^£Dsrx
I _ \
I w-'
Regulatory Impact Analysis for the Final National
Emission Standards for Hazardous Air Pollutants:
Coal- and Oil-Fired Electric Utility Steam Generating
Units Review of the Residual Risk and Technology
Review
U.S. Environmental Protection Agency
Office of Air and Radiation
Office of Air Quality Planning and Standards
Research Triangle Park, NC 27711
-------
EPA-452/R-24-005
April 2024
Regulatory Impact Analysis for the Final National Emission Standards for Hazardous Air
Pollutants: Coal- and Oil-Fired Electric Utility Steam Generating Units Review of the Residual
Risk and Technology Review
U.S. Environmental Protection Agency
Office of Air Quality Planning and Standards
Health and Environmental Impacts Division
Research Triangle Park, NC
-------
CONTACT INFORMATION
This document has been prepared by staff from the Office of Air and Radiation, U.S.
Environmental Protection Agency. Questions related to this document should be addressed to the
Air Economics Group in the Office of Air Quality Planning and Standards, U.S. Environmental
Protection Agency, Office of Air and Radiation, Research Triangle Park, North Carolina 27711
(email: OAQPSeconomics@epa.gov).
-------
TABLE OF CONTENTS
Table of Contents i
List of Tables iv
List of Figures vii
Executive Summary ES-1
ES.l Introduction ES-1
ES.2 Regulatory Requirements ES-2
ES.3 Baseline and Analysis Years ES-5
ES.4 Emissions Impacts ES-6
ES.5 Compliance Costs ES-9
0.6 Benefits ES-12
ES.6.1 Health Benefits ES-12
ES.6.2 Climate Benefits ES-13
ES.6.3 Additional Unqualified Benefits ES-14
ES.6.4 Total Benefits ES-15
ES.7 Environmental Justice Impacts ES-16
ES.8 Comparison of Benefits and Costs ES-19
ES.9 References ES-22
1 Introduction and Background 1-1
1.1 Introduction 1-1
1.2 Legal and Economic Basis for Rulemaking 1-2
1.2.1 Statutory Requirement 1-2
1.2.2 Regulated Pollutants 1-3
1.2.3 The Potential Need for Regulation 1-4
1.3 Overview of Regulatory Impact Analysis 1-4
1.3.1 Regulatory Options 1-4
1.3.2 Baseline and Analysis Years 1-8
1.4 Organization of the Regulatory Impact Analysis 1-9
1.5 References 1-9
2 Industry Profile 2-1
2.1 Background 2-1
2.2 Power Sector Overview 2-1
2.2.1 Generation 2-1
2.2.2 Transmission 2-13
2.2.3 Distribution 2-13
2.3 Sales, Expenses, and Prices 2-14
2.3.1 Electricity Prices 2-15
2.3.2 Prices of Fossil Fuel Used for Generating Electricity 2-17
2.3.3 Changes in Electricity Intensity of the U.S. Economy from 2010 to 2021 2-18
3 Costs, Emissions, and Energy Impacts 3-1
3.1 Introduction 3-1
3.2 EPA's Power Sector Modeling Platform 2023 using IPM 3-1
3.3 Baseline 3-4
3.4 Regulatory Options Analyzed 3-5
3.5 Power Sector Impacts 3-8
3.5.1 Emissions 3-8
3.5.2 Compliance Costs 3-11
3.5.3 Projected Compliance Actions for Emissions Reductions 3-16
3.5.4 Generating Capacity 3-17
l
-------
3.5.5 Generation Mix 3 -20
3.5.6 Coal and Natural Gas Use for the Electric Power Sector 3-22
3.5.7 Fuel Price, Market, and Infrastructure 3-23
3.5.8 Retail Electricity Prices 3 -24
3.6 Limitations of Analysis and Key Areas of Uncertainty 3-29
3.7 References 3-31
4 Benefits Analysis 4-1
4.1 Introduction 4-1
4.2 Hazardous Air Pollutant Benefits 4-3
4.2.1 Hg 4-4
4.2.2 Non-Hg HAP Metal 4-6
4.2.3 Additional HAP Benefits 4-12
4.3 Criteria Pollutant Benefits 4-12
4.3.1 Air Quality Modeling Methodology 4-14
4.3.2 Selecting Air Pollution Health Endpoints to Quantify 4-14
4.3.3 Calculating Counts of Air Pollution Effects Using the Health Impact Function 4-17
4.3.4 Calculating the Economic Valuation of Health Impacts 4-19
4.3.5 Benefits Analysis Data Inputs 4-19
4.3.6 Quantifying Cases of Ozone-Attributable Premature Death 4-25
4.3.7 Quantifying Cases of PM2 s-Attributable Premature Death 4-27
4.3.8 Characterizing Uncertainty in the Estimated Benefits 4-28
4.3.9 Estimated Number and Economic Value of Health Benefits 4-30
4.3.10 Additional Unqualified Benefits 4-39
4.4 Climate Benefits 4-45
4.5 Total Benefits 4-63
4.6 References 4-65
5 Economic Impacts 5-1
5.1 Overview 5-1
5.2 Small Entity Analysis 5-1
5.2.1 Methodology 5-2
5.2.2 Results 5-7
5.2.3 Conclusion 5-8
5.3 Labor Impacts 5-8
5.3.1 Overview of Methodology 5-10
5.3.2 Overview of Power Sector Employment 5-11
5.3.3 Projected Sectoral Employment Changes due to the Final Rule 5-12
5.3.4 Conclusions 5-13
5.4 References 5-14
6 Environmental Justice Impacts 6-1
6.1 Introduction 6-1
6.2 Analyzing EJ Impacts in this Final Rule 6-3
6.3 Qualitative Assessment of HAP Impacts 6-4
6.4 Demographic Proximity Analyses of Existing Facilities 6-6
6.5 EJ PM2.5 and Ozone Exposure Impacts 6-9
6.5.1 Populations Predicted to Experience PM2 5 and Ozone Air Quality Changes 6-12
6.5.2 PM25 EJ Exposure Analysis 6-12
6.5.3 Ozone EJ Exposure Analysis 6-16
6.6 GHG Impacts on Environmental Justice and other Populations of Concern 6-20
6.7 Summary 6-26
6.8 References 6-28
7 Comparison of Benefits and Costs 7-1
11
-------
7.1 Introduction 7-1
7.2 Methods 7-2
7.3 Results 7-3
7.4 Uncertainties and Limitations 7-10
7.5 References 7-11
Appendix A: Air Quality Modeling A-l
A.l Introduction A-l
A.2 Air Quality Modeling Simulations A-l
A.3 Applying Modeling Outputs to Create Spatial Fields A-14
A. 4 Scaling Factors Applied to Source Apportionment Tags A-22
A. 5 Air Quality Surface Results A-3 6
A.6 Uncertainties and Limitations of the Air Quality Methodology A-38
A. 7 References A-3 9
Appendix B: Climate Benefits Appendix B-l
B. 1 Climate Benefits Estimated using the Interim SC-CO2 values used in the Proposal B-l
B.2 References B-2
iii
-------
LIST OF TABLES
Table ES-1 Summary of Regulatory Options Examined in this RIA ES-4
Table ES-2 Projected EGU Emissions and Emissions Changes for the Baseline and under the Final Rule for 2028,
2030, and 2035 ES-8
Table ES-3 Total Compliance Cost Estimates for the Final Rule and the Less Stringent Alternative (millions of
2019 dollars, discounted to 2023) ES-11
Table ES-4 Total Benefits for the Final Rule from 2028 through 2037 (millions of 2019 dollars, discounted to
2023) ES-15
Table ES-5 Projected Net Benefits of the Final Rule (millions of 2019 dollars, discounted to 2023) ES-21
Table 1-1 Summary of Regulatory Options Examined in this RIA 1-7
Table 2-1 Total Net Summer Electricity Generating Capacity by Energy Source, 2010-2022 2-3
Table 2-2 Total Net Summer Electricity Generating Capacity by Energy Source, 2015-2022 2-4
Table 2-3 Net Generation by Energy Source, 2010 to 2022 (Trillion kWh = TWh) 2-6
Table 2-4 Net Generation by Energy Source, 2015 to 2022 (Trillion kWh = TWh) 2-6
Table 2-5 Coal and Natural Gas Generating Units, by Size, Age, Capacity, and Average Heat Rate in 2023 ... 2-9
Table 2-6 Total U.S. Electric Power Industry Retail Sales, 2010-22 and 2014-22 (billion kWh) 2-15
Table 3-1 Summary of Final Regulatory Options Examined in this RIA 3-6
Table 3-2 PM Control Technology Modeling Assumptions 3-7
Table 3-3 EGU Emissions and Projected Emissions Changes for the Baseline and the Final Rule for 2028, 2030,
and 2035 3-9
Table 3-4 Cumulative Projected Emissions Reductions for the Final Rule, 2028 to 2037 3-10
Table 3-5 Power Sector Annualized Compliance Cost Estimates under the Final Rule in 2028, 2030, and 2035
(millions of 2019 dollars) 3-12
Table 3-6 Incremental Cost of Final Continuous Emissions Monitoring (PM CEMS) Requirement 3-13
Table 3-7 Stream of Projected Compliance Costs for the Final Rule and Less Stringent Regulatory Alternative
(millions of 2019 dollars) 3-16
Table 3-8 Projected PM Control Strategies under the Final Rule in 2028 (GW) 3-17
Table 3-9 2028, 2030, and 2035 Projected U.S. Capacity by Fuel Type for the Baseline and the Final Rule.. 3-18
Table 3-10 2028, 2030, and 2035 Projected U.S. Retirements by Fuel Type for the Baseline and the Final Rule
3-19
Table 3-11 2028, 2030, and 2035 Projected U.S. New Capacity Builds by Fuel Type for the Baseline and the
Final Rule 3-20
Table 3-12 2028, 2030, and 2035 Projected U.S. Generation by Fuel Type for the Baseline and the Final Rule
3-21
Table 3-13 2028, 2030, and 2035 Projected U.S. Power Sector Coal Use by Coal Supply Region for the Baseline
and the Final Rule 3-22
Table 3-14 2028, 2030, and 2035 Projected U.S. Power Sector Coal Use by Rank for the Baseline and the Final
Rule 3-23
Table 3-15 2028, 2030, and 2035 Projected U.S. Power Sector Natural Gas Use for the Baseline and the Final
Rule 3-23
Table 3-16 2028, 2030, and 2035 Projected Minemouth and Power Sector Delivered Coal Price (2019 dollars) for
the Baseline and the Final Rule 3-24
Table 3-17 2028, 2030, and 2035 Projected Henry Hub and Power Sector Delivered Natural Gas Price (2019
dollars) for the Baseline and the Final Rule 3-24
Table 3-18 Projected Average Retail Electricity Price by Region for the Baseline and under the Final Rule, 2028
3-26
Table 3-19 Projected Average Retail Electricity Price by Region for the Baseline and under the Final Rule, 2030
3-27
Table 3-20 Projected Average Retail Electricity Price by Region for the Baseline and under the Final Rule, 2035
3-28
Table 4-1 Health Effects of PM2 5, Ambient Ozone, and Climate Effects 4-16
Table 4-2 Estimated Avoided Ozone-Related Premature Respiratory Mortalities and Illnesses for the Final Rule
for 2028, 2030, and 2035 (95 percent confidence interval) 4-31
-------
Table 4-3 Estimated Avoided PM2 s-Related Premature Mortalities and Illnesses for the Final Rule in 2028,
2030, and 2035 (95 percent confidence interval) 4-32
Table 4-4 Estimated Discounted Economic Value of Avoided Ozone and PM2 s-Attributable Premature
Mortality and Illness for the Final Rule 2028, 2030, and 2035 (95 percent confidence interval;
millions of 2019 dollars) 4-34
Table 4-5 Stream of Estimated Human Health Benefits from 2028 through 2037: Monetized Benefits Quantified
as Sum of Long-Term Ozone Mortality and Long-Term PM2 5 Mortality (discounted at 2 percent to
2023; millions of 2019 dollars) 4-35
Table 4-6 Stream of Estimated Human Health Benefits from 2028 through 2037: Monetized Benefits Quantified
as Sum of Long-Term Ozone Mortality and Long-Term PM25 Mortality (discounted at 3 percent to
2023; millions of 2019 dollars) 4-36
Table 4-7 Stream of Estimated Human Health Benefits from 2028 through 2037: Monetized Benefits Quantified
as Sum of Long-Term Ozone Mortality and Long-Term PM25 Mortality (discounted at 7 percent to
2023; millions of 2019 dollars) 4-37
Table 4-8 Additional Unqualified Benefit Categories 4-40
Table 4-9 Estimates of the Social Cost of C02 Values, 2028-2037 (2019 dollars per metric tonne C02) 4-56
Table 4-10 Stream of Projected Climate Benefits under the Final Rule from 2028 through 2037 (discounted to
2023, millions of 2019 dollars 4-58
Table 4-11 Stream of Monetized Benefits under the Final Rule from 2028 through 2037 (discounted to 2023,
millions of 2019 dollars) 4-64
Table 5-1 SBA Size Standards by NAICS Code 5-4
Table 5-2 Projected Impacts of Final Rule on Small Entities in 2028 5-8
Table 5-3 Projected Changes in Labor Utilization: Construction-Related (Number of Job-Years of Employment
in a Single Year) 5-13
Table 5-4 Projected Changes in Labor Utilization: Recurring Non-Construction (Number of Job-Years of
Employment in a Single Year) 5-13
Table 6-1 Proximity Demographic Assessment Results Within 10 km of Coal-Fired Units Greater than 25 MW
Without Retirement or Gas Conversion Plans Before 2029 Affected by this Rulemaking 6-9
Table 6-2 Demographic Populations Included in the PM2 5 and Ozone EJ Exposure Analyses 6-12
Table 7-1 Cumulative Projected Emissions Reductions for the Final Rule, 2028 to 2037 7-2
Table 7-2 Projected Net Benefits of the Final Rule in 2028 (millions of 2019 dollars) 7-4
Table 7-3 Projected Net Benefits of the Final Rule in 2030 (millions of 2019 dollars) 7-5
Table 7-4 Projected Net Benefits of the Final Rule in 2035 (millions of 2019 dollars) 7-6
Table 7-5 Projected Monetized Benefits, Costs, and Net Benefits of the Less Stringent Option in 2028 (millions
