, JS, Ezzati, M, Higbee, JD, Marshall, JD, Kim, S-Y, Bechle, M, Gilliat, KS,
Vernon, SE and Robinson, AL (2019). Mortality risk and fine particulate air pollution in
a large, representative cohort of US adults. Environmental health perspectives 127(7):
077007.
U.S. EPA (2023b). Technical Support Document (TSD) for the 2022 PM NAAQS
Reconsideration Proposal RIA: Estimating PM2.5- and Ozone-Attributable Health
Benefits U.S. Environmental Protection Agency. Durham, NC. Office of Air Quality
Planning and Standards. January 2023. Available at:
https://www.epa.gov/system/files/documents/2023-01/Estimating%20PM2.5-
%20and%200zone-Attributable%20Health%20Benefits%20TSD_0.pdf.
Wu, X, Braun, D, Schwartz, J, Kioumourtzoglou, M and Dominici, F (2020). Evaluating the
impact of long-term exposure to fine particulate matter on mortality among the elderly.
Science advances 6(29): eaba5692.
Woodruff, TJ, Darrow, LA and Parker, JD (2008). Air pollution and postneonatal infant
mortality in the United States, 1999-2002. Environmental Health Perspectives 116(1):
110-115.
325
-------
CHAPTER 6: ENVIRONMENTAL JUSTICE
6.1 Analyzing EJ Impacts in This Final Action
In addition to the benefits assessment (Chapter 5), the EPA considers potential EJ
concerns of this rulemaking. An EJ concern is defined as the actual or potential lack of fair
treatment or meaningful involvement on the basis of income, race, color, national origin,
Tribal affiliation, or disability in the development, implementation and enforcement of
environmental laws, regulations and policies. For analytic purposes, this concept refers
more specifically to disproportionate and adverse impacts that may exist prior to or be
created by the proposed regulatory action. (U.S. EPA, 2015). Although EJ concerns for each
rulemaking are unique and should be considered on a case-by-case basis, the EPA's EJ
Technical Guidance (U.S. EPA, 2015) states that "[t]he analysis of potential EJ concerns for
regulatory actions should address three questions:
1. Are there potential EJ concerns associated with environmental stressors affected by
the regulatory action for population groups of concern in the baseline?
2. Are there potential EJ concerns associated with environmental stressors affected by
the regulatory action for population groups of concern for the regulatory option(s)
under consideration?
3. For the regulatory option(s) under consideration, are potential EJ concerns created
[, exacerbated,] or mitigated compared to the baseline?"
To address these questions, the EPA developed an analytical approach that
considers the purpose and specifics of this rulemaking, as well as the nature of known and
potential exposures and health impacts. The purpose of this Regulatory Impact Analysis
(RIA) is to provide estimates of the potential costs and benefits of the illustrative national
control strategies in 2032 for the revised and alternative standard levels analyzed. As
revised and alternative standard levels evaluated in the RIA are more stringent than the
current standards, when reducing emissions to reach lower standard levels, some areas
above or near the current standards are expected to experience greater air quality
improvements, and thus health improvements, than other areas already at or below lower
alternative standard levels. As differences in both exposure and susceptibility (i.e., intrinsic
326
-------
individual risk factors) contribute to environmental impacts, the analytical approach used
here first determines whether exposure (Section 6.3) and health effect (Section 6.4)
disparities exist under the baseline scenario. The approach then evaluates if and how
disparities are impacted when illustrative emissions control strategies are analyzed. Both
the exposure and health effects analyses were developed using available scientific evidence
from the current PM NAAQS reconsideration, for the future year 2032, and are associated
with various uncertainties. Consistent with the methods the EPA uses to fully characterize
the benefits of a regulatory action, these EJ analyses evaluate the full set of exposure and
health outcome distributions resulting from this action at the national scale.
Since NAAQS RIAs are national-level assessments and air quality issues are complex
and local in nature, the RIA presents costs and benefits of PM2.5 emission reductions
associated with illustrative control strategies. Correspondingly, the main EJ analyses in this
chapter also evaluate implications of air quality surfaces associated with the illustrative
emission control strategies for current (i.e., baseline), revised, and alternative standard
levels. However, the illustrative control strategies do not result in all counties identifying
emissions reductions needed to meet the revised or alternative standard levels (Chapter 3).
As such, the appendix to this chapter provides EJ implications of air quality scenarios
associated with meeting the standard levels (labelled in some Section 6.6 figures as
"Standards") and also repeats results associated with the illustrative emissions control
strategies (labelled in some Section 6.6 figures as "Controls"), allowing for direct
comparison across the two air quality adjustment methods.
As only some areas of the U.S. are projected to exceed alternative standard levels, in
the proposal RIA the EPA conducted a case study examining the subset of areas in which air
quality is projected to change after the application of controls. In this final RIA, instead of
the case study analysis, we illustrate how air quality improvements in the areas with the
highest starting concentrations might be distributed by enlarging the portions of the
distributional exposure figures that cover relatively high PM2.5 exposures.1 This permits
visualization of the sometimes-nuanced exposure disparities.
1 Input data (e.g., air quality surfaces, configuration files, and command line scripts) used to prepare the EJ
analysis described in this chapter are in the docket and available upon request with the Docket Office.
327
-------
The EJ exposure assessment portion of the analysis focuses on associating ambient
PM2.5 concentrations with various demographic variables. Because this type of analysis
requires less a priori information, we were able to include a broad array of demographic
characteristics (e.g., race/ethnicity, poverty status, educational attainment, etc.). In
contrast, estimating health outcomes modified by demographic population requires
additional scientific information, which constrained the scope of the second portion of the
assessment. We focused the EJ health effects analysis on populations and health outcomes
with the strongest scientific support (U.S. EPA, 2019, 2020, 2022a). However, the EJ health
effects analysis does not include information about differences in other factors that could
affect the likelihood of adverse impacts (e.g., access to health care, body mass index, genetic
susceptibilities, etc.) across groups, due to limitations in the underlying data.2
Complex analyses using estimated parameters and inputs from numerous models
are likely to include multiple sources of uncertainty. As this analysis is based on the same
PM2.5 spatial fields as the benefits assessment (Appendix 2A), it is subject to similar types
of uncertainty (Chapter 5, Section 5.3). A particularly germane limitation is the illustrative
nature of the emission reductions in NAAQS RIAs; as a result, the EJ analyses in this chapter
illustrate the estimated EJ impacts of the illustrative control strategies and may not reflect
state-level implementation decisions. Relatedly, while proximity analyses can sometimes
provide limited EJ information regarding the demographics of populations living near
emissions sources, in this case state-level implementation decisions are unknown.
Therefore, proximity analyses of populations living near individual sources that could
potentially install controls would be highly uncertain and were not conducted in this EJ
assessment. However, the EJ exposure and health analyses included in this chapter provide
more relevant and high-confidence information than a proximity analysis, since these
analyses relate actual PM2.5 concentrations (not just emissions) to various demographic
populations.
Both the EJ exposure and health effects analyses are subject to uncertainties related
to input parameters and assumptions. For example, both analyses focus on annual PM2.5
2 We do not ascribe differential health effects to be caused by race or ethnicity. Instead, race and ethnicity
likely serve as proxies for a variety of environmental and social stressors.
328
-------
concentrations and do not evaluate whether concentrations experienced by different
groups persist across the distribution of daily PM2.5 exposures. Additionally, air quality
simulation input information is at a 12 km grid resolution, population information is either
at the Census tract- or county-level, and baseline mortality rates are mostly at the county-
level. The resolution of the input data may potentially mask impacts at geographic scales
more highly resolved than the input information. The EJ health effects analysis is also
subject to additional uncertainties related to concentration-response relationships and
baseline incidence data.
As with all EJ analyses, data limitations make it quite possible that there exist
additional disparities unidentified in this analysis. This is especially relevant for potential
EJ characteristics and more granular spatial resolutions that were not evaluated. For
example, results are provided here at national-, regional- and tract-levels, potentially
masking block group- or block-level EJ impacts. Additional uncertainties are briefly
discussed in the summary of this analysis (Section 6.5).
6.2 Introduction
Executive Order 12898 directs the EPA to "achiev[e] environmental justice (EJ) by
identifying and addressing, as appropriate, disproportionately high and adverse human
health or environmental effects" (59 FR 7629, February 16,1994), termed
disproportionate impacts in this chapter. Additionally, Executive Order 13985 was signed
to advance racial equity and support underserved communities through Federal
government actions (86 FR 7009, January 20, 2021). Recently, Executive Order 14096 (88
FR 25251, April 26, 2023) strengthens the directives for achieving environmental justice
that are set out in Executive Order 12898.
Executive Order 14096 defines EJ as the just treatment and meaningful involvement
of all people, regardless of income, race, color, national origin, Tribal affiliation, or
disability, in agency decision-making and other Federal activities that affect human health
and the environment. The EPA further defines the term fair treatment to mean that "no
group of people should bear a disproportionate burden of environmental harms and risks,
including those resulting from the negative environmental consequences of industrial,
329
-------
governmental, and commercial operations or programs and policies".3 Meaningful
involvement means that: (1) potentially affected populations have an appropriate
opportunity to participate in decisions about an activity that will affect their environment
and/or health; (2) the public's contribution can influence the regulatory Agency's decision;
(3) the concerns of all participants involved will be considered in the decision-making
process; and (4) the rule-writers and decision-makers seek out and facilitate the
involvement of those potentially affected.
The term "disproportionate impacts" refers to differences in impacts or risks that
are extensive enough that they may merit Agency action.4 In general, the determination of
whether a disproportionate impact exists is ultimately a policy judgment which, while
informed by analysis, is the responsibility of the decision-maker. The terms "difference" or
"differential" indicate an analytically discernible distinction in impacts or risks across
population groups. It is the role of the analyst to assess and present differences in
anticipated impacts across population groups of concern for both the baseline and final
regulatory options, using the best available information (both quantitative and qualitative)
to inform the decision-maker and the public.
A regulatory action may involve potential EJ concerns if it could: (1) create new
disproportionate impacts on populations or communities of concern based on income, race,
color, national origin, Tribal affiliation, or disability; (2) exacerbate existing
disproportionate impacts on populations or communities of concern; or (3) present
opportunities to address existing disproportionate impacts on populations or communities
of concern through the action under development.
Executive Order 14094 (88 FR 21879, April 11, 2023) directs Federal agencies to
recognize distributive impacts and equity in regulatory analysis, to the extent permitted by
law, as practicable and appropriate. For purposes of analyzing regulatory impacts, the EPA
relies upon its June 2016 "Technical Guidance for Assessing Environmental Justice in
3 See, e.g., "Environmental Justice." Epa.gov, U.S. Environmental Protection Agency, 4 Mar. 2021,
https: //www.epa.gov/ environmentaljustice.
4 See https://www.epa.gov/environmentaljustice/technical-guidance-assessing-environmental-justice-
regulatory-analysis.
330
-------
Regulatory Analysis,"5 which provides recommendations that encourage analysts to
conduct the highest quality analysis feasible, recognizing that data limitations, time,
resource constraints, and analytical challenges will vary by media and circumstance.
A reasonable starting point for assessing the need for a more detailed EJ analysis is
to review the available evidence from the published literature and from community input
on what factors may make population groups of concern more vulnerable to adverse effects
(e.g., underlying risk factors that may contribute to higher exposures and/or impacts). It is
also important to evaluate the data and methods available for conducting an EJ analysis. EJ
analyses can be grouped into two types, both of which are informative, but not always
feasible for a given rulemaking:
1. Baseline: Describes the current (pre-control) distribution of exposures and risk
under control strategies associated with the current standard, identifying
potential disparities.
2. Policy: Describes the distribution of exposures and risk after the regulatory
option(s) have been applied (post-control), identifying how potential disparities
change in response to the rulemaking.
EPA's 2016 Technical Guidance does not prescribe or recommend a specific
approach or methodology for conducting EJ analyses, though a key consideration is
consistency with the assumptions underlying other parts of the regulatory analysis when
evaluating the baseline and regulatory options.
6.3 EJ Analysis of Exposures Under Current, Revised, and Alternative Standard
Levels
This EJ PM2.5 exposure6 analysis aims to evaluate the potential for EJ concerns
related to PM2.5 exposures7 among potentially vulnerable populations8 from three
5 See https://www.epa.gov/environmentaljustice/technical-guidance-assessing-environmental-justice-
regulatory-analysis.
6 The term exposure is used here to describe estimated PM2.5 concentrations and not individual dosage.
7 Air quality surfaces used to estimate exposures are based on 12 x 12 km grids. Additional information on air
quality modeling can be found in Chapter 2.
8 Race, ethnicity, sex, and age population input information is at the tract level, whereas poverty status and
educational attainment population input information is at the county level.
331
-------
perspectives, which correspond to the three EJ questions listed in Section 6.1. Specifically,
the following questions are addressed:
1) Are there disparate PM2.5 exposures under baseline/current PM NAAQS standard
levels (question 1)?
2) Are there disparate PM2.5 health effects under illustrative alternative PM NAAQS
standard levels (question 2)?
3) Are PM2.5 exposure disparities created, exacerbated, or mitigated under illustrative
alternative PM NAAQS standard levels as compared to the baseline (question 3)?
Population variables considered in this EJ exposure assessment are listed in Table
6-1).9 EPA has expanded the populations evaluated in this exposure EJ assessment as
compared to those included in the proposal RIA EJ assessment to evaluate potential
exposure disparities more comprehensively, regardless of any pre-existing biases. Several
populations added to this final action have been included in other rulemaking RIAs
evaluating PM2.5 emission changes (e.g., employment status, health insurance status, and
linguistic isolation). Additionally, this RIA assesses exposure in communities with a legacy
of discriminatory land use designations and siting decisions (i.e., historically redlined
areas, using the Home Owners' Loan Corporation (HOLC) gradings A-D, with grade D areas
being defined as "redlined"). EPA believes all population groups added to this exposure
assessment provide additional insight into community-level vulnerability, as all of these
factors can reasonably be understood to lead to increased susceptibility and vulnerability
(U.S. EPA, 2023). Additional information on the populations can be found in Section 6.6.1.
There is substantial research demonstrating that structural or historic racism can
have lasting effects. For example, examiners for the Home Owners' Loan Corporation
(HOLC) rated neighborhoods in the 1930s and 1940s based on a long list of criteria,
including the age and condition of housing, access to highways and other forms of
transportation, proximity to parks and polluting industries, and residents' economic class,
employment status, and ethnic and racial composition. Neighborhoods that received a D
grade were designated hazardous or undesirable (so called "redlined'). On this basis,
9 Information on the input population data used in the exposure EJ analysis is available in Section 6.6.1.
332
-------
potential homeowners were denied access to credit or lent money at higher rates than
residents of other neighborhoods. Redlining was in effect across 239 cities and, although
illegal now for many decades, has had lasting effects on investments in these
neighborhoods, where greater proportions of low income and people of color still reside.
Residents of neighborhoods that were redlines also tend to have poorer health outcomes
(Mitchell et al., 2018, Swope et al., 2022, Lee etal., 2022).
Table 6-1 Populations Included in the PM2.5 Exposure Analysis
Population
Groups
Ethnicity
Hispanic; Non-Hispanic
Race
Asian; American Indian; Black; White
Educational Attainment
High school degree or more; No high school degree
Employment Status
Employed; Unemployed; Not in the labor force
Health Insurance Status
Insured; Uninsured
Linguistic Isolation
English "well or better"; English < "well"
Poverty Status
Above the poverty line; Below the poverty line
Redlined Areas
HOLC Grades A-C; HOLC Grade D; Not graded by HOLC
Age
Children (0-17); Adults (18-64); Older Adults (65-99)
Sex
Female; Male
Results presented in the main chapter are associated with identified emissions
controls, whereas the EJ appendix, Section 6.6, includes results for air quality scenarios
associated with identified emissions controls ("Controls") and with fully meeting the
current and alternate standard levels ("Standards"). Additional air quality information
regarding identified controls, as well as areas where air quality has been adjusted, is
available in Chapters 2 and 3.
6.3.1 National
We begin by considering the first two questions from EPA's EJ Technical Guidance
(i.e., are there potential EJ concerns 1) in the baseline, and 2) for the regulatory option(s)
under consideration) with respect to PM2.5 exposures across the contiguous U.S., which we
term "national." We also consider the extent to which exposures change for each
demographic population, to compare improvements in air quality among populations. We
then address the third question from EPA's EJ Technical Guidance, how disparities
333
-------
observed between demographic groups in the baseline scenario (12/35) are impacted (e.g.,
exacerbated/mitigated/unchanged) under alternative standard levels.
6.3.1.1 Absolute Exposures Under Current and Alternative Standard Levels and
Exposure Changes When Moving from Current to Revised and Alternative
Standard Levels
As NAAQS are national rules, we begin by evaluating annual average PM2.5
concentrations in absolute terms projected to be experienced by various demographic
groups that may be of EJ concern, averaged across the contiguous U.S. Figure 6-1 shows the
national average annual PM2.5 exposure concentration burdens of various population
groups under current standard (i.e., baseline) and alternative, more stringent standard
scenarios as a heat map. Each scenario is a combination of an annual standard, listed first,
and a 24-hr standard, listed after the slash, both presented in micrograms per cubic meter
(|ig/m3). Higher estimated annual total PM2.5 concentrations are shown in darker shades of
blue for convenience. Populations with potential EJ concerns can be compared to the
reference/overall population and/or other populations (i.e., White, Non-Hispanic, above
the poverty line, more educated, etc.). It is also noteworthy that the national average
annual exposures for all demographic groups are well below the current annual NAAQS.
Relatedly, to display these results as a heat map, only the national population-weighted
average is provided, with distributional information provided subsequently.
Figure 6-1 also shows the average PM2.5 concentration reductions for each
population when moving from the current standard to alternative standard levels, in
columns with a dash (e.g., 12/35-10/35). Please note that the magnitude of the PM2.5
concentration reductions is small (i.e., tenths of a [ig/m3) because they are national
averages and include individuals residing in the relatively small number of areas predicted
to experience PM2.5 concentration reductions when moving to lower alternative standards,
as well as in the relatively large number of areas already below the lower alternative PM2.5
standard levels where no PM2.5 concentration reductions are predicted under more
stringent standards.
On average, under the current and alternative scenarios evaluated, Asian, Black,
Hispanic, less educated, unemployed, uninsured, linguistically isolated, below the poverty
334
-------
line populations live in areas with higher annual average PM2.5 concentrations than the
reference population. Residents of both HOLC Grade D (i.e., redlined) and HOLC Grades A,
B, and C (A-C) neighborhoods in urban areas also have higher annual average PM2.5
concentrations compared to populations living outside of these urban areas (i.e., not
graded by HOLC). Linguistically isolated, Hispanic, and Asian populations are projected to
experience the highest relative concentrations under each scenario for the demographic
groups examined. While residents of neighborhoods with HOLC grades of A-C in urban
areas also experience some of the highest relative concentrations, concentrations in
neighborhoods with a HOLC grade of D are even higher.
Regarding PM2.5 concentration reductions (columns showing average annual PM2.5
concentration reductions when shifting from the current to alternative standards), most
demographic groups are projected to experience similar reductions under more stringent
standards, in absolute terms. While certain populations are predicted to experience slightly
greater annual PM2.5 concentration reductions, these populations also often began with
higher baseline concentrations, making the absolute change alone insufficient to determine
if disparities are being mitigated. EJ guiding question 3, regarding the impacts of baseline
disparities, is directly evaluated in Section 6.4.1.2.
335
-------
Population
Group
Population (Ages)
Number
of People
12/35
10/35
12/35-
10/35
10/30
1235-
10/30
9/35
12/35-
9/35
8/35
12/35-
8/35
Reference
All (0-99)
371M
7.2
7.1
0.1
7.1
0.1
7.0
0.1
6.9
0.3
Race
White (0-99)
287M
7.1
7.0
0.1
7.0
0.1
6.9
0.1
6.8
0.3
American Indian (0-99)
4M
6.8
6.7
0.1
6.7
0.1
6.6
0.1
6.5
0.3
Asian (0-99)
27M
7.8
7-7
0.1
7.7
0.1
7.6
0.2
7.4
0.4
Black (0-99)
52M
7.4
7.3
0.0
7.3
0.0
7.2
0.1
7.0
0.3
Ethnicity
Non-Hispanic (0-99)
287M
6.9
6.9
0.0
6.9
0.0
6.8
0.1
6.7
0.2
Hispanic (0-99)
84M
7.9
7.8
0.1
7.8
0.1
7.7
0.2
7.5
0.4
Educational
More educated (25-99)
219M
7.1
7.0
0.0
7.0
0.1
7.0
0.1
6.8
0.3
Attainment
Less educated (25-99)
37M
7.5
7.4
0.1
7.4
0.1
7.3
0.2
7.1
0.3
Employment Employed (0-99)
174M
7.2
7.1
0.1
7.1
0.1
7.0
0.1
6.9
0.3
Status
Unemployed (0-99)
9M
7.3
7.2
0.1
7.2
0.1
7.1
0.2
7.0
0.3
Not in the labor force (0-99)
188 M
7.2
7.1
0.1
7.1
0.1
7.0
0.1
6.9
0.3
Insurance
Insured (0-64)
264M
7.2
7.1
0.1
7.1
0.1
7.1
0.1
6.9
0.3
Status
Unisured (0-64)
32M
7.3
7.3
0.1
7.2
0.1
7.2
0.1
7.0
0.3
Linguistic
English "well or better" (0-99) 354M
7.1
7.1
0.0
7.1
0.1
7.0
0.1
6.8
0.3
Isolation
English < "well" (0-99)
17 M
8.1
8.0
0.1
8.0
0.2
7.9
0.3
7.7
0.5
Poverty
Above the poverty line (0-99)
312M
7.1
7.1
0.1
7.1
0.1
7.0
0.1
6.9
0.3
Status
Below poverty line (0-99)
58IVI
7.3
7 2
0.1
7.2
0.1
7.2
0.1
7.0
0.3
Redlined
HOLC Grades A-C (0-99)
44M
8.0
7.8
0.1
7.8
0.2
7.7
0.3
7.5
0.5
Areas
HOLC Grade D (0-99)
16M
8.2
8.1
0.1
8.1
0.1
7.9
0.3
7-7
0.5
Ungraded by HOLC (0-99)
311M
7.0
7.0
0.0
7.0
0.0
6.9
0.1
6.8
0.2
Age
Adults (18-64)
212M
7.2
7.2
0.1
7.1
0.1
7.1
0.1
6.9
0.3
Children (0-17)
84M
7.2
7.2
0.1
7.2
0.1
7.1
0.1
6.9
0.3
Older Adults (65-99)
75M
7.0
6.9
0.0
6.9
0.1
6.9
0.1
6.7
0.3
Sex
Females (0-99)
188 M
7.2
7.1
0.1
7.1
0.1
7.0
0.1
6.9
0.3
Males (0-99)
184 M
7.2
7.1
0.1
7.1
0.1
7.0
0.1
6.9
0.3
Figure 6-1 Heat Map of National Average Annual PM2.5 Concentrations and
Concentration Reductions (|Ag/m3) by Demographic for Current,
Revised, and Alternative PM NAAQS Levels (annual/24-hr) After
Application of Controls in 2032
While average PM2.5 concentrations can provide some insight when comparing
across population impacts, information on the full distribution of concentrations affords a
more comprehensive understanding. This is because both demographic groups and
ambient concentrations can be unevenly distributed across the spectrum of exposures,
meaning that average exposures may mask important disparities that occur on a more
spatially localized basis. To evaluate how the distribution of annual exposures varies within
and across demographic groups at the county level, we plot the full array of exposures
(including very high and very low exposures) projected to be experienced by the entirety of
each population. Distributional figures present the running sum of each population,
converted to a percentage, on the y-axes (i.e., cumulative percent of population).
336
-------
Conversion of each total population to a percent of the total permits direct comparison of
annual PM2.5 exposures across demographic populations with different absolute numbers.
The x-axes show annual PM2.5 concentrations ([ig/m3) from low to high. PM2.5
concentrations are tract-level averages from all Census tracts in the contiguous U.S.10 In
other words, plots compare the running sum of each population against increasing annual
PM2.5 concentrations.
Information on the distribution of tract-level PM2.5 concentrations associated with
the illustrative control strategies for the current and alternative PM standard levels across
and within populations can be found in Figure 6-2.11 Information on the distribution of
tract-level PM2.5 concentration reductions associated with the illustrative control strategies
for the current and alternative PM standard levels across and within populations can be
found in Figure 6-4. The reference population including everyone ages 0-99 is in the top
row of both figures. In Figure 6-2, the reference row shows that the emissions reductions
associated with the current or alternative standard levels yields a fairly smooth S-curve,
with the majority of the population experiencing annual PM2.5 concentrations between 4
and 10 |ig/m3 under air quality scenarios associated with controls applied to meet the
current standards (12/35). Lower PM2.5 concentrations remain similar across lower
alternative standard levels, while higher concentrations are reduced (Figure 6-24).
To evaluate differential exposure distributions under current, revised, and alternate
standard levels, populations of potential EJ concern are shown with colored lines and can
be compared to the respective reference population shown with a black line. Colored lines
to the right of a black line in Figure 6-2 indicate that the potential EJ population is
experiencing disproportionately higher PM2.5 concentrations. Notably, at exposures below
~8.5 |ig/m3, Black population exposures are substantially higher than White population
exposures. This could suggest that exposure disparities in the Black population occur in
10 Distributional figures in the proposal RIA EJ exposure assessment were based on county-level averages.
While tract-level averages are preferable due to the higher resolution, they required substantial additional
computing power (~10-fold] and generate similar results. Therefore, EPA will select the geographic
resolution that is most reasonable in future EJ assessments.
11 Unemployed, uninsured, and the two sexes experience virtually identical distributions of exposure of all
standard levels so were not included in these distributional figures.
337
-------
more rural areas with lower PM2.5 concentrations. Black population exposures above
|ig/m3 are considered further in the discussion of Figure 6-3.
338
-------
Ethnicity
c 100%
Educational to .
. • ~. "5 50%
Attainment g,
n
o
Employment %
Cf'afnc 3
Status
Insurance
Status
Linguistic
Isolation
Poverty
Status
Redlined
Areas
Age
I All (0-99)
I White (0-99)
I Black (0-99)
American Indian (0-99)
I Asian (0-99)
Non-Hispanic (0-99)
Hispanic (0-99)
I More educated (25-99)
I Less educated (25-99)
I Employed (0-99)
I Unemployed (0-99)
I Insured (0-54)
Unisured (0-64)
I English "well or better" (0-99)
English < "well" (0-99)
I Above the poverty line (0-99)
I Below poverty line (0-99)
I HOLC Grades A-C (0-99)
I HOLC Grade D (0-99)
Ungraded by HOLC (0-99)
I Adults (18-S4)
I Children (0-17)
I Older Adults ($5-99)
456789 45 6789 456789
PMj.s (ug/m3)
PMj.s ((ig/m3) PMj.s (ug/m3) PM
Figure 6-2
456789 456789
>s s (Mg/m3) PMj,.s (ug/m3)
National Distributions of Annual PM2.5 Concentrations by
Demographic for Current, Revised, and Alternative PM NAAQS Levels
After Application of Controls in 2032
339
-------
Figure 6-3 hones in on the distributional graphics for the portion of the population
predicted to experience the highest exposures (>7 [ig/m3) for a subset of the population
groups. This allows for improved visualization of tract-level exposure changes under
emissions control strategies associated with the current, revised and alternative regulatory
options at levels where changes are expected to occur. This figure replaces the case study
included in the proposal RIA, which only provided average exposures in areas expected to
experience changes when moving from 12/35-9/35. Figure 6-3 permits visualization of the
sometimes-nuanced absolute disparities occuring at relatively high PM2.5 exposures. For
example, under all scenarios evaluated, there are proportionally more American Indian
exposures than White exposures over 9 |ig/m3, whereas the proportion of the Black
population experiencing exposures over 9 |ig/m3 is actually smaller than the proportion of
White population exposures.
340
-------
Population
Group
12/35
10/35
10/30
9/35
8/35
^ 100%
¦| 80%
Race
p 60%
Q. |
40% 1
f
/
Zj
¦ White (0-99)
¦ Black (0-99)
American Indian (0-99)
¦ Asian (0-99)
^ 100%
£
•2 80%
Ethnicity -5
§" 60% .
lL 1
40%
ff
1
r
¦ Non-Hispanic (0-99)
¦ Hispanic (0-99)
^ 100%
, 1 80%
Educational rc
Attainment 3
0 60%
1
40% 1
r
r
Q
r
¦ More educated (25-99)
¦ Less educated (25-99)
+ 100%
. . J 80%
Linguistic S
Isolation 2
0 60%
Q.
40%
/
/
/
/
/
J
w\.
f
/
/
/ 1
f
J
/
/
f
/
}
¦ English "well or better" (0-99)
English < "well" (0-99)
^ 100%
•| 80%
Poverty
Status
0 60%
40% '
0
r
r
r
r
¦ Above the poverty line (0-99)
¦ Below poverty line (0-99)
^ 100%
£
•2 80%
Redlined S
Areas 2
0 60%
Q.
40%
r
r
f
r
¦ HOLC Grades A-C (0-99)
¦ HOLC Grade D (0-99)
Ungraded by HOLC (0-99)
8 9 10 11
PM2.5 (ng/m3)
8 9 10 11
PM2.s (ng/m3)
8 9 10 11
PM2.5 (Mg/rn3)
8 9 10 11
PM2.5 (ng/m3)
8 9 10 11
PM2.5 (ng/in3)
Figure 6-3 National Distributions of High Annual PM2.5 Concentrations by
Demographic for Current, Revised, and Alternative PM NAAQS Levels
After Application of Controls in 2032
When moving from controls associated with the current standards to controls
associated with the alternative standards, most potential EJ populations of concern with
higher baseline PM2.5 concentrations are predicted to experience greater absolute PM2.5
concentration reductions than the reference groups (e.g., Hispanic populations). Colored
lines again represent potential populations of E] concern and black lines the respective
reference population; however, in these figures, colored lines to the right of the black line
341
-------
now indicate greater relative air quality improvements. For example, Figure 6-4 shows that
~25% of the non-Hispanic population are predicted to experience PM2.5 concentration
reductions when moving from the baseline of control strategies associated with the current
standards (12/35) to control strategies associated with the revised standard levels of 9/35,
whereas ~45% of the Hispanic population are predicted to experience PM2.5 concentration
reductions under control strategies when moving from 12/35-9/35. Figure 6-4 also shows
that greater reductions are expected in the ~45% of the Hispanic population projected to
experience PM2.5 concentration reductions than the ~25% of the non-Hispanic population
projected to experience PM2.5 concentration reductions.
In general, populations with higher absolute national PM2.5 exposures (Section
6.3.1) are also expected to see the greatest reductions in average PM2.5 concentrations
under the alternative standard levels. Populations of Asians, Hispanics, less educated, and
linguistically isolated, are predicted to experience substantially greater PM2.5 concentration
reductions under air quality scenarios associated with control strategies for all alternative
standard levels as compared to the reference population. However, Black populations are
predicted to experience slightly smaller absolute PM2.5 concentration reductions than
White populations when moving to alternative standard levels of 10/35,10/30, and the
revised standards 9/35.
342
-------
Population
Group
12/35-10/35
12/35-10/30
12/35-9/35
12/35-8/35
c 100%
Reference 50%
¦ All (0-99)
c 100%
Q
Race 50%
Q.
^ 0%
T*
¦ White (0-99)
¦ Black (0-99)
American Indian (0-99)
¦ Asian (0-99)
c 100%
_o
Ethnicity 3 50%
Q.
O
0%
r~
¦ Non-Hispanic (0-99)
¦ Hispanic (0-99)
c 100%
O
Educational %
~ 50%
Attainment 5,
P
0%
¦ More educated (25-99)
¦ Less educated (25-99)
¦ Employed (0-99)
¦ Unemployed (0-99)
c 100%
,0
Employment ^
Status g. 50%
4 0*
C 100%
0
Insurance %
c. . "5 50%
Status
* 0%
¦ Insured (0-64)
Unisured (0-64)
c 100%
0
Linguistic %
, , . -S 50%
Isolation £
4 m
Y\
¦ English "well or better" (0-99)
¦ English < "well" (0-99)
C 100%
'°vry 1»%
Status q.
4 0*
rn
¦ Above the poverty line (0-99)
¦ Below poverty line (0-99)
C 100%
.2
Redlined % _
"5 50%
Areas 5,
0
n
0%
\ \
¦ HOLC Grades A-C (0-99)
¦ HOLC Grade D (0-99)
¦1 Ungraded by HOLC (0-99)
c 100%
0
Age -5 50%
t-J
¦ Adults (18-64)
¦ Children (0-17)
¦ Older Adults (65-99)
0 12
PM2.5 Reduction
(pg/m3)
012
PM2.E Reduction
(pg/m3)
0 12
PM2.s Reduction
((ig/rn3)
012
PM2.5 Reduction
(pg/m3)
Figure 6-4 National Distributions of Annual PM2.5 Concentration Reductions by
Demographic from Current to Revised and Alternative PM NAAQS
Levels After Application of Controls in 2032
343
-------
6.3.1.2 Proportional Exposure Changes When Moving from Current to Revised and
Alternative Standard Levels
To put the changes in exposure discussed in section 6.3.1.1in perspective, especially
in light of the disparities in the exposure baseline across population groups also discussed
in section 6.3.1.1, it helps to consider whether the absolute changes represent equivalent
(proportional) reductions in exposure. In some cases, moving to more stringent control
strategies could both reduce total average exposures and reduce disparities in exposure
across groups. However, it can be difficult to determine the relative proportionality of
changes in PM2.5 concentrations for demographic populations using just the absolute
exposure changes when moving from the current standards to a potential alternative
standard level, like those shown above.
In this section, the proportionality of PM2.5 concentration changes when moving
from the current (baseline) to revised and alternative standard levels under air quality
scenarios associated with the illustrative emissions control strategies is directly
calculated.12 To compare air quality improvements on a percentage basis, first exposures
under the alternative standard levels are divided by exposures under the current standards
at the national and regional levels. Those results are then subtracted from 1 to get the
remainder, and then multiplied by 100 to get the percent change. For example, if the
average annual PM2.5 concentration in population A is 6 |ig/m3 under an alternative
standard level and 7 |ig/m3 under control strategies associated with the current standards,
the proportional change would be (l-(6/7)) x 100 = (1-0.857) x 100 = 0.143 x 100 = 14.3%.
If the average annual PM2.5 concentration in population B is 5 |ig/m3 under an alternative
standard level and 6 |ig/m3 under the current standards, the proportional change would be
(l-(5/6)) x 100 = (1-0.833) x 100 = 0.167 x 100 = 16.7%. Therefore, even though the
absolute reduction is equivalent, population B would experience a proportionally larger
reduction under controls strategies associated with the alternate standard level because
the starting concentration was lower. As average PM2.5 concentrations have generally been
12 Results for air quality scenarios associated with meeting the revised and alternative standard levels can be
found in the Appendix to this chapter.
344
-------
representative of the distributions, for simplicity we only present the average proportional
reduction for each population and scenario.
Alternative PM standard levels associated with control strategies reduce the
national average PM2.5 exposure concentrations experienced by the reference population
by an increasing percentage as the alternative standards become lower, with a 0.7%
improvement for 12/35-10/35 and a 3.9% improvement for 12/35-8/35 (Figure 6-5).
Non-Hispanics experience slightly smaller proportional reductions, 0.5% for 12/35-10/35
and 3.5% for 12/35-8/35. Hispanics, linguistically isolated, HOLC Grade D, and HOLC
Grades A-C populations are predicted to experience the relatively largest proportional
reductions in PM2.5 concentrations under all alternative standard levels evaluated, followed
by Asians populations and those less educated. Black populations are predicted to
experience smaller proportional PM2.5 concentration improvements than Whites when
moving from 12/35-10/35,12/35-10/30, and 12/35-9/35, but greater proportional PM2.5
concentration improvements than Whites when moving from 12/35-8/35. This is likely
because disparities between the PM2.5 concentrations experienced by Black and White
populations in the baseline occur at lower ambient PM2.5 concentrations (Figure 6-2 and
Figure 6-4). This leads to proportionally greater improvements for Black populations (i.e., a
narrowing of disparities as compared to White populations) at the lowest alternative PM2.5
standards evaluated. This phenomenon is due to the standards needing to be reduced to
the level at which the disparities are occuring (i.e., PM2.5 concentrations below ~8.5 [ig/m3).
Certain populations (e.g., Native Americans and older adults) are estimated to
experience proportionally smaller reductions in PM2.5 concentrations under all alternative
standard levels evaluated, but it should be noted that these populations are predicted to
experience lower baseline PM2.5 concentrations under air quality scenarios associated with
control strategies (Figure 6-1 through Figure 6-10).
345
-------
Population Groups
Populations (Ages)
12/35-
10/35
12/35-
10/30
12/35-
9/35
12/35-
8/35
Reference
All (0-99)
0.7
0.9
1.9
3.9
Race
White (0-99)
American Indian (0-99)
0.7
0.8
0.9
1.0
1.8
1.7
3.7
3.7
Asian (0-99)
1.3
1.5
3-1
5.7
Black (0-99)
0.5
0.5
1.7
4.2
Ethnicity
Ko."-Hispanic (0-99)
0.5
0.6
1.5
3.5
Hispanic (0-99)
1.5
1.7
2.8
5.1
Educational Attainment More educated (25-99)
0.7
0.8
1.8
3.8
Less educated (25-99)
1.2
1.3
2.4
4.5
Employment Status
Employed (0-99)
Not in the labor force (0-99)
Unemployed (0-99)
0.7
0.7
1.0
0.8
0.9
1.1
1.8
1.9
2.2
4.0
3.9
4.4
Insurance Status
Insured (0-64)
Unisured (0-64)
0.8
0.8
0.9
0.9
1.9
1.9
4.0
4.3
Linguistic Isolation
English "well or better" (0-99)
0.7
0.8
1.8
3.8
English < "well" (0-99)
1.8
1.9
3.2
5.6
Poverty Status
Above the poverty line (6-99)
Below poverty line (0-99)
0.7
0.9
0.8
1.0
1.8
2.0
3.9
4.1
Red lined Areas
HOLC Grades A-C (0-99)
HOLC Grade D (0-99)
1.8
1.7
1.9
1.8
3.5
3.4
5.9
6.2
Ungraded by HOLC (0-99)
0.5
0.6
1.5
3.5
Age
Adults (18-64)
Children (0-17)
Older Adults (65-99)
0.8
0.7
0.7
0.9
0.9
0.8
1.9
1.9
1.7
4.0
4.0
3.6
Sex
Females (0-99)
Males (0-99)
0.7
0.7
0.9
0.9
1.9
1.8
3.9
Figure 6-5 Heat Map of National Percent Reductions (%) in Average Annual PM2.5
Concentrations for Demographic Groups When Moving from Current
to Revised and Alternative PM NAAQS Levels After Application of
Controls in 2032
6.3.2 Regional
Like the national analysis, we evaluate if there are potential EJ concerns 1) in the
baseline, and 2) for the regulatory option(s) under consideration with respect to PMzs
exposures within each of the four regions used in this RIA. We also consider the extent to
which exposures change for each demographic population and how disparities observed
between demographic groups in the baseline scenario (12/35) are impacted (e.g.,
exacerbated/mitigated) under alternative standard levels within each region.
6.3.2.1 Absolute Exposures Under Current, Revised, and Alternative Standard Levels
As both emissions changes and the proportion of people/communities of color
(POC/COC) vary with respect to location, we also parse the aggregated and distributional
absolute PM2.5 concentration by geographic region (northeast [NE], southeast [SH], west
346
-------
[W], and California [CA]) (Figure 6-6 through Figure 6-10).1314 Beginning with total
exposure burdens, average annual reference PM2.5 concentrations are highest in CA across
all current, revised, and alternative standard levels, followed by the SE, then and NE, and
are lowest in the W (Figure 6-6 and Figure 6-7). Comparing populations of potential EJ
concern with their respective references within each region, disparities are observed in all
four regions, although not all for the same demographic populations.
Regarding racial and ethnic disparities, annual PM2.5 concentrations for Black
populations are higher in all four regions than either the overall reference population or
the White population (Figure 6-6, Figure 6-7, and Figure 6-8). These increases are most
visible in the distributional figures for the NE, W, and CA (Figure 6-7). PM2.5 concentrations
among Hispanics are higher than concentrations for Non-Hispanic populations in all four
regions, with disparities being largest at higher PM2.5 concentrations in the SE under
controls associated with the current standards (Figure 6-6 and Figure 6-7). Total PM2.5
concentrations for Asian populations are higher than White population concentrations in
all four regions.
People living below the poverty level, less educated, unemployed, children, and
those living in areas previously designated as redlined areas are predicted to experience
higher annual PM2.5 concentrations than the overall reference population to varying
degrees within certain regions. Older adults (65-99) and those living in urban areas that
were not graded by HOLC are predicted to experience lower PM2.5 concentrations than the
overall reference population in all regions.
13 The regions defined here and used throughout this chapter are consistent with the areas used to present
the costs and benefits in this RIA.
14 Some potential EJ population groups and scenarios were excluded from the distributional figures for visual
clarity.
347
-------
Population
Group
Population (Ages)
NE
12/35
SE W
CA
NE
10/35
SE W
CA
NE
10/30
SE W
CA
NE
9/35
SE W
CA
NE
8/35
SE W
CA
Reference
All (0-99)
6.7
7.1
6.7
3.3
6.7
7.1
6.7
8.9
6.7
7.1
6.7
8.9
6.6
7.0
6.7
8.8
6.5
6.9
6.5
8.6
Race
White (0-99)
American Indian (0-99)
6.6
6.6
7.1
7.1
6.7
5.5
9.2
9 1
6.6
6.5
7.0
7.1
6.7
5.5
8.9
8.8
6.6
6.5
7.0
7.1
6.7
5.5
8.8
8.7
6.5
6.5
7.0
7.1
6.7
5.5
8.7
8.6
6.4
6.4
6.8
6.9
6.5
5.4
8.6
8.5
Asian (0-99)
Black (0-99)
7.1
7.3
7.5
7.2
7.0
7.1
9.7
7.1
7.2
7.5
7.2
7.0
7.1
9.1
9.3
7.1
7.2
7.5
7.2
7.0
7.1
9.0
9.2
7.0
7.1
7.4
7.1
7.0
7.0
8.8
9.1
6.8
6.9
7.1
6.9
6.8
6.7
8.6
9.0
Ethnicity
Non-Hispanic (0-99)
6.7
6.9
6.6
9.1
6.6
6.9
6.6
8.7
6.6
6.9
6.6
8.7
6.6
6.9
6.6
8.5
6.4
6.7
6.4
8.4
Hispanic (0-99)
7.1
7.7
7.0
9.6
7.1
7 6
7.0
9.1
7.1
7.6
7.0
9.1
7.0
7.5
7.0
9.0
6.8
7.3
6.7
8.9
Educational
Attainment
Less educated (25-99)
More educated (25-99)
6.8
6.7
7.2
7.0
6.9
6.7
9.6
9.2
6.8
6.7
7.2
7.0
6.9
6.7
9.1
8.9
6.8
6.7
7.2
7.0
6.9
6.7
9.1
8.8
6.7
6.6
7.1
7.0
6.8
6.6
9.0
8.7
6.6
6.5
7.0
6.8
6.6
6.5
8.9
8.5
Employment Employed (0-99)
Status Unemployed (0-99)
Not in the labor force (0-99)
6.7
6.8
6.7
7.1
7.2
7.1
6.7
6.8
6.7
9.3
94
; 3
6.7
6.8
6.7
7.1
7.2
7.1
6.7
6.8
6.7
9.0
8.9
6.7
6.8
6.7
7.1
7.2
7.1
6.7
6.8
6.7
8.9
9.0
8.9
6.6
6.7
6.6
7.1
7.1
7.0
6.7
6.7
6.7
8.8
89
8.8
6.5
6.5
6.5
6.9
6.9
6.8
6.5
6.5
6.5
8.6
88
8.6
Insurance
Status
Insured (0-64)
Unisured (0-64)
6.7
6.9
7.1
7.3
6.8
6.7
: :
9.4
6.7
6.8
7.1
7.3
6.8
6.7
9.0
9.0
6.7
6.8
7.1
7.3
6.8
6.7
8.9
8.9
6.6
6.7
7.1
7.2
6.7
6.6
8.8
8.8
6.5
6.6
6.9
7.0
6.5
6.4
8.6
8.7
Linguistic
English "well or better" (0-99)
6.7
7.1
6.7
9 2
6.7
7.1
6.7
8.9
6.7
7.1
6.7
8.8
6.6
7.0
6.7
8.7
6.5
6.8
6.5
8.6
Isolation
English < "well" (0-99)
7.2
7.8
7.3
9.8
7.2
7.8
7.3
9.3
7.2
7.8
7.3
9.2
7.1
7.7
7.2
9.1
7.0
7.4
6.9
9.0
Poverty
Status
Above 200% of the poverty line (0-99) 6.7
Below 200% of the poverty line (0-9.. 6.8
7.1
7.2
6.7
6.8
9.2
9.5
6.7
6.8
7.0
7.2
6.7
6.8
8.9
9.0
6.7
6.8
7.0
7.2
6.7
6.8
B a
9.0
6.6
6.7
7.0
7.1
6.6
6.8
8.7
8.9
6.5
6.6
6.8
6.9
6.5
6.6
8.5
8.8
Redlined
Areas
HOLC Grades A-C (0-99)
HOLC Grade D (0-99)
7.4
7.7
7.9
8.0
7.5
7.7
10.5
10.4
7.4
7.7
7.9
8.0
7.5
7.7
9.6
9.7
7.4
7.7
7.9
8.0
7.5
7.7
9.6
9.7
7.2
7.5
7.8
7.9
7.5
7.7
9.5
9 5
7.0
7.2
7.6
7.7
7.4
7.6
94
94
Ungraded by HOLC (0-99)
6.5
7.0
6.7
9.0
6.4
7.0
6.7
8.7
6.4
7.0
6.6
8.7
6.4
7.0
6.6
8.6
6.3
6.8
6.4
8.4
Age
Adults (18-64)
Children (0-17)
Older Adults (65-99)
6.8
6.7
6.6
7.1
7.2
6.9
6.8
6.8
6.6
9.3
9.3
3:
6.7
6.7
6.6
7.1
7.2
6.9
6.8
6.8
6.6
9.0
9.0
8.8
6.7
6.7
6.6
7.1
7.2
6.9
6.7
6.7
6.5
8.9
8.9
8.8
6.6
6.6
6.5
7.1
7.1
6.9
6.7
6.7
6.5
8.8
8.8
8.6
6.5
6.5
6.4
6.9
6.9
6.7
6.5
6.5
6.3
8.7
8.7
8.5
Sex
Females (0-99)
Males (0-99)
6.7
6.7
7.1
7.1
6.7
6.7
9.3
6.7
6.7
7.1
7.1
6.7
6.7
8 9
8.9
6.7
6.7
7.1
7.1
6.7
6.7
8.9
8.9
6.6
6.6
7.0
7.0
6.7
6.7
8.8
8.8
6.5
6.5
6.9
6.9
6.5
6.5
8.6
8.6
Figure 6-6 Heat Map of Regional Average Annual PM2.5 Concentrations (^g/m3) by
Demographic for Current, Revised, and Alternative PM NAAQS Levels
After Application of Controls in 2032
348
-------
NE
SE
w
CA
c 100%
0
Lf>
ro m
^3 50%
rH CL
O
Q.
S 0%
J
¦ White (0-99)
¦ Black (0-99)
2 c 100%
0
£1 50%
<7> CL
O
Cl
0%
f
f
f
J
American Indian (0-99)
¦ Asian (0-99)
c 100%
0
Si 50%
Hi cl
>¦, O
:i a 0 %
f
f
f
J
¦ Non-Hispanic (0-99)
| c 100%
LLI O
£! 50%
Cn cl
O
0.
0%
f
f
L
J
¦ Hispanic (0-99)
c 100%
J&| 50%
r°
0%
f
/
J
¦ English "well or better" (0-99)
¦c _ 100%
'1 .1
.£1 50%
—1 01 g.
O
Cl
0%
/
f
/
J
English < "well" (0-99)
c 100%
0
in 'z
ro ro
W 3 50 /O
rj rl CL
15 a
CO 0%
y
f
/
J
¦ Above the poverty 1 i ne (0-99)
£ c 100%
§ 1
O in w
50%
of a
O
a
0%
¦ Below poverty line (0-99)
c 100%
0
if) 'Z
50%
I" &
-------
The highest burden PM2.5 concentrations are again enlarged in Figure 6-8, for each
region under control strategies associated with 12/35 and 9/35, making it clearer which
potential populations of EJ concern are being impacted by lowering the PM2.5 standard
levels.
350
-------
I White (0-99)
I Black (0-99)
American Indian (0-99)
I Asian (0-99)
Non-Hispanic (0-99)
Hispanic (0-99)
9/35
Population
M
00 §
r
/
r
r
c 100%
0
u"i Z
m ra
= 80%
0 1
Q.
60%
/
//
f
/ 1
i
(
I More educated (25-99)
I Less educated (25-99)
I English "well or better" (0-99)
English < "well" (0-99)
I Above the poverty Iine (0-99)
I Below poverty line (0-99)
I HOLC Grades A-C (0-99)
HOLC Grade D (0-99)
Ungraded by HOLC (0-99)
S 9 101112 8 9 10 1112 8 9 10 1112 8 9 10 1112
PM2.S (ng/m3) PM2.s (ng/ms) PM2.s(|jg/ms) PM;.s(ng/m3)
Figure 6-8 Regional Distributions of High Annual PM2.5 Concentration Reductions
for Demographic Groups for Current PM NAAQS Levels and the 9/35
Revised Standard Scenario After Application of Controls in 2032
(Revised Scale)
351
-------
6.3.2.2 Absolute Exposure Changes When Moving from Current to Revised and
Alternative Standard Levels
Next, we consider how average exposures change across different demographic
groups at the regional level. Information on average and distributional exposure changes
by region when moving from control strategies associated with the current standards to
control strategies associated with alternative standard levels are available in Figure 6-9
and Figure 6-10, respectively.15 Similar to the average annual PM2.5 concentrations going
from highest in CA, followed by the SE and NE, and being lowest in the W (Section 6.3.1.2),
average PM2.5 concentration reductions also follow the same order. Comparing how these
reductions affect populations of potential EJ concern within each region, we note that there
are differences across regions in terms of which demographic populations benefit the most
(or least), particularly for 12/35-9/35 or 12/35-8/35.
Going through each region, the largest regional PM2.5 concentration reductions
occur in CA, where populations of HOLC Grade D, HOLC Grades A-C, Blacks, Hispanics, those
below the poverty line, unemployed, uninsured, and those less educated are expected to
experience greater PM2.5 concentration reductions when moving from the baseline to
alternative standard levels. In the SE, there are greater PM2.5 concentration reductions for
Asians, Hispanics, and those less educated under all alternative standard levels. Asian and
Black populations in CA experience greater PM2.5 concentration reductions when moving
from 12/35-8/35. In the NE for 12/35-9/35 and 12/35-8/35, there are greater PM2.5
concentration reductions for Blacks, and slightly greater PM2.5 concentration reductions for
Asians. This is similar to the W, where Blacks, Hispanics, and those less educated are
predicted to see greater PM2.5 concentration reductions for 12/35-9/35 and 12/35-8/35.
15 Distributions for the reference, male, and female populations were excluded from Figure 6-8 as they closely
reflect overall distributions.