of 2019 dollars) 7-7
Table 7-6 Projected Monetized Benefits, Costs, and Net Benefits of the Less Stringent Option in 2030 (millions
of 2019 dollars) 7-7
Table 7-7 Projected Monetized Benefits, Costs, and Net Benefits of the Less Stringent Option in 2035 (millions
of 2019 dollars) 7-7
Table 7-8 Stream of Projected Monetized Benefits, Costs, and Net Benefits of the Final Rule, 2028 to 2037
(discounted to 2023, millions of 2019 dollars) 7-8
Table 7-9 Stream of Projected Monetized Benefits, Costs, and Net Benefits of the Less Stringent Option, 2028
to 2037 (millions of 2019 dollars, discounted to 2023) 7-9
Table A-l Future-Year Emissions Allocated to Each Modeled Coal EGU State Source Apportionment Tag... A-5
Table A-2 Future-Year Emissions Allocated to Each Modeled Natural Gas EGU State Source Apportionment
Tag A-7
Table A-3 Future-Year Emissions Allocated to the Modeled Other EGU Source Apportionment Tag A-8
Table A-4 Ozone Seasonal NOx Scaling Factors for Coal EGU Tags in the Baseline and the Final Rule A-22
Table A-5 Ozone Seasonal NOx Scaling Factors for Gas EGU Tags in the Baseline and the Final Rule A-24
Table A-6 Nitrate Scaling Factors for Coal EGU Tags in the Baseline and the Final Rule A-26
Table A-7 Nitrate Scaling Factors for Gas EGU Tags in the Baseline and the Final Rule A-28
Table A-8 Sulfate Scaling Factors for Coal EGU Tags in the Baseline and the Final Rule A-30
Table A-9 Primary PM2 5 Scaling Factors for Coal EGU Tags in the Baseline and the Final Rule A-32
Table A-10 Primary PM2 5 Scaling Factors for Gas EGU Tags in the Baseline and the Final Rule A-34
Table A-l 1 Scaling Factors for Other EGU Tags in the Baseline and the Final Rule A-36
Table B-l Interim SC-C02 Values, 2028 to 2037 (2019 dollars per metric ton) B-l
v
-------
Table B-2 Stream of Projected Climate Benefits under the Final Rule from 2028 to 2037 (millions of 2019
dollars, discounted to 2023) B-2
vi
-------
LIST OF FIGURES
Figure 2-1 National Coal-fired Capacity (GW) by Age of EGU, 2023 2-5
Figure 2-2 Average Annual Capacity Factor by Energy Source 2-8
Figure 2-3 Cumulative Distribution in 2021 of Coal and Natural Gas Electricity Capacity and Generation, by Age
2-10
Figure 2-4 Fossil Fuel-Fired Electricity Generating Facilities, by Size 2-11
Figure 2-5 Selected Historical Mean LCOE Values 2-12
Figure 2-6 Real National Average Electricity Prices (including taxes) for Three Major End-Use Categories .. 2-16
Figure 2-7 Relative Real Prices of Fossil Fuels for Electricity Generation; Change in National Average Real
Price per MMBtu Delivered to EGU 2-17
Figure 2-8 Relative Growth of Electricity Generation, Population and Real GDP Since 2010 2-18
Figure 2-9 Relative Change of Real GDP, Population and Electricity Generation Intensity Since 2010 2-19
Figure 3-1 Electricity Market Module Regions 3 -29
Figure 4-1 Data Inputs and Outputs for the BenMAP-CE Tool 4-20
Figure 6-1 Heat Map of the National Average PM2 5 Concentrations in the Baseline and Reductions in
Concentrations Due to the Final Regulatory Option Across Demographic Groups in 2028, 2030, and
2035 (ng/m3) 6-14
Figure 6-2 Heat Map of the National Average Ozone Concentrations in the Baseline and Reductions in
Concentrations under the Final Rule Across Demographic Groups in 2028, 2030, and 2035 (ppb)
6-18
Figure 6-3 Heat Map of the State Average Ozone Concentrations Reductions (Green) and Increases (Red) under
the Final Rule Across Demographic Groups in 2028 (ppb) 6-19
Figure A-l Air Quality Modeling Domain A-3
Figure A-2 Maps of California EGU Tag Contributions to a) April-September Seasonal Average MDA8 Ozone
(ppb); b) Annual Average PM2 5 Nitrate (|ig/m3); c) Annual Average PM25 Sulfate (|ig/m3): d)
Annual Average PM25 Organic Aerosol (ng/m3) A-10
Figure A-3 Maps of Georgia EGU Tag Contributions to a) April-September Seasonal Average MDA8 Ozone
(ppb); b) Annual Average PM2 5 Nitrate |ig/m3): c) Annual Average PM25, Sulfate (|ig/m3): d) Annual
Average PM2 5 Organic Aerosol (ng/m3) A-l 1
Figure A-4 Maps of Iowa EGU Tag contributions to a) April-September Seasonal Average MDA8 Ozone (ppb);
b) Annual Average PM25 Nitrate (|ig/m3): c) Annual Average PM2 5 Sulfate (|ig/m3): d) Annual
Average PM2 5 Organic Aerosol (ng/m3) A-12
Figure A-5 Maps of Ohio EGU Tag Contributions to a) April-September Seasonal Average MDA8 Ozone (ppb);
b) Annual Average PM2 5 Nitrate (|ig/m3): c) Annual Average PM2 5 Sulfate (|ig/m3): d) Annual
Average PM2 5 Organic Aerosol (ng/m3) A-13
Figure A-6 Maps of national EGU Tag Contributions to April-September Seasonal Average MDA8 Ozone (ppb)
by fuel for a) coal EGUs; b) natural gas EGUs; c) all other EGUs A-14
Figure A-7 Maps of national EGU Tag Contributions Annual Average PM2 5 (|ig/m3) by fuel for a) coal EGUs; b)
natural gas EGUs; c) all other EGUs A-14
Figure A-8 Maps of change in ASM-03 for the final rule compared to baseline values (ppb) shown in 2028 (right
panel), 2030 (middle panel) and 2035 (right panel) A-38
Figure A-9 Maps of change in PM2 5 for the final rule compared to baseline values (|ig/m3) shown in 2028 (right
panel), 2030 (middle panel) and 2035 (right panel) A-38
Vll
-------
EXECUTIVE SUMMARY
ES.l Introduction
Exposure to hazardous air pollutants ("HAP," sometimes known as toxic air pollution,
including mercury (Hg), chromium, arsenic, and lead) can cause a range of adverse health effects
including harming people's central nervous system; damage to their kidneys; and cancer. These
adverse effects can be particularly acute for communities living near sources of HAP.
Recognizing the dangers posed by HAP, Congress enacted Clean Air Act (CAA) section 112.
Under CAA section 112, EPA is required to set standards based on maximum achievable control
technology (known as "MACT" standards) for major sources1 of HAP that "require the
maximum degree of reduction in emissions of the hazardous air pollutants . . . (including a
prohibition on such emissions, where achievable) that the Administrator, taking into
consideration the cost of achieving such emission reduction, and any nonair quality health and
environmental impacts and energy requirements, determines is achievable." 42 U.S.C.
7412(d)(2). EPA is further required to "review, and revise" those standards every eight years "as
necessary (taking into account developments in practices, processes, and control technologies)."
Id. 7412(d)(6).
On January 20, 2021, President Biden signed Executive Order (E.O.) 13990, "Protecting
Public Health and the Environment and Restoring Science to Tackle the Climate Crisis" (86 FR
7037; January 25, 2021). The executive order, among other things, instructed EPA to review the
2020 final rule titled National Emission Standards for Hazardous Air Pollutants: Coal- and Oil-
Fired Electric Utility Steam Generating Units—Reconsideration of Supplemental Finding and
Residual Risk and Technology Review (85 FR 31286; May 22, 2020) (2020 Final Action) and to
consider publishing a notice of proposed rulemaking suspending, revising, or rescinding that
action. The 2020 Final Action included two parts: (1) a finding that it is not appropriate and
necessary to regulate coal- and oil-fired electric utility steam generating units (EGUs) under
CAA section 112; and (2) the risk and technology review (RTR) for the 2012 Mercury and Air
Toxics (MATS) Final Rule.
1 The term "major source" means any stationary source or group of stationary sources located within a contiguous
area and under common control that emits or has the potential to emit considering controls, in the aggregate, 10 tons
per year or more of any hazardous air pollutant or 25 tons per year or more of any combination of hazardous air
pollutants. 42 U.S.C. 7412(a)(1).
ES-1
-------
EPA reviewed both parts of the 2020 Final Action. The results of EPA's review of the
first part, finding it is appropriate and necessary to regulate EGUs under CAA section 112, was
proposed on February 9, 2022 (87 FR 7624) (2022 Proposal) and finalized on March 6, 2023 (88
FR 13956). In the 2022 Proposal, EPA also solicited information on the performance and cost of
new or improved technologies that control HAP emissions, improved methods of operation, and
risk-related information to further inform EPA's review of the second part, the 2020 MATS
RTR. EPA proposed amendments to the RTR on April 24, 2023 (88 FR 24854) (2023 Proposal)
and this action finalizes those amendments and presents the final results of EPA's review of the
MATS RTR. This RIA presents the expected economic consequences of EPA's final MATS
RTRRTR. As EPA determined not to reopen the 2020 Residual Risk Review, and accordingly
did not propose or finalize any revisions to that review, no projected impacts are associated with
the residual risk review.
This RIA is prepared in accordance with E.O. 12866 and 14904, the guidelines of OMB
Circular A-4, and EPA's Guidelines for Preparing Economic Analyses (2014).T. The RIA
analyzes the benefits and costs associated with the projected emissions reductions under the final
requirements to inform EPA and the public about these projected impacts. The projected benefits
and costs of the final rule and less stringent regulatory alternative are presented for the period
from 2028 to 2037.2
ES.2 Regulatory Requirements
For coal-fired EGUs, the 2012 MATS rule established standards to limit emissions of Hg,
acid gas HAP, non-Hg HAP metals (e.g., nickel, lead, chromium), and organic HAP (e.g.,
formaldehyde, dioxin/furan). For oil-fired EGUs, the 2012 MATS rule established standards to
limit emissions of hydrogen chloride (HC1) and hydrogen fluoride (HF), total HAP metals (e.g.,
Hg, nickel, lead), and organic HAP (e.g., formaldehyde, dioxin/furan).
This RIA focuses on evaluating the benefits, costs, and other impacts of four amendments
to the 2012 MATS rule:
2 Circular A-4 was recently revised. The effective date of the revised Circular A-4 (2023) is March 1, 2024, for
regulatory analyses received by OMB in support of proposed rules, interim final rules, and direct final rules, and
January 1, 2025, for regulatory analyses received by OMB in support of other final rules. For all other rules, Circular
A-4 (2003) is applicable until those dates.
ES-2
-------
• Lowering the Standard for Non-Hg HAP Metals Emissions for Existing Coal-fired
EGUs: Existing coal-fired EGUs are subject to numeric emission limits for fPM, a
surrogate for the total non-Hg HAP metals. MATS currently requires existing coal-fired
EGUs to meet a fPM emission standard of 0.030 pounds per million British thermal units
(lb/MMBtu) of heat input. After reviewing updated information on the current emission
levels of fPM from existing coal-fired EGUs and the costs of meeting a standard more
stringent than 0.030 lb/MMBtu, EPA is finalizing a fPM emission standard for existing
coal-fired EGUs of 0.010 lb/MMBtu. Additionally, EPA is finalizing updated limits for
non-Hg HAP metals and total non-Hg HAP metals that have been reduced proportional to
the reduction of the fPM emission limit. EGU owners or operators who would choose to
comply with the non-Hg HAP metals emission limits instead of the surrogate fPM limit
must request and receive approval to use a non-Hg HAP metal continuous monitoring
system as an alternative test method (e.g., multi-metal continuous monitoring system)
under the provisions of 40 CFR 63.7(f).
• Hg Emission Standard for Lignite-fired EGUs: EPA is also finalizing a revision to the
Hg emission standard for existing lignite-fired EGUs. Until this final rule, lignite-fired
EGUs must meet a Hg emission standard of 4.0 pounds per trillion British thermal units
(lb/TBtu) or 4.0E-2 pounds per gigawatt hour (lb/GWh). EPA is finalizing the
requirement that lignite-fired EGUs meet the same standard as existing EGUs firing other
types of coal, which is 1.2 lb/TBtu or 1.3E-2 lb/GWh.
• Continuous Emissions Monitoring Systems: After considering updated information on
the costs for performance testing compared to the cost of PM CEMS and capabilities of
PM CEMS measurement abilities, as well as the benefits of using PM CEMS, which
include increased transparency, compliance assurance, and accelerated identification of
anomalous emissions, EPA is finalizing the requirement that coal- and oil-fired units
demonstrate compliance with the fPM emission standard by using PM CEMS. Prior to
this final rule, EGUs had a choice of demonstrating compliance with the non-Hg HAP
metals by monitoring fPM with quarterly sampling or using PM CEMS. EPA proposed to
require PM CEMS for existing integrated gasification combined cycle (IGCC) EGUs but
is not finalizing this requirement due to technical issues calibrating CEMS on these types
of EGUs and the related fact that fPM emissions from IGCCs are very low.
• Startup Definitions: Separate from the technology review, EPA is finalizing the removal
of one of the two options for defining the startup period for EGUs. The first option
defines startup as either the first-ever firing of fuel in a boiler for the purpose of
producing electricity, or the firing of fuel in a boiler after a shutdown event for any
purpose. In the second option, startup is defined as the period in which operation of an
EGU is initiated for any purpose. EPA is removing the second option, which is currently
being used by fewer than 10 EGUs.
More detail regarding these amendments can be found in the preamble of the final rule and in
Section 1.3.1 of this document.
ES-3
-------
Table ES-1 summarizes how we have structured the regulatory options to be analyzed in
this RIA. The finalized regulatory option includes the amendments just discussed in this section:
the revision to the fPM standard to 0.010 lb/MMBtu, in which PM is a surrogate for non-Hg
HAP metals, the revision to the Hg standard for lignite-fired EGUs to 1.2 lb/TBtu, the
requirement to use PM CEMS to demonstrate compliance, and the removal of the startup
definition number two. The less stringent regulatory option examined in this RIA assumed the
fPM and Hg limits remain unchanged and examines just the finalized PM CEMS requirement
and removal of startup definition number two.
Table ES-1 Summary of Regulatory Options Examined in this RIA
Regulatory Options Examined in this RIA
Provision
Finalized
Less Stringent
FPM Standard (Surrogate
Standard for Non-Hg HAP
metals)
Revised fPM standard of 0.010
lb/MMBtu
Retain existing fPM standard of
0.030 lb/MMBtu
Hg Standard
Revised Hg standard for lignite-
fired EGUs of 1.2 lb/TBtu
Retain Hg standard for lignite-fired
EGUs of 4.0 lb/TBtu
Continuous Emissions
Monitoring Systems (PM CEMS)
Require installation of PM CEMS
to demonstrate compliance
Require installation of PM CEMS
to demonstrate compliance
Startup Definition
Remove startup definition #2
Remove startup definition #2
The compliance date for affected coal-fired sources to comply with the revised fPM limit
of 0.010 lb/MMBtu and for lignite-fired sources to meet with the lower Hg limit of 1.2 lb/Tbtu is
three years after the effective date of the final rule. EPA is finalizing the requirement that
affected sources use PM CEMS for compliance demonstration by three years after the effective
date of the final rule. The compliance date for existing affected sources to comply with
amendments pertaining to the startup definition is 180 days after the effective date of the final
rule.
Both the finalized and less stringent options described in Table ES-1 have not been
changed from the final rule and less stringent options examined in the RIA for the proposal of
this action. The proposal RIA included a more stringent regulatory option that projected the
impacts of a lowering the fPM standard to 0.006 lb/MMBtu, while holding the other three
proposed amendments unchanged from the proposed option. EPA solicited comment on this
ES-4
-------
more stringent fPM standard in the preamble of the proposed rule. As explained in the preamble
of the final rule, EPA determined not to pursue a more stringent standard for fPM emissions,
such as a limit of 0.006 lb/MMBtu. After considering comments to the proposed rule and
conducting additional analysis, EPA determined that a fPM standard lower than 0.010 lb/MMBtu
would not currently be compatible with PM CEMS due to measurement uncertainty. While a
fPM emission limit of 0.006 lb/MMBtu paired with the use of quarterly stack testing may appear
to be more stringent than the 0.010 lb/MMBtu standard paired with the use of PM CEMS that the
EPA is finalizing in this rule, there is no way to confirm emission reductions during periods in
between quarterly stack tests when emission rates may be higher. Therefore, the Agency is
finalizing a fPM limit of 0.010 lb/MMBtu with the use of PM CEMS as the only means of
compliance demonstration. EPA has determined that this combination of fPM limit and
compliance demonstration represents the most stringent option taking into account the statutory
considerations.
ES.3 Baseline and Analysis Years
The impacts of regulatory actions are evaluated relative to a modeled baseline that
represents expected behavior in the electricity sector under market and regulatory conditions in
the absence of a regulatory action. EPA frequently updates the power sector modeling baseline to
reflect the latest available electricity demand forecasts from the U.S. Energy Information
Administration (EIA) as well as expected costs and availability of new and existing generating
resources, fuels, emission control technologies, and regulatory requirements.
The baseline for this final rule includes the Good Neighbor Plan (GNP), the Revised
Cross-State Air Pollution Rule (CSAPR) Update, CSAPR Update, and CSAPR, MATS, the 2015
Effluent Limitation Guidelines (ELG) and the 2015 Coal Combustion Residuals (CCR), and the
recently finalized 2020 ELG and CCR rules.3 This version of the model also includes recent
updates to state and federal legislation affecting the power sector, including Public Law 117-169,
136 Stat. 1818 (August 16, 2022), commonly known as the Inflation Reduction Act of 2022 (the
IRA). The modeling documentation includes a summary of all legislation reflected in this version
3 For a full list of modeled policy parameters, please see: https://www.epa.gov/power-sector-modeling.
ES-5
-------
of the model as well as a description of how that legislation is implemented in the model.4 Also,
see Section 3.3 for additional detail about the power sector baseline for this RIA.
The year 2028 is the first year of detailed power sector modeling for this RIA and
approximates when the impacts of the final rule on the power sector will begin.5'6 In addition, the
regulatory impacts are evaluated for the specific analysis years of 2030 and 2035. These results
are used to estimate the present value (PV) and equivalent annualized value (EAV) of the 2028
through 2037 period, discounted to 2023.
ES.4 Emissions Impacts
EPA estimated emission reductions under the final rule for the years 2028, 2030, and
2035 based upon IPM projections. The quantified emissions estimates were developed with the
EPA's Power Sector Modeling Platform 2023 using IPM, a state-of-the-art, peer-reviewed
dynamic, deterministic linear programming model of the contiguous U.S. electric power sector.
IPM provides forecasts of least-cost capacity expansion, electricity dispatch, and emission
control strategies while meeting electricity demand and various environmental, transmission,
dispatch, and reliability constraints. IPM's least-cost dispatch solution is designed to ensure
generation resource adequacy, either by using existing resources or through the construction of
new resources. IPM addresses reliable delivery of generation resources for the delivery of
electricity between the 78 IPM regions, based on current and planned transmission capacity, by
setting limits to the ability to transfer power between regions using the bulk power transmission
system. The model includes state-of-the-art estimates of the cost and performance of air pollution
control technologies with respect to Hg and other HAP controls.
The quantified emission estimates presented in the RIA include changes in pollutants
directly covered by this rule, such as Hg and non-Hg HAP metals, and changes in other
pollutants emitted from the power sector as a result of the compliance actions projected under
4 Documentation for EPA's Power Sector Modeling Platform 2023 using IPM can be found at
https://www.epa.gov/power-sector-modeling and is available in the docket for this action.
5 Note that the Agency has granted the maximum time allowed for compliance under CAA section 112(i)(3) of three
years, and individual facilities may seek, if warranted, an additional 1-year extension of the compliance from their
permitting authority pursuant to CAA section 112(i)(3)(B). Facilities may also request, if warranted, emergency
authority to operate through the Department of Energy under section 202(c) of the Federal Power Act.
6 We note that, while the compliance date of the rule will likely be mid- to late-2027 and all compliance costs are
accounted for, any emissions reductions and benefits that in occur over a few months in 2027 are omitted from this
analysis.
ES-6
-------
this final rule. The model projections capture the emissions changes associated with
implementation of HAP mitigation measures at affected sources as well as the resulting effects
on dispatch as the relative operating costs for some affected units have changed. Table ES-2
presents the estimated impact on power sector emissions resulting from compliance with the
final rule in the contiguous U.S. As the incremental cost of operating PM CEMS relative to
baseline requirements is small relative to the ongoing costs of operation, it is not necessary to
model the less stringent regulatory alternative using IPM. The estimation of impacts outside of
the model is a reasonable approach given the relatively small costs.
ES-7
-------
Table ES-2 Projected EGU Emissions and Emissions Changes for the Baseline and under
the Final Rule for 2028, 2030, and 2035"
Total Emissions
Year
Baseline
Final Rule
Change from
Baseline
% Change
under Final
Rule
2028
6,129
5,129
-999.1
-16.3%
Hg (lbs.)
2030
5,863
4,850
-1,013
-17.3%
2035
4,962
4,055
-907.0
-18.3%
2028
70.5
69.7
-0.77
-1.09%
PM2.5 (thousand tons)
2030
66.3
65.8
-0.53
-0.79%
2035
50.7
50.2
-0.47
-0.93%
2028
79.5
77.4
-2.07
-2.60%
PMio (thousand tons)
2030
74.5
73.1
-1.33
-1.79%
2035
56.0
54.8
-1.18
-2.11%
2028
454.3
454.0
-0.290
-0.06%
SO2 (thousand tons)
2030
333.5
333.5
0.025
0.01%
2035
239.9
239.9
-0.040
-0.02%
Ozone-season NOx
(thousand tons)
2028
2030
2035
189.0
174.99
116.99
188.8
175.4
119.1
-0.165
0.488
2.282282
-0.09%
0.28%
1.95%
Annual NOx (thousand
tons)
2028
2030
460.55
392.88
460.3
392.7
-0.283
-0.022
-0.06%
-0.01%
2035
253.44
253.5
0.066
0.03%
2028
2.474
2.474
0.000
0.01%
HC1 (thousand tons)
2030
2.184
2.184
0.000
0.01%
2035
1.484
1.485
0.001
0.06%
2028
1,158.8
1,158.7
-0.0655
-0.01%
CO2 (million metric tons)
2030
1,098.3
1,098.3
0.0361
0.00%
2035
724.2
724.1
-0.099
-0.01%
a This analysis is limited to the geographically contiguous lower 48 states. Values are independently rounded and
may not sum.
We also estimate that the final rule will reduce at least seven tons of non-Hg HAP metals
in 2028, five tons of non-Hg HAP metals in 2030, and four tons of non-Hg HAP metals in 2035.7
7 The estimates on non-mercury HAP metals reductions were obtained by multiplying the ratio of non-mercury HAP
metals to fPM by estimates of PMio reductions under the rule, as we do not have estimates of fPM reductions using
IPM, only PMio. The ratios of non-mercury HAP metals to fPM were based on analysis of 2010 MATS Information
Collection Request (ICR) data. As there may be substantially more fPM than PMio reduced by the control
techniques projected to be used under this rule, these estimates of non-mercury HAP metals reductions are likely
underestimates. More detail on the estimated reduction in non-mercury HAP metals can be found in the docketed
memorandum Estimating Non-Hg HAP Metals Reductions for the 2024 Technology Review for the Coal-Fired EGU
Source Category.
ES-8
-------
These reductions are composed of reductions in emissions of antimony, arsenic, beryllium,
cadmium, chromium, cobalt, lead, manganese, nickel, and selenium.
Importantly, the continuous monitoring of fPM required in this rule will likely induce
additional emissions reductions that we are unable to quantify. Continuous measurements of
emissions accounts for unforeseeable changes to processes and fuels, fluctuations in load,
operations of pollution controls, and equipment malfunctions. By measuring emissions across all
operations, power plant operators and regulators can use the data to ensure controls are operating
properly and to assess compliance with relevant standards. Because CEMS enable power plant
operators to quickly identify and correct problems with pollution control devices, it is possible
that fPM emissions could be lower than they otherwise would have been for up to three
months—or up to three years if testing less frequently under the LEE program— at a time. This
potential reduction in fPM and non-Hg HAP metals emission resulting from the information
provided by continuous monitoring coupled with corrective actions by plant operators could be
sizeable over the existing coal-fired fleet and is not quantified in this rulemaking. Further
discussion of the emissions transparency provided by PM CEMS is available in the "2024
Update to the 2023 Proposed Technology Review for the Coal- and Oil-Fired EGU Source
Category" memorandum, available in the docket.
As we are finalizing the removal of paragraph (2) of the definition of "startup," the time
period for engaging fPM or non-Hg HAP metal controls after non-clean fuel use, as well as for
full operation of fPM or non-Hg HAP metal controls, is expected to be reduced when
transitioning to paragraph (1). The reduced time period for engaging controls therefore increases
the duration in which pollution controls are employed and lowers emissions.