352
-------
12/35-10/35 12/35-10/30 12/35-9/35 12/35-8/35
KUpUldUOIl
Group
Population (Ages)
NE
SE
w
CA
NE
SE
w
CA
NE
SE
w
CA
NE
SE
W
CA
Reference
All (0-99)
0.0
0.0
0.0
0.4
0.0
0.0
0.0
0.4
0.1
0.1
0.1
0.5
0.2
0.2
0.2
0.7
Race
White (0-99)
0.0
0.0
0.0
0.4
0.0
0.0
0.0
0.4
0.1
0.1
0.1
0.5
0.2
0.2
0.2
0.6
American Indian (0-99)
0.0
0.0
0.0
0.4
0.0
0.0
0.0
0.4
0.1
0.0
0.0
0.5
0.2
0.2
0.2
0.6
Asian (0-99)
0.0
0.0
0.0
0.4
0.0
0.0
0.0
0.4
0.1
0.1
0.0
0.6
0.3
0.4
0.3
0.8
Black (0-99)
0.0
0.0
0.0
0.4
0.0
0.0
0.0
0.5
0.2
0.0
0.1
0.6
0.3
0.2
0.3
0.7
Ethnicity
Non-Hispanic (0-99)
0.0
0.0
0.0
0.3
0.0
0.0
0.0
0.4
0.1
0.0
0.0
0.5
0.2
0.2
0.2
0.7
Hispanic (0-99)
0.0
0.0
0.0
0.4
0.0
0.0
0.0
0.5
0.1
0.1
0.1
0.6
0.3
0.4
0.3
0.7
Educational
More educated (25-99)
0.0
0.0
0.0
0.3
0.0
0.0
0.0
0.4
0.1
0.0
0.1
0.5
0.2
0.2
0.2
0.7
Attainment
Less educated (25-99)
0.0
0.0
0.0
0.5
0.0
0.0
0.0
0.5
0.1
0.1
0.1
0.6
0.2
0.3
0.3
0.7
Employment Employed (0-99)
0.0
0.0
0.0
0.4
0.0
0.0
0.0
0.4
0.1
0.1
0.1
0.5
0.2
0.2
0.2
0.7
Status
Unemployed (0-99)
0.0
0.0
0.0
0.4
0.0
0.0
0.0
0.5
0.1
0.1
0.1
0.6
0.2
0.3
0.3
0.7
Mot in the labor force (0-99)
0.0
0.0
0.0
0.4
0.0
0.0
0.0
0.4
0.1
0.1
0.1
0.5
0.2
0.2
0.2
0.7
Insurance
Insured (0-64)
0.0
0.0
0.0
0.4
0.0
0.0
0.0
0.4
0.1
0.1
0.1
0.5
0.2
0.2
0.2
0.7
Status
Unisured (0-64)
0.0
0.0
0.0
0.4
0.0
0.0
0.0
0.5
0.1
0.1
0.1
0.6
0.3
0.3
0.3
0.7
Linguistic
English "well or better" (0-99)
0.0
0.0
0.0
0.4
0.0
0.0
0.0
0.4
0.1
0.1
0.1
0.5
0.2
0.2
0.2
0.7
Isolation
English < "well" (0-99)
0.0
0.0
0.0
0.5
0.0
0.0
0.0
0.5
0.1
0.1
0.1
0.6
0.3
0.4
0.4
0.7
Poverty
Above the poverty line (0-99)
0-0
0.0
0.0
0.4
0.0
0.0
0.0
0.4
0.1
0.1
0.1
0.5
0.2
0.2
0.2
0.7
Status
Below poverty line (0-99)
0.0
0.0
0.0
0.4
0.0
0.0
0.0
0.5
0.1
0.1
0.1
0.6
0.2
0.2
0.3
0.7
Redlined
HOLC Grades A-C (0-99)
0.0
0.0
0.0
0.9
0.0
0.0
0.0
0.9
0.2
0.1
0.0
1.0
0.4
0.4
0.1
1.1
Areas
HOLC Grade D (0-99)
0.0
0.0
0.0
0,8
0.0
0.0
0.0
OS
0.2
0.0
0.0
0.9
0.4
0.3
0.1
1.1
Ungraded by HOLC (0-99)
0.0
0.0
0.0
0.2
0.0
0.0
0.0
0.3
0.1
0.1
0.1
0.4
0.2
0.2
0.2
0.6
Age
Adults (18-64)
0.0
0.0
0.0
0.4
0.0
0.0
0.0
0.4
0.1
0.1
0.1
0.5
0.2
0.2
0.2
0.7
Children (0-17)
0.0
0.0
0.0
0.4
0.0
0.0
0.0
0.4
0.1
0.1
0.1
0.5
0.2
0.3
0.2
0.7
Older Adults (65-99)
0.0
0.0
0.0
0.4
0.0
0.0
0.0
0.4
0.1
0.0
0.1
0.5
0.2
0.2
0.2
0.7
Sex
Females (0-99)
0.0
0.0
0.0
0.4
0.0
0.0
0.0
0.4
0.1
0.1
0.1
0.5
0.2
0.2
0.2
0.7
Males (0-99)
0.0
0.0
0.0
0.4
0.0
0.0
0.0
0.4
0.1
0.1
0.1
0.5
0.2
0.2
0.2
0.7
Figure 6-9 Heat Map of Regional Reductions in Average Annual PM2.5
Concentrations (|ig/m3) for Demographic Groups When Moving from
Current to Revised and Alternative PM NAAQS Levels After Application
of Controls in 2032
353
-------
Population
Group
NE
12/35-9/35
SE W
CA
g 100%
Race -5 50%
0%
r~
f
¦ White (0-99)
¦ Black (0-99)
American Indian (0-99)
¦ Asian (0-99)
g 100%
Ethnicity -5 50%
Q.
a 0%
¦ Non-Hispanic (0-99)
¦ Hispanic (0-99)
= 100%
O
Educational m „„
.... .3 50%
Attainment gL
0
a 0%
¦ More educated (25-99)
¦ Less educated (25-99)
= 100%
0
Employmentl 50%
Status q,
0
0%
1 Employed (0-99)
I Unemployed (0-99)
c 100%
O
Insurance H
Status | 50%
0%
r^
¦ Insured (0-64)
Unisured (0-64)
j= 100%
Linguistic | 50%
Isolation q.
a 0%
/"
¦ English "well or better" (0-99)
¦ English < "well" (0-99)
= 100%
Poverty | 5Q%
Status q.
a 0%
r
¦ Above the poverty line (0-99)
¦ Below poverty line (0-99)
c 100%
O
fdlined 1 50%
Areas q.
0
a 0%
(
r"
j
¦ HOLC Grades A-C (0-99)
¦ HOLC Grade D (0-99)
Ungraded by HOLC (0-99)
0 1
pm2.s
Reduction
(ng/m3)
0 1
pm2.5
Reduction
(ng/m3)
0 1
pm2.5
Reduction
(pg/m3)
0 1
PM2.s
Reduction
(pg/m3)
Figure 6-10 Regional Distributions of Annual PM2.5 Concentration Reductions for
Demographic Groups When Moving from Current PM NAAQS Levels to
9/35 After Application of Controls in 2032
354
-------
6.3.2,3 Proportional Exposure Changes When Moving from Current to Revised and
Alternative Standard Levels
Regionally the greatest proportional reductions are estimated for CA when moving
from the current standards to all alternative standard levels associated with the illustrative
emission control strategies (Figure 6-11). Like the national analysis, percent reductions get
larger as alternative standard levels decrease.
Population , , .
Populations (Aqes)
Groups
12/35-10/35
NE SE W CA
12/35-10/30
NE SE W CA
12/35-9/35
NE SE W CA
12/35-8/35
NE SE W CA
Reference All (0-99)
0.2 0.1 0.0
4.0
0.2 0.1 0.4
4.5
1.5 0.8 0.8
3.3 3.3 3.5
m
Race White (0-99)
American Indian (0-99)
0.2 0.1 0.0
0.1 0.0 0.0
4.0
3.9
0.2 0.1 0.4
0.1 0.0 0.3
4.5
4.5
1.4 0.8 0.8
1.1 0.6 0.9
5.5
5.2
3.0 3.2 3.4
2.6 3.3 3.1
6.9
6.4
Asian (0-99)
Black (0-99)
0.2 0.0 0.0
0.2 0.0 0.0
3.8
4.5
0.2 0.0 0.2
0.2 0.0 0.2
4.2
5.1
1.6 1.4 0.6
2.2 0.6 1.0
6.5
6.3
3.6 3.6
4.6 3.3 4.9
8.7
7.6
Ethnicity Non-Hispanic (0-99)
Hispanic (0-99)
0.2 0.0 0.0
0.1 0.3 0.0
3.5
4.6
0.2 0.0 0.4
0.1 0.3 0.3
4.1
4.9
1.5 0.5 0.7
1.5 1.6 1.2
5.6
5.8
3.2 2.8 3.0
3.7 4.6 4.9
7.5
6.9
Educational More educated (25-99)
Attainment Less educated (25-99)
0.2 0.1 0.0
0.1 0.2 0.0
3.8
4.9
0.2 0.1 0.4
0.1 0.2 0.4
4.3
H
1.5 0.7 0.8
1.5 1.1 1.1
5.6
6.2
3.2 3.1 3.3
3.5 3.6 4.6
7.3
7.4
Employment Employed (0-99)
Status Unemployed (0-99)
Not in the labor force (0-99)
0.2 0.1 0.0
0.2 0.1 0.0
0.2 0.1 0.0
4.0
4.6
4.0
0.2 0.1 0.4
0.2 0.1 0.3
0.2 0.1 0.4
4.5
5.0
4.5
1.5 0.8 0.8
1.7 1.0 0.8
1.5 0.8 0.8
5.8
6.0
5.6
3.2 3.4 3.4
3.6 3.7 3.8
3.3 3.2 3.5
7.3
7.3
7.1
Insurance Insured (0-64)
Status Unisured (0-64)
0.2 0.1 0.0
0.2 0.2 0.0
4.0
4.6
0.2 0.1 0.4
0.2 0.2 0.4
4.5
5.1
1.5 0.8 0.8
1.7 1.2 1.0
5.7
5.9
3.3 3.4 3.4
3.7 4.0 4.1
7.2
7.1
Linguistic English "well or better" (0-99)
Isolation English < "well" (0-99)
0.2 0.1 0.0
0.1 0.3 0.0
3.9
5.1
0.2 0.1 0.4
0.1 0.3 0.2
4.4
i
1.5 0.8 0.8
1.5 1.8 1.1
5.6
6.5
3.2 3.2
3.8 5.1
m
5.1
7.2
7.6
Poverty Above the poverty line (0-99)
Status Below poverty line (0-99)
0.2 0,1 0.0
0.2 0.2 0.0
3.9
4.6
0.2 0.1 0.4
0.2 0.2 0.4
m
1.5 0.8 0.8
1.7 0.9 1.0
5.6
6.0
3.2 3.3 3.4
3.6 3.3 4.0
7.2
7.2
Redlined HOLC Grades A-C (0-99)
Areas HOLC Grade D (0-99)
Ungraded by HOLC (0-99)
0.2 0.0 0.0
0.2 0.0 0.0
0.2 0.1 0.0
8.2
7.3
0.2 0.0 0.1
0.2 0.0 0.2
0.2 0.1 0.4
1
2.3 1.0 0.3
2.4 0.5 0.3
5.2
5.8 |
1.4
1.8
10.2
10.2
2.8
3.4
1.2 0.8 0.9
4.6
2.4 3.2 3.6
6.3
Age Adults (18-64)
Children (0-17)
Older Adults (65-99)
0.2 0.1 0.0
0.2 0.1 0.0
0.2 0.1 0.0
4.0
4.0
3.9
0.2 0.1 0.4
0.2 0.1 0.4
0.2 0.1 0.4
4.5
4.5
4.4
1.5 0.8 0.8
1.5 0.9 0.8
1.4 0.6 0.8
5.8
5.6
5.6
3.3 3.4 3.5
3.3 3.5 3.4
3.1 2.7 3.3
7.3
7.1
7.2
Sex Females (0-99)
Males (0-99)
0.2 0.1 0.0
0.2 0.1 0.0
4.0
4.0
0.2 0.1 0.4
0.2 0.1 0.4
4.5
4.5
1.5 0.8 0.8
1.5 0.8 0.8
5.7
5.7
3.3 3.3 3.5
3.2 3.3 3.5
7.2
7.2
Figure 6-11 Heat Map of Regional Percent Reductions (%) in Average Annual PM2.5
Concentrations for Demographic Groups When Moving from Current
to Revised and Alternative PM NAAQS Levels After Application of
Controls in 2032
6.4 EJ Analysis of Health Effects under Current, Revised, and Alternative Standard
Levels
In addition to comparing PM2.5 concentrations for potential demographic
populations of concern in the EJ exposure analysis (Section 6.3.1), we conducted an EJ
analysis of health effects. This analysis aims to evaluate the potential for EJ concerns
355
-------
related to PM2.5 health outcomes among populations potentially at increased risk of or to
higher PM2.5 exposures from three perspectives, which correspond to the three EJ
questions listed in Section 6.1. Specifically, the following questions are addressed:
1) Are there disparate PM2.5 health effects (e.g., mortality) under baseline/current PM
NAAQS standard levels (question 1)?
2) Are there disparate PM2.5 health effects under illustrative alternative PM NAAQS
standard levels (question 2)?
3) Are disparities in PM2.5 health effects created, exacerbated, or mitigated under
illustrative alternative PM NAAQS standard levels as compared to the baseline
(question 3)?
There is considerable scientific evidence that specific populations and lifestages are
at increased risk of PM2.5-related health effects (Section 1.5.5 and Chapter 12 of U.S. EPA,
2019). Factors that may contribute to increased risk of PM2.5-related health effects include
lifestage (e.g., children), pre-existing diseases (e.g., cardiovascular disease and respiratory
disease), race/ethnicity, and socioeconomic status.16 Of these factors, the ISA found
"adequate evidence" indicating that children and some races are at increased risk of
PM2.5-related health effects, in part due to disparities in exposure. However, we lack
associated epidemiologic information that would enable us to conduct a health effects
analysis for children.
Due to the limited availability of both new scientific evidence in this NAAQS review
and input information (U.S. EPA, 2019, 2022a), the one health endpoint for which we
evaluate EJ implications is premature mortality. The PM ISA and PM ISA Supplement
provided evidence that there are consistent racial and ethnic disparities in PM2.5 exposure
across the U.S., particularly for Black/African Americans, as compared to non-Hispanic
White populations. Additionally, some studies provided evidence of increased PM2.5-related
mortality and other health effects from long-term exposure to PM2.5 among Black
16 As described in the 2019 ISA, other factors that have the potential to contribute to increased risk include
obesity, diabetes, genetic factors, smoking status, sex, diet, and residential location (U.S. EPA, 2019, chapter
12).
356
-------
populations. Taken together, the 2019 PM ISA concluded that the evidence was adequate to
conclude that race and ethnicity modify PIVh.s-related risk, and that non-White individuals,
particularly Black individuals, are at increased risk for PIVh.s-related health effects, in part
due to disparities in exposure (U.S. EPA, 2019, 2022a).
As such, this EJ health analysis assesses long-term PIVh.s-attributable mortality rates
stratified by racial and ethnic demographic populations.17 Mortality is presented as a rate
per 100,000 (100k) individuals to permit direct comparisons between population
demographics with different total population counts.18 Additional information on the
concentration-response functions and baseline incidence rates used as input information in
this health EJ analysis can be found in Section 6.6.2 and Appendix C of the draft PM Policy
Assessment (U.S. EPA, 2021).
In the proposal RIA for this rulemaking, a single epidemiological study (i.e., Di et al.,
2017) providing race/ethnicity-stratified exposure-mortality relationships was identified
as best characterizing risk across the U.S. While this was the largest study of race/ethnicity-
stratified exposure-mortality relationships to date, the Pope III et al., 2019 study of the
National Health Interview Survey (NHIS) cohort also provided national, high-quality
race/ethnicity-stratified exposure-mortality relationships. However, at the time of the PM
NAAQS proposal RIA development, BenMAP-CE was unable to appropriately include the
combined race and ethnicity populations used by Pope III et al., 2019. When preparing for
this final RIA EJ assessment, we were able to update the input parameters necessary to also
include concentration-response relationships from Pope III et al., 2019. This allows for two
independent estimates of race/ethnicity-stratified mortality impacts of different age
ranges, similar to the approach currently used for assessing benefits (Chapter 5).
17 As the ISA and ISA Supplement found that mortality studies evaluated continued to support a linear,
no-threshold concentration-response relationship, mortality rates are calculated here using exposure
estimates across all PM2.5 concentrations (U.S. EPA, 2019, 2022a). However, uncertainties remain regarding
the shape of mortality concentration-response functions, particularly at low concentrations. Additional
uncertainties are related to this analysis, as a single epidemiologic study was used to relate exposure to
mortality health effects that applies only to older adults aged 65 and over (Di et al., 2017).
18 Current Agency VSL practices do not differentiate based on race or ethnicity, so the health analysis did not
include monetization. Separately, although the valuation of morbidity outcomes may differ by race or
ethnicity (e.g., someone without insurance may delay seeing seen by a medical professional until the
situation requires more expensive treatment), available scientific evidence for race/ethnicity-stratified
valuation estimates is lacking.
357
-------
Additional study information, including the specific hazard ratios (HRs), beta coefficients,
and standard errors are available in appendix Table 6-2.
National and regional absolute mortality rates (per 100k individuals) under air
quality scenarios associated with control strategies for the current standards and revised
and alternative lower standard levels and changes in mortality rates when moving from air
quality scenarios associated with control strategies for the current standards to potential
alternative lower standard levels across the demographic-specific mortality rates are
provided in Sections 6.4.1 and 6.4.1.2, respectively.
Similar to what was done for the exposure analysis above, we address the guiding EJ
questions with respect to PIVh.s-attributable mortality impacts first across the contiguous
U.S. in Section 6.4.1, and then at the regional level in Section 6.4.2.
6.4.1 National
National absolute PIVh.s-attributable mortality impacts are provided in Section
6.4.1.1 and national proportional PIVh.s-attributable mortality impacts are provided in
Section 6.4.1.2.
6.4.1.1 Absolute Mortality Rates Under Current, Revised, and Alternative Standard
Levels and Mortality Rate Changes When Moving from Current to Revised
and Alternative Standard Levels
Figure 6-12 and Figure 6-13 show the national averages and distributions of
estimated mortality rates per 100k individuals for each race/ethnicity evaluated. It is
important to note that Di et al., 2017 and Pope III et al., 2019 evaluate different age ranges.
Di et al., 2017 evaluated Medicare enrollees aged 65-99 and Pope III et al., 2019 evaluated
ages 18-99. In this assessment, hazard ratios (HRs) derived from each study-specific age
range were applied to populations of the same ages in this assessment, across the
contiguous U.S. in 2032.
Mortality rate estimates are calculated using additional inputs as compared to
exposure estimates, specifically HRs and baseline incidence. In general, while the greater
magnitude HR for the Black population of older adults found by Di et al., 2017 resulted in
higher estimated mortality rates in Black than White populations, the greater magnitude
HR for the Black population of all adults from Pope III et al., 2019 did not lead to higher
358
-------
estimated mortality rates in Black than White populations, due to a lower ratio of baseline
deaths (Figure 6-12). However, HRs from the Pope III et al, 2019 study did lead to higher
estimated mortality rates in both Black and White populations, as compared to the overall
reference population.
Study „ , .
(Ages) P°Pu'atlon GrouP
Number
of
People
Baseline
Mortality
Ratio of
Baseline
Mortality
HR
(Beta)
12/35
10/35
12/35-
10/35
10/30
12/35-
10/30
9/35
12/35-
9/35
8/35
12/35-
8/35
Di 2017 Reference
75M
2,926K
3.9
0.0070
186
185
1
185
1
183
3
180
7
(65-99) White
62M
2,479K
4.0
0.0061
163
162
1
162
1
161
2
158
5
American Indian
1M
14K
2.6
0.0095
151
150
1
150
2
149
2
147
5
Asian
4M
76K
2.0
0.0092
145
142
4
142
4
139
7
136
10
Black
8M
298K
3.6
0.0189
464 462
3
462
3
456
9
445
22
Hispanic
9M
216K
2.4
0.0110
206
203
4
203
4
201
6
197
10
Pope Reference
287M
3,414K
1.2
0.0113
90
90
1
90
1
89
2
87
3
2019 NH White
1G7M
2,577K
1.5
0.0104
104
104
0
104
1
103
1
101
3
(18-99) NH American Indian
2M
17 K
0.8
0.0095
45
45
0
45
0
45
0
44
1
NH Asian
21M
96K
0.5
0.0095
34
34
1
34
1
33
1
32
2
NH Black
37M
397K
1.1
0.0140
103
103
1
103
1
102
2
99
5
Hispanic
59M
298K
0.5
0.0182
69
67
1
67
1
67
2
65
4
Figure 6-12 Heat Map of National Average Annual Total Mortality Rates and Rate
Reductions (per 100K) for Demographic Groups for Current, Revised,
and Alternative PM NAAQS Levels After Application of Controls in
2032 (NH, Non-Hispanic)
359
-------
Di 2017
Race/Ethnicity
Pope 2019
12/35
10/35
10/30
9/35
Population Group
¦ Reference
¦ White
American Indian
111 Asian
¦ Black
¦ Hispanic
¦ NH White
NH American Indian
¦ NH Asian
¦ NH Black
8/35
100%"
50%-
100%"
50%-
100%"
50%-
1 1 1 1 1 1 1 1 1 r
200 400 600 800 0 50 100 150 200 250
Mortality Rate (per 100k) Mortality Rate (per 100k)
100%"
50%-
Figure 6-13 National Distributions of Total Annual Mortality Rates (per 100k) for
Demographic Groups for Current, Revised, and Alternative PM NAAQS
Levels After Application of Controls in 2032 (NH, Non-Hispanic)
While mortality rate reductions are small when averaged across the contiguous U.S.,
due to the inclusion of areas with no air quali ty improvements (Figure 6-12), reductions
can be substantial in individual tracts (Figure 6-14). Nationally, the rate of PM2.5-
attributable mortality is estimated to decrease for most races and ethnicities when moving
from current standards to alternative standard levels, and more so under lower alternative
360
-------
standard levels (Figure 6-14), In addition, reductions in mortality rates are often larger for
other races as compared to White or non-Hispanic White populations.
100%"
80%-
12/35-
10/35 60%"
40%-
100%"
80%-
12/35-
10/30 60%"
40%-
100%"
80%-
12/35-
9/35 60%"
40%-
100%"
80%-
12/35-
8/35 60%"
40%-
0 50 100 150 0 10 20 30 40 50
Mortality Rate Reduction Mortality Rate Reduction
(per 100k) (per 100k)
Population Group
¦ Reference
¦ White
American Indian
¦ Asian
¦ Black
¦ Hispanic
¦ NH White
¦ NH Asian
NH American Indian
¦ NH Black
Race/Ethnicity
Di 2017 Pope 2019
Figure 6-14 National Distributions of Annual Mortality Rate Reductions for
Demographic Groups When Moving from Current to Revised and
Alternative PM NAAQS Levels After Application of Controls in 2032
(NH, Non-Hispanic)
6.4.1.2 Proportional Mortality Rate Changes When Moving from Current to Revised
and Alternative Standard Levels
To put the changes in exposure discussed in Section 6.3.1.1 in perspective,
especially in light of disparities in the exposure baseline across population groups, it helps
to consider whether the absolute changes represent equivalent (i.e., proportional)
361
-------
reductions in exposure. In some cases, moving to more stringent control strategies could
both reduce total average exposures and reduce disparities in exposure across groups.
However, it can be difficult to determine the relative proportionality of changes in PM2.5
concentrations for demographic populations using just the absolute exposure changes
when moving from the current standards to a potential alternative standard level, like
those shown in Section 6.3.1.1.
In this section, the proportionality of PM2.5 concentration changes when moving
from the current standard (i.e., baseline) to alternative lower standard levels under air
quality scenarios associated with the illustrative emissions control strategies is directly
calculated, using a similar approach to calculating the proportionality of exposures in
Sections 6.3.1.2 and 6.3.2.3. As average PM2.5 concentrations have generally been
representative of the distributions, for simplicity we only present the average national
proportional reduction for each population and scenario.
Results from both studies estimate Hispanics and Asian populations experience
proportionally larger reductions in mortality rates when moving from the current
standards to all lower alternative standard levels associated with control strategies,
thereby mitigating disparities (Figure 6-15). However, national mortality rate disparities in
Black/non-Hispanic Black populations are only mitigated when moving from the current
standards to alternative standard levels of 9/35 or 8/35.
362
-------
Study
Race/Ethnicity
12/35-
10/35
12/35-
10/30
12/35-
9/35
12/35-
8/35
Di 2017
Reference
0.6
0.8
1.7
3.6
(65-99)
White
0.6
0.7
1.5
3.3
American Indian
0.8
1.0
1.5
3.2
Asian
2.6
2.9
4.7
i.i
Black
0.6
0.6
2.0
4.7
Hispanic
1.8
2.0
3.0
5.0
Pope 2019 Reference
0.7
0.8
1.7
3.7
(18-99)
NH White
0.4
0.5
1.3
3.1
NH American Indian
0.4
0.6
0.9
2.3
NH Asian
2.4
2.7
MEM
6.9
NH Black
0.6
0.6
2.0
4.7
Hispanic
1.9
2.0
3.1
5.3
Figure 6-15 Heat Map of National Average Percent Mortality Rate Reductions (per
100k People) for Demographic Groups When Moving from Current to
Revised and Alternative PM NAAQS Levels After Application of
Controls in 2032 (NH, Non-Hispanic)
6.4.2 Regional
Regional absolute PIVh.s-attributable mortality impacts are provided in Sections
6.4.2.1 and 6.4.2.2, whereas regional proportional PIVh.s-attributable mortality impacts are
provided in Section 6.4.2.3.
6.4.2.1 Absolute Mortality Rates Under Current, Revised, and Alternative Standard
Levels
Regionally, the highest mortality rates for reference populations are in CA under air
quality scenarios associated with control strategies for both current standards and
alternative PM standard levels, followed by the NE, SE, and then the W (Figure 6-16 and
Figure 6-17). Total mortality rates in the reference populations decrease slightly under
alternative standard levels in all regions, and by the greatest absolute number in CA. Within
each of the four regions, average and distributional mortality rates are highest among Black
populations, although there are differences in the ordinality of other races and ethnicities
across regions. Interestingly, the distribution of Hispanic mortality rates in the SE suggests
there may be a subset of locations in which Hispanics have higher baseline incidence rates,
as the PM2.5 concentration differentials between Hispanic and non-Hispanic populations
remained fairly constant across PM2.5 concentration distributions (Figure 6-7).
363
-------
12/35
NE SE W
CA
10/35
NE SE W
CA
10/30
NE SE W
CA
9/35
NE SE W
CA
8/35
NE SE W
CA
Di Reference
2017 White
American Indian
Asian
Black
Hispanic
184 183 172
160 161153
130 161133
106 103 128
220
198
202
199
184 183 172 212 184 183 172 210
160 161153 191160 161152 190
129 161133 195 129 161133 193
106 103 128 191106 103 127 190
181182 171 208 178 179 167 205
158 160 152 188 156 157 148 185
128 160 133 192 126 157 131190
105 101127 186 102 97 124183
467 447 417
596 466 447 417
569 466 447 416
567
456 445 413
559
444 435 396
552
157 211181
249
157 210 181
238 157 210 181
237
155 208 179
235
151 204 174 233
Pope Reference
2019 NH White
91 89 82
100106 97
97 91 89 82
138100106 97
94 91 89 82
134100106 97
93 90 89 82
133 99 105 97
92
132
88 87 80
97 104 95
91
130
NH Asian
22 21 31
60
22 21 31
57
22 21 31
57
22 21 31
56
21 20 30
55
NH Black
103 99 78
155
103 99 78
148103 99 78
147101 98 77
145
98 96 74
144
NH American Indian
Hispanic
35 47 38
49 73 58
60
86
35 47 38
49 73 58
58
82
35 47 38
49 73 57
57
82
35 47 38
48 72 57
57
81
35 46 38
47 71 55
57
80
Figure 6-16 Heat Map of Regional Average Annual Total Mortality Rates (per
100K) for Demographic Groups for Current, Revised, and Alternative
PM NAAQS Levels After Application of Controls in 2032 (NH, Non-
Hispanic)
364
-------
12/35
10/35
10/30
9/35
8/35
80%-
NE
40%-
0%_
u
V
u
u
u
80%-
SE
40%-
Di o%.