To the extent that the CEMS requirement and removal of the second definition of startup
leads to actions that may otherwise not occur absent the amendments to those provisions in this
final rule, there may be emissions impacts we are unable to estimate.
ES.5 Compliance Costs
The power industry's compliance costs are represented in this analysis as the change in
electric power generation costs between the baseline and policy scenarios. In other words, these
costs are an estimate of the increased power industry expenditures required to implement the
ES-9
-------
final requirements of this rule. The compliance cost estimates were mainly developed using the
EPA's Power Sector Modeling Platform 2023 using IPM. The incremental costs of the final
rule's PM CEMS requirement were estimated outside of IPM and added to the IPM-based cost
estimate presented here and in Section 3 of the RIA.
The baseline includes approximately 5 GW of operational EGU capacity designed to burn
low rank virgin coal (i.e., lignite) in 2028. All of this capacity is currently equipped with
Activated Carbon Injection (ACI) technology, which is designed to reduce Hg emissions, and
operation of this technology for compliance with existing Hg emissions limits (e.g., MATS and
other enforceable state regulations) is reflected in the baseline. In the final rule modeling
scenario, each of these EGUs projected to consume lignite is assigned an additional variable
operating cost that is consistent with improvements in sorbent that EPA assumes are necessary to
achieve the finalized lower limit. In the final rule, this additional cost does not result in
incremental retirements for these units, nor does it result in a significant change to the projected
generation level for these units.
In 2028, the baseline projection also includes 11.6 GW of operational coal capacity that,
based on the analysis documented in the EPA memorandum titled "2024 Update to the 2023
Proposed Technology Review for the Coal- and Oil-Fired EGU Source Category," EPA assumes
would either need to improve existing PM controls or install new PM controls to comply with the
final rule. With the exception of one facility (Colstrip, located in Montana), all of that 11.6 GW
is currently operating existing electrostatic precipitators (ESPs) and/or fabric filters, and all of
that capacity is projected to install control upgrades and remain operational in 2028 in the IPM
policy scenario.
Table ES-3 below summarizes the PV and EAV of the total national compliance cost
estimates for EGUs for the final rule and the less stringent alternative. We present the PV of the
costs over the 10-year period of 2028 to 2037. We also present the EAV, which represents a flow
of constant annual values that, had they occurred annually, would yield a sum equivalent to the
PV. The EAV represents the value of a typical cost for each year of the analysis.
We note that IPM provides EPA's best estimate of the costs of the rules to the electricity
sector. These compliance cost estimates are used as a proxy for the social cost of the rule.
ES-10
-------
Chapter 3 reports how annual power costs are projected to change over the time period of
analysis.8
Table ES-3 Total Compliance Cost Estimates for the Final Rule and the Less Stringent
Alternative (millions of 2019 dollars, discounted to 2023)
2% Discount Rate
3% Discount Rate
7% Discount Rate
Regulatory Option
PV
EAV
PV
EAV
PV EAV
Final Rule
860
96
790
92
560 80
Less Stringent
19
2.3
18
2.1
13 1.8
Note: Values have been rounded to two significant figures.
Additionally, to the extent that the CEMS requirement and removal of the second definition of
startup lead to actions that may otherwise not occur absent the amendments to those provisions
in this final rule, there may be cost impacts we are unable to estimate. With respect to the
finalized removal of the startup definition, as the majority of EGUs currently rely on work
practice standards under paragraph (1) of the definition of "startup," we believe this change is
achievable by all EGUs and would result in little to no additional expenditures, especially since
the additional reporting and recordkeeping requirements associated with use of paragraph (2)
would no longer apply.
The compliance costs for the final rule are higher than the estimates in the RIA for the
proposal of this action, largely due to changes in fPM control assumptions. At proposal, EPA
estimated that incremental fPM controls would be required for about 5 GW of operational coal
capacity. Based on public comments, the Agency reevaluated the unit-level data and now
estimates that nearly three times more capacity would require incremental fPM controls (14 GW
of operational coal capacity). It is also important to note that EPA also updated the IPM baseline
power sector modeling.
8 Results using the 2 percent discount rate were not included in the proposal for this action. The 2003 version of
OMB 's Circular A-4 had generally recommended 3 percent and 7 percent as default rates to discount social costs
and benefits. The analysis of the proposed rule used these two recommended rates. In November 2023, OMB
finalized an update to Circular A-4, in which it recommended the general application of a 2 percent rate to discount
social costs and benefits (subject to regular updates), which is an estimate of consumption-based discount rate.
Given the substantial evidence supporting a 2 percent discount rate, we include cost and benefits results calculated
using a 2 percent discount rate consistent with the update to Circular A-4.
ES-11
-------
ES.6 Benefits
ES.6.1 Health Benefits
ES.6.1.1 Hazardous Air Pollutants
This final rule is projected to reduce emissions of Hg and non-Hg HAP metals. Hg
emitted from U.S. EGUs can deposit to watersheds and associated waterbodies where it can
accumulate as Methylmercury (MeHg) in fish. MeHg is formed by microbial action in the top
layers of sediment and soils, after Hg has precipitated from the air and deposited into
waterbodies or land. Once formed, MeHg is taken up by aquatic organisms and bioaccumulates
up the aquatic food web. MeHg in fish, originating from U.S. EGUs, is consumed both as self-
caught fish by subsistence fishers and as commercial fish by the general population. Exposure to
MeHg is known to have adverse impacts on neurodevelopment and the cardiovascular system.
MeHg is known to exert some genotoxic activity and EPA has classified MeHg as a "possible"
human carcinogen. The projected reductions in Hg are expected to reduce the bioconcentration
of MeHg in fish. As part of the 2020 risk review, EPA examined risk to subsistence fishers from
MeHg exposure at a lake near three U.S. EGU lignite-fired facilities (U.S. EPA, 2020). While
the analysis that EPA completed suggests that exposures associated with Hg emitted from EGUs,
including lignite-fired EGUs, are below levels of concern from a public health standpoint, further
reductions in these emissions should further decrease fish burden and exposure through fish
consumption including exposures to subsistence fishers to MeHg.
In addition, U.S. EGUs are a major source of HAP metals emissions including arsenic,
beryllium, cadmium, chromium, cobalt, lead, nickel, manganese, and selenium. Some HAP
metals emitted by U.S. EGUs are known to be persistent and bioaccumulative and others have
the potential to cause cancer. Exposure to these HAP metals, depending on exposure duration
and levels of exposures, is associated with a variety of adverse health effects. These adverse
health effects may include chronic health disorders (e.g., irritation of the lung, skin, and mucus
membranes; decreased pulmonary function, pneumonia, or lung damage; detrimental effects on
the central nervous system; damage to the kidneys; and alimentary effects such as nausea and
vomiting. The emissions reductions projected under this final rule from the use of PM controls
are expected to reduce exposure of individuals residing near these facilities to non-Hg HAP
metals, including carcinogenic HAP.
ES-12
-------
ES. 6.1.1 Criteria Pollutants
This rule is expected to reduce emissions of directly emitted PM2.5, NOx and SO2
throughout the year. Because NOx and SO2 are also precursors to secondary formation of
ambient PM2.5, reducing these emissions would reduce human exposure to ambient PM2.5
throughout the year and would reduce the incidence of PIVfo.s-attributable health effects.
This final rule is expected to reduce ozone season NOx emissions. In the presence of
sunlight, NOx, and volatile organic compounds (VOCs) can undergo a chemical reaction in the
atmosphere to form ozone. Reducing NOx emissions generally reduces human exposure to ozone
and the incidence of ozone-related health effects, though the degree to which ozone is reduced
will depend in part on local concentration levels of VOCs.
In this RIA, EPA reports estimates of the health benefits of changes in PM2.5 and ozone
concentrations. The health effect endpoints, effect estimates, benefit unit-values, and how they
were selected, are described in the Technical Support Document (TSD) titled Estimating PM2.5-
and Ozone-Attributable Health Benefits (U.S. EPA, 2023). This document, hereafter referred to
as the "Health Benefits TSD," can be found in the docket for this rulemaking. Our approach for
updating the endpoints and to identify suitable epidemiologic studies, baseline incidence rates,
population demographics, and valuation estimates is summarized in Section 4.3.
ES. 6.2 Climate Benefits
Elevated concentrations of carbon dioxide (CO2) and other greenhouse gases (GHGs) in
the atmosphere have been warming the planet, leading to changes in the Earth's climate
including changes in the frequency and intensity of heat waves, precipitation, and extreme
weather events, rising seas, and retreating snow and ice. The well-documented atmospheric
changes due to anthropogenic GHG emissions are changing the climate at a pace and in a way
that threatens human health, society, and the natural environment. Climate change touches nearly
every aspect of public welfare in the U.S. with resulting economic costs, including: changes in
water supply and quality due to changes in drought and extreme rainfall events; increased risk of
storm surge and flooding in coastal areas and land loss due to inundation; increases in peak
electricity demand and risks to electricity infrastructure; and the potential for significant
agricultural disruptions and crop failures (though offset to some extent by carbon fertilization).
ES-13
-------
There will be important climate benefits associated with the CO2 emissions reductions
expected from this final rule. Climate benefits from reducing emissions of CO2 can be monetized
using estimates of the social cost of carbon (SC-CO2). See Section 4.4 for more discussion of the
approach to monetization of the climate benefits associated with this rule.
ES. 6.3 Additional Unquantified Benefits
As stated above, EPA is unable to quantify and monetize the potential benefits of
requiring facilities to utilize CEMS rather than continuing to allow the use stack testing, but the
requirement has been considered qualitatively. Relative to periodic testing practices, continuous
monitoring of fPM will result in increased transparency, as well as potential emissions reductions
from identifying problems more rapidly. Hence, the final rule may induce further reductions of
fPM and non-Hg HAP metals than we project in this RIA, and these reductions would likely lead
to additional health benefits. However, due to data and methodological challenges, EPA is
unable to quantify these potential additional reductions. The continuous monitoring of fPM
required in this rule is also likely to provide several additional important benefits to the public
which are not quantified in this rule, including greater certainty, accuracy, transparency, and
granularity in fPM emissions information than exists today. Additionally, to the extent that the
CEMS requirement and removal of the second definition of startup leads to actions and
emissions impacts that may otherwise not occur absent the amendment in this final rule, there
may be beneficial impacts we are unable to estimate.
Regarding the potential health and ecological benefits from HAP emission reductions,
data, time, and resource limitations prevent us from quantifying these potential benefits.
Additionally, data, time, and resource limitations prevented EPA from quantifying the estimated
health impacts or monetizing estimated benefits associated with direct exposure to NO2 and SO2
(independent of the role NO2 and SO2 play as precursors to PM2.5 and ozone), as well as
ecosystem effects, and visibility impairment due to the absence of air quality modeling data for
these pollutants in this analysis. While all health benefits and welfare benefits were not able to be
quantified, it does not imply that there are not additional benefits associated with reductions in
exposures to HAP, ozone, PM2.5, NO2 or SO2.
ES-14
-------
ES.6.4 Total Benefits
Table ES-4 presents the total monetized health and climate benefits for the final rule.9
Note the less stringent regulatory alternative only describes the benefits associated with the
requirements for PM CEMS qualitatively. As a result, there are no quantified benefits associated
with this regulatory option.
Table ES-4 Total Benefits for the Final Rule from 2028 through 2037 (millions of 2019
dollars, discounted to 2023)"
Health Benefits
Health Benefits
All Values
Calculated using
2% Discount Rate
Calculated using
Calculated using
3% Discount Rate,
Climate Benefits
7% Discount Rate,
Climate Benefits
Calculated using
2% Discount Rate
Calculated using
2% Discount Rate
PV
Health Benefits
EAV
300
33
260
31
180
25
PV
Climate Benefits0
EAV
130
14
130
14
130
14
Total Monetized PV
420
390
300
Benefits EAV
47
45
39
Non-Monetized Benefits'1
Benefits from reductions of about 900 to 1000 pounds of Hg annually
Benefits from reductions of about 4 to 7 tons of non-Hg HAP metals annually
Benefits from the increased transparency, compliance assurance, and accelerated identification of anomalous
emission anticipated from requiring PM CEMS
a Values have been rounded to two significant figures. Rows may not appear to add correctly due to rounding.
b The estimated value of the air quality-related health benefits reported here are from Table 4-5,
Table 4-6, and Table 4-7. Monetized benefits include those related to public health associated with reductions in
PM2 5 and ozone concentrations. For discussions of the uncertainty associated with these health benefits estimates,
see Section4.3.8.
0 Monetized climate benefits are based on reductions in CO2 emissions and are calculated using three different
estimates of the social cost of CO2 (SC-CO2) (under 1.5 percent, 2.0 percent, and 2.5 percent near-term Ramsey
discount rates). For the presentational purposes of this table, we show the climate benefits associated with the SC-
CO2 at the 2 percent near-term Ramsey discount rate. See Table 4-10 for the full range of monetized climate benefit
estimates. See Section 4.3.10 for a discussion of the uncertainties associated with the climate benefit estimates.
d The list of non-monetized benefits does not include all potential non-monetized benefits. See Table 4-8 for a more
complete list.