If
IT
'ir
If
!ir
2017
80%-
W
40%-
0%.
ff
» /
jf f
ff- s
a /
80%-
CA
40%-
0%.
jy
U
u
V
u
80%-
NE
40%-
0%_
i
80%-
SE
40%-
Pope o%_
!
1
2019
80%-
W 40%-
0%.
f
jr
r
r
ff
TV
ff
80%-
CA
40%-
0%.
II
0 400 800
Mortality Rate
(per 100k)
0 400 800
Mortality Rate
(per 100k)
0 400 800
Mortality Rate
(per 100k)
0 400 800
Mortality Rate
(per 100k)
0 400 800
Mortality Rate
(per 100k)
Population Group
¦ Reference
¦ White
American Indian
¦ Asian
¦ Black
¦ Hispanic
¦ NH White
NH American Indian
¦ NH Asian
¦ NH Black
Figure 6-17 Regional Distributions of Total Annual Mortality Rates (per 100k) for
Demographic Groups for Current, Revised, and Alternative PM NAAQS
Levels After Application of Controls in 2032 (NH, Non-Hispanic)
365
-------
6.4.2,2 Absolute Mortality Rate Changes When Moving from Current to Revised and
Alternative Standard Levels
Of the four regions, the largest absolute mortality rate reductions for each
population are estimated in CA when moving from the current to alternative standard
levels (Figure 6-18 and Figure 6-19). Reductions are smaller in the other three regions,
although reductions become more substantial for 12/35-9/35 or 12/35-8/35, especially in
the NE. When comparing across race and ethnicities. Black and non-Hispanic Black
populations are predicted to see the largest mortality rate reductions when moving to
lower standard levels.
12/35-10/35
12/35-10/30
12/35-9/35
12/35-8/35
NE SE W
CA
NE SE W
CA
NE SE W
CA
NE SE W
CA
Di Reference
0.4 0.1 0.0
8.7
0.4 0.1 0.7
10.0
2.9 1.0 1.3
12.4
6.0 4.8 5.6
16.0
2017 White
0.3 0.1 0.0
7.3
0.3 0.1 0.7
8.6
2.3 0.8 1.1
10.4
4.8 3.9 4.9
13.5
American Indian
0.2 0.0 0.0
8.2
0.2 0.0 0.4
9.6
1.6 0.6 0.7
10.7
3.6 3.9 2.7
13.0
Asian
0.1 0.0 0.0
9.5
0.1 0.0 0.2
10.3
2.0 2.3 0.6
14.3
4.5 6.9 4.2
18.1
Black
1.3 0.0 0.0
31.7
1.3 0.0 0.6
35.0
12.7 2.4 3.9
43.7
26.7 13.6 24.0,r51.9
Hispanic
0.1 0.9 0.0
12.7
0.1 0.9 0.4
13.6
2.3 3.1 1.9
15.5
6.3 7.9 8.3
18.3
Pope Reference
0.2 0.0 0.0
4.0
0.2 0.0 0.3
4.5
1.5 0.5 0.6
5.7
3.1 2.4 2.8
7.2
2019 NH White
0.2 0.0 0.0
4.5
0.2 0.0 0.4
5.5
1.4 0.4 0.7
6.9
3.0 2.4 3.0
9.4
NH Asian
0.0 0.0 0.0
2.8
0.0 0.0 0.1
3.0
0.4 0.4 0.2
4.2
0.9 1.4 1.0
5.4
NH Black
0.3 0.0 0.0
8.1
0.3 0.0 0.1
9.0
2.8 0.5 0.8
11.3
5.9 3.0 4.6
13.3
NH American Indian
0.0 0.0 0.0
1.9
0.0 0.0 0.1
2.5
0.3 0.1 0.2
2.6
0.7 1.0 0.6
3.2
Hispanic
0.0 0.3 0.0
4.5
0.0 0.3 0.2
4.8
0.8 1.2 0.7
5.5
2.0 3.0 2.8
6.5
Figure 6-18 Heat Map of Regional Average Annual Mortality Rate Reductions (per
100k) for Demographic Groups When Moving from Current to Revised
and Alternative PM NAAQS Levels After Application of Controls in
2032 (NH, Non-Hispanic)
366
-------
Di
2017
Pope
2019
80%-
NE
40%-
0%~
12/35-10/35
12/35-10/30
r~
12/35-9/35
12/35-8/35
r~
80%-
SE
40%-
0%"
r
r
80%-
W
40%-
0%~
F
80%-
CA
40%-
0%~
F
I
y
F
80%-
NE
40%-
0%
r
r
f
r
80%-
SE
40%-
0%"
-r
rr
r
f
80%-
W
40%-
0%"
r
r
f
80%-
CA
40%-
0%"
f
I
[
i i i i
0 50 100
Mortality Rate
(per 100k)
i i i i
0 50 100
Mortality Rate
(per 100k)
0 50 100
Mortality Rate
(per 100k)
0 50 100
Mortality Rate
(per 100k)
Population Group
¦ Reference
¦ White
American Indian
¦ Asian
¦ Black
¦ Hispanic
¦ NH White
NH American Indian
¦ NH Asian
¦ NH Black
Figure 6-19 Regional Distributions of Annual Mortality Rate Reductions (per
100k) for Demographic Groups When Moving from Current to Revised
and Alternative PM NAAQS Levels After Application of Controls in
2032 (NH, Non-Hispanic)
367
-------
6.4.2.3 Proportional Mortality Rate Changes When Moving from Current to Revised
and Alternative Standard Levels
Proportionally, CA is also predicted to have the largest mortality rate reductions for
each population when moving from current to all lower alternative standard levels
evaluated. Reductions are proportionally larger for both Black and non-Hispanic Black
populations in CA and the NE under all lower standards in which baseline mortality rate
disparities were observed in Section 6.4.2.1, Further, we see a mitigation of mortality rate
disparities in Hispanic populations aged 65-99 in the SE and CA, where baseline disparities
were present in Figure 6-16.
When moving from 12/35-9/35 and 12/35-8/35, the proportional mitigation of
disparities in Black and non-Hispanic Black populations in CA and the NE is predicted to be
notably larger in magnitude. Also, mortality rate disparities in Black populations aged 65-
99 in the W are expected to be reduced when moving from 12/35-9/35 and 12/35-8/35,
where baseline disparities were also found using the Di et al., 2017 study in Section 6.4.2.1.
Finally, mortality rate disparities in non-Hispanic Black and Hispanic populations aged 18-
99 in the W are expected to be reduced when moving from 12/35-9/35 and 12/35-8/35.
12/35-10/35
12/35-10/30
12/35-9/35
12/35-8/35
Study Race/Ethnicity NE SE W
CA
NE SE W
CA
NE SE W
CA
NE SE W
CA
Di 2017 Reference 0.2 0.1 0.0
3.9
0.2 0.1 0.4
1.6 0.5 0.7
5.7
3.3 2.6 3.3
7.3
(65-99) white 0.2 0.1 0.0
3.7
0.2 0.1 0.4
4.3
1.4 0.S 0.7
5.2
3.0 2.4 3.2
6.8
American Indian 0.1 0.0 0.0
4.1
0.1 0.0 0.3
4.8
1.3 0.4 0.6
5.3
2.8 2.4 2.1
6.4
Asian 0.1 0.0 0.0
4.7
0.1 0.0 0.1
5.1
1.8 2.2 0.5
7.1
4.2fS 3.3
9.1
Black 0.3 0.0 0.0
5.3
0.3 0.0 0.1
5.9
2.7 0.5 0.9
7.3
3 3.0 c
8.7
Hispanic 0.1 0.4 0.0
5.1
0.1 0.4 0.2
5.5
1.5 1.5 1.1
6.2
4.6
7.3
Pope 2019 Reference 0.2 0.1 0.0
4.1
0.2 0.1 0.4
m
1.6 0.6 0.8
5.8
3.4 2.7 3.4
7.4
(18-99) NH White 0.2 0.0 0.0
3.2
0.2 0.0 0.5
4.0
1.4 0.4 0.7
5.0
2.9 2.3 3.1
6.8
NH American Indian 0.1 0.0 0.0
3.2
0.1 0.0 0.3
4.2
1.0 0.2 0.5
4.3
2.0 2.1 1.7
5.4
NH Asian 0.1 0.0 0.0
4.6
0.1 0.0 0.2
5.1
1.7 2.0 0.6
7.1
4.003 3.3
9.0
NH Black 0.3 0.0 0.0
5.2
0.3 0.0 0.1
5.8
2.7 0.6 1.0
7.3
gQ 3.i E
8.6
Hispanic 0.1 0.4 0.0
5.3
0.1 0.4 0.3
5.6
1.6 1.6 1.2
6.5
4.1 4.0 2
7.6
Figure 6-20 Heat Map of Regional Average Proportional Mortality Rate Reductions
(per 100k) for Demographic Groups When Moving from Current to
Revised and Alternative PM NAAQS Levels After Application of
Controls in 2032 (NH, Non-Hispanic)
368
-------
6.5 Summary
For this final rulemaking, we quantitatively evaluate the potential for disparities in
PM2.5 concentrations and mortality effects across different demographic populations for
the current standards (12/35, i.e., baseline) and potential alternative PM2.5 NAAQS levels
(10/35,10/30, 9/35, and 8/35) under air quality scenarios associated with illustrative
emissions control strategies. Specifically, we provide information on totals, changes, and
proportional changes in 1) annual average PM2.5 concentrations and 2) premature
mortality as rates per 100,000 individuals across and within various demographic
populations, at national and regional scales. Each type of analysis has strengths and
weaknesses, but when taken together can respond to the three questions from EPA's EJ
Technical Guidance. Total concentration/mortality rate analyses provide information on
absolute PM2.5 concentrations and mortality rates; however, it can be difficult to compare
improvements in air quality/mortality rates among populations from total information. In
contrast, comparing changes in concentration/mortality rates provides information on how
improvements compare across and within populations, but does not provide information
on which populations experience higher total concentration/mortality rates. Proportional
changes are provided as a percent of the total concentration/mortality rate information, so
although they are similar to absolute changes, they are more closely related to total
concentration/mortality rate information.
EJ analyses were performed both at national and regional scales, as geography is
relevant both to PM NAAQS attainment and population demographics. We also conducted
an analysis focusing on relatively high PM2.5 exposures, to examine the subset of areas in
which air quality is projected to change after the application of controls outlined in Chapter
3. For all of these analyses, we note that the results should be considered illustrative only,
Further, as with all EJ analyses, data limitations make it possible that disparities may exist
that our analysis did not identify. This is especially relevant for potential EJ characteristics,
environmental impacts, and more granular spatial resolutions that were not evaluated. We
note again that this analysis is based on air quality modeling conducted on a 12 x 12 km
grid scale, which may mask more local disparities in exposure and risk. Additionally, EJ
concerns for each rulemaking are unique and should be considered on a case-by-case basis.
369
-------
Whereas all populations experience reductions in PM2.5 concentrations and health
effects at lower PM standard levels, to make conclusions regarding EJ impacts of this final
rule we refer back to the three questions that EPA's EJ Technical Guidance (U.S. EPA, 2015)
states should be addressed, which for purposes of the PM NAAQS RIA EJ analysis are:
1) Are there disparate PM2.5 exposures/mortality rates under baseline/current PM
NAAQS standard levels?
2) Are there disparate PM2.5 exposures/ mortality rates under illustrative alternative
PM NAAQS standard levels?
3) Are PM2.5 exposure/ mortality rates disparities created, exacerbated, or mitigated
under illustrative alternative PM NAAQS standard levels as compared to the
baseline?
Considering results of both the EJ exposure analysis (Section 6.3) and the EJ health
effects analysis (Section 6.4), responses to the above three questions can be summarized
as:
1) Disparities in the baseline: Regarding exposures, under air quality scenarios
associated with control strategies for the baseline (12/35) PM NAAQS scenario,
some populations are predicted to experience disproportionately higher annual
PM2.5 concentrations nationally than the reference (overall) population, both in
terms of aggregated average concentrations and across the distribution of air
quality (Figure 6-1 through Figure 6-5). Specifically, populations living in redlined
areas or not redlined areas, linguistically isolated, Hispanic, Asian, Black, less
educated, unemployed, uninsured, and below the poverty line have higher national
annual concentrations, on average and across the distributions, than both the
overall reference population or other populations (e.g., non-Hispanic, White, and
more educated). In particular, those living in redlined areas are estimated to
experience the highest concentrations, both on average and across PM2.5
concentration distributions, of all demographic groups analyzed. These
disproportionalities are also observed in the baseline at the regional level, though to
different extents (Figure 6-6 through Figure 6-11).
370
-------
In terms of mortality rates, some racial/ethnic populations are also predicted to
experience notably higher rates of premature mortality than reference populations
(Figure 6-12 through Figure 6-20). Black populations are estimated to have the
highest national and regional mortality rates in older adults aged 65-99, both on
average and across population distributions. This may be partly due to higher PM2.5
concentrations for this population, which could contribute to the higher magnitude
concentration-response relationship between exposure concentrations and
premature mortality (Di et al., 2017), as well as other underlying health factors that
may increase susceptibility to adverse outcomes among Black populations. There
may be additional baseline disparities in some populations in certain regions, to
varying degrees.
2) Disparities under alternative policy options: While more stringent control
strategies reduce PM2.5 concentrations and mortality rates across all demographic
groups, disparities seen in the baseline also persist in the alternative policy options
and revised standard levels. Higher PM2.5 concentrations and health effects remain
for Asian, Black, Hispanic, less educated, unemployed, uninsured, linguistically
isolated, those below the poverty line, HOLC Grade D, and HOLC Grades A-C
populations, as compared to the reference population under air quality scenarios
associated with the illustrative control strategies (10/35,10/30, 9/35, and 8/35)
(Figure 6-1 through Figure 6-20). Nationally and regionally, these patterns and the
populations affected are similar to those seen in the baseline.
3) Relative impact of alternative policy options on disparities in the baseline: For
most populations assessed, PM2.5 exposure disparities are mitigated in the
illustrative air quality scenarios associated with control strategies for more
stringent PM2.5 control strategies (10/35,10/30, 9/35, and 8/35) as compared to
the baseline (12/35), to differing degrees (Figure 6-5 and Figure 6-11). More
specifically, increasing portions of certain populations of potential EJ concern are
expected to experience greater PM2.5 concentration reductions as the illustrative
control strategies become more stringent (Figure 6-4 and Figure 6-9). At the
national scale, linguistically isolated, redlined areas, not redlined areas, Hispanic,
371
-------
Asian, those less educated, and unemployed populations are estimated to
experience greater proportional reductions in PM2.5 concentrations than reference
populations under all lower standard levels evaluated, with proportional reductions
increasing as the standard levels decrease. In addition, exposure disparities in
baseline Black and uninsured populations are estimated to be mitigated when
moving to an alternative standard level of 8/35. Average concentration reductions
were also similar across Black and White populations when evaluating the relatively
high PM2.5 exposures that are most affected by the illustrative control strategies.
Considering the four geographic regions, PM2.5 baseline exposure disparities were
mitigated in CA under all lower standard levels for many populations in other
regions when moving to either 9/35 or 8/35.
In general, more stringent control strategies are also associated with reductions in
mortality rate disparities. Specifically, the analysis shows that as the illustrative
PM2.5 control strategies become increasingly stringent, differences in mortality rates
across demographic groups decline, particularly for the lowest alternatives
evaluated (12/35-9/35 and 12/35-8/35). Nationally, Black and non-Hispanic Black
populations are predicted to experience proportionally similar mortality rate
reductions to the reference populations under control strategies associated with
12/35-10/35 or 12/35-10/30, but greater reductions in mortality rates under
control strategies associated with 12/35-9/35 or 12/35-8/35. Similar to the
estimated changes in PM2.5 concentrations following reductions in PM2.5
concentrations under alternative standard levels, disparities in national PM2.5
mortality rates across demographic groups are mitigated for Hispanics in all the
alternative PM standard levels (10/35,10/30, 9/35, and 8/35) as compared to the
baseline (Figure 6-15).
6.6 Environmental Justice Appendix
6.6.1 EJ Exposure Analysis Input Data
In Appendix 2A, the exposure assessment involves demographic population data projected
out to the future year 2032. We use population projections based on economic forecasting
models developed by Woods and Poole, Inc. (Woods & Poole, 2015). The Woods and Poole
372
-------
database contains county-level projections of population by age, sex, and race/ethnicity out
to 2060, relative to a baseline using the 2010 Census data. Projections in each county are
determined simultaneously with every other county in the U.S. to consider patterns of
economic growth and migration. The sum of growth in county-level populations is
constrained to equal a previously determined national population growth, based on Bureau
of Census estimates (Hollmann et al., 2000). According to Woods and Poole, linking county-
level growth projections together and constraining to a national-level total growth avoids
potential errors introduced by forecasting each county independently (Woods & Poole,
2015).
We also include data on educational attainment, poverty status, unemployment,
health insurance coverage, occupational status, linguistic isolation, and redlining status.
These datasets are included in BenMAP-CE. The data sources and processing methodology
for each dataset is described below. County-level datasets were generated for 3,109
counties in the contiguous U.S. Census tract level datasets were generated for 72,538
census tracts in the contiguous U.S.
6.6.1.1 Educational Attainment
We use data from the ACS to provide tract-level summaries of educational
attainment. These data represent 5-year average ACS estimates from 2015 to 2019 and
span the same two education categories as county-level demographic variables: no high
school diploma (termed "no_hs_degree_tract") and high school diploma (or equivalency)
and above (termed "hs_degree_plus_tract"). For both education groups (with/without high
school diploma), we estimate the fraction of the total tract population (ages 25 years and
above) in each education group. Thus, the two estimates sum to one for each tract. All
estimates were generated at the tract level for 72,538 tracts in the contiguous United
States. For each estimate, we generate a coefficient of variation (CV) equal to the ratio of
the standard error to the point estimate. For tracts with a CV greater than 0.3, we impute
the tract-level estimate with a county-level estimate following Census guidance, which
defines any estimate with a CV greater than 0.3 as low reliability and to be used with
extreme caution (King et al. 2015). In cases of counties with a CV greater than 0.3, we
further impute to the state-level.
373
-------
6.6.1.2 Poverty Status
To determine the poverty status at the tract level, we utilize ACS 5-year estimates
from 2015 to 2019. The resulting datasets represent the fraction of the total population in
the tract that falls below the federal poverty line (termed "below_poverty_line_tract") and
the fraction of the population that falls above the poverty line (termed
"above_poverty_line_tract"). The EPA Standard Variables dataset also includes two
variables representing the fraction of the tract-level population below and above 200% of
the poverty line (termed "below_2x_poverty_line_tract" and "above_2x_poverty_line_tract").
We followed the same imputation procedures described in Section 6.6.1.1 to process
poverty variables.
6.6.1.3 Unemployment Status
BenMAP includes county-level employment variables representing average rates
from 2017 to 2021 (termed "adj_employment_rate" and "adj_unemployment_rate").
Importantly, the employment variables are adjusted for use in BenMAP to use total
population as the denominator rather than labor force. This allows users to multiply the
rates by the total population (in a health impact function) to assess populations that are (a)
employed, (b) unemployed, and (c) not part of the labor force (e.g., retirees, students,
discouraged workers). County-level unemployment rates are from the Bureau of Labor
Statistics. We adjusted these rates using county-level population estimates from the U.S.
Census Bureau from 2017 to 2021. The calculations were done for each year individually
and then averaged together to create a five-year average. The value calculated in BenMAP
represents the average rate across all months within this period. The rate of residents not
in the labor force (termed "not_in_labor_force_rate") was calculated by subtracting the
labor force from the total population and then dividing the remainder by the total
population.
6.6.1.4 Health Insurance Status
We use data from the Small Area Health Insurance Estimates (SAHIE) collected by
the U.S. Census Bureau from 2015 to 2019 to calculate the percentage of individuals with
and without health insurance in each county. The SAHIE date provides the number of
374
-------
individuals with and without health insurance by county. Calculations were done for each
year individually and then averaged together over the five-year period.
6.6.1.5 Linguistic Isolation
We use 5-year ACS estimates from 2015-2019 to estimate linguistic isolation at the
tract level. The resulting datasets represent the fraction of the total population age 5 and
over in the tract that are linguistically isolated (defined by speaking English "less than very
well", termed "english_less_than_veiywell_tract") and the fraction of the population age 5
and over that speak English "very well or better" (termed
"english_veiywell_or_better_tract"). This definition follows as closely as possible the
definition of linguistic isolation used in EJ Screen and in other regulatory analyses.
Additionally, we produced an alternative definition in which the linguistically isolated
population is more restrictive, i.e., those who speak English less than "well" (termed
"english_less_than_well_tract"), and inversely, those who speak English "well or better"
(termed "english_well_or_better_tract"). Instead of imputing in cases of uncertainty, we
generate tract-level point estimates of linguistic isolation in an effort to prioritize
geographic specificity in these estimates.
6.6.1.6 Redlined Areas
We use graded census tracts developed by Noelke et al., 2022 from digitized Home
Owners' Loan Corporation (HOLC) residential security maps overlaid onto 2010 Census
tracts. Each census tract is classified as being covered by "Mainly A," "Mainly B," "Mainly C,"
and "Mainly D" grading, corresponding to coverage of different hazard ratings from original
HOLC maps. The dataset covers 14,818 census tracts, since HOLC maps only covered
certain urban areas. This dataset was adapted to cover 72,538 census tracts for use in
BenMAP, with the remaining census tracts categorized as "redlined_na" since they were not
covered by HOLC grading. Census tracts labeled as "Mainly D" were categorized as
"redlined" and census tracts that were mainly A-C were categorized as "not_redlined."
6.6.2 EJ Health Effects Analysis Input Data
The health assessment requires input data in addition to the information used in the
exposure assessment (Section 6.3). As such, there are additional uncertainties related to
375
-------
the HRs and baseline incidence data used, albeit similar to the benefits assessment results
(Section 5.3).
When selecting HRs to use when estimating race/ethnicity-stratified mortality
effects, we first evaluated the available studies and concentration-response functions to
determine if sufficient information exists for use in a quantitative analysis and to determine
which study or studies best characterizes at-risk nonwhite populations across the U.S. from
the PM ISAs. Of the available studies, Di et al., 2017 and Pope III et al., 2019 were identified
as best characterizing populations potentially at increased risk of long-term exposure and
all-cause mortality, in order to provide two independent estimates of race- and ethnicity-
stratified all-cause mortality impacts of this final rulemaking. Factors contributing to their
selection include that were nationwide studies, included substantial study sizes (i.e.,
hundreds of thousands to millions of individuals), used sophisticated exposure estimation
techniques over fairly recent time spans, and provided sufficient information to apply risk
models quantifying increased risks to non-White/non-Hispanic White groups.
Health impact functions used in this EJ health effects analysis, including beta
parameters and standard errors (SE), were developed for each at-risk population
demographic described by Di et al., 2017 and Pope III et al., 2019 are provided in Table 6-2.
Di et al., 2017 effect estimates were derived from a cohort aged 65 and older and provided
HRs for reference, White, Hispanic, Black, Asian, and Native American populations. The
Pope III et al., 2019 effect estimates were derived from a cohort aged 18 and over and
provided HRs for reference, non-Hispanic White, non-Hispanic Black, Hispanic, and
other/unknown populations. In this analysis, the other/unknown HRs were applied to non-
Hispanic Asian, and non-Hispanic Native American populations.
376
-------
Table 6-2 Hazard Ratios, Beta Coefficients, and Standard Errors (SE) from (Di et
al, 2017)
Study
Demographic Population
Risk of Death Associated with
3
a 10 jig/m Increase in PM^
Beta Coefficient (SE)
Di et al.,
All
1.073 (1.071,1.075)
0.0070 (0.0001)
2017
White
1.063 (1.060,1.065)
0.0061 (0.0001)
Hispanic
1.116 (1.100,1.133)
0.0110 (0.0008)
Black
1.208 (1.199,1.217)
0.0189 (0.0004)
Asian
1.096 (1.075,1.117)
0.0092 (0.0010)
Native American
1.100 (1.060,1.140)
0.0095 (0.0019)
Pope III
et al.,
2019
All
Non-Hispanic White
Non-Hispanic Black
Hispanic
Other/Unknown
1.12 (1.08-1.15)
1.11 (1.07-1.15)
1.15 (1.05-1.27)
1.20 (1.11-1.30)
1.10 (0.94-1.28)
0.0113 (0.0016)
0.0104 (0.0018)
0.0139 (0.0048)
0.0182 (0.0040)
0.0095 (0.0078)
BenMAP-CE includes baseline incidence rates at the most geographically- and age-
specific levels available for each health endpoint assessed. For many locations within the
U.S., these data are resolved at the county- or state-level, providing a better
characterization of the geographic distribution of mortality rates than the national-level
rates. Race- and ethnicity-stratified baseline incidence rates from 2007-2016 Census data
were recently improved for the all-cause mortality health endpoint, by adding the
geographic level option of rural/urban state between county-level and state-level. Both
overall and race/ethnicity-stratified baseline rates are used in this analysis of EJ health
impacts analysis.
To estimate race-stratified and age-stratified incidence rates at the county level, we
downloaded all-cause mortality data from 2007 to 2016 from the CDC WONDER mortality
database.19 Race-stratified incidence rates were calculated for the following age groups: < 1
year, 1-4 years, 5-14 years, 15-24 years, 25-34 years, 35-44years, 45-54 years, 55-64
years, 65-74 years, 75-84 years, and 85+ years. To address the frequent county-level data
suppression for race-specific death counts, we stratified the county-level data into two
broad race categories, White and Non-White populations. In a later step, we stratified the
19 https://wonder.cdc.gov/
377
-------
non-White incidence rates by race and ethnicity using the relative magnitudes of incidence
values by race at the regional level, described in more detail below.
We followed methods outlined in Section D.l.l of the BenMAP-CE User Manual with
one notable difference in methodology; we included an intermediate spatial scale between
county and state for imputation purposes.20 We designated urban and rural counties within
each state using CDC WONDER and, where possible, imputed missing data using the state-
urban and state-rural classifications before relying on broader statewide data. We followed
methods for dealing with suppressed and unreliable data at each spatial scale as described
in Section D.l.l.