9 Monetized climate benefits are discounted using a 2 percent discount rate, consistent with EPA's updated estimates
of the SC-CO2. OMB has long recognized that climate effects should be discounted only at appropriate
consumption-based discount rates. Because the SC-CO2 estimates reflect net climate change damages in terms of
reduced consumption (or monetary consumption equivalents), the use of the social rate of return on capital (7
percent under OMB Circular A-4 (2003)) to discount damages estimated in terms of reduced consumption would
inappropriately underestimate the impacts of climate change for the purposes of estimating the SC-CO2. See Section
4.4 for more discussion.
ES-15
-------
The estimates of monetized benefits under the final rule are lower than estimated at
proposal. While the estimated Hg reductions are higher under the final rule than at proposal, it is
important to note that the EPA is unable to quantify the potential benefits of any HAP reductions
for this rule. Additionally, while EPA is assuming more filterable PM controls in the final rule,
the EPA is unable to quantify the potential benefits of any reductions of non-Hg HAP metals that
are expected to result from these controls. Furthermore, because the EPA is no longer projecting
any significant change in utilization or capacity at facilities that install additional fPM controls,
we do not project major changes in emissions of the criteria and GHG pollutants monetized in
the benefit-cost analysis. Consequently, the monetized benefits of the rule are lower than
previously projected.
ES.7 Environmental Justice Impacts
EE.O. 12898 directs EPA to identify the populations of concern who are most likely to
experience unequal burdens from environmental harms; specifically, minority populations, low-
income populations, and Indigenous peoples.10 Additionally, EE.O. 13985 is intended to advance
racial equity and support underserved communities through federal government actions.11 Most
recently, E.O. 14096 (88 FR 25251, April 26, 2023) strengthens the directives for achieving
environmental justice that are set out in E.O. 12898. EPA defines environmental justice (EJ) as
the fair treatment and meaningful involvement of all people regardless of race, color, national
origin, or income, with respect to the development, implementation, and enforcement of
environmental laws, regulations, and policies. 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,
governmental, and commercial operations or programs and policies."12 In recognizing that
minority and low-income populations often bear an unequal burden of environmental harms and
risks, EPA continues to consider ways of protecting them from adverse public health and
environmental effects of air pollution.
10 59 FR 7629, February 16, 1994.
11 86 FR 7009, January 20, 2021.
12 https://www.epa.gov/environmentaljustice.
ES-16
-------
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 (2015)13 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
disproportionate and adverse exposures and impacts. For the rule, we quantitatively evaluate 1)
the proximity of affected facilities to potentially vulnerable and/or overburdened populations for
consideration of local pollutants impacted by this rule but not modeled here (Section 6.3) and 2)
the distribution of ozone and PM2.5 concentrations in the baseline and changes due to the
rulemaking across different demographic groups on the basis of race, ethnicity, educational
attainment, employment status, health insurance status, life expectancy, linguistic isolation,
poverty status, redlined areas, tribal land, age, and sex (Section 6.5). It is important to note that
due to the small magnitude of underlying emissions changes, and the corresponding small
magnitude of the ozone and PM2.5 concentration changes, the rule is expected to have only a
small impact on the distribution of exposures across each demographic group. We also
qualitatively discuss potential EJ HAP and climate impacts (Sections 6.3 and 6.6). Each of these
analyses was performed to answer separate questions and is associated with unique limitations
and uncertainties. Baseline demographic proximity analyses provide information as to whether
there may be potential EJ concerns associated with environmental stressors, such as noise,
traffic, and emissions such as NO2 and SO2 covered by the regulatory action for certain
population groups of concern (Section 6.4). The baseline demographic proximity analyses
examined the demographics of populations living within 10 km of the following sources: lignite
plants with units potentially impacted by the final Hg standard revision and coal plants with units
13 https://www.epa.gov/environmentaljustice/technical-guidance-assessing-environmental-justice-regulatory-
analysis.
ES-17
-------
potentially impacted by the final fPM standard revision. We evaluated a 5 km radius for the
demographic analysis and found it yielded several facilities with zero population within 5 km
(i.e., no data) and over 10 percent of the facilities had less than 100 people within 5 km. At a 10-
km radius, all facilities but one have population data and only two percent of facilities had less
than 100 people within 10 km. Therefore, the 10-km distance was used on the basis that it
captures large enough populations to avoid excessive demographic uncertainty.
The baseline analysis indicates that on average the population living within 10 km of coal
plants potentially impacted by the final fPM standards shas a higher percentage of people living
below two times the poverty level than the national average. In addition, on average the
percentage of the Native American population living within 10 km of lignite plants potentially
impacted by the final Hg standard is higher than the national average. Relating these results to
question 1, above, we conclude that there may be potential EJ concerns associated with directly
emitted pollutants that are affected by the regulatory action (e.g., PM2.5 and HAP) for certain
population groups of concern in the baseline. However, as proximity to affected facilities does
not capture variation in baseline exposure across communities, nor does it indicate that any
exposures or impacts will occur, these results should not be interpreted as a direct measure of
exposure or impact.
As HAP exposure results generated as part of the 2020 MATS RTR were below both the
presumptive acceptable cancer risk threshold and the reference dose (RfD), and this final
regulation should further reduce exposure to HAP, there is no evidence of 'disproportionate and
adverse effects' of potential EJ concern. Therefore, we did not perform a quantitative EJ
assessment of HAP risk.
In contrast, ozone and PM2.5 precursor emission changes that influence ambient
concentrations of ozone and PM2.5 are also expected from this action, and exposure analyses that
evaluate demographic variables are better able to evaluate any potentially disproportionate
pollution impacts of this rulemaking. The baseline ozone and PM2.5 exposure analyses respond to
question 1 from EPA's EJ Technical Guidance document more directly than the proximity
analyses, as they evaluate a form of the environmental stressor affected by the regulatory action
(see Section 6.5). PM2.5 and ozone exposure analyses show that certain populations, such as
residents of redlined census tracts, those who are linguistically isolated, Hispanic individuals,
ES-18
-------
Asian individuals, those without a high school diploma, and the unemployed may experience
disproportionately higher ozone and PM2.5 exposures in the baseline as compared to the national
average. American Indian individuals, residents of Tribal Lands, populations with higher life
expectancy or with life expectancy data unavailable, children, and insured populations may also
experience disproportionately higher ozone concentrations in the baseline than the reference
group. Hispanic individuals, Black individuals, those below the poverty line, and uninsured
populations may also experience disproportionately higher PM2.5 concentrations in the baseline
than the reference group. Therefore, there likely are potential EJ concerns associated with
environmental stressors affected by the regulatory action for population groups of concern in the
baseline.
Finally, we evaluate how the final rule may be expected to differentially impact
demographic populations, informing questions 2 and 3 from EPA's EJ Technical Guidance with
regard to ozone and PM2.5 exposure changes. Due to the small magnitude of the exposure
changes across population demographics associated with the rulemaking relative to the
magnitude of the baseline disparities, we infer that disparities in the ozone and PM2.5
concentration burdens in the baseline are likely to remain after implementation of the regulatory
action or alternatives under consideration. This is due to the small magnitude of the
concentration changes associated with this rulemaking across population demographic groups,
relative to the magnitude of the baseline disparities (question 2). Also, due to the very small
differences observed in the distributional analyses of post-policy ozone and PM2.5 exposure
impacts across population groups, we do not find evidence that potential EJ concerns related to
ozone and PM2.5 concentrations will be created or mitigated as compared to the baseline
(question 3).
ES.8 Comparison of Benefits and Costs
In this RIA, the regulatory impacts are evaluated for the specific years of 2028, 2030, and
2035. Comparisons of benefits to costs for these snapshot years are presented in Section 7.3 of
this RIA. Here we present the PV of costs, benefits, and net benefits, calculated for the years
2028 to 2037 from the perspective of 2023, using two percent, three percent, and seven percent
end-of-period discount rate. All dollars are in 2019 dollars. We also present the EAV, which
represents a flow of constant annual values that, had they occurred in each year from 2028 to
ES-19
-------
2037, 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 year-specific estimates reported in the
costs and benefits sections of this RIA. The comparison of benefits and costs in PV and EAV
terms for the final rule is presented in Table ES-5. The benefits associated with the less stringent
regulatory alternative, from the final requirements for PM CEMS are only described
qualitatively. As a result, there are no quantified benefits associated with this regulatory option,
and we do not include a table reporting the quantified net benefits of that option (the quantified
costs are reporting in Table ES-3).
ES-20
-------
Table ES-5 Projected Net Benefits of the Final Rule (millions of 2019 dollars, discounted
to 2023)a,b
Health
Benefitsb
Climate
Benefits0'"1
Compliance
Costs
Net
Benefits0
Year
2%
3%
7%
2%
2%
3%
7%
2%
3%
7%
2028
79
71
52
13
100
99
82
-12
-15
-16
2029
79
71
50
13
100
96
77
-10
-13
-13
2030
27
24
16
-7.1
100
95
73
-82
-78
-64
2031
27
24
16
-7.1
100
92
68
-80
-76
-60
2032
14
13
8
19
79
73
52
-46
-41
-24
2033
14
13
8
19
78
71
48
-44
-39
-21
2034
14
12
7.3
19
76
69
45
-43
-37
-19
2035
14
12
7.0
19
75
67
42
-41
-35
-16
2036
14
12
6.7
19
73
65
39
-40
-33
-14
2037
14
12.0
6.4
19
72
63
37
-39
-32
-11
Health
Benefitsb
Climate
Benefits0
Compliance
Costs
Net
Benefits0
Discount Rate
2%
3%
7%
2%
2%
3%
7%
2%
3%
7%
PV
300
260
180
130
860
790
560
-440
-400
-260
EAV
33
31
25
14
96
92
80
-49
-47
-41
Non-Monetized Benefits®
Benefits from reductions of about 900 to 1000 pounds of Hg annually
Benefits from reductions of about 4 to 7 tons of non-Hg HAP metals annually
Benefits from the increased transparency, compliance assurance, and accelerated identification of anomalous
emission anticipated from requiring PM CEMS
a Values have been rounded to two significant figures. Rows may not appear to add correctly due to rounding.
b The estimated value of the air quality-related health benefits reported here are the larger of the two estimates
presented in Table 4-5, Table 4-6, and Table 4-7. Monetized benefits include those related to public health
associated with reductions in PM2 5 and ozone concentrations. For discussions of the uncertainty associated with
these health benefits estimates, see Section 4.3.8.
0 Monetized climate benefits are based on reductions in C02 emissions and are calculated using three different
estimates of the social cost of CO2 (SC-CO2) (under 1.5 percent, 2.0 percent, and 2.5 percent near-term Ramsey
discount rates). For the presentational purposes of this table, we show the climate benefits associated with the SC-
CO2 at the 2 percent near-term Ramsey discount rate. See Table 4-10 for the full range of monetized climate benefit
estimates.
d The small increases and decreases in climate and health benefits and related EJ impacts result from very small
changes in fossil dispatch and coal use relative to the baseline. For context, the projected increase in CO2 emission
of less than 40,000 tons in 2030 is roughly one percent of the emissions of a mid-size coal plant operating at
availability (about 4 million tons).
e The list of non-monetized benefits does not include all potential non-monetized benefits. See Table 4-8 for a more
complete list.
The monetized estimates of benefits presented in this section are underestimated because
important categories of benefits, including benefits from reducing Hg and non-Hg HAP metals
emissions and the increased transparency, compliance assurance, and accelerated identification
ES-21
-------
of anomalous emissions anticipated from requiring PM CEMS, were not monetized in our
analysis. Simultaneously, the estimates of compliance costs used in the net benefits analysis may
provide an incomplete characterization of the true costs of the rule. We nonetheless consider
these potential impacts in our evaluation of the net benefits of the rule. As the EPA no longer
projects incremental facility retirement and expects less change in capacity and utilization,
higher compliance costs are expected along with smaller monetized benefits than in the proposal
analysis of this rulemaking. The result of combining those updated estimates is a lower estimate
of net benefits than in the proposal analysis.
ES.9 References
OMB. (2003). Circular A-4: Regulatory Analysis. Washington DC.
https://www.whitehouse.gov/wp-
content/uploads/legacy_drupal_files/omb/circulars/A4/a-4.pdf
OMB. (2023). Circular A-4: Regulatory Analysis. Washington DC.
https://www.whitehouse.gov/wp-content/uploads/2023/ll/CircularA-4.pdf
U.S. EPA. (2014). Guidelines for Preparing Economic Analyses. (EPA 240-R-10-001).
Washington DC: U.S. Environmental Protection Agency, Office of Policy, National
Center for Environmental Economics, https://www.epa.gov/environmental-
economics/guidelines-preparing-economic-analyses
U.S. EPA. (2015). Guidance on Considering Environmental Justice During the Development of
Regulatory Actions, https://www.epa.gov/sites/default/files/2015-
06/documents/considering-ej-in-rulemaking-guide-final.pdf
U.S. EPA. (2020). Residual Risk Assessment for the Coal- and Oil-Fired EGU Source Category
in Support of the 2020 Risk and Technology Review Final Rule. Research Triangle Park,
NC: U.S. Environmental Protection Agency, Office of Air Quality Planning and
Standards. https://www.regulations.gov/document/EPA-HQ-OAR-2018-0794-4553
U.S. EPA. (2023). Estimating PM2.5- and Ozone-Attributable Health Benefits. Research
Triangle Park, NC: U.S. Environmental Protection Agency, Office of Air Quality
Planning and Standards, Health and Environmental Impact Division.
https://www.epa.gov/system/files/documents/2023-01/Estimating%20PM2.5-
%20and%200zone-Attributable%20Health%20Benefits%20TSD_0.pdf
ES-22
-------
INTRODUCTION AND BACKGROUND
1.1 Introduction
Exposure to hazardous air pollution ("HAP," sometimes known as toxic air pollution,
including Hg, chromium, arsenic, and lead) can cause a range of adverse health effects including
harming people's central nervous system; damaging their kidneys; and causing cancer.