A pooled non-White incidence rate masks important differences in mortality risks
by race. To estimate county-level mortality rates by individual race (Black, Asian, Native
American), we applied regional race-specific incidence relationships to the county-level
pooled non-White incidence rates. We calculated a weighted average of race-specific
incidence rates using regional incidence rates for each region/age/race group normalized
to one reference population (the Asian race group) and county population proportions
based on race-specific county populations from CDC WONDER where available. In cases of
population suppression across two or more races per county, we replaced all three race-
specific population proportions derived from CDC WONDER with population proportions
derived from 2010 Census data in BenMAP-CE (e.g., 50 percent Black, 30 percent Asian, 20
percent Native American).
To estimate ethnicity-stratified and age-stratified incidence rates at the county level,
we downloaded all-cause and respiratory mortality data from 2007 to 2016 from the CDC
WONDER mortality database.21 Ethnicity-stratified incidence rates were calculated for the
following age groups: < lyear, 1-4 years, 5-14 years, 15-24 years, 25-34 years, 35-44 years,
45-54 years, 55-64 years, 65-74 years, 75-84 years, and 85+ years. We stratified county-
level data by Hispanic origin (Hispanic and non-Hispanic). We followed the methods
outlined in Section D.l.l to deal with suppressed and unreliable data. We also included an
intermediate spatial scale between county and state designating urban and rural counties
20 https://www.epa.gov/sites/default/files/2015-04/documents/benmap-ce_user_manual_march_2015.pdf
21 https://wonder.cdc.gov/
378
-------
for imputation purposes, described in detail in Section D.1.3 of the BenMAP-CE User
Manual.22
6.6.3 National EJ Analysis of Total Exposures and Exposure Changes Associated with
Meeting the Revised and Alternative Standard Levels
In addition to air quality surfaces associated with the illustrative emissions control
strategies evaluated in the main EJ chapter, PM2.5 air quality surfaces associated with
meeting the current standards and alternative standard levels were also developed. Air
quality associated with meeting the standards was based on assumptions that emissions
controls could be identified to obtain the estimated emissions reductions needed
(Appendix 2A). Results for both air quality scenarios (termed "controls" when associated
with the illustrative control strategies and "standards" when it is assumed areas are fully
meeting the current standards or alternate standard levels) are included in this appendix,
to allow for direct comparisons. In general, for populations experiencing higher baseline
PM2.5 concentrations and mortality rates, air quality scenarios associated with meeting the
current and alternative standard levels reduce disparities more so than air quality
scenarios associated with the control strategies, especially for populations in CA.
National and regional PM2.5 concentrations and concentration changes by
demographic populations for air quality scenarios associated with both the control
strategies and meeting the standards are provided in Sections 6.6.5 and 6.6.6.
National absolute PM2.5 exposure impacts are provided in Section 6.6.3.1 and
national proportional PM2.5 exposure impacts are provided in Section 6.6.3.2
6.6.3.1 Absolute National Exposures Under Current, Revised, and Alternative
Standard Levels and Exposure Changes When Moving from Current Standard
to Revised and Alternative Standard Levels
At the national level, air quality scenarios associated with meeting the standards led
to similar and/or slightly lower PM2.5 concentrations under the current standards and
lower alternative standard levels than air quality scenarios associated with control
strategies (Figure 6-21 and Figure 6-22). This may narrow disproportionate PM2.5
concentrations for certain populations, such as Hispanics, under air quality associated with
22 https://www.epa.gov/sites/default/files/2015-04/documents/benmap-ce_user_manual_march_2015.pdf
379
-------
more stringent alternative standard levels. Please note, some populations are excluded
from the distributional figure for visual clarity,
Population
Group Population (Ages)
Controls
00
Standards oi
Controls g
CO
Standards ^
12/35-
10/35
CO
J/> T3
O
4-> T3
C C
o u
£ ™
4J "O
c c
o
-------
Ethnicity
Educational Linguistic
Attainment
Poverty
Status
100%
Population (Ages)
¦ White (0-99)
¦ Black(0-99)
American Indian (0-99)
M Asian (0-99)
¦ Non-Hispanic (0-99)
¦ Hispanic (0-99)
¦ More educated (25-99)
¦ Less educated (25-99)
¦ English "well or better" (0-99)
English < "well" (0-99)
¦ Above the poverty I ine (0-99)
¦ Below poverty line (0-99)
¦ HOLC Grades A-C (0-99)
¦ HOLC Grade D (0-99)
Ungraded by HOLC (0-99)
3 5 7 9 113 5 7 9 llj3 5
PM2.5 (^g/m3)PM2.s (ng/m3)PM2.
7 9 llj3 5 7 9 1135 7 9 11,3 5 7 9 11
(ng/m3)PM2.5 (ng/m3)PM2.5 (tig/m3)PM2.s (ug/m3)
Figure 6-22 National Distributions of Annual PM2.5 Concentrations Associated
Either with Control Strategies or with Meeting the Standards by
Demographic for Current, Revised, and Alternative PM NAAQS Levels
in 2032
381
-------
Race
Ethnicity
Educational
Attainment
Linguistic
Isolation
Poverty
Status
Redlined
Areas
-10/35
Controls
Population
K
cn 0
000
vO -P N0
12/35
Standards
Population
M
in 0
000
V.O vO -0
0s 0s 0s
-10/30
Controls
Population
l-»
in 0
000
•S.O sO sp
O"" o""-
r*
12/35
Standards
Population
H
in 0
000
--9 2? -9
cy* 0s
La
5-9/35
Controls
Population
R
in 0
000
vO >P
| -3 50%
05 Q.
-M O
-------
shift in higher annual PM exposures associated with "Standards" as opposed to associated
with "Controls" (Figure 6-24).
Reference
100%
80%
.o 60%
Controls -5
Q.
8. 40%
20%
0%
100%
Standards -s
.0 60%
+->
ro
Q.
£ 40%
12/35
10/35
10/30
9/35
8/35
4.0 5.0 6.0
7.0 8.0 9.0 10.0 11.0 12.0
PM2.5 (ng/m3) *
Figure 6-24 National Distributions of Annual Concentrations Experienced by the
Reference Population Associated Either with Control Strategies or
with Meeting the Standards for Current PM NAAQS of 12/35 in 2032
383
-------
6.6.3.2 Proportional Regional Exposure Changes When Moving from Current to
Revised and Alternative Standard Levels
Proportionally, national air quality scenarios associated with meeting the standards
reduce PM2.5 concentrations in the reference population by a larger amount than air quality
scenarios associated with the illustrative control strategies as alternative standard levels
are lowered (Figure 6-25).
384
-------
Scenario
Population
Groups
Populations (Ages)
Attainment
12/35-10/35 12/35-10/30
12/35-9/35
12/35-8/35
Reference
All (0-99)
Controls
0.7
0.9
1.9
3.9
Standards
1.4
1.7
3.3
6.7
Race
White (0-99)
Controls
0.7
0.9
1.8
3.7
Standards
1.4
1.7
3.2
6.5
American Indian (0-99)
Controls
0.8
1.0
1.7
3.7
Standards
1.7
2.1
3.5
6.9
Asian (0-99)
Controls
1.3
1.5
3.1
5.7
Standards
2.4
2.8
5.5
10.4 m
Black (0-99}
Controls
0.5
0.5
1.7
4.2
Standards
0.9
0.9
2.3
5.7
Ethnicity
Non-Hispanic (0-99)
Controls
0.5
0.6
1.5
3.5
Standards
0.9
1.1
2.3
5.2
Hispanic (0-99)
Controls
1.5
1.7
2.8
5.1
Standards
3.2
3.6
6.3
11.2
Educational
More educated (25-99)
Controls
0.7
0.8
1.8
3.8
Attainment
Standards
1.3
1.5
3.0
6.2
Less educated (25-99)
Controls
1.2
13
2.4
4.5
Standards
2.3
2.7
4.8
8.9 ¦
Employment
Status
Employed (0-99)
Controls
0.7
0.8
1.8
4.0
Standards
1.4
1.6
3.2
6.6
Unemployed (0-99)
Controls
1.0
1.1
2.2
4.4
Standards
1.9
2.2
4.1
7.9
Not in the labor force (0-99)
Controls
0.7
0.9
1.9
3.9
Standards
1.5
1.8
3.4
6.8
Insurance
Insured (0-64)
Controls
0.8
0.9
1.9
4.0
Status
Standards
1.5
1.8
3.4
6.9
Unisured (0-64)
Controls
0.8
0.9
1.9
4.3
Standards
1.4
1.7
3.4
Linguistic
Isolation
English "well or better" (0-99)
Controls
0.7
0.8
1.8
3.8
Standards
1.3
1.6
3.1
6.4
English < "well" (0-99)
Controls
1.8
1.9
3.2
5.6
Standards
3.3
3.7
6.7
11.8 ¦
Poverty
Status
Above the poverty line (0-99)
Controls
0.7
0.8
1.8
3.9
Standards
1.4
1.7
3.2
6.6
Below poverty line (0-99)
Controls
0.9
1.0
2.0
4.1
Standards
1.7
1.9
3.7
7.3
Red lined
HOLC Grades A-C (0-99)
Controls
1.8
1.9
3.5
5.9
Areas
Standards
2.8
2.9
5.6
io.o m
HOLC Grade D (0-99)
Controls
1.7
1.8
3.4
6.2
Standards
2.6
2.6
5.2
10.0
Ungraded by HOLC (0-99)
Controls
0.5
0.6
1.5
3.5
Standards
1.2
1.5
2.8
6.0
Age
Adults (18-64)
Controls
0.8
0.9
1.9
4.0
Standards
1.5
1.7
3.4
6.9
Children (0-17)
Controls
0.7
0.9
1.9
4.0
Standards
1.5
1.8
3.4
6.9
Older Adults (65-99)
Controls
0.7
0.8
1.7
3.6
_Standards
1.3
1.5
2.9
6.0
Sex
Females (0-99)
Controls
0.7
0.9
1.9
3.9
Standards
1.4
1.7
3.3
6.7
Males (0-99)
Controls
0.7
0.9
1.8
3.9
Standards
1.4
1.7
3.3
6.7
Figure 6-25 Heat Map of National Percent Reductions (%) in Average Annual PM2.5
Concentrations Associated Either with Control Strategies or with
Meeting the Standard Levels by Demographic When Moving from
Current to Revised and Alternative PM NAAQS Standard Levels in 2032
385
-------
6.6.4 Regional EJ Analysis of Total Exposures and Exposure Changes Associated
with Meeting the Standards
Regional absolute exposure burdens, changes, and proportional changes are in
Sections 6.6.4.1, 6.6.4.2, and 6.6.4.3, respectively.
6.6.4.1 Absolute Regional Exposures Under Current, Revised, and Alternative
Standard Levels
Comparing 'Controls' and 'Standards' in CA associated with the lower alternative
standard levels allows for some insight into areas without identified emissions control
strategies. Regionally, air quality scenarios associated with meeting the standards also led
to similar or slightly lower PM2.5 exposure burdens as air quality scenarios associated with
the current standards or more stringent standard levels, except for in CA, where air quality
associated with the standards resulted in lower PM2.5 concentrations (Figure 6-26 and
Figure 6-27).23 In other words, disparities between Hispanics and non-Hispanics predicted
with controls at 9/35 are mitigated to a greater extent if CA were to fully meet the revised
standard level of 9/35.
23 The overall reference, ages, and sex population groups were excluded from Figure 6-30 to so that the figure
could fit on a single page.
386
-------
Population
Group
12/35 10/35
Population (Ages) Attainment NE SE W Ca|nE SE W CA NE
10/30 9/35 8/35
SE W CA NE SE W CA NE SE W CA
Reference
All (0-99)
Controls
6.?
7.1
6.7
9.3
6.7
7.1
6.7
8.9
6.7
7.1
6.7
8.9
6.6
7.0
6.7
8.8
6.5
5.9
Standards
6.7
7.1
6.7
9.3
6.7
7.1
6.7
8.5
6.7
7.1
6.7
8.4
6.6
7.0
6.7
7.9
6.5
6.8
Race
White (0-99)
Controls
6.6
7.1
6.7
9.2
6.6
7.0
6.7
8.9
6.6
7.0
6.7
3.8
6.5
7.0
6.7
8.7
6.4
6.8
Standards
6.6
7.1
6.7
9.2
6.6
7.0
6.7
8.4
6.6
7.0
6.7
8.3
6.5
7.0
6.7
7.8
6.4
6,8
American Indian
Controls
6.6
7.1
5.5
9.1
6.5
7.1
5.5
8.8
6.5
7.1
5.5
8.7
6.5
7.1
5.5
8.6
6.4
6.9
(0-99)
Standards
6.6
7.1
5.5
9.1
6.5
7.1
5.5
8.3
6.5
7.1
5.5
8.2
6.5
7.1
5.5
7.7
6.4
6.9
Asian (0-99)
Controls
7.1
7.5
7.0
9.4
7,
7.5
7.0
9.1
7.1
7.5
7.0
9.0
7.0
7.4
7.0
8.8
6.8
7.1
Standards
7.1
7.5
7.0
9.4
7,
7.5
7.0
8.7
7.1
7.5
7.0
8.7
7.0
7.4
7.0
8.1
6.8
7.1
Black (0-99)
Controls
7.3
7.2
7.1
7.2
7.2
7.1
9.3
7.2
7.2
7.1
9.2
7.1
7.1
7.0
9.1
6.9
6.9
Standards
7.3
7.2
7.1
9.7
7.2
7.2
7.1
8.8
7.2
7.2
7.1
8.7
7.1
7.1
7.0
8.1
6.9
6.9
Ethnicity
Non-Hispanic
Controls
6.7
6.9
6.6
9.1
6.6
6.9
6.6
8.7
6.6
6.9
6.6
8.7
6.6
6.9
6.6
a.
6.4
6.7
(0-99)
Standards
6.7
6.9
6.6
9.0
6.6
6.9
6.6
8.4
6.6
6.9
6.6
8.3
6.6
6.9
6.6
7.9
6.4
6.7
Hispanic (0-99)
Controls
7.1
7.7
7.0
9.6
7.1
7.6
7.0
9.1
7.1
7.6
7.0
9.1
7.0
7.5
7,0
9.0
6.8
7.3
Standards
7.1
7.7
7.0
9.5
7.1
7.6
7.0
8.6
7.1
7.6
7.0
8.5
7.0
7.5
7.0
7.9
6.8
7.2
Educational
More educated
Controls
6.7
7.0
6.7
9*
6.7
7.0
6.7
8.9
6.7
7.0
6.7
8.8
6.6
7.0
6.6
8.7
6.5
6.8
Attainment
(25-99)
Standards
6.7
7.0
6.7
9.2
6.7
7.0
6.7
8.5
6.7
7.0
6.6
8.3
6.6
7.0
6.6
7.9
6.4
6.8
Less educated
Controls
6.B
7.2
6.9
9.6
6.8
7.2
6.9
9.1
6.8
7.2
6.9
9.1
6.7
7.1
6.8
9.0
6.6
7.0
(25-99)
Standards
68
7.2
6.9
9.5
6.8
7.2
6.9
8.6
6.8
7.2
6.9
8.5
6.7
7.1
6.8
7.9
6.6
6.9
Employment Employed (0-99)
Controls
6.7
7.1
6.7
9.j
6.7
7.1
6.7
8.S
6.7
7.1
6.7
8.9
6.6
7.1
6.7
8.3
6.5
6.9
Status
Standards
6.7
7.1
6.7
9.3
6.7
7.1
6.7
8.5
6.7
7.1
6.7
8.4
6.6
7.1
6.7
7.9
6.5
6.9
Unemployed
Controls
6.8
7.2
6.8
9.4
6.8
7.2
6.8
9.0
6.8
7.2
6.8
9.0
6.7
7.1
6.7
8.9
6.5
6.9
(0-99)
Standards
6.8
7.2
6.8
9.4
6.8
7.2
6.8
8.5
6.8
7.2
6.7
8.4
6.7
7.1
6.7
7.9
6.5
6.9
Not in the labor
Controls
6.7
7.1
6.7
9.3
6.7
7.1
6.7
8.9
6.7
7.1
6.7
8.9
6.6
7.0
6.7
8.8
6.5
6.8
force(0-99)
Standards
6.7
7.1
6.7
93
6.7
7JL
6.7
8.5
6.7
7.1
6.7
8.4
6.6
7.0
6.7
7.9
6.5
6.8
Insurance
Insured (0-64)
Controls
6.7
7.1
6.8
9.3
6.7
7.1
6.8
9.0
6.7
7.1
6.8
8.9
6.6
7.1
6.7
3.8
6.5
6.9
Status
Standards
6.7
7.1
6.8
9.3
6.7
7.1
6.8
8.5
6.7
7.1
6.7
8.4
6.7
7.1
6.7
7.9
6.5
6.9
Unisured (0-64)
Controls
6.9
7.3
6.7
9.4
6.8
7.3
6.7
9.0
6.8
7.3
6.7
8.9
6.7
7.2
6.6
3.8
6.6
7.0
Standards
6.9
7.3
6.7
B
6.8
7.3
6.7
8.4
6.8
7.3
6.7
8.3
6.7
7.2
6.6
7.8
6.6
6.9
Linguistic
English "well or
Controls
6.7
7JL
6.7
9.2
6.7
7.1
6.7
8.9
6.7
7.1
6.7
8.8
6.6
7.0
6.7
8.7
6.5
6.8
Isolation
better" (0-99)
Standards
6.7
7.1
6.7
M
6.7
1A
6.7
8.5
6.7
7.1
6.7
8.3
6.6
7.0
6.7
7.9
6.5
6.8
English < "well"
Controls
7.2
7.8
7.3
9.8
7.2
7.8
7.3
9.3
7.2
7.8
13
9.2
7.1
7.7
7.2
9.1
7.0
7.4
(0-99)
Standards
7.2
7.8
7.3
9.7
7.2
7.8
7.3
8.8
7.2
7.8
12
8.7
7.2
7.6
7.2
8.1
6.9
7.3
Poverty
Above the poverty Controls
6.7
7.1
6.7
S.3
6.7
7.1
6.7
8.9
6.7
7.1
6.7
8.9
6.6
7.0
6.7
8.7
6.5
6.8
Status
line (0-99)
Standards
6.7
7.1
6.7
9.2
6.7
7.1
6.7
8.5
6.7
7.1
6.7
8.4
6.6
7.0
6.7
7.9
6.5
6.8
Below poverty
Controls
6.9
7.2
6.8
9.5
6.9
12.
6.8
9.1
6.9
7.2
6.8
9.0
6.8
7.1
6.7
8.9
6.6
7.0
line (0-99)
Standards
6.9
12
6.8
9.5
6.9
12
6.8
8.6
6.9
7.2
6.8
8.4
6.8
7.1
6.7
7.9
6.6
6.9
Red lined
HOLC Grades A-C
Controls
7.4
7.9
7.5
10.5
7.4
7.9
7.5
9.6
7.4
7.9
7.5
9.5
7.2
7.8
9.5
7.0
7.6
Areas
(0-99)
Standards
7.4
7.9
7.5
10.5
7.4
7.9
7.5
9.1
7.4
7.9
7.5
9.1
7.2
7.8
7.5
8,
7.0
7.6
HOLC Grade D
Controls
7.7
3.0
7.7
10.4
7.7
8.0
9.7
8.0
i
7.5
7.9
7.7
9.5
7.2
7.7
(0-99)
Standards
7.7
8.0
7.7
10.4
7.7
8.0
9.3
,7
8.0
77
9.2
7.6
7.9
7.7
8.5
12
7.7
Ungraded by
Controls
6.5
7.0
6.7
9.0
6.4
7.0
6.7
8.7
6.4
7.0
6.6
8.7 6.4
7.0
6.6
8.6
6.3
6.8
HOLC (0-99)
Standards
6.5
7.0
6.7
8.9
6.4
7.0
6.7
8.3
6.4
7.0
6.6
8.2
6.4
7.0
6.6
7.7
6.3
6.8
Age
Adults (18-64)
Controls
6.8
IX
6.8
9.3
6.7
7.1
6.8
9.0
6.7
7.1
6.7
3.9
6.6
7.1
6.7
8.8
6.5
6.9
Standards
6.8
7.1
6.8
|
6.7
7.1
6.8
8.5
6.7
7.1
6.7
8.4
6.7
7.1
6.7
7.9
6.5
6.9
Children (0-17)
Controls
6.7
7.2
6.8
9-3
6.7
7.2
6.8
9.0
6.7
7.2
6.7
8.9
6.6
7.1
6.7
8.3
6.5
6.9
Standards
6.7
7.2
6.8
9.3
6.7
12
6.8
8.5
6.7
7.2
6.7
8.4
6.7
7.1
6.7
7.9
6.5
6.9
Older Adults
Controls
6.6
6.9
6.6
9.2
6.6
6.9
6.6
8.8
6.6
6.9
6.5
8.8
6.5
6.9
6.5
3.6
6.4
6.7
(65-99)
Standards
6.6
6.9
6.6
9.1
6.6
6.9
6.6
8.4
6.6
6.9
6.5
8.3
6.5
6.8
6.5
7.8
6.4
6.7
Sex
Females (0-99)
Controls
6.7
7.1
6.7
9.3
6.7
7.1
6.7
8.9
6.7
7.1
6.7
8.9
6.6
7.0
6.7
8.8
6.5
6.9
Standards
6.7
7.1
6.7
9.3
6.7
7.1
6.7
8.5
6.7
7.1
6.7
8.4
6.6
7.0
6.7
7.9
6.5
6.8
Males (0-99)
Controls
6.7
7.1
6.7
9.3
6.7
7.1
6.7
8.5
6.7
7.1
6.7
8.9
6.6
7.0
6.7
8.8
6.5
6.9
Standards
6.7
7.1
6.7
9.2
6.7
7.1
6.7
8.5
6.7
7.1
6.7
8.4
6.6
7.0
6.7
7.9
6.5
6.8
Figure 6-26 Heat Map of Regional Average Annual PM2.5 Concentrations (|ig/m3)
Associated Either with Control Strategies or with Meeting the
Standards by Demographic for Current, Revised, and Alternative PM
NAAQS Standard Levels in 2032
387
-------
Educational
Attainment
Redlined
Areas
Linguistic
Isolation
Poverty
Status
Population (Ages)
¦ White (0-99)
¦ Black (0-99)
American Indian (0-99)
¦ Asian (0-99)
¦ Non-Hispanic (0-99)
¦ Hispanic (0-99)
¦ More educated (25-99)
¦ Less educated (25-99)
¦ English "well or better" (0-99)
English < "well" (0-99)
¦ Above the poverty line (0-99)
¦ Below poverty line (0-99)
¦ HOLC Grades A-C (0-99)
¦ HOLC Grade D (0-99)
. Ungraded by HOLC (0-99)
Ethnicity
Ln
on
c 100%
jrt O
s i
NE £ 3 50%
° 8-
u s.
0%
» c 100%
¦o 2
V- ¦—
TJ OJ
3 50%
CD Q.
-M o
CO Q.
0%
c 100%
JS) o
SE £ -| 50%
° R-
u a.
0%
. - 100%
to ra
"o 3 50%
ru Cl
-M O
CO Q.
o '
100%
W £ 3 50%
° s-
u a
0%
_ 100%
1/5 £
"O o
v_ •—
ro ro
P 3 50%
c 100%
o
CA £ s 50%
0%
100%
50%
6 8 10 6 8 10 | 6 8 10 6 8 10 |6 8 10 6 8 10
PM2.5 (ng/m3)PM2.5 (ng/m3)PM25 (ng/m3)PM2.5 (ng/m3)PM2.5 (|jg/m3)PM2.5 (ng/m3)
Figure 6-27 Regional Distributions of Annual PM2.5 Concentrations Associated
Either with Control Strategies or with Meeting the 12/35 and Revised
9/35 Standard Levels in 2032
6.6.4.2 Absolute Regional Exposure Changes When Moving from Current Standard to
Revised and Alternative Standard Levels
Regional absolute exposure changes are shows as averages (Figure 6-28) and as
distributions (Figure 6-29). Again, absolute changes in exposure are greater associated
with the "Standards" than with the "Controls" associated with 9/35 in CA.
388
-------
12/35-10/35 12/35-10/30 12/35-9/35 12/35-8/35
Group tl0n P°Pulation (A9es) Attainment NE SE W CA NE SE W CA NE SE W CA NE SE W CA
Reference All (0-99) Controls
Standards
0.0 0.0 0.0
0.0 0.0 0.0
0.4
0.8
0.0 0.0 0.0
0.0 0.0 0.0
0.4
0.3
0.1 0.1 0.1
0.1 0.1 0.1
0.5
1.4
0.2 0.2 0.2
0.2 0.3 0.3
0.7
,i
Race White (0-99) Controls
Standards
0.0 0.0 0.0
0.0 0.0 0.0
0.4
0.8
0.0 0.0 0.0
0.0 0.0 0.0
0.4
0.3
0.1 0.1 0.1
0.1 0.1 0.1
0.5
1.4
0.2 0.2 0.2
0.2 0.3 0.3
0.6
,i
American Indian Controls
(0-99) Standards
0.0 0.0 0.0
0.0 0.0 0.0
0.4
0.8
0.0 0.0 0.0
0.0 0.0 0.0
0.4
0.3
0.1 0.0 0.0
0.1 0.0 0.1
0.5
1.3
0.2 0.2 0.2
0.2 0.2 0.2
0.6
,0
Asian (0-99) Controls
Standards
0.0 0.0 0.0
0.0 0.0 0.0
0.4
0.7
0.0 0.0 0.0
0.0 0.0 0.0
0.4
0.7
0.1 0.1 0.0
0.1 0.1 0.0
0.6
1.3
0.3 0.4 0.3
0.3 0.4 0.3
0.8
,i
Black (0-99) Controls
Standards
0.0 0.0 0.0
0.0 0.0 0.0
0.4
0.3
0.0 0.0 0.0
0.0 0.0 0.0
0.5
1.0
0.2 0.0 0.1
0.1 0.0 0.1
0.6
1.6
0.3 0.2 0.3
0.4 0.2 0.4
0.7
«
Ethnicity Non-Hispanic Controls
(0-99) Standards
0.0 0.0 0.0
0.0 0.0 0.0
0.3
0.6
0.0 0.0 0.0
0.0 0.0 0.0
0.4
0.7
0.1 0.0 0.0
0.1 0.0 0.0
0.5
1.2
0.2 0.2 0.2
0.2 0.2 0.2
0.7
1.9
Hispanic (0-99) Controls
Standards
0.0 0.0 0.0
0.0 0.0 0.0
0.4
1.0
0.0 0.0 0.0
0.0 0.0 0.0
0.5
1.1
0.1 0.1 0.1
0.1 0.2 0.1
0.6
1.6
0.3 0.4 0.3
0.3 0.5 0.4
0.7
«
Educational More educated Controls
Attainment (25-99) Standards
0.0 0.0 0.0
0.0 0.0 0.0
0.3
0.7
0.0 0.0 0.0
0.0 0.0 0.0
0.4
0.8
0.1 0.0 0.1
0.1 0.1 0.1
0.5
1.3
0.2 0.2 0.2
0.2 0.2 0.3
0.7
,0
Less educated Controls
(25-99) Standards
0.0 0.0 0.0
0.0 0.0 0.0
0.5
0.3
0.0 0.0 0.0
0.0 0.0 0.0
0.5
1.0
0.1 0.1 0.1
0.1 0.1 0.1
0.6
1.6
0.2 0.3 0.3
0.3 0.3 0.4
0.7
2,
Employment Employed (0-99) Controls
Status Standards
0.0 0.0 0.0
0.0 0.0 0.0
0.4
0.8
0.0 0.0 0.0
0.0 0.0 0.0
0.4
0.3
0.1 0.1 0.1
0.1 0.1 0.1
0.5
1.4
0.2 0.2 0.2
0.2 0.3 0.3
0.7
,i
Unemployed Controls
(0-99) Standards
0.0 0.0 0.0
0.0 0.0 0.0
0.4
0.3
0.0 0.0 0.0
0.0 0.0 0.0
0.5
1.0
0.1 0.1 0.1
0.1 0.1 0.1
0.6
1.5
0.2 0.3 0.3
0.3 0.3 0.3
0.7
,3
Not in the labor Controls
force (0-99) Standards
0.0 0.0 0.0
0.0 0.0 0.0
0.4
0.8
0.0 0.0 0.0
0.0 0.0 0.0
0.4
0.3
0.1 0.1 0.1
0.1 0.1 0.1
0.5
1.4
0.2 0.2 0.2
0.2 0.3 0.3
0.7
,i
Insurance Insured (0-64) Controls
Status Standards
0.0 0.0 0.0
0.0 0.0 0.0
0.4
0.8
0.0 0.0 0.0
0.0 0.0 0.0
0.4
0.3
0.1 0.1 0.1
0.1 0.1 0.1
0.5
1.4
0.2 0.2 0.2
0.2 0.3 0.3
0.7
,i
Unisured (0-64) Controls
Standards
0.0 0.0 0.0
0.0 0.0 0.0
0.4
0.3
0.0 0.0 0.0
0.0 0.0 0.0
0.5
1.0
0.1 0.1 0.1
0.1 0.1 0.1
0.6
1.5
0.3 0.3 0.3
0.3 0.3 0.3
0.7
,3
Linguistic English "well or Controls
Isolation better" (0-99) Standards
0.0 0.0 0.0
0.0 0.0 0.0
0.4
0.8
0.0 0.0 0.0
0.0 0.0 0.0
0.4
0.3
0.1 0.1 0.1
0.1 0.1 0.1
0.5
1.3
0.2 0.2 0.2
0.2 0.2 0.3
0.7
,i
English < "well" Controls
(0-99) Standards
0.0 0.0 0.0
0.0 0.0 0.0
0.5
1.0
0.0 0.0 0.0
0.0 0.0 0.0
0.5
1.1
0.1 0.1 0.1
0.1 0.2 0.1
0.6
1.6
0.3 0.4 0.4
0.3 0.5 0.4
0.7
«
Poverty Above the poverty Controls
Status line (0-99) Standards
0.0 0.0 0.0
0.0 0.0 0.0
0.4
0.8
0.0 0.0 0.0
0.0 0.0 0.0
0.4
0.3
0.1 0.1 0.1
0.1 0.1 0.1
0.5
1.3
0.2 0.2 0.2
0.2 0.3 0.3
0.7
,i
Below poverty Controls
line (0-99) Standards
0.0 0.0 0.0
0.0 0.0 0.0
0.4
0.3
0.3
1.3
0.0 0.0 0.0
0.0 0.0 0.0
0.5
1.0
0.3
1.4
0.1 0.1 0.1
0.1 0.1 0.1
0.6
,5
1.0
¦
0.3
H
0.2 0.2 0.3
0.3 0.3 0.3
0.7
Redlined HOLC Grades A-C Controls
Areas (0-99) Standards
0.0 0.0 0.0
0.0 0.0 0.0
0.0 0.0 0.0
0.0 0.0 0.0
0.2 0.1 0.0
0.2 0.1 0.0
0.4 0.4 0.1
0.4 0.4 0.2
1.1
,0
HOLC Grade D Controls
(0-99) Standards
0.0 0.0 0.0
0.0 0.0 0.0
0.8
1.2
0.0 0.0 0.0
0.0 0.0 0.0
0.8
1.2
0.2 0.0 0.0
0.1 0.0 0.0
0.4 0.3 0.1
0.5 0.3 0.3
1.1
,3
Ungraded by Controls
HOLC (0-99) Standards
0.0 0.0 0.0
0.0 0.0 0.0
0.2
0.6
0.0 0.0 0.0
0.0 0.0 0.0
0.3
0.8
0.1 0.1 0.1
0.1 0.1 0.1
0.4
1.2
0.2 0.2 0.2
0.2 0.3 0.3
0.6
1.9
Age Adults (18-64) Controls
Standards
0.0 0.0 0.0
0.0 0.0 0.0
0.4
0.8
0.0 0.0 0.0
0.0 0.0 0.0
0.4
0.3
0.1 0.1 0.1
0.1 0.1 0.1
0.5 0.2 0.2 0.2
1.4 0.2 0.3 0.3
0.7
,i
Children (0-17) Controls
Standards
0.0 0.0 0.0
0.0 0.0 0.0
0.4
0.8
0.0 0.0 0.0
0.0 0.0 0.0
0.4
0.3
0.1 0.1 0.1
0.1 0.1 0.1
0.5
1.4
0.2 0.3 0.2
0.2 0.3 0.3
0.7
,2
Older Adults Controls
(65-99) Standards
0.0 0.0 0.0
0.0 0.0 0.0
0.4
0.7
0.0 0.0 0.0
0.0 0.0 0.0
0.4
0.8
0.1 0.0 0.1
0.1 0.0 0.1
0.5
1.3
0.2 0.2 0.2
0.2 0.2 0.2
0.7
,0
Sex Females (0-99) Controls
Standards
0.0 0.0 0.0
0.0 0.0 0.0
0.4
0.8
0.0 0.0 0.0
0.0 0.0 0.0
0.4
0.3
0.1 0.1 0.1
0.1 0.1 0.1
0.5
1.4
0.2 0.2 0.2
0.2 0.3 0.3
0.7
2,
Males (0-99) Controls
Standards
0.0 0.0 0.0
0.0 0.0 0.0
0.4
0.8
0.0 0.0 0.0
0.0 0.0 0.0
0.4
0.3
0.1 0.1 0.1
0.1 0.1 0.1
0.5
1.4
0.2 0.2 0.2
0.2 0.3 0.3
0.7
2,
Figure 6-28 Heat Map of National Average Annual Reductions in PM2.5
Concentrations (|ig/m3) Associated Either with Control Strategies or
with Meeting the Standards by Demographic When Moving from
Current to Revised and Alternative PM NAAQS Levels in 2032
389
-------
Race
Ethnicity
Educational
Attainment
Linguistic
Isolation
Poverty
Status
Redlined
Areas
c 100%
0
° m
NE £ -5 50%
0 9-
u s.
0%
r
H
Standards
Population
M
cn 0
000
¦vp Vp Vp
0s 0s 0s
c 100%
CO 0
0
SE £ -5 50%
0 £¦
0 o.
0%
P
f
.-9/35
Standards
Population
M
cn 0
000
sja vp -p
cr- o~- cr-
W=
ro c 100%
CM CO O
H O *£
W £ *5 50%
0 S-
U 0
0%
f
/
/
r
Standards
Population
cn 0
0 0 0
--P vO vp
a-
r>
J
r
c 100%
ui 0
O '+-j
CA £ "5 50%
0 s-
u s.
0%
(
f
r
/
f
71
_ 100%
1/1 5
"O .0
re
"o 3 50%
03 CI
•M O
CO Q_
0%
/
/
/
/
0 1 2 3:0 1 2 3 0 1 2 3 0 1 2 3 0 1 2 30 1 2 3
PM2.5 (|ig/m3)PM2.5 (ng/m3)PM2.5 (ng/m3)PM2.5 (iig/m3)PM2.5 (ng/m3)PM2.s (ng/m3)
Population (Ages)
¦ White (0-99)
¦ Black (0-99)
American Indian (0-99)
¦ Asian (0-99)
¦ Non-Hispanic (0-99)
¦ Hispanic (0-99)
¦ More educated (25-99)
¦ Less educated (25-99)
¦ English "well or better" (0-99)
English < "well" (0-99)
¦ Above the poverty I i ne (0-99)
¦ Below poverty line (0-99)
¦ HOLC Grades A-C (0-99)
¦ HOLC Grade D (0-99)
Ungraded by HOLC (0-99)
Figure 6-29
Regional Distributions of Annual PM2.5 Concentration Reductions
When Moving From 12/35-9/35 Associated Either with Control
Strategies or Meeting the Standards in 2032
6.6.4.3 Proportional Regional Exposure Changes When Moving from Current to
Revised and Alternative Standard Levels
The proportionality of regional changes in demographic-specific PM2.5
concentrations when moving from air quality scenarios associated with meeting the
standards, as opposed to air quality scenarios associated with control strategies, when
390
-------
moving from current standards to more stringent alternative standard levels are provided
in Figure 6-30. Dividing the country into the four regions shows that air quality associated
with the "Standards" in CA would lead to substantially greater proportional PM2.5
concentration reductions under all scenarios evaluated. Also, differences between air
quality scenarios associated with "Controls" are slightly larger than differences associated
with the "Standards" when moving to lower alternative standard levels.
Scenario / Region union
Population
Groups
Reference
Race
Ethnicity
Insurance
Status
Linguistic
Isolation
Poverty
Status
Redlined
Areas
Age
Sex
12/35-10/35
12/35-10/30
12/35-9/35
12/35-8/35
Populations (Ages)
Attainment
NE
SE
w
CA
NE
SE
w
CA
NE
SE
W
CA
NE
SE
W
CA
All (0-99)
Controls
0.2
0.1
0.0
4.0
0.2
0.1
0.4
4.5
1.5
0.8
0.8
5.7
3.3
3.3
3.5
7.2
Standards
0.2
0.1
0.0
8.4
0.2
0.1
0.6
9.6
1.3
1.0
0.8
14.8
3.5
3.6
4.0 1
23.0
White (0-99)
Controls
0.2
0.1
0.0
4.0
0.2
0.1
0.4
4.5
1.4
0.8
0.8
5.5
3.0
3.2
3.4
6.9
Standards
0.2
0.1
0.0
8.6
0.2
0.1
0.6
10.0
1.2
1.0
0.9
15.0
3.2
3.6
3.9 |
23.0|
American Indian (0-99)
Controls
0.1
0.0
0.0
3.9
0.1
0.0
0.3
4.5
1.1
0.6
0.9
5.2
2.6
3.3
3.1
6.4
Standards
0.1
0.0
0.0
8.5
0.1
0.0
0.7
10.0
1.0
0.7
0.9
X4.7|
2.8
3.4
3.5 !
22.3
Asian (0-99)
Controls
0.2
0.0
0.0
3.8
0.2
0.0
0.2
4.2
1.6
1.4
0.6
6.5
3.6
5.3
3.6
8.7
Standards
0.2
0.0
0.0
7.0
0.2
0.0
0.2
7.9
1.3
1.4
0.6
13.6
3.8
5.3
4.0
Black (0-99)
Controls
0.2
0.0
0.0
4.5
0.2
0.0
0.2
5.1
2.2
0.6
1.0
6.3
4.6
3.3
4.9
7.6
Standards
0.2
0.0
0.0
9.4
0.2
0.0
0.3
10.3
1.7
0.6
1.0
16.2
4.9
3.3
5.5 i
24.5
Non-Hispanic (0-99)
Controls
0.2
0.0
0.0
3.5
0.2
0.0
0.4
4.1
1.5
0.5
0.7
5.6
3.2
2.8
3.0
7.5
Standards
0.2
0.0
0.0
6.9
0.2
0.0
0.6
8.3
1.3
0.5
0.7
13.0
3.4
2.8
3.4
21.1
Hispanic (0-99)
Controls
0.1
0.3
0.0
4.6
0.1
0.3
0.3
4.9
1.5
1.6
1.2
5.8
3.7
4.6
4.9
6.9
Standards
0.1
0.3
0.0
10.0
0.1
0.3
0.6
11.2
1.3
2.1
1.2
16.93
4.0
5.9
5.5 |
25.1
More educated (25-99)
Controls
0.2
0.1
0.0
3.8
0.2
0.1
0.4
4.3
1.5
0.7
0.8
5.6
3.2
3.1
3.3
7.3
Standards
0.2
0.1
0.0
7.8
0.2
0.1
0.6
9.1
1.3
0.8
0.8
14.1
3.4
3.4
3.8
22 3
Less educated (25-99)
Controls
0.1
0.2
0.0
4.9
0.1
0.2
0.4
5.3
1.5
1.1
1.1
6.2
3.5
3.6
4.6
7.4
Standards
0.1
0.2
0.0
9.8
0.1
0.2
0.7
11.0
1.3
1.4
1.1
16.7
3.7
4.3
5.2 24.9
Employed (0-99)
Controls
0.2
0.1
0.0
4.0
0.2
0.1
0.4
4.5
1.5
0.8
0.8
5.8
3.2
3.4
3.4
7.3
Standards
0.2
0.1
0.0
8.2
0.2
0.1
0.6
9.5
1.3
0.9
0.8
14.6
3.5
3.6
4.0
Unemployed (0-99)
Controls
0.2
0.1
0.0
4.6
0.2
0.1
0.3
5.0
1.7
1.0
0.8
6.0
3.6
3.7
3.8
7.3
Standards
0.2
0.1
0.0
9.4
0.2
0.1
0.6
10.6
1.4
1.2
0.8
16.0
3.8
4.1
4.4 1
24.2
Not in the labor force {0-99)
Controls
0.2
0.1
0.0
4.0
0.2
0.1
0.4
4.5
1.5
0.8
0.8
5.6
3.3
3.2
3.5
7.1
Standards
0.2
0.1
0.0
8.4
0.2
0.1
0.6
9.7
1.3
1.0
0.9
14.9
3.5
3.6
4.0 |
23.1
Insured (0-64)
Controls
0.2
0.1
0.0
4.0
0.2
0.1
0.4
4.5
1.5
0.8
0.8
5.7
3.3
3.4
3.4
7.2
Standards
0.2
0.1
0.0
8.4
0.2
OJ.
0.6
9.6
1.3
0.9
0.8
14.9
3.5
3.7
4.0 23.0
Unisured (0-64)
Controls
0.2
0.2
0.0
4.6
0.2
0.2
0.4
5.1
1.7
1.2
1.0
5.9
3.7
4.0
4.1
7.1
Standards
0.2
0.2
0.0
9.6
0,2
0.2
0.7
10.9
1.5
1.5
1.0
16.3
4.0
4.7
4.7 j
24.5
English "well or better" (0-99)
Controls
0.2
0.1
0.0
3.9
0.2
0.1
0.4
4.4
1.5
0.8
0.8
5.6
3.2
3.2
3.4
7.2
Standards
0.2
0.1
0.0
8.2
0.2
0.1
0.6
9.5
1.3
0.9
0.8
14.6
3.5
3.5
3.9 22.7
English < "well" (0-99)
Controls
0.1
0.3
0.0
5.1
0.1
0.3
0.2
5.4
1.5
1.8
1.1
6.5
3.8
5.1
5.1
7.6
Standards
0.1
0.3
0.0
10.0
0.1
0.3
0.6
11.0
1.3
2.4
1.1 |
sua
4.1
6.4
5.8 |
¦¦
Above the poverty line (0-99)
Controls
0.2
0.1
0.0
3.9
0.2
0.1
0.4
4.4
1.5
0.8
0.8
5.6
3.2
3.3
3.4
7.2
Standards
0.2
0.1
0.0
8.1
0.2
0.1
0.6
9.4
1.3
0.9
0.8
14.5
3.4
3.6
3.9 |
22.7
Below poverty line (0-99)
Controls
0.2
0.2
0.0
4.6
0.2
0.2
0.4
5.1
1.7
0.9
1.0
6.0
3.6
3.3
4.0
7.2
Standards
0.2
0.2
0.0
9.5
0.2
0.2
0.6
10.7
1.4
1.2
1.0
16.3
3.8
4.0
4.5 |
24.4
HOLC Grades A-C (0-99)
Controls
0.2
0.0
0.0
8.2
0.2
0.0
0.1
8.4
2.3
1.0
0.3
9.4
5.2
4.5
1.4
10.2
Standards
0.2
0.0
0.0
12.7
0.2
0.0
0.1
13.0
2.1
1.0
0.3
20.2
5.5
4.5
2.4 |
28.6
HOLC Grade D (0-99)
Controls
0.2
0.0
0.0
7.3
0.2
0.0
0.2
7.4
2.4
0.5
0.3
9.0
5.8
3.4
1.8
10.2
Standards
0.2
0.0
0.0
11.2
0.2
0.0
0.2
11.5
1.9
0.4
0.3
18.4
5.9
3.4
3.7 ¦¦
Ungraded by HOLC (0-99)
Controls
0.2
0.1
0.0
2.8
0.2
0.1
0.4
3.4
1.2
0.8
0.9
4.6
2.4
3.2
3.6
6.3
Standards
0.2
0.1
0.0
7.1
0.2
0.1
0.6
8.7
1.0
1.0
0.9
13.3
2.6
3.6
4.1 21.3
Adults (18-64)
Controls
0.2
0.1
0.0
4.0
0.2
0.1
0.4
4.5
1.5
0.8
0.8
5.8
3.3
3.4
3.5
7.3
Standards
0.2
O-l
0.0
8.4
0.2
0.1
0.6
9.6
1.3
1.0
0.8
14.9
3.5
3.8
4.1 23.1
Children (0-17)
Controls
0.2
0.1
0.0
4.0
0.2
0.1
0.4
4.5
1.5
0.9
0.8
5.6
3.3
3.5
3.4
7.1
Standards
0.2
0.1
0.0
8.6
0.2
0.1
0.6
9.9
1.3
1.1
0.8
15.2
3.5
4.0
3.9 1
23.4
Older Adults (65-99)
Controls
0.2
0.1
0.0
3.9
0.2
0.1
0.4
4.4
1.4
0.6
0.8
5.6
3.1
2.7
3.3
7.2
Standards
0.2
0.1
0.0
7.9
0.2
0.1
0.7
9.2
1.3
0.7
0.8
—
3.3
2.9
3.8 22.1
Females (0-99)
Controls
0.2
0.1
0.0
4.0
0.2
0.1
0.4
4.5
1.5
0.8
0.8
5.7
3.3
3.3
3.5
7.2
Standards
0.2
0.1
0.0
8.4
0.2
0.1
0.6
9.7
1.3
1.0
0.8
14.9
3.5
3.6
4.0
Males (0-99)
Controls
0.2
0.1
0.0
4.0
0.2
0.1
0.4
4.5
1.5
0.8
0.8
5.7
3.2
3.3
3.5
7.2
Standards
0.2
0.1
0.0
8.3
0.2
0.1
0.6
9.6
1.3
1.0
0.8
14.8
3.4
3.7
4.0 22.9
Figure 6-30 Heat Map of Regional Percent Reductions (%) in Average Annual PM2.5
Concentrations Associated Either with Control Strategies or with
Meeting the Standards by Demographic When Moving from Current to
Revised and Alternative PM NAAQS Standard Levels in 2032
391
-------
6.6.5 National EJ Analysis of Total Mortality Rates and Rate Changes Associated with
Meeting the Standards
National absolute mortality rate burdens and changes in absolute mortality rates for
air quality associated with the "Standards" and "Controls" are provided in Section 6.6.5.1.
National proportional changes in mortality rate burdens are provided in Section 6.6.5.2.
National and regional changes in demographic-specific mortality rates when moving
from current to alternative standard levels under air quality surfaces associated with either
control strategies or meeting the standards levels are provided in Sections respectively.
6.6.5.1 Absolute Mortality Rates Under Current and Alternative Standard Levels and
Mortality Rate Reductions When Moving from Current to Revised and
Alternative Standard Levels
Using concentration-response relationships derived from Di et al., 2017 and Pope III
et al., 2019, absolute mortality rates and mortality rate changes were similar under air
quality scenarios associated with the "Standards" as opposed to the "Controls", except for
mortality rate reductions in the distributional figure for older Asian and Hispanic
populations when moving to 9/35 or 8/35 (Figure 6-31, Figure 6-32, and Figure 6-33). This
is because while reductions can be small when averaged across the contiguous U.S., due to
the inclusion of areas with no PM2.5 air quality improvements, reductions can be
substantial within certain tracts.
392
-------
Study Number
(Ages) Population Group of People
12/35
t/>
-0
2 1
¦M T3
C C
o u
u &
Controls g
U)
Standards t-n
12/35-
10/35
L0
-Q
2 -g
-M T3
C C
o n?
u tn
Controls g
"uT
Standards o
12/35-
10/30
LO
in -o
2 -g
4-> T3
C C
O H3
u &
Controls
LU
Standards 1/1
Controls fo
Standards V
Controls co
ijj
Standards 01
12/35-
8/35
cn
10 -Q
2 -g
-M T3
C C
O H3
u «
Di 2017 Reference 75M
186 186
185 184
1 2
185 184
1 3
183 181
3 5
180 176
7 11
(65-99) White 62M
163 163
162 162
1 2
162 161
1 2
161159
2 4
158 155
5 9
American Indian 1M
151151150 149
1 3
150 148
2 3
149 146
2 5
147 142
5 10
Asian 4M
145 145
142 139
4 7
142 138
4 7
139 133
7 13
136 124
10 23
Black 8M
464 464 462 460
3 5
462 460
3 5
456 453
9 12
445 438
22 29
Hispanic 9M
206 206
203 199
4 8
203 198
4 9
201193
6 15
197 183
10 25
Pope Reference 287M
90 90
90 89
1 1
90 89
1 1
89 88
2 3
87 85
3 5
2019 NH White 167M
104 104 104 103
0 1
104 103
1 1
103 102
1 2
101100
3 5
(18-99) NH American Indian 2M
42 42
42 42
0 0
42 42
0 1
42 42
0 1
41 41
1 2
NH Asian 21M
34 34
33 33
1 1
33 32
1 2
32 31
1 3
32 29
2 5
NH Black 37IVI
103 103 102 102
1 1
102 102
1 1
101100
2 3
99 97
5 6
Hispanic 59M
69 68
67 66
1 3
67 66
1 3
67 64
2 5
65 61
4 9
Figure 6-31 Heat Map of National Average Annual Total Mortality Rates (per 100K
People) Associated Either with Control Strategies or with Meeting the
Standards by Demographic for Current, Revised, and Alternative PM
NAAQS Levels in 2032
393
-------
Di 2017
Pope
2019
12/35
10/35
10/30
9/35
8/35
j/> 80%-
2 60%-
§ 40%-
u 20%-
H
a
/I
Jl
fl
-D 80%-
_g 60%-
S 40%-
ffi 20%-
11
ji
n
U
a
j/> 80%-
2 60%-
§ 40%-
u 20%-
a
i
M
3
-p 80%-
_g 60%-
^ 40%-
tr, 20%-
f
a
a
3
a
10 100
Mortality Rate
(per 100k)
10 100
Mortality Rate
(per 100k)
10 100
Mortality Rate
(per 100k)
10 100
Mortality Rate
(per 100k)
10 100
Mortality Rate
(per 100k)
I White
American Indian
I Asian
I Black
I Hispanic
I Reference
I Reference
I NH Asian
I NH Black
NH American Indian
I NH White
Figure 6-32 National Distributions of Total Mortality Rates (per 100k) Associated
Either with Control Strategies or with Meeting the Standards by
Demographic for Current, Revised, and Alternative PM NAAQS Levels
in 2032 (NH, Non-Hispanic)
394
-------
12/35-10/35
12/35-10/30
12/35-9/35
12/35-8/35
_v> 80%_
Di 2017 £ 60%"
o 40%-
KJ
20%-
r~
r~
r
r
-o 80%-
^ 1 60%-
*± m 40%-
U -M
'E 20%-
JZ
rp
r-
r~
LLJ
u J2 80%-
ra o
ac £ 60%-
Pope £
2019 o 40%-
20%-
r
r
r
r
(/i
-o 80%-
-g 60%-
m 40%-
& 20%-
r~
r
r
r
0 50 100 150
Mortality Rate
(per 100k)
0 50 100 150
Mortality Rate
(per 100k)
0 50 100 150
Mortality Rate
(per 100k)
0 50 100 150
Mortality Rate
(per 100k)
Population Group
¦ White
American Indian
¦ Asian
¦ Black
¦ Hispanic
¦ Reference
¦ NH Asian
¦ NH Black
NH American Indian
¦ NH White
Figure 6-33
National Distributions of Annual Total Mortality Rate Reductions (per
100k) Associated Either with Control Strategies or with Meeting the
Standards by Demographic When Moving from Current to Revised and
Alternative PM NAAQS Levels in 2032
6.6.5.2 Proportional Regional Exposure Changes When Moving from Current to
Revised and Alternative Standard Levels
Proportional reductions when moving to air quality scenarios associated with the
"Standards" led to approximately twice the level of disparity mitigation for most
populations when moving to most lower standard levels (Figure 6-34). Some populations
and scenarios led to even great disparity mitigations, such as Hispanic populations for
12/35-9/35 and 12/35-8/35.