Recognizing the dangers posed by HAP, Congress enacted Clean Air Act (CAA) section 112.
Under CAA section 112, the Environmental Protection Agency (EPA) is required to set
standards (known as "MACT" (maximum achievable control technology) standards) for major
sources of HAP that "require the maximum degree of reduction in emissions of the hazardous air
pollutants . . . (including a prohibition on such emissions, where achievable) that the
Administrator, taking into consideration the cost of achieving such emission reduction, and any
non-air quality health and environmental impacts and energy requirements, determines is
achievable." 42 U.S.C. 7412(d)(2). On January 20, 2021, President Biden signed EE.O. 13990,
"Protecting Public Health and the Environment and Restoring Science to Tackle the Climate
Crisis" (86 FR 7037; January 25, 2021). The executive order, among other things, instructed
EPA to review the 2020 final rule titled "National Emission Standards for Hazardous Air
Pollutants: Coal- and Oil- Fired Electric Utility Steam Generating Units—Reconsideration of
Supplemental Finding and Residual Risk and Technology Review" (85 FR 31286; May 22,
2020) and to consider publishing a notice of proposed rulemaking suspending, revising, or
rescinding that action. The 2020 Final Action included a finding that it is not appropriate and
necessary to regulate coal- and oil-fired EGUs under CAA section 112 as well as the RTR for the
MATS rule. The results of EPA's review of the 2020 appropriate and necessary finding were
proposed on February 9, 2022 (87 FR 7624) (2022 Proposal) and finalized on March 6, 2023 (88
FR 13956). In the 2022 Proposal, EPA also solicited information on the performance and cost of
new or improved technologies that control HAP emissions, improved methods of operation, and
risk-related information to further inform EPA's review of the 2020 MATS RTR. The review of
the RTR was proposed on April 24, 2023 (88 FR 24854) and this action presents the final results
of EPA's review of the MATS RTR. This RIA presents the expected economic consequences of
EPA's final MATS Risk and Technology Review.
1-1
-------
Several statutes and executive orders apply to federal rulemakings. In accordance with
E.O. 12866 and E.O. 14094 and the guidelines of OMB Circular A-4, the RIA presents the
benefits and costs associated with the projected emissions reductions under the final rule.14 The
benefits and costs of the final rule and regulatory alternative are presented for the 2028 to 2037
time period. The estimated monetized benefits are those health benefits expected to arise from
reduced PM2.5 and ozone concentrations and the climate benefits from reductions in GHGs.
Several categories of benefits remain unmonetized including important benefits from reductions
in Hg and non-Hg HAP metal emissions. The estimated monetized costs for EGUs are the costs
of installing and operating controls and the increased costs of producing electricity. Unquantified
benefits and costs are described qualitatively. This section contains background information
relevant to the rule and an outline of the sections of this RIA.
1.2 Legal and Economic Basis for Rulemaking
In this section, we summarize the statutory requirements in the CAA that serve as the
legal basis for the final rule and the economic theory that supports environmental regulation as a
mechanism to enhance social welfare. The CAA requires EPA to prescribe regulations for new
and existing sources. In turn, those regulations attempt to address negative externalities created
when private entities fail to internalize the social costs of air pollution.
1.2.1 Statutory Requirement
The statutory authority for this action is provided by sections 112 and 301 of the CAA, as
amended (42 U.S.C. 7401 et seq.). Section 112 of the CAA establishes a two-stage regulatory
process to develop standards for emissions of HAP from stationary sources. Generally, the first
stage involves establishing technology-based standards and the second stage involves evaluating
those standards that are based on maximum achievable control technology (MACT) to determine
whether additional standards are needed to address any remaining risk associated with HAP
emissions. This second stage is commonly referred to as the "residual risk review." In addition to
the residual risk review, the CAA also requires EPA to review standards set under CAA section
14 Circular A-4 was recently revised. The effective date of the revised Circular A-4 (2023) is March 1, 2024, for
regulatory analyses received by OMB in support of proposed rules, interim final rules, and direct final rules, and
January 1, 2025, for regulatory analyses received by OMB in support of other final rules. For all other rules, Circular
A-4 (2003) is applicable until those dates.
1-2
-------
112 no less than every eight years and revise the standards as necessary taking into account any
"developments in practices, processes, or control technologies." This review is commonly
referred to as the "technology review," and is the subject of this rulemaking.
1.2.2 Regulated Pollutants
For coal-fired EGUs, the 2012 MATS rule established standards to limit emissions of Hg,
acid gas HAP, non-Hg HAP metals (e.g., nickel, lead, chromium), and organic HAP (e.g.,
formaldehyde, dioxin/furan). Standards for hydrochloric acid (HC1) serve as a surrogate for the
acid gas HAP, with an alternate standard for sulfur dioxide (SO2) that may be used as a surrogate
for acid gas HAP for those coal-fired EGUs with flue gas desulfurization (FGD) systems and
SO2 CEMS installed and operational. Standards for fPM serve as a surrogate for the non-Hg
HAP metals, with standards for total non-Hg HAP metals and individual non-Hg HAP metals
provided as alternative equivalent standards. Work practice standards limit formation and
emission of the organic HAP.
For oil-fired EGUs, the 2012 MATS rule established standards to limit emissions of HC1
and hydrogen fluoride (HF), total HAP metals (e.g., Hg, nickel, lead), and organic HAP (e.g.,
formaldehyde, dioxin/furan). Standards for fPM serve as a surrogate for total HAP metals, with
standards for total HAP metals and individual HAP metals provided as alternative equivalent
standards. Work practice standards limit formation and emission of the organic HAP.
1.2.2.1 Definition of Affected Source
The source category that is the subject of this final rule is Coal- and Oil-Fired EGUs
regulated under 40 CFR 63, subpart UUUUU. The North American Industry Classification
System (NAICS) codes for the Coal- and Oil-fired EGU industry are 221112, 221122, and
921150. This list of categories and NAICS codes is not intended to be exhaustive, but rather
provides a guide for readers regarding the entities that this action is likely to affect. The final
standards will be directly applicable to the affected sources. Federal, state, local, and tribal
government entities that own and/or operate EGUs subject to 40 CFR part 63, subpart UUUUU
would be affected by this action. The Coal- and Oil-Fired EGU source category was added to the
list of categories of major and area sources of HAP published under section 112(c) of the CAA
on December 20, 2000 (65 FR 79825). CAA section 112(a)(8) defines an EGU as: any fossil fuel
fired combustion unit of more than 25 MW that serves a generator that produces electricity for
1-3
-------
sale. A unit that cogenerates steam and electricity and supplies more than one-third of its
potential electric output capacity and more than 25 MW electrical output to any utility power
distribution system for sale is also considered an EGU.
1.2.3 The Potential Need for Regulation
OMB Circular A-4 indicates that one of the reasons a regulation may be issued is to
address a market failure. The major types of market failure include externalities, market power,
and inadequate or asymmetric information. Correcting market failures is one reason for
regulation; it is not the only reason. Other possible justifications include improving the function
of government, correcting distributional unfairness, or securing privacy or personal freedom.
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. For the regulatory action analyzed in this RIA, the good produced is electricity
from coal- and oil-fired EGUs. If these electricity producers pollute the atmosphere when
generating power, the social costs will not be borne exclusively by the polluting firm but rather
by society as a whole. Thus, the producer is imposing a negative externality, or a social cost of
emissions, on society. The equilibrium market price of electricity may fail to incorporate the full
opportunity cost to society of these products. Consequently, absent a regulation on emissions,
producers will not internalize the social cost of emissions and social costs will be higher as a
result. This regulation will work towards addressing this market failure by causing affected
producers to begin internalizing the negative externality associated with HAP emissions from
electricity generation by coal- and oil-fired EGUs.
1.3 Overview of Regulatory Impact Analysis
1.3.1 Regulatory Options
This RIA focuses on four amendments to the MATS rule, which are described in more
detail in this section.
1-4
-------
1.3.1.1 Filterable Particulate Matter Standards for Existing Coal-fired EGUs
Existing coal-fired EGUs are subject to numeric emission limits for fPM, a surrogate for
the total non-Hg HAP metals.15 Before this final rule, MATS required existing coal-fired EGUs
to meet a fPM emission standard of 0.030 pounds per million British thermal units (lb/MMBtu)
of heat input. The standards for fPM serve as a surrogate for standards for non-Hg HAP metals.
After reviewing updated information on the current emission levels of fPM from existing coal-
fired EGUs and the costs of meeting a standard more stringent than 0.030 lb/MMBtu, EPA is
revising the fPM emission standard for existing coal-fired EGUs to 0.010 lb/MMBtu.
Additionally, EPA is finalizing updated limits for non-Hg metals and total non-Hg metals that
have been reduced proportional to the reduction of the fPM emission limit. EGU owners or
operators who would choose to comply with the non-Hg HAP metals emission limits instead of
the fPM limit must request and receive approval of a non-Hg HAP metal continuous monitoring
system as an alternative test method (e.g., multi-metal continuous monitoring system) under the
provisions of 40 CFR 63.7(f).
1.3.1.2 Hg Emission Standard for Lignite-fired EGUs
EPA is revising the Hg emission standard for lignite-fired EGUs. Before this final rule,
lignite-fired EGUs were required to meet a Hg emission standard of 4.0 pounds per trillion
British thermal units (lb/TBtu) or 4.0E-2 pounds per gigawatt hour (lb/GWh). EPA recently
collected information on current emission levels and Hg emission controls for lignite-fired EGUs
using the authority provided under CAA section 114.16 That information showed that many units
are able to achieve a Hg emission rate that is much lower than the current standard, and there are
cost-effective control technologies and methods of operation that are available to achieve a more
15 As described in section III of the preamble to 2023 proposal, EGUs in seven subcategories are subject to numeric
emission limits for specific HAP or fPM, a surrogate for the total non-mercury HAP metals. The fPM was chosen as
a surrogate in the original rulemaking because the non-mercury HAP metals are predominantly a component of PM,
and control of PM will also result in co-reduction of non-mercury HAP metals. Additionally, not all fuels emit the
same type and amount of HAP metals, but most generally emit PM that include some amount and combination of all
the HAP metals. Lastly, the use of fPM as a surrogate eliminates the cost of performance testing to comply with
numerous standards for individual non-mercury HAP metals (Docket ID No. EPA-HQ-OAR-2009-0234). For these
reasons, EPA focused its review on the fPM emissions of coal-fired EGUs as a surrogate for the non-mercury HAP
metals.
16 For further information, see EPA memorandum titled "2024 Update to the 2023 Proposed Technology Review for
the Coal- and Oil-Fired EGU Source Category" which is available in the docket.
1-5
-------
stringent standard. EPA is finalizing a standard for lignite-fired EGUs of 1.2 lb/TBtu or 1.3E-2
lb/GWh, the same standard applied to EGUs firing other types of coal.
1.3.1.3 Require that All Coal- and Oil-FiredEGUs Demonstrate Compliance with the jPM
Emission Standard by Using PM CEMS
In addition to revising the PM emission standard for existing coal-fired EGUs, EPA is
revising the requirements for demonstrating compliance with the PM emission standard for coal-
and oil-fired EGUs. Before this final rule, EGUs that were not part of the low-emitting EGU
(LEE) program could demonstrate compliance with the fPM standard either by conducting
performance testing quarterly or by using PM CEMS. After considering updated information on
the costs for performance testing, the costs of PM CEMS, the capabilities of PM CEMS
measurement abilities, and the benefits of using PM CEMS, including increased transparency,
compliance assurance, and accelerated identification of anomalous emissions, EPA is requiring
that all coal- and oil-fired fired EGUs demonstrate compliance with the PM emission standard by
using PM CEMS. EPA proposed to require PM CEMS for existing IGCC EGUs but is not
finalizing this requirement due to technical issues calibrating CEMS on these types of EGUs and
the related fact that fPM emissions from IGCCs are very low.
1.3.1.4 Startup Definitions
Finally, separate from the technology review, EPA is removing one of the two options for
defining the startup period for EGUs. The first option defines startup as either the first-ever firing
of fuel in a boiler for the purpose of producing electricity, or the firing of fuel in a boiler after a
shutdown event for any purpose. Startup ends when any of the steam from the boiler is used to
generate electricity for sale over the grid or for any other purpose (including on-site use). In the
second option, startup is defined as the period in which operation of an EGU is initiated for any
purpose. Startup begins with either the firing of any fuel in an EGU for the purpose of producing
electricity or useful thermal energy (such as heat or steam) for industrial, commercial, heating, or
cooling purposes (other than the first-ever firing of fuel in a boiler following construction of the
boiler) or for any other purpose after a shutdown event. Startup ends four hours after the EGU
generates electricity that is sold or used for any other purpose (including on-site use), or four
hours after the EGU makes useful thermal energy (such as heat or steam) for industrial,
1-6
-------
commercial, heating, or cooling purposes, whichever is earlier. EPA is removing the second
option, which is currently being used by fewer than 10 EGUs.