395
-------
Di 2017
(65-99)
(18-99)
Race/Ethnicity
Attainment
12/35-
10/35
12/35-
10/30
12/35-
9/35
12/35-
8/35
Reference
Controls
0.6
0.8
1.7
3.6
Standards
1.2
1.4
2.8
5.8
White
Controls
0.6
0.7
1.5
3.3
Standards
1.1
1.3
2.5
5.3
American Indian
Controls
0.8
1.0
1.5
3.2
Standards
1.7
2.3
3.4
6.4
Asian
Controls
2.6
2.9
4.7
7.1
Standards
4.6
5.1
9.2
Black
Controls
Standards
0.6
1.0
0.6
1.1
2.0
2.7
4.7
6.3
Hispanic
Controls
1.8
2.0
3.0
5.0
Standards
3.8
4.2
7.1
12.2
» Reference
Controls
0.7
0.8
1.7
3.7
Standards
1.2
1.5
2.9
5.9
NH White
Controls
Standards
0.4
0.7
0.5
1.0
1.3
2.0
3.1
4.5
NH American Indian
Controls
Standards
0.4
0.8
0.6
1.3
0.9
1.7
2.3
3.8
NH Asian
Controls
2.4
2.7
4.4
6.9
Standards
4.3
4.7
8.6
14.8
NH Black
Controls
Standards
0.6
1.0
0.6
1.0
2.0
2.6
4.7
6.2
Hispanic
Controls
1.9
2.0
3.1
5.3
Standards
3.8
4.2
7.2
12.5
Figure 6-34 Heat Map of National Percent Reductions (%) in Average Mortality
Rate Reductions Associated Either with Control Strategies or with
Meeting the Standards by Demographic When Moving from Current to
Revised and Alternative PM NAAQS Levels in 2032
6.6.6 Regional National EJ Analysis of Total Mortality Rates and Mortality Rate
Changes Associated with Meeting the Standards
Absolute regional mortality rate burdens are shown in Section 6.6.4.1, absolute
regional mortality rate changes when moving to lower standard levels are in Section
6.6.4.2, and proportional regional mortality rate changes when moving to lower standard
levels are in Section 6.6.4.3.
6.6.6.1 Absolute Regional Mortality Rates Under Current, Revised, and Alternative
Standard Levels
Regional trends are virtually identical when comparing across air quality scenarios
associated with the "Controls " and the "Standards" for the current standards, although at
more stringent standard levels, lower mortality rates are observed in air quality scenarios
associated with the "Standards" (Figure 6-35 and Figure 6-36).
396
-------
12/35
NE SE W
CA
10/35
NE SE W
CA
10/30
NE SE W CA
9/35
NE SE W
CA
8/35
NE SE W CA
Di 2017 Reference Controls
Standards
184 183 172 220
184 183 172 219
184 183 172 212 184 183 172 210
184 183 172 203 184 183 171 200
181 182 171 208 178 179 167 205
182 182 171 189 178 178 166 172
White Controls
Standards
160 161153 198
160 161153 197
160 161 153 191 160 161 152 190 158 160 152 188 156 157 148 185
160 161153 183 160 161152 180 158 160 152 171 156 157 147 156
American Controls
Indian Standards
130 161133
130 161133
202
201
129 161133 195
129 161133 185
129 161 133 193
129 161 132 182
128 160 133 192 126 157 131190
128 160 133 173 126 157 130 159
Asian Controls
Standards
Black Controls
Standards
106 103 128
106 103 128
199
199
106 103 128
106 103 128
191
184
106 103 127 190
106 103 127 182
105 101 127
105 101127
186
171
102 97 124 183
102 97 123 154
467 447 417
467 447 417
596
593
466 447 417
466 447 417
569
541
466 447 416 567
466 447 416 536
456 445 413
458 445 413
559
503
444 435 396 552
443 435 393 456
Hispanic Controls
Standards
157 211181
157 211181
249
248
157 210 181 238
157 210 181 223
157 210 181 237
157 210 180 221
155 208 179 235
155 207 179 207
151 204 174 233
151 200 172 187
Pope Reference Controls
2019 Standards
91 89 82
91 89 82
97 91 89 82
97 91 89 82
94 91 89 82 93 90 89 82
90 91 89 82 89 90 89 82
92 88 87 80 91
84 88 87 79 77
NH White Controls
Standards
100 106 97
100 106 97
138 100 106 97
138 100 106 97
134 100 106 97 133 99 105 97
130 100 106 97 127 99 105 97
132 97 104 95 130
122 97 104 94 112
NH Asian Controls
Standards
23 23 31
23 23 31
60
60
23 23 31
23 23 31
57
55
23 23 31 57
23 23 31 55
22 22 31
22 22 31
56
51
22 21 30 55
22 21 30 46
NH Black Controls
Standards
104 99 79
104 99 79
155 104 99 79
155 104 99 79
148 104 99 79 148 101 99 78
141 104 99 79 140 102 99 78
146 99 96 75 144
131 98 96 74 118
NH American Controls
Indian Standards
40 50 40
40 50 40
60
60
40 50 40
40 50 40
59
56
40 50 40 58
40 50 40 55
40 50 40
40 50 40
58
53
40 49 40 57
40 49 39 50
Hispanic Controls
Standards
49 73 58
49 73 58
86
86
49 73 58
49 73 58
82
77
49 73 57 82
49 73 57 76
48 72 57
48 72 57
81
72
47 71 55 80
47 69 55 65
Figure 6-35 Heat Map of Regional Average Annual Total Mortality Rates (per 100K
People) Associated Either with Control Strategies or with Meeting the
Standards by Demographic for Current, Revised, and Alternative PM
NAAQS Levels in 2032 (NH, Non-Hispanic)
397
-------
Di 2017
12/35 9/35
Pope
12/35
2019
9/35
100%"
80%-
JO
£ 60%-
¦M
o 40%-
20%-
J,
I
Jl
ki
0
ioo%-
v) 80%-
T3
60%-
j§ 40%-
co
20%-
.1
ii
w
m
100%"
80%-
£ 60%-
-M
O 40%-
20%-
S 100%"
w 80%-
¦o
™ 60%-
>
±L <3 40%-
M &
C 20%-
_c
I)
II
fj
u
ft
II
fj
u
LU 100%"
$ „ 80%-
(2 £ 60%-
o 40%-
20%-
fj
r
J
Mf
IT
100%_
to 80%-
"O
.g 60%-
S 40%-
IS)
20%-
fj
r
J]
w
f
100%"
„ 80%-
£ 60%-
¦M
o 40%-
20%-
f)
7
#
100%"
« 80%-
"o
^ 60%-
j§ 40%-
20%-
U
#
#
10 100
Mortality Rate
Reduction (per 100k)
10 100
Mortality Rate
Reduction (per 100k)
10 100
Mortality Rate
Reduction (per 100k)
10 100
Mortality Rate
Reduction (per 100k)
Population Group
¦ White
American Indian
¦ Asian
¦ Black
¦ Hispanic
¦ Reference
¦ NH Asian
¦ NH Black
NH American Indian
¦ NH White
Figure 6-36 Regional Distributions of Total Mortality Rates (per 100k) Associated
Either with Control Strategies or with Meeting the Standards by
Demographic for Current, Revised, and Alternative PM NAAQS Levels
in 2032
398
-------
6.6.6,2 Absolute Regional Exposure Changes When Moving from Current to Revised
and Alternative Standard Levels
Absolute mortality rate reductions per 100k individuals are most notably larger
under air quality scenarios associated with the "Standards" as opposed to with the
"Controls" in CA (Figure 6-37 and Figure 6-38).
Di Reference
2017
White
American
Indian
Asian
Black
Hispanic
Pope Reference
2019
NH White
NH Asian
NH Black
NH American
Indian
Hispanic
Figure 6-37
Controls
Standards
Controls
Standards
Controls
Standards
Controls
Standards
Controls
Standards
Controls
Standards
Controls
Standards
Controls
Standards
Controls
Standards
Controls
Standards
Controls
Standards
Controls
Standards
NE
0.4
0.4
0.3
0.3
0.2
0.2
0.1
0.1
1.3
1.3
0.1
0.1
0.2
0.2
0.2
0.2
0.0
0.0
0.3
0.3
0.0
0.0
0.0
0.0
12/35
SE
0.1
0.1
0.1
0.1
0.0
0.0
0.0
0.0
0.0
0.0
0.9
0.9
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.3
0.3
10/35
w
0.0
0.1
0.0
0.1
0.0
0.0
0.0
0.0
0.0
0.0 I
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
CA
8.7
17.5
7.3
15.2
8.2
17.7
9.5
16.6
31.7
61.9
12.7
27 5
4.0
8.0
4.5
9.0
2.8
4.9
8.1
15.9
1.9
4.1
4.5
9.7
NE
0.4
0.4
0.3
0.3
0.2
0.2
0.1
0.1
1.3
1.3
0.1
0.1
0.2
0.2
0.2
0.2
0.0
0.0
0.3
0.3
0.0
0.0
0.0
0.0
12/35
SE
0.1
0.1
0.1
0.1
0.0
0.0
0.0
0.0
0.0
0.0
0.9
0.9
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.3
0.3
10/30
w
0.7
1.2
0.7
1.1
0.4
1.0
0.2
0.3
0.6
1.1
0.4
0.9
0.3
0.6
0.4
0.7
0.1
0.1
0.1
0.2
0.1
0.3
0.2
0.3
CA
10.0
20.6
8.6
18.4
9.6
20.9
10.3
18.2
35.0
67.3
13.6
30.3
4.5
9.4
5.5
11.6
3.0
5.4
9.0
17.2
2.5
5.3
4.8
10.7
NE
2.9
2.5
2.3
2.0
1.6
1.5
2.0
1.5
12.7
10.2
2.3
2.0
1.5
1.3
1.4
1.3
0.4
0.3
2.8
2 3
0.3
0.3
0.8
0.7
12/35-9/35
12/35-8/35
SE
1.0
1.2
0.8
1.0
0.6
0.7
2.3
2.3
2.4
2.4
3.1
4.6
0.5
0.6
0.4
0.4
0.4
0.4
0.5
0.6
0.1
0.1
1.2
1.7
W
1.3
1.4
1.1
1.3
0.7
0.8
0.6
0.6
3.9
4.0
1.9
1.9
0.6
0.7
0.7
0.8
0.2
0.2
0.8
0.8
0.2
0.2
0.7
0.7
CA
12.4
31.3
10.4
27.2
10.7
30.1
14.3
30.7
43.7
105.5
15.5
45.4
5.7
14.3
6.9
16.9
4.2
9.1
11.3
27.1
2.6
7.2
5.5
16.1
NE
6.0
6.4
4.8
5.0
3.6
3.8
4.5
4.7
26.7
27.7
6.3
6.6
3.1
3.2
3.0
3.1
0.9
0.9
5.9
6.1
0.7
0.8
2.0
2.1
SE
4.8
5.2
3.9
4.3
3.9
3.9
6.9
6.9
13.6
13.5
7.9
12.1
2.4
2.6
2.4
2.4
1.4
1.4
3.0
3.0
1.0
1.0
3.0
4.4
W
5.6
6.5
4.9
5.7
2.7
3.2
4.2
4.6
24.0
27.5
8.3
9.7
2.8
3.2
3.0
3.5
1.0
1.2
4.6
5.3
0.6
0.7
2.8
3.2
CA
16.0
49.1
13.5
42.8
13.0
45.4
18.1
48.4
51.9
18.3
66.4
7.2
22.3
9.4
27.7
5.4
14.4
13.4
40.5
3.3
11.1
6.5
23.4
Heat Map of Regional Average Annual Total Mortality Rate Reductions
(per 100K People) Associated Either with Control Strategies or with
Meeting the Standards by Demographic When Moving from Current to
Revised and Alternative PM NAAQS Levels in 2032 (NH, Non-Hispanic)
399
-------
12/35-9/35
Di 2017 Pope
Controls Standards Controls
2019
Standards
100%"
80%-
NE 60%"
40%-
20%-
r
r
100%"
80%-
60%-
>> SE
40%-
u
"E 20%-
JZ
+j
r
r
r
r
100%"
g 80%"
S. w 60%-
40%-
20%-
V
r ~
r
f
100%"
80%-
CA 60%"
40%-
20%-
y
y
f
f \
0 100 200
Mortality Rate
Reduction (per 100k)
0 100 200
Mortality Rate
Reduction (per 100k)
0 100 200
Mortality Rate
Reduction (per 100k)
0 100 200
Mortality Rate
Reduction (per 100k)
Population Group
¦ White
American Indian
¦ Asian
¦ Black
¦ Hispanic
¦ Reference
¦ NH Asian
¦ NH Black
NH American Indian
¦ NH White
Figure 6-38 Regional Distributions of Average Annual Total Mortality Rate
Reductions (per 100k) Associated Either with Control Strategies or
with Meeting the Standards by Demographic for When Moving from
Current to Revised and Alternative PM NAAQS Levels in 2032
6.6.6.3 Proportional Regional Exposure Changes When Moving from Current to
Revised and Alternative Standard Levels
Proportional changes also demonstrate that mortality rates disparities are expected
to be further reduced on average for all populations in CA under the "Standards", as
opposed to the "Controls", especially under more stringent alternative standard (Figure
6-39).
400
-------
12/35-10/35 12/35-10/30 12/35-9/35 12/35-8/35
Study
Race/Ethnicity
Attainment
NE
SE
w
CA
NE
SE
w
CA
NE
SE
W
CA NE
SE
W
CA
Di 2017
Reference
Controls
0.2
0.1
0.0
3.9
0.2
0.1
0.4
4.5
1.6
0.5
0.7
5.7 3.3
2.6
3.3
7.3
(65-99)
Standards
0.2
0.1
0.0
8.0
0.2
0.1
0.7
9.4
1.4
0.6
0.8
14.3 3.5
2.8
3.8
22.4
White
Controls
0.2
0.1
0.0
3.7
0.2
0.1
0.4
4.3
1.4
0.5
0.7
5.2 3.0
2.4
3.2
6.8
Standards
0.2
0.1
0.0
7.7
0.2
0.1
0.7
9.3
1.3
0.6
0.8
13.8 3.1
2.7
3.7
21.7
American Indian
Controls
0.1
0.0
0.0
4.1
0.1
0.0
0.3
4.8
1.3
0.4
0.6
5.3 2.8
2.4
2.1
6.4
Standards
0.1
0.0
0.0
8.8
0.1
0.0
0.8
10.4
1.2
0.4
0.6
15.0 2.9
2.4
2.4
22.5
Asian
Controls
0.1
0.0
0.0
4.7
0.1
0.0
0.1
5.1
1.8
2.2
0.5
7.1 4.2
6.7
3.3
9.1
Standards
0.1
0.0
0.0
8.4
0.1
0.0
0.2
9.1
1.4
2.2
0.5
15.4 4.4
6.7
3.6
¦24.3
Black
Controls
0.3
0.0
0.0
5.3
0.3
0.0
0.1
5.9
2.7
0.5
0.9
7.3 5.7
3.0
5.8
Standards
0.3
0.0
0.0
10.4
0.3
0.0
0.-3
11.3
2.2
0.5
1.0
17.8 5.9
3.0
6.6
BBS
Hispanic
Controls
0.1
0.4
0.0
S.l
0.1
0.4
0.2
5 5
1.5
1.5
1.1
6.2 4.0
3.7
4.6
Standards
0.1
0.4
0.0
11.1
0.1
0.4
0.5
12.2
1.3
2.2
1.1
18.3 4.2
5.7
5.3
m
Pope 2019
(18-99)
Reference
Controls
0.2
0.1
0.0
4.1
0.2
0.1
0.4
4.7
1.6
0.6
0.8
5.8 3.4
2.7
3.4
7.4
Standards
0.2
0.1
0.0
8.2
0.2
0.1
0.7
9.7
1.4
0.7
0.8
14.7 3.5
2.9
3.9
22.9
Hispanic
Controls
0.1
0.4
0.0
S.3
0.1
0.4
0.3
5.6
1.6
1.6
1.2
6.5 4.1
4.0
4.8
7.6
Standards
0.1
0.4
0.0
11.4
0.1
0.4
0.5
12.6
1.3
2.3
1.2
18.8 4.3
6.0
5.6
27.4
NH American Indian
Controls
0.1
0.0
0.0
3.2
0.1
0.0
0.3
4.2
1.0
0.2
0.5
4.3 2.0
2.1
1.7
5.4
Standards
0.1
0.0
0.0
6.9
0.1
0.0
0.8
8.9
0.8
0.2
0.5
12.1 2.1
2.1
1.9
18.7
NH Asian
Controls
0.1
0.0
0.0
4.6
0.1
0.0
0.2
5.1
1.7
2.0
0.6
7.1 4.0
6.5
3.3
9.0
Standards
0.1
0.0
0.0
8.2
0.1
0.0
0.3
9.1
1.3
2.1
0.6
15.3 4.2
6.5
3.7
24.1
NH Black
Controls
0.3
0.0
0.0
5.2
0.3
0.0
0.1
5.8
2.7
0.6
1.0
7.3 5.7
3.1
5.9
8.6
Standards
0.3
0.0
0.0
10.3
0.3
0.0
0.2
11.1
2.2
0.6
1.0
17.5 5.9
3.1
6.8
NH White
Controls
0.2
0.0
0.0
3.2
0.2
0.0
0.5
4.0
1.4
0.4
0.7
5.0 2.9
2.3
3.1
6.8
Standards
0.2
0.0
0.0
6.5
0.2
0.0
0.7
8.4
1.3
0.4
0.8
12.2 3.1
2.3
3.6
20.0
Figure 6-39 Heat Map of Regional Percent Reductions (%) in Average Mortality
Rate Reductions Associated Either with Control Strategies or with
Meeting the Standards by Demographic When Moving from Current to
Revised and Alternative PM NAAQS Levels in 2032 (NH, Non-Hispanic)
401
-------
6.7 References
Di, Q, Wang, Y, Zanobetti, A, Wang, Y, Koutrakis, P, Choirat, C, Dominici, F and Schwartz, JD
(2017). Air pollution and mortality in the Medicare population. New England Journal of
Medicine 376(26): 2513-2522.
Hollmann, F, Mulder, T and Kalian, JJW, DC: US Bureau of the Census (2000). Methodology
and assumptions for the population projections of the United States: 1999 to 2100
(Population Division Working Paper No. 38). 338.
Mitchell, B and Franco J. (2018). HOLC "redlining" maps: The persistent structure of
segregation and economic inequality. National Community Reinvestment Coalition,
https://ncrc.org/?s=redlining+maps, https://ncrc.org/wp-
content/uploads/dlm_uploads/2018/02/NCRC-Research-HOLC-10.pdf.
Lee, EK, Donley, G, Cielielski, T, Gill, I, Yamoah, O, Roche, A, Martinez, R, and Freedman, D.
Health outcomes in redlined versus non-redlined neighborhoods: A systematic review
and meta-analysis. Social Science & Medicine. 294 (2022) 114696.
Noelke, C., Outric, M., Baek, M., Reece, J., Osypuk, T.L., McArdle, N., Ressler, R.R.,
AcevedoGarcia, D. 2022. Connecting past to present: Examining different approaches to
linking historical redlining to present day health inequities. PloS One, 17(5),
e0267606e0267606. https://doi.org/10.1371/journal.pone.0267606
Swope, C, Hernandez, D, and Cushing, L. (2022). The Relationship of Historical Redlining
with Present-Day Neighborhood Environmental and Health Outcomes: A Scoping
Review and Conceptual Model. Journal of Urban Health 99: 959-983.
U.S. EPA (2015). Guidance on Considering Environmental Justice During the Development
of Regulatory Actions, https://www.epa.gov/sites/default/files/2016-
06/documents/ejtg_5_6_16_v5.1.pdf.
U.S. EPA (2019). Integrated Science Assessment (ISA) for Particulate Matter (Final Report).
U.S. Environmental Protection Agency, Office of Research and Development, Center for
Public Health and Environmental Assessment. Research Triangle Park, NC. U.S. EPA.
EPA/600/R-19/188. December 2019. Available at:
https://www.epa.gov/naaqs/particulate-matter-pm-standards-integrated-science-
assessments-current-review.
U.S. EPA (2020). Policy Assessment for the Review of the National Ambient Air Quality
Standards for Particulate Matter. U.S. Environmental Protection Agency, Office of Air
Quality Planning and Standards, Health and Environmental Impacts Division. Research
Triangle Park, NC. U.S. EPA. EPA-452/R-20-002. January 2020. Available at:
https://www.epa.gov/naaqs/particulate-matter-pm-standards-policy-assessments-
current-review-O.
402
-------
U.S. EPA (2021). Draft Policy Assessment for the Reconsideration of the National Ambient
Air Quality Standards for Particulate Matter. U.S. Environmental Protection Agency,
Office of Air Quality Planning and Standards, Health and Environmental Impacts
Division. Research Triangle Park, NC. U.S. EPA. EPA-452/P-21-001. October 2021.
Available at: https://www.epa.gov/system/files/documents/2021-10/draft-policy-
assessment-for-the-reconsideration-of-the-pm-naaqs_october-2021_0.pdf.
U.S. EPA (2022a). Supplement to the 2019 Integrated Science Assessment for Particulate
Matter (Final Report). U.S. Environmental Protection Agency, Office of Research and
Development, Center for Public Health and Environmental Assessment. Research
Triangle Park, NC. U.S. EPA. EPA/600/R-22/028. May 2022. Available at:
https://www.epa.gov/isa/integrated-science-assessment-isa-particulate-matter.
U.S. EPA (2023). Environmental Justice Mapping and Screening Tool, EJScreen Technical
Documentation for Version 2.2. U.S Environmental Protection Agency, Office of
Environmental Justice and External Civil Rights. Washington, DC. June 2023. Available
at: https://www.epa.gov/system/files/documents/2023-06/ejscreen-tech-doc-
version-2-2.pdf.
Woods & Poole (2015). Complete Demographic Database.
https://www.woodsandpoole.com/our-databases/united-states/cedds/
403
-------
CHAPTER 7: LABOR IMPACTS
Overview
This chapter discusses baseline employment in some of the industries potentially
affected by this rule. As economic activity shifts in response to a regulation, typically there
will be a mix of declines and gains in employment in different parts of the economy over
time and across regions. To present a complete picture, an employment impact analysis
will describe the potential positive and negative changes in employment levels. Significant
challenges arise, however, when trying to evaluate the employment effects due to an
environmental regulation and separate them from employment effects due to a wide
variety of other concurrent economic changes, including such important macroeconomic
events as the coronavirus pandemic, or the state of the macroeconomy generally. Despite
these challenges, the economics literature provides a constructive framework and
empirical evidence that sheds light on the labor impacts of environmental regulation. To
simplify, we focus on potential impacts on labor demand related to compliance behavior.
Environmental regulation may also have important effects on labor supply through
changes in worker health and productivity (Graff Zivin and Neidell, 2018).
7.1 Labor Impacts
Economic theory of labor demand indicates that employers affected by
environmental regulation may increase their demand for some types of labor, decrease
demand for other types, or for still other types, not change it at all (Morgenstern et al. 2002,
Deschenes 2018, Berman and Bui 2001). To study labor demand impacts empirically, a
growing literature has compared employment levels at facilities subject to an
environmental regulation to employment levels at similar facilities not subject to that
environmental regulation; some studies find no employment effects, and others find
significant differences. For example, see (Berman and Bui 2001), (Greenstone 2002),
(Ferris, Shadbegian and Wolverton 2014), and (Curtis 2018, 2020).
A variety of conditions can affect employment impacts of environmental regulation,
including baseline labor market conditions and employer and worker characteristics such
as occupation and industry. This baseline labor analysis is illustrative and focused on
404
-------
potential labor impacts in the emissions inventory sectors and industries that may apply
end-of-pipe and area source controls, as identified in Chapter 3. We present information on
baseline characteristics of labor markets in the affected emissions inventory sectors: non-
electric generating unit (non-EGU) point, oil and gas point, non-point (area), residential
wood combustion, and area fugitive dust. Baseline information presented includes
employment levels, recent trends in employment, and labor intensity of production. We do
not have detailed information on the industries that may require pollution controls, and in
which states they may be required. Thus, the presentation of nationwide baseline
information is merely suggestive of employment conditions in the industries that might be
affected.
Table 7-1 presents baseline employment for industries that fall into the emissions
inventory sectors of non-EGU point, oil and gas point, residential wood combustion, and
area fugitive dust. The table shows national employment levels in 2021 and the percent
change in employment over the ten years from 2011 to 2021 for the industries, and their
corresponding North American Industry Classification System (NAICS) codes, identified as
potentially affected industries under each emissions inventory sector. Non-EGU point
sources include emissions units in the cement and concrete product manufacturing, basic
chemical manufacturing, pulp, paper, and paperboard mills, iron and steel mills and
ferroalloy manufacturing, non-ferrous metals production and processing, petroleum and
coal products manufacturing, and mining industries. The oil and gas point emissions
inventory sector includes oil and gas extraction. The residential wood combustion
emissions inventory sector reflects HVAC and commercial refrigeration equipment
manufacturing, and hardware, and plumbing and heating equipment supplies merchant
wholesalers as both of those industries include establishments engaged in manufacturing
and repairing heating equipment, including wood stoves, fireplaces, and wood furnaces.
Because potential controls that could reduce fugitive road dust include applying asphalt or
concrete to roadbeds or roadsides, we include asphalt paving, roofing, and saturated
materials under the area fugitive dust emissions inventory sector.
405
-------
Table 7-1 Baseline Industry Employment
Potentially Affected Industries by Emissions
Inventory Sector and by Industry
NAICS
Employment in
2021
(thousands)
Percent Change
in Employment
2011-2021
Non-EGU Point
Cement and Concrete Product Manufacturing
3273
194.6
18
Basic Chemical Manufacturing
3251
147.9
4
Pulp, Paper, and Paperboard Mills
3221
88.2
-19
Iron and Steel Mills and Ferroalloy Manufacturing
3311
80.5
-14
Non-ferrous Metal (except Aluminum) Production and
Processing
3314
57.3
-6
Petroleum and Coal Products Manufacturing
3241
105.4
-6
Mining (except Oil and Gas)
212
177.6
-20
Oil and Gas Point
Oil and Gas Extraction
211
118.3
-32
Residential Wood Combustion
Ventilation, Heating, Air Conditioning and Commercial
Refrigeration Equipment Manufacturing
3334
136.2
5
Hardware, and Plumbing and Heating Equipment
Supplies Merchant Wholesalers
4237
294.5
24
Area Fugitive Dust
Asphalt Paving, Roofing, and Saturated Materials
Manufacturing
32412
N/Aa
N/A
Note: NAICS is North American Industry Classification System. The source of the information is the U.S.
Bureau of Labor Statistics and is available at https://www.bls.gov/emp/data/industry-out-and-emp.htm and
Detailed industries: hours and employment at https://www.bls.gov/productivity/tables/.
a N/A - not available. The U.S. Bureau of Labor Statistics only provides information at the 4-digit NAICS code.
By Standard Industrial Classification (SIC) code, we located information on employment for paving, surfacing
and tamping equipment operators (47-2071), which is briefly discussed below.
Cement and concrete product manufacturing, hardware and plumbing and heating
equipment supplies merchant wholesalers, and mining are the largest industries in terms
of number people employed. The basic chemical manufacturing and ventilation, heating, air
conditioning and commercial refrigeration equipment manufacturing industries also have
high employment. Each of the industries has had different trends in employment over the
past decade. Cement and concrete product manufacturing and hardware and plumbing and
heating equipment supplies merchant wholesalers have had sizable increases in
employment over the past decade, while pulp, paper, and paperboard mills, oil and gas
extraction, and mining have experienced a decline in employment over the last decade.
Under the area fugitive dust emissions inventory sector, potential controls that
could reduce fugitive road dust include applying asphalt or concrete to roadbeds or
406
-------
roadsides, i.e., shoulders. Associated with these controls, the overall employment for
paving, surfacing and tamping equipment operators in 2022 was 41,470.1 The industry
with the highest concentration of employment in paving, surfacing and tamping equipment
operators is highway, street and bridge construction which employs 14,480 workers.
Texas, New York, Florida, California, and Illinois are the states with the highest
employment levels in paving, surfacing and tamping equipment operators.
Understanding the relative use of labor and capital in potentially affected industries
can shed light on potential labor impacts. Many of these manufacturing industries are
capital intensive. We rely on three public sources to get a range of estimates of employment
per output by industry. Two of the public sources are provided by the U.S. Census Bureau:
the Economic Census (EC) and the Annual Survey of Manufacturers (ASM). The EC is
conducted every 5 years and although it was most recently conducted in 2022, the data are
not available yet. The latest set of data from the EC is 2017. The ASM is an annual subset of
the EC and is based on a sample of establishments. The latest set of data from the ASM is
from 2021. Both sets of U.S. Census Bureau data provide detailed industry data, providing
estimates at the 4-digit NAICS level. The data sets provide separate estimates of the
number of employees and the value of shipments at the 4-digit NAICS, which we convert to
a ratio in this analysis. The third public source that gives an estimate of employment per
output by industry is the U.S. Bureau of Labor Statistics (BLS). Table 7-2 provides estimates
of employment per $1 million of products sold by the industry for each data source in
2017$. While the ratios are not the same, they are similar across time for each data source.
Cement and concrete products manufacturing and ventilation, heating, air conditioning and
commercial refrigeration equipment manufacturing appear to be the most labor-intensive
industries.
1 The source of the information is the U.S. Bureau of Labor Statistics and is available at
(https://www.bls.gov/oes/current/oes472071.htm).
407
-------
Table 7-2 Employment per $1 Million Output (2017$) by Industry (4-digit
NAICS)
Source of Estimate
Emissions Inventory Sector and Industry Sector
BLS
Economic
Census
ASM
Non-EGU Point
Cement and Concrete Product Manufacturing
3.22
2.92
2.81
Basic Chemical Manufacturing
0.81
0.68
0.64
Pulp, and Paper, and Paperboard Mills
1.15
1.24
1.33
Iron and Steel Mills and Ferroalloy Manufacturing
0.98
0.97
0.67
Non-ferrous Metals (except Aluminum) Production and
Processing
1.25
1.21
0.94
Petroleum and Coal Products Manufacturing
N/A
0.20
0.31
Mining (except Oil and Gas)
N/A
2.02
N/A
Oil and Gas Point
Oil and Gas Extraction
N/A
0.54
N/A
Residential Wood Combustion
Ventilation, Heating, Air Conditioning and Commercial
Refrigeration Equipment Manufacturing
3.08
3.04
3.23
Hardware, and Plumbing and Heating Equipment
Supplies Merchant Wholesalers
N/A
1.39
N/A
Area Fugitive Dust
Asphalt Paving, Roofing, and Saturated Materials
Manufacturing
N/A
1.12
1.19
Note: N/A - not available. The source of the information is the U.S. Bureau of Labor Statistics: BLS and is
available at https://www.bls.gov/emp/data/industry-out-and-emp.htm.
In general, there are significant challenges when evaluating the employment effects
due to an environmental regulation. Employment effects associated with a regulation must
be assessed while considering a wide variety of dynamic economic and social factors that
also influence employment in the U.S. economy. In addition to these challenges, the EPA
does not have detailed information on the industries that may require pollution controls
for this rule. Thus, the EPA did not estimate potential employment impacts associated with
this rule. However, to provide information about baseline conditions in relevant
employment markets that might experience incremental impacts, this chapter presented
employment levels, trends, and labor intensities of production in potentially affected
industries.
408
-------
7.2 References
Berman, E. and L. T. M. Bui (2001). Environmental regulation and labor demand: evidence
from the South Coast Air Basin. Journal of Public Economics. 79(2): 265-295.
Curtis, E. M. (2018). Who loses under cap-and-trade programs? The labor market effects of
the NOx budget trading program. Review of Economics and Statistics 100 (1): 151-66.
Curtis, E.M. (2020). Reevaluating the ozone nonattainment standards: evidence from the
2004 expansion. Journal of Environmental Economics and Management, 99: 102261.
Deschenes, O. (2018). Environmental regulations and labor markets. IZA World of Labor:
22: 1-10;
Ferris, A. E., R. Shadbegian, A. Wolverton (2014). The effect of environmental regulation on
power sector employment: phase I of the Title IV S02 trading program. Journal of the
Association of Environmental and Resource Economics 1(4): 521-553.
Graff Zivin, J. and M. Neidell (2018). Air pollution's hidden impacts. Science. 359(6371). 39-
40.
Greenstone, M. (2002). The impacts of environmental regulations on industrial activity:
evidence from the 1970 and 1977 Clean Air Act Amendments and the Census of
Manufactures. Journal of Political Economy 110(6): 1175-1219.
Morgenstern, R.D., W.A. Pizer, and J. Shih (2002). Jobs versus the environment: an industry-
level perspective. Journal of Environmental Economics and Management 43: 412-436.
409
-------
CHAPTER 8: COMPARISON OF BENEFITS AND COSTS
Overview
As discussed in Chapter 1, the Agency is revising the current annual PM2.5 standard
to a level of 9 |~ig/m3. The Agency is also retaining the current 24-hour standard of 35
Hg/m3. OMB Circular A-4 requires analysis of one potential alternative standard level more
stringent than the revised standard and one less stringent than the revised standard. In this
Regulatory Impact Analysis (RIA), we are analyzing the revised annual and current 24-hour
standard levels of 9/35 ng/m3, as well as the following less and more stringent alternative
standard levels: (1) a less stringent alternative annual standard level of 10 |~ig/m3 in
combination with the current 24-hour standard (i.e.,10/35 |j,g/m3), (2) a more stringent
alternative annual standard level of 8 |~ig/m3 in combination with the current 24-hour
standard (i.e., 8/35 |j,g/m3), and (2) a more stringent alternative 24-hour standard level of
30 ng/m3 in combination with an annual standard level of 10 ng/m3 (i.e., 10/30 |j,g/m3).
Because the EPA is not changing the current secondary PM2.5 standards at this time, as well
as retaining the primary and secondary PM10 standards, we did not evaluate alternative
levels of those standards in this RIA. The docket for the rulemaking is EPA-HQ-OAR-2015-
0072.
The analyses in this RIA rely on national-level data (emissions inventory and control
measure information) for use in national-level assessments (air quality modeling, control
strategies, environmental justice, and benefits estimation). However, the ambient air
quality issues being analyzed are highly complex and local in nature, and the results of
these national-level assessments therefore contain uncertainty. It is beyond the scope of
this RIA to develop detailed local information for the areas being analyzed, including
populating the local emissions inventory information, obtaining local information to
increase the resolution of the air quality modeling, and obtaining local information on
emissions controls, all of which would reduce some of the uncertainty in these national-
level assessments. For example, having more refined data would be ideal for agricultural
dust and burning, prescribed burning, and non-point (area) sources due to their large
410
-------
contribution to primary PM2.5 emissions and the limited availability of emissions controls.1
In addition, for residential wood combustion emissions, people will respond differently to
the various regulations and incentives offered for controlling PM2.5 emissions from wood
burning, making it important to identify the right balance of controls for each area. The
estimated benefits and costs associated with applying emissions controls are incremental
to a baseline of attaining the current primary annual and 24-hour PM2.5 standards of 12/35
Hg/m3 in ambient air and incorporate air quality improvements achieved through the
projected implementation of existing regulations.
8.1 Results
The EPA prepared an illustrative control strategy analysis to estimate the costs and
human health benefits associated with the control strategies applied toward reaching the
revised and less and more stringent alternative PM2.5 standard levels. The control
strategies presented in this RIA are an illustration of one possible set of control strategies
states might choose to implement toward meeting the revised standard levels. States, not
the EPA, will implement the revised NAAQS and will ultimately determine appropriate
emissions control strategies. This section summarizes the results of the analyses.
As shown in Chapter 4, the estimated costs associated with the control strategies for
the revised standard levels of 9/35 |~ig/m3 are approximately $594 million in 2032 (2017$,
7 percent interest rate).2 As shown in Chapter 5, the estimated monetized benefits
associated with these control strategies for the revised standard levels of 9/35 |~ig/m3 are
approximately $20 billion and $42 billion in 2032 (2017$, based on a real discount rate of 7
percent).3 The benefits are associated with two point estimates from two different
epidemiologic studies discussed in more detail in Chapter 5, Section 5.2.3. It is expected
1 Examples of area source emissions include area fugitive dust, residential wood combustion, and commercial
cooking emissions.
2 When calculating the annualized costs, we prefer to use the interest rates faced by firms; however, we do not
know what those rates are.
3 As indicated in Chapter 5, we assume that there is a "cessation" lag between the change in PM exposures and
the total realization of changes in mortality effects. Specifically, we assume that some of the incidences of
premature mortality related to PM2.5 exposures occur in a distributed fashion over the 20 years following
exposure, which affects the valuation of mortality benefits at different discount rates. Similarly, we assume
there is a cessation lag between the change in PM exposures and both the development and diagnosis of
lung cancer.
411
-------
that some costs and benefits will begin occurring before 2032, as states begin
implementing controls to attain earlier or to show progress towards attainment.
As discussed in Chapter 3, Section 3.2.4, the estimated PM2.5 emissions reductions
from control applications do not fully account for all the emissions reductions needed to
reach the revised and less and more stringent alternative standard levels in some counties
in the northeast, southeast, west, and California. In Chapter 2, Section 2.4 and Chapter 3,
Section 3.2.5, we discuss the remaining air quality challenges for areas in the northeast and
southeast, as well as in the west and California for the revised standard levels of 9/35 |Lxg/-
m3. The EPA calculates the monetized net benefits of the revised and alternative standard
levels by subtracting the estimated monetized compliance costs from the estimated
monetized benefits in 2032. The estimates of costs and benefits do not fully account for all
of the emissions reductions needed to reach the revised and less and more stringent
alternative standard levels. In 2032, the monetized net benefits of the revised standard
levels of 9/35 |~ig/m3 are approximately $22 billion and $46 billion using a 3 percent real
discount rate for the benefits estimates (in 2017$). The benefits are associated with two
point estimates from two different epidemiologic studies discussed in more detail in
Chapter 5, Section 5.2.3. Table 8-1 presents a summary of these impacts for the revised
standard levels and the less and more stringent alternative standard levels for 2032.
Table 8-1 Estimated Monetized Benefits, Costs, and Net Benefits of the Control
Strategies Applied Toward Primary Revised and Alternative Standard
Levels of 10/35 (j,g/m3,10/30 |ig/m3,9/35 (j,g/m3, and 8/35 |ig/m3 in
2032 for the U.S. (millions of 2017$)
10/35
10/30
9/35
8/35
Benefits3
$8,500 and $17,000
$10,000 and $21,000
$22,000 and $46,000
$48,000 and $99,000
Costsb
$200
$340
$590
$1,500
Net Benefits
$8,300 and $17,000
$9,900 and $21,000
$22,000 and $46,000
$46,000 and $97,000
Notes: Rows may not appear to add correctly due to rounding.
We focus results to provide a snapshot of costs and benefits in 2032, using the best available information to
approximate social costs and social benefits recognizing uncertainties and limitations in those estimates.
The estimated PM2.5 emissions reductions from the control strategies do not fully account for all the emissions
reductions needed to reach the revised and less and more stringent alternative standard levels in some
counties in the northeast, southeast, west, and California.
a We assume that there is a cessation lag between the change in PM exposures and the total realization of
changes in mortality effects. Specifically, we assume that some of the incidences of premature mortality related
to PM2.5 exposures occur in a distributed fashion over the 20 years following exposure, which affects the
valuation of mortality benefits at different discount rates. Similarly, we assume there is a cessation lag between
412
-------
the change in PM exposures and both the development and diagnosis of lung cancer. The benefits are associated
with two point estimates from two different epidemiologic studies, and we present the benefits calculated at a
real discount rate of 3 percent. The benefits exclude additional health and welfare benefits that could not be
quantified (see Chapter 5, Sections 5.2.4 and 5.2.5).
b The costs are annualized using a 7 percent interest rate.
As part of fulfilling analytical guidance with respect to E.0.12866, the EPA presents
estimates of the present value (PV) of the monetized benefits and costs over the twenty-
year period 2032 to 2051. To calculate the present value of the social net benefits of the
revised and alternative standard levels, annual benefits and costs are discounted to 2023 at
3 percent and 7 percent discount rates as recommended by OMB's Circular A-4. The EPA
also presents the equivalent annualized value (EAV), which represents a flow of constant
annual values that, had they occurred in each year from 2032 to 2051, would yield a sum
equivalent to the PV. The EAV represents the value of a typical cost or benefit for each year
of the analysis, in contrast to the 2032-specific estimates.
For the twenty-year period of 2032 to 2051, for the revised standard levels of 9/35
Hg/m3 the PV of the costs, in 2017$ and discounted to 2023, is $7 billion when using a 3
percent discount rate and $3.7 billion when using a 7 percent discount rate. The EAV is
$470 million per year when using a 3 percent discount rate and $350 million when using a
7 percent discount rate. The costs in PV and EAV terms for the revised and alternative
standard levels can be found in Table 8-2 and Table 8-3.
413
-------
Table 8-2 Summary of Present Values and Equivalent Annualized Values for
Estimated Monetized Compliance Costs of the Control Strategies
Applied Toward the Primary Revised and Alternative Standard Levels
of 10/35 ng/m3,10/30 (ig/m3, 9/35 ng/m3 8/35 ng/m3 (millions of
2017$, 2032-2051, discounted to 2023, 3 percent discount rate)
Year
10/35
10/30
9/35
8/35
2032
$160
$260
$460
$1,200
2033
$150
$250
$440
$1,100
2034
$150
$250
$430
$1,100
2035
$140
$240
$420
$1,100
2036
$140
$230
$400
$1,000
2037
$130
$220
$390
$990
2038
$130
$220
$380
$960
2039
$130
$210
$370
$940
2040
$120
$210
$360
$910
2041
$120
$200
$350
$880
2042
$120
$190
$340
$860
2043
$110
$190
$330
$830
2044
$110
$180
$320
$810
2045
$110
$180
$310
$780
2046
$100
$170
$300
$760
2047
$100
$170
$290
$740
2048
$97
$160
$280
$720
2049
$94
$160
$280
$700
2050
$91
$150
$270
$680
2051
-ee-
00
vo
$150
$260
$660
Present Value
$2,400
$4,000
$7,000
$18,000
Equivalent
Annualized Value
$160
$270
$470
$1,200
414
-------
Table 8-3 Summary of Present Values and Equivalent Annualized Values for
Estimated Monetized Compliance Costs of the Control Strategies
Applied Toward the Primary Revised and Alternative Standard Levels
of 10/35 ng/m3,10/30 (ig/m3, 9/35 ng/m3 8/35 ng/m3 (millions of
2017$, 2032-2051, discounted to 2023, 7 percent discount rate)
Year
10/35
10/30
9/35
8/35
2032
$110
$180
$320
$820
2033
$100
$170
$300
$760
2034
$96
$160
$280
$710
2035
$90
$150
$260
$670
2036
$84
$140
$250
$620
2037
$79
$130
$230
$580
2038
$73
$120
$220
$540
2039
$69
$120
$200
$510
2040
$64
$110
$190
$480
2041
$60
$100
$180
$440
2042
$56
$94
$160
$420
2043
$52
$88
$150
$390
2044
$49
$82
$140
$360
2045
$46
$77
$130
$340
2046
$43
$72
$130
$320
2047
$40
$67
$120
$300
2048
$37
$63
$110
$280
2049
$35
$59
$100
$260
2050
$33
$55
$96
$240
2051
$30
$51
$89
$230
Present Value
$1,200
$2,100
$3,700
$9,300
Equivalent
Annualized Value
$120
$200
$350
$870
For the twenty-year period of 2032 to 2051, for the revised standard levels of 9/35
Hg/m3 the PV of the benefits, in 2017$ and discounted to 2023, is $540 billion when using a
3 percent discount rate and $290 billion when using a 7 percent discount rate. The EAV is
$36 billion per year when using a 3 percent discount rate and $27 billion when using a 7
percent discount rate. The benefits in PV and EAV terms for the revised and alternative
standard levels can be found in Table 8-4 and Table 8-5.
415
-------
Table 8-4 Summary of Present Values and Equivalent Annualized Values for
Estimated Monetized Benefits of the Control Strategies Applied
Toward the Primary Revised and Alternative Standard Levels of 10/35
(ig/m3,10/30 (ig/m3, 9/35 ng/m3 8/35 ng/m3 (millions of 2017$,
2032-2051, discounted to 2023, 3 percent discount rate)
Year
10/35
10/30
9/35
8/35
2032
$13,000
$16,000
$35,000
$76,000
2033
$13,000
$16,000
$34,000
$73,000
2034
$13,000
$15,000
$33,000
$71,000
2035
$12,000
$15,000
$32,000
$69,000
2036
$12,000
$14,000
$31,000
$67,000
2037
$12,000
$14,000
$31,000
$65,000
2038
$11,000
$13,000
$30,000
$63,000
2039
$11,000
$13,000
$29,000
$62,000
2040
$11,000
$13,000
$28,000
$60,000
2041
$10,000
$12,000
$27,000
$58,000
2042
$10,000
$12,000
$26,000
$56,000
2043
$9,700
$12,000
$26,000
$55,000
2044
$9,400
$11,000
$25,000
$53,000
2045
$9,100
$11,000
$24,000
$52,000
2046
$8,900
$11,000
$23,000
$50,000
2047
$8,600
$10,000
$23,000
$49,000
2048
$8,300
$10,000
$22,000
$47,000
2049
$8,100
$9,700
$21,000
$46,000
2050
$7,900
$9,400
$21,000
$44,000
2051
$7,600
$9,200
$20,000
$43,000
Present Value
Equivalent
Annualized Value
$210,000
$14,000
$250,000
$17,000
$540,000
$36,000
$1,200,000
$78,000
416
-------
Table 8-5 Summary of Present Values and Equivalent Annualized Values for
Estimated Monetized Benefits of the Control Strategies Applied
Toward the Primary Revised and Alternative Standard Levels of 10/35
(ig/m3,10/30 (ig/m3, 9/35 ng/m3 8/35 ng/m3 (millions of 2017$,
2032-2051, discounted to 2023, 7 percent discount rate)
Year
10/35
10/30
9/35
8/35
2032
$9,500
$11,000
$25,000
$54,000
2033
$8,900
$11,000
$24,000
$50,000
2034
$8,300
$10,000
$22,000
$47,000
2035
$7,800
$9,300
$21,000
$44,000
2036
$7,300
$8,700
$19,000
$41,000
2037
$6,800
$8,100
$18,000
$38,000
2038
$6,300
$7,600
$17,000
$36,000
2039
$5,900
$7,100
$16,000
$33,000
2040
$5,500
$6,600
$15,000
$31,000
2041
$5,200
$6,200
$14,000
$29,000
2042
$4,800
$5,800
$13,000
$27,000
2043
$4,500
$5,400
$12,000
$26,000
2044
$4,200
$5,100
$11,000
$24,000
2045
$3,900
$4,700
$10,000
$22,000
2046
$3,700
$4,400
$9,800
$21,000
2047
$3,400
$4,100
$9,100
$19,000
2048
$3,200
$3,900
$8,500
$18,000
2049
$3,000
$3,600
$8,000
$17,000
2050
$2,800
$3,400
$7,400
$16,000
2051
$2,600
$3,200
$7,000
$15,000
Present Value
$110,000
$130,000
$290,000
$610,000
Equivalent
Annualized Value
$10,000
$12,000
$27,000
$57,000
For the twenty-year period of 2032 to 2051, for the revised standard levels of 9/35
|j,g/m3the PV of the net benefits, in 2017$ and discounted to 2023, is $540 billion when
using a 3 percent discount rate and $280 billion when using a 7 percent discount rate. The
EAV is $36 billion per year when using a 3 percent discount rate and $27 billion when
using a 7 percent discount rate. The comparison of benefits and costs in PV and EAV terms
for the revised standard levels can be found in Table 8-6. Estimates in the table are
presented as rounded values.
417
-------
Table 8-6 Summary of Present Values and Equivalent Annualized Values for
Estimated Monetized Compliance Costs, Benefits, and Net Benefits of
the Control Strategies Applied Toward the Revised Primary
Alternative Standard Levels of 9/35 (ig/m3 (millions of 2017$, 2032-
2051, discounted to 2023 using 3 and 7 percent discount rates)
Benefits3
Costsb
Net Benefits
Year
3%
7%
3%
7%
3%
7%
2032
$35,000
$25,000
$460
$320
$35,000
$25,000
2033
$34,000
$24,000
$440
$300
$34,000
$23,000
2034
$33,000
$22,000
$430
$280
$33,000
$22,000
2035
$32,000
$21,000
$420
$260
$32,000
$20,000
2036
$31,000
$19,000
$400
$250
$31,000
$19,000
2037
$31,000
$18,000
$390
$230
$30,000
$18,000
2038
$30,000
$17,000
$380
$220
$29,000
$17,000
2039
$29,000
$16,000
$370
$200
$28,000
$15,000
2040
$28,000
$15,000
$360
$190
$28,000
$14,000
2041
$27,000
$14,000
$350
$180
$27,000
$14,000
2042
$26,000
$13,000
$340
$160
$26,000
$13,000
2043
$26,000
$12,000
$330
$150
$25,000
$12,000
2044
$25,000
$11,000
$320
$140
$25,000
$11,000
2045
$24,000
$10,000
$310
$130
$24,000
$10,000
2046
$23,000
$9,800
$300
$130
$23,000
$9,600
2047
$23,000
$9,100
$290
$120
$22,000
$9,000
2048
$22,000
$8,500
$280
$110
$22,000
$8,400
2049
$21,000
$8,000
$280
$100
$21,000
$7,900
2050
$21,000
$7,400
$270
$96
$21,000
$7,300
2051
$20,000
$7,000
$260
-ee-
00
vo
$20,000
$6,900
Present Value
$540,000
$290,000
$7,000
$3,700
$540,000
$280,000
Equivalent
Annualized Value
$36,000
$27,000
$470
$350
$36,000
$27,000
Notes: Rows may not appear to add correctly due to rounding. The annualized present value of costs and
benefits are calculated over a 20-year period from 2032 to 2051.
¦'The benefits values use the larger of the two avoided premature deaths estimates presented in Chapter 5,
Table 5-5, and are discounted at a rate of 3 percent over the SAB-recommended 20-year segmented lag. The
benefits exclude additional health and welfare benefits that could not be quantified (see Chapter 5, Sections
5.2.4 and 5.2.5).
b The costs are annualized using a 7 percent interest rate.
8.2 Limitations of Present Value Estimates
The net present value (NPV) estimates presented reflect the costs and benefits
associated with the illustrative control strategies; as discussed in Chapter 3, Section 3.2.4,
some areas still need emissions reductions after control applications for the alternative
standards analyzed. Additionally, there are methodological complexities associated with
calculating the NPV of a stream of costs and benefits for national ambient air quality
standards. The estimated NPV can better characterize the stream of benefits and costs over
a multi-year period; however, calculating the PV of improved air quality is generally quite
418
-------
data-intensive and costly. While NPV analysis allows evaluation of alternatives by summing
the present value of all future costs and benefits, insights into how costs will occur over
time are limited by underlying assumptions and data. Calculating a PV of the stream of
future benefits also poses special challenges, which we describe below. Further, the results
are sensitive to assumptions regarding the time period over which the stream of benefits is
discounted.
To estimate engineering costs, the EPA employs the equivalent uniform annual cost
(EUAC) method, which annualizes costs over varying lifetimes of controls applied in the
analysis. Using the EUAC method results in a stream of annualized costs that is equal for
each year over the lifetime of control, resulting in a value similar to the value associated
with an amortized mortgage or other loan payment. Control equipment is often purchased
by incurring debt rather than through a single up-front payment. Recognizing this led the
EPA to estimate costs using the EUAC method instead of a method that mimics firms paying
up front for the future costs of installation, maintenance, and operation of pollution control
devices.
Further, because we do not know when a facility will stop using a control or change
to another measure based on economic or other reasons, the EPA assumes the controls
applied in the illustrative control strategies remain in service for their full useful life. As a
result, the annualized cost of controls in a single future year is the same throughout the
lifetimes of controls analyzed, allowing the EPA to compare the annualized control costs
with the benefits in a single year for consistent comparison.
The theoretically appropriate approach for characterizing the PV of benefits is the
life table approach. The life table, or dynamic population, approach explicitly models the
year-to-year influence of air pollution on baseline mortality risk, population growth and
the birth rate—typically for each year over the course of a 50- to 100-year period (U.S. EPA
SAB, 2010; Miller, 2003). In contrast to the pulse approach that is employed in this
analysis4, a life table models these variables endogenously by following a population cohort
over time. For example, a life table will "pass" the air pollution-modified baseline death rate
4 The pulse approach assumes changes in air pollution in a single year and affects mortality estimates over a
20-year period.
419
-------
and population from year to year; impacts estimated in year 50 will account for the
influence of air pollution on death rates and population growth in the preceding 49 years.
Calculating year-to-year changes in mortality risk in a life table requires some
estimate of the annual change in air quality levels. It is both impractical to model air quality
levels for each year and challenging to account for changes in federal, state, and local
policies that will affect the annual level and distribution of pollutants. For each of these
reasons the EPA does not always report the PV of benefits for air rules but has instead
pursued a pulse approach.
420
-------
8.3 References
Miller BG (2003). Life table methods for quantitative impact assessments in chronic
mortality. Journal of Epidemiology & Community Health, 57(3):200-206.
U.S. EPA Science Advisory Board (2010). Review of EPA's DRAFT Health Benefits of the
Second Section 812 Prospective Study of the Clean Air Act. Washington, DC.
421
-------
United States Office of Air Quality Planning and Standards Publication No. EPA-452/R-24-006
Environmental Protection Health and Environmental Impacts Division January 2024
Agency Research Triangle Park, NC
------- |