1.3.1.5 Summary of Regulatory Options Examined in this RIA
Table 1-1 summarizes how we have structured the regulatory options to be analyzed in
this RIA. The final regulatory option includes the amendments just discussed in this section: the
revision to the fPM standard to 0.010 lb/MMBtu, in which fPM is a surrogate for non-Hg HAP
metals, the revision to the Hg standard for lignite-fired EGUs to 1.2 lb/TBtu, the requirement to
use PM CEMS to demonstrate compliance, and the removal of the startup definition number two.
The less stringent regulatory option examined in this RIA assumed the PM and Hg limits remain
unchanged and examines just the PM CEMS requirement and removal of startup definition
number two.
Table 1-1 Summary of Regulatory Options Examined in this RIA
Regulatory Options Examined in this RIA
Provision
Less Stringent
Final Rule
FPM Standard (Surrogate
Standard for Non-Hg HAP
Metals)
Retain existing fPM standard of
0.030 lb/MMBtu
Revised fPM standard of 0.010
lb/MMBtu
Hg Standard
Retain Hg standard for lignite-fired
EGUs of 4.0 lb/TBtu
Revised Hg standard for lignite-
fired EGUs of 1.2 lb/TBtu
Continuous Emissions
Monitoring Systems (PM CEMS)
Require installation of PM CEMS
to demonstrate compliance
Require installation of PM CEMS
to demonstrate compliance
Startup Definition
Remove startup definition #2
Remove startup definition #2
The compliance date for affected coal-fired sources to comply with the revised fPM limit
of 0.010 lb/MMBtu and for lignite-fired sources to meet with the lower Hg limit of 1.2 lb/Tbtu is
three years after the effective date of the final rule. EPA is finalizing the requirement that
affected sources use PM CEMS for compliance demonstration by three years after the effective
date of the final rule. The compliance date for existing affected sources to comply with
amendments pertaining to the startup definition is 180 days after the effective date of the final
rule.
Both the final rule and less stringent options described in Table 1-1 have not been
changed from the proposed and less stringent options examined in the RIA for the proposal of
1-7
-------
this action. The proposal RIA included a more stringent regulatory option that projected the
impacts of lowering the fPM standard to 0.006 lb/MMBtu, while holding the other three
proposed amendments unchanged from the proposed option. As explained in the preamble of the
final rule, EPA determined not to pursue a more stringent standard for fPM emissions, such as a
limit of 0.006 lb/MMBtu. After considering comments to the proposed rule and conducting
additional analysis, EPA determined that a fPM standard lower than 0.010 lb/MMBtu would not
be compatible with PM CEMS due to measurement uncertainty. While a fPM emission limit of
0.006 lb/MMBtu may appear to be more stringent than the 0.010 lb/MMBtu standard that the
EPA is finalizing in this rule, there is no way to confirm emission reductions during periods
where emission rates may be higher. Therefore, the Agency is finalizing a fPM limit of 0.010
lb/MMBtu with the use of PM CEMS as the only means of compliance demonstration. The EPA
has determined that this combination of fPM limit and compliance demonstration represents the
most stringent option taking into account the statutory considerations.
1.3.2 Baseline and Analysis Years
The impacts of regulatory actions are evaluated relative to a baseline that represents the
world without the action. This version of the model ("EPA's Power Sector Modeling Platform
2023") used for the baseline in this RIA includes recent updates to state and federal legislation
affecting the power sector, including Public Law 117-169, 136 Stat. 1818 (August 16, 2022),
commonly known as the Inflation Reduction Act of 2022 (IRA). The modeling documentation
includes a summary of all legislation reflected in this version of the model as well as a
description of how that legislation is implemented in the model.17 Also, see Section 3.3 for
additional detail about the power sector baseline for this RIA.
The year 2028 is the first year of detailed power sector modeling for this RIA and
approximates when the regulatory impacts of the final rule on the power sector will begin.18'19 In
17 Documentation for EPA's Power Sector Modeling Platform 2023 using IPM can be found at
https://www.epa.gov/power-sector-modeling and is available in the docket for this action. For information regarding
inclusion of the IRA in the baseline, see section 3.10.4 and 4.5.
18 Note that the Agency has granted the maximum time allowed for compliance under CAA section 112(i)(3) of
three years, and individual facilities may seek, if warranted, an additional 1-year extension of the compliance from
their permitting authority pursuant to CAA section 112(i)(3)(B). Facilities may also request, if warranted,
emergency authority to operate through the Department of Energy under section 202(c) of the Federal Power Act.
19 We note that, while the compliance date of the rule will likely be mid- to late-2027 and all compliance costs are
accounted for, any emissions reductions and benefits that in occur over a few months in 2027 are omitted from this
analysis.
1-8
-------
addition, the regulatory impacts are evaluated for the specific analysis years of 2030 and 2035.
These results are used to estimate the PV and EAV of the 2028 through 2037 period.
1.4 Organization of the Regulatory Impact Analysis
This RIA is organized into the following remaining sections:
• Section 2: Power Sector Industry Profile. This section describes the electric power
sector in detail.
• Section 3: Cost, Emissions, and Energy Impacts. The section summarizes the projected
compliance costs and other energy impacts associated with the regulatory options.
• Section 4: Benefits Analysis. The section presents the projected health and
environmental benefits of reductions in emissions of HAP, direct PM2.5, and PM2.5 and
ozone precursors and the climate benefits of CO2 emissions reductions across regulatory
options.
• Section 5: Economic Impacts. The section includes a discussion of potential small
entity, economic, and labor impacts.
• Section 6: Environmental Justice Impacts. This section includes an assessment of
potential impacts to potential EJ populations.
• Section 7: Comparison of Benefits and Costs. The section compares of the total
projected benefits with total projected costs and summarizes the projected net benefits of
the three regulatory options examined. The section also includes a discussion of potential
benefits that EPA is unable to quantify and monetize.
1.5 References
OMB. (2003). Circular A-4: Regulatory Analysis. Washington DC.
https://www.whitehouse.gov/wp-
content/uploads/legacy_drupal_files/omb/circulars/A4/a-4.pdf
OMB. (2023). Circular A-4: Regulatory Analysis. Washington DC.
https://www.whitehouse.gov/wp-content/uploads/2023/ll/CircularA-4.pdf
1-9
-------
INDUSTRY PROFILE
2.1 Background
In the past decade, there have been substantial structural changes in both the mix of
generating capacity and in the share of electricity generation supplied by different types of
generation. These changes are the result of multiple factors in the power sector, including
replacements of older generating units with new units, changes in the electricity intensity of the
U.S. economy, growth and regional changes in the U.S. population, technological improvements
in electricity generation from both existing and new units, changes in the prices and availability
of different fuels, and substantial growth in electricity generation from renewable energy
sources. Many of these trends will likely continue to contribute to the evolution of the power
sector.20 The evolving economics of the power sector, specifically the increased natural gas
supply and subsequent relatively low natural gas prices, have resulted in more natural gas being
used to produce both base and peak load electricity. Additionally, rapid growth in the
deployment of wind and solar technologies has led to their now constituting a significant share of
generation. The combination of these factors has led to a decline in the share of electricity
generated from coal. This section presents data on the evolution of the power sector over the past
two decades from 2010 through 2022, as well as a focus on the period 2015 through 2022.
Projections of future power sector behavior and the projected impacts of the final rule are
discussed in more detail in Section 3 of this RIA.
2.2 Power Sector Overview
The production and delivery of electricity to customers consists of three distinct
segments: generation, transmission, and distribution.
2.2.1 Generation
Electricity generation is the first process in the delivery of electricity to consumers. There
are two important aspects of electricity generation: capacity and net generation. Generating
Capacity refers to the maximum amount of production an EGU is capable of producing in a
20 For details on the evolution of EPA's power sector projections, please see archive of IPM outputs available at:
epa.gov/power-sector-modeling.
2-1
-------
typical hour, typically measured in megawatts (MW) for individual units, or gigawatts (1 GW =
1000 MW) for multiple EGUs. Electricity Generation refers to the amount of electricity actually
produced by an EGU over some period of time, measured in kilowatt-hours (kWh) or gigawatt-
hours (1 GWh = 1 million kWh). Net Generation is the amount of electricity that is available to
the grid from the EGU (i.e., excluding the amount of electricity generated but used within the
generating station for operations). Electricity generation is most often reported as the total annual
generation (or some other period, such as seasonal). In addition to producing electricity for sale
to the grid, EGUs perform other services important to reliable electricity supply, such as
providing backup generating capacity in the event of unexpected changes in demand or
unexpected changes in the availability of other generators. Other important services provided by
generators include facilitating the regulation of the voltage of supplied generation.
Individual EGUs are not used to generate electricity 100 percent of the time. Individual
EGUs are periodically not needed to meet the regular daily and seasonal fluctuations of
electricity demand. Units are also unavailable during routine and unanticipated outages for
maintenance. Furthermore, EGUs relying on renewable resources such as wind, sunlight, and
surface water to generate electricity are routinely constrained by the availability of adequate
wind, sunlight, or water at different times of the day and season. These factors result in the share
of potential generating capacity being substantially different from the share of actual electricity
produced by each type of EGU in a given season or year.
Most of the existing capacity generates electricity by creating heat to create high pressure
steam that is released to rotate turbines which, in turn, create electricity. Natural gas combined
cycle (NGCC) units have two generating components operating from a single source of heat. The
first cycle is a gas-fired combustion turbine, which generates electricity directly from the heat of
burning natural gas. The second cycle reuses the waste heat from the first cycle to generate
steam, which is then used to generate electricity from a steam turbine. Other EGUs generate
electricity by using water or wind to rotate turbines, and a variety of other methods including
direct photovoltaic generation also make up a small, but growing, share of the overall electricity
supply. The most common generating capacity includes fossil-fuel-fired units, nuclear units, and
hydroelectric and other renewable sources (see Table 2-1 and Table 2-2). Table 2-1 and Table
2-2 also show the comparison between the generating capacity in 2010 to 2022 and 2015 to
2022, respectively.
2-2
-------
In 2022 the power sector comprised a total capacity21 of 1,201 GW, an increase of 162
GW (or 16 percent) from the capacity in 2010 (1,039 GW). The largest change over this period
was the decline of 127 GW of coal capacity, reflecting the retirement/rerating of close to 40
percent of the coal fleet. This reduction in coal capacity was offset by increases in natural gas,
solar, and wind capacities of 95 GW, 72 GW, and 102 GW respectively. Substantial amounts of
distributed solar (40 GW) were also added.
These trends persist over the shorter 2015-21 period as well; total capacity in 2022 (1,201
GW) increased by 127 GW (or 12 percent). The largest change in capacity was driven by a
reduction of 90 GW of coal capacity. This was offset by a net increase of 63 GW of natural gas
capacity, an increase of 69 GW of wind, and an increase of 59 GW of solar. Additionally, 30
GW of distributed solar were also added over the 2015-22 period.
Table 2-1 Total Net Summer Electricity Generating Capacity by Energy Source, 2010-
2022
2010
2022
Change Between '10
and '22
Energy Source
Net
Summer
Capacity
(GW)
% Total
Capacity
Net
Summer
Capacity
(GW)
Net
Summer
Capacity
(GW)
% Total
Capacity
Net
Summer
Capacity
(GW)
Coal
317
30%
189
16%
-40%
-127
Natural Gas
407
39%
502
42%
23%
95
Nuclear
101
10%
95
8%
-6%
-7
Hydro
101
10%
103
9%
2%
2
Petroleum
56
5%
31
3%
-45%
-25
Wind
39
4%
141
12%
261%
102
Solar
1
0%
73
6%
8310%
72
Distributed Solar
0
0%
40
3%
40
Other Renewable
14
1%
15
1%
7%
1
Misc
4
0%
12
1%
239%
9
Total
1,039
100%
1,201
100%
16%
162
Source: EIA. Electric Power Annual 2022, Table 3. l.A and 3.l.B
21 This includes generating capacity at EGUs primarily operated to supply electricity to the grid and combined heat
and power facilities classified as Independent Power Producers (IPP) and excludes generating capacity at
commercial and industrial facilities that does not operate primarily as an EGU. Natural Gas information in this
section (unless otherwise stated) reflects data for all generating units using natural gas as the primary fossil heat
source. This includes Combined Cycle Combustion Turbine, Gas Turbine, steam, and miscellaneous (< 1 percent).
2-3
-------
Table 2-2 Total Net Summer Electricity Generating Capacity by Energy Source, 2015-
2022
2015
2022
Change Between '15
and '22
Energy Source
Net
Summer
Capacity
(GW)
% Total
Capacity
Net
Summer
Capacity
(GW)
% Total
Capacity
%
Increase
Capacity
Change
(GW)
Coal
280
26%
189
16%
-32%
-90
Natural Gas
439
41%
502
42%
14%
63
Nuclear
99
9%
95
8%
-4%
-4
Hydro
102
10%
103
9%
1%
1
Petroleum
37
3%
31
3%
-16%
-6
Wind
73
7%
141
12%
95%
69
Solar
14
1%
73
6%
433%
59
Distributed Solar
10
1%
40
3%
307%
30
Other Renewable
17
2%
15
1%
-11%
-2
Misc
4
0%
12
1%
182%
8
Total
1,074
100%
1,201
100%
12%
127
Source: EIA. Electric Power Annual 2022, Table 3. l.A and 3.l.B
The average age of coal-fired power plants that retired between 2015 and 2023 was over
50 years. Older power plants tend to become uneconomic over time as they become more costly
to maintain and operate, and as newer and more efficient alternative generating technologies are
built. As a result, coal's share of total U.S. electricity generation has been declining for over a
decade, while generation from natural gas and renewables has increased significantly.22 As
shown in Figure 2-1 below, 70 percent of the coal fleet in 2023 had an average age of over 40
years.
22 EIA, Today in Energy (April 17, 2017) available at https://www.eia.gov/todayinenergy/detail.php?id=30812.
2-4
-------
0-10yrs 10 - 20yrs 20-30yrs 30-40yrs 40-50yrs 50-60yrs 60 + yrs
Figure 2-1 National Coal-fired Capacity (GW) by Age of EGU, 2023
Source: NEEDS v6
In 2022, electric generating sources produced a net 4,292 TWh to meet national
electricity demand, which was around 4 percent higher than 2010. As presented in Table 2-2, 60
percent of electricity in 2022 was produced through the combustion of fossil fuels, primarily coal
and natural gas, with natural gas accounting for the largest single share. The total generation
share from fossil fuels in 2022 (60 percent) was 10 percent less than the share in 2010 (70
percent). Moreover, the share of fossil generation supplied by coal fell from 65 percent in 2010
to 33 percent by 2022, while the share of fossil generation supplied by natural gas rose from 35
percent to 67 percent over the same period. In absolute terms, coal generation declined by 55
percent, while natural gas generation increased by 71 percent. This reflects both the increase in
natural gas capacity during that period as well as an increase in the utilization of new and
existing gas EGUs during that period. The combination of wind and solar generation also grew
from 2 percent of the mix in 2010 to 14 percent in 2022.
2-5
-------
Table 2-3 Net Generation by Energy Source, 2010 to 2022 (Trillion kWh = TWh)
2010
2022
Change Between '10
and '22
Net
Fuel
Net
Fuel
Generation
Energy Source
Generation
Source
Generation
Source
Change
(TWh)
Share
(TWh)
Share
111 1
(TWh)
Coal
1,847
45%
832
19%
-55%
-1,016
Natural Gas
988
24%
1,687
39%
71%
699
Nuclear
807
20%
772
18%
-4%
-35
Hydro
255
6%
249
6%
-2%
-6
Petroleum
37
1%
23
1%
-38%
-14
Wind
95
2%
434
10%
359%
340
Solar
1
0%
144
3%
11764%
143
Distributed Solar
0
0%
61
1%
61
Other Renewable
71
2%
68
2%
-5%
-3
Misc
24
1%
23
1%
-6%
-1
Total
4,125
100%
4,292
100%
4%
167
Table 2-4 Net Generation by Energy Source, 2015 to 2022 (Trillion kWh =
TWh)
2015
2022
Change Between '15
and '22
Net
Fuel
Net
Fuel
OA
Generation
Energy Source
Generation
Source
Generation
Source
Change
(TWh)
Share
(TWh)
Share
X11V1
(TWh)
Coal
1,352
33%
832
19%
-39%
-521
Natural Gas
1,335
33%
1,687
39%
27%
354
Nuclear
797
19%
772
18%
-3%
-26
Hydro
249
6%
249
6%
2%
5
Petroleum
28
1%
23
1%
-19%
-5
Wind
191
5%
434
10%
128%
244
Solar
25
1%
144
3%
478%
119
Distributed Solar
14
0%
61
1%
333%
47
Other Renewable
80
2%
68
2%
-15%
-12
Misc
27
1%
23
1%
-16%
-4
Total
4,092
100%
4,292
100%
5%
200
2-6
-------
Coal-fired and nuclear generating units have historically supplied "base load" electricity,
meaning that these units operate through most hours of the year and serve the portion of
electricity load that is continually present. Although much of the coal fleet has historically
operated as base load, there can be notable differences in the design of various facilities (see
Table 2-3 and Table 2-4) which, along with relative fuel prices, can impact the operation of coal-
fired power plants. As one example of design variations, coal-fired units less than 100 MW in
size comprise 17 percent of the total number of coal-fired units, but only 2 percent of total coal-
fired capacity, and they tend to have higher heat rates. Gas-fired generation is generally better
able to vary output, is a primary option used to meet the variable portion of the electricity load
and has historically supplied "peak" and "intermediate" power, when there is increased demand
for electricity (for example, when businesses operate throughout the day or when people return
home from work and run appliances and heating/air-conditioning), versus late at night or very
early in the morning, when demand for electricity is reduced. Over the last decade, however, the
generally low price of natural gas and the growing age of the coal fleet has resulted in increasing
capacity factors for many gas-fired plants and decreasing capacity factors for many coal-fired
plants. As shown in Figure 2-2, average annual coal capacity factors have declined from 67
percent to 50 percent over the 2010 to 2022 period, indicating that a larger share of units are
operating in non-baseload fashion. Over the same period, natural gas combined cycle capacity
factors have risen from an annual average of 44 percent to 57 percent.
2-7
-------
70%
3
£
< 20%
10%
0%
2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022
— — — Coal — — — Natural Gas
Figure 2-2 Average Annual Capacity Factor by Energy Source
Source: EIA. Electric Power Annual 2022 Table 4.08 A
Table 2-5 also shows comparable data for the capacity and age distribution of coal and
natural gas units. Compared with the fleet of coal EGUs, the natural gas fleet of EGUs is
generally smaller and newer. While 69 percent of the coal EGU fleet capacity is over 500 MW
per unit, 82 percent of the gas fleet is between 50 and 500 MW per unit.
2-8
-------
Table 2-5 Coal and Natural Gas Generating Units, by Size, Age, Capacity, and Average
Heat Rate in 2023
Unit Size
Grouping
(MW)
No. Units
% of All
Units
Avg. Age
Avg. Net
Summer
Capacity
(MW)
Total Net
Summer
Capacity
(MW)
% Total
Capacity
Avg. Heat
Rate
(Btu/kWh)
COAL
0-24
25-49
50-99
100 - 149
150-249
250-499
500 - 749
750 - 999
1000 - 1500
17
27
20
24
38
95
104
44
9
4%
7%
5%
6%
10%
25%
28%
12%
2%
56
37
32
52
47
42
41
39
46
13
36
76
120
195
379
612
818
1,264
218
978
I,510
2,869
7,394
36,008
63,604
35,979
II,380
0%
1%
1%
2%
5%
23%
40%
22%
7%
12,103
11,739
11,858
11,195
10,809
10,660
10,243
10,167
9,813
Total Coal
378
100%
42
423
159,940
100%
10,722
NATURAL GAS
0-24
25-49
50-99
100 - 149
150-249
250-499
500 - 749
750 - 999
1000 - 1500
4,679
899
1,000
391
1,037
309
47
8
0
56%
11%
12%
5%
12%
4%
1%
0%
0%
30
26
29
26
20
21
30
47
4
41
72
125
180
330
585
838
20,963
36,619
71,611
48,863
186,503
101,969
27,495
6,706
0
4%
7%
14%
10%
37%
20%
5%
1%
0%
13,006
11,545
12,194
9,548
8,194
8,072
9,374
11,366
Total Gas
8,362
100%
27
60
500,730
100%
11,790
Source: National Electric Energy Data System (NEEDS) v.6
Note: The average heat rate reported is the mean of the heat rate of the units in each size category (as opposed to a
generation-weighted or capacity-weighted average heat rate.) A lower heat rate indicates a higher level of fuel
efficiency.
In terms of the age of the generating units, almost 67 percent of the total coal generating
capacity has been in service for more than 40 years, while nearly 81 percent of the natural gas
capacity has been in service less than 40 years. Figure 2-3 presents the cumulative age
distributions of the coal and gas fleets, highlighting the pronounced differences in the ages of the
fleets of these two types of fossil-fuel generating capacity. Figure 2-3 also includes the
distribution of generation, which is similar to the distribution of capacity.
2-9
-------
100%
90%
80%
70%
60%
50%
40%
30%
20%
10%
0%
— *
/
//
i
/
I
4
/,
//
V M
1
1
//
/ /
//
If
//
#i
u
ll
/ /
//
//
/ /
~ /
~ /
i
/i
//
//
//
*
/
t
~ >
0 4
10 20 30 40
Age of EGU (years)
50
60
70
¦ Gas Cap
¦ Gas Gen
¦ Coal Cap
1 Coal Gen
Figure 2-3 Cumulative Distribution in 2021 of Coal and Natural Gas Electricity
Capacity and Generation, by Age
Source: eGRID 2021 (November 2023 release from EPA eGRID website). Figure presents data from generators that
came online between 1950 and 2021 (inclusive); a 71-year period. Full eGRID data include generators that came
online as far back as 1915. Full data from 1915 onward are used in calculating cumulative distributions; figure
truncation at 70 years is merely to improve visibility of diagram.
2-10
-------
The locations of existing fossil units in EPA's National Electric Energy Data System
(NEEDS) v.6 are shown in Figure 2-4.
Figure 2-4 Fossil Fuel-Fired Electricity Generating Facilities, by Size
Source: National Electric Energy Data System (NEEDS) v.6
Note: This map displays fossil capacity at facilities in the NEEDS v.6 IPM frame. NEEDS v.6 reflects generating
capacity expected to be on-line at the end of 2023. This includes planned new builds already under construction and
planned retirements. In areas with a dense concentration of facilities, some facilities may be obscured.
The costs of renewable generation have fallen significantly due to technological
advances, improvements in performance, and local, state, and federal incentives such as the
recent extension of federal tax credits. According to Lazard, a financi al advisory and asset
management firm, the current unsubsidized levelized cost of electricity for wind and solar energy
technologies is lower than the cost of technologies like coal, natural gas or nuclear, and in some
cases even lower than just the operating cost, which is expected to lead to ongoing and
significant deployment of renewable energy. Levelized cost of electricity is only one metric used
to compare the cost of different generating technologies. It contains a number of uncertainties
including utilization and regional factors.23 While this chart illustrates general trends, unit
specific build decisions will incorporate many other variables. These trends of declining costs
23 Lazard, Levelized Cost of Energy Analysis-Version 16.0,2023. https://wmiJazardxom/media/iypdgxnmi/lazards-
Icoephis-april-2023.pdf
£
o
2-11
-------
and cost projections for renewable resources are borne out by a range of other studies including
the NREL Annual Technology Baseline,24 DOE's Land-Based Wind Market Report,25 LBNL's
Utility Scale solar report,26 EIA's Annual Energy Outlook,27 and DOE's 2022 Grid Energy
Storage Technology Cost and Performance Assessment,-8
Selected Historical Mean Unsubsidized LCOE Values11'
Mean LCOE
(VMWh)
\
Figure 2-5 Selected Historical Mean LCOE Values
Source: Lazard, Levelized Cost of Energy Analysis-Version 16.0. April 2023
The broad trends away from coal-fired generation and toward lower-emitting generation
are reflected in the recent actions and recently announced plans of many power plants across the
industry — spanning all types of companies in all locations. Throughout the country, utilities
have included commitments towards cleaner energy in public releases, planning documents, and
integrated resource plans (IRPs). For strategic business reasons and driven by the economics of
different supply options, most major utilities plan to increase their renewable energy holdings
and continue reducing GFIG emissions, regardless of what federal regulatory requirements might
exist.
241 Available at: httpszOkitb.rireI.gov/.
25 Available at: https'JMww.energy.gov/eere/wind'articles/land-based-wind-market-report-2022-edition.
26 Available at: https:IkmpJbl.gov/UtUity-scale-solar/:
27 Available at: https.'Mvww. eia.gov/outlooks/aeo/pdf/electricity_generation.pdf.
28 Available at: https:Mvww.energv.gov/eere/anaIysis/2022-gnd-energy-storage-technobgy-cost-and-performance-
assessment.
2-12
-------
While EPA does not account for future planning statements from utility providers in the
economic modeling since they are not legally enforceable, the number and scale of these
announcements is significant on a systemic level. These statements are part of long-term
planning processes that cannot be easily revoked due to considerable stakeholder involvement in
the planning process, including the involvement of regulators. The direction to which these
utility providers have publicly stated they are moving is consistent across the sector and
undergirded by market fundamentals lending economic credibility to these commitments and
confidence that that most plans will be implemented.
2.2.2 Transmission
Transmission is the term used to describe the bulk transfer of electricity over a network
of high voltage lines, from electric generators to substations where power is stepped down for
local distribution. In the U.S. and Canada, there are three separate interconnected networks of
high voltage transmission lines,29 each operating synchronously. Within each of these
transmission networks, there are multiple areas where the operation of power plants is monitored
and controlled by regional organizations to ensure that electricity generation and load are kept in
balance. In some areas, the operation of the transmission system is under the control of a single
regional operator;30 in others, individual utilities31 coordinate the operations of their generation,
transmission, and distribution systems to balance the system across their respective service
territories.
2.2.3 Distribution
Distribution of electricity involves networks of lower voltage lines and substations that
take the higher voltage power from the transmission system and step it down to lower voltage
levels to match the needs of customers. The transmission and distribution system is the classic
example of a natural monopoly, in part because it is not practical to have more than one set of
29 These three network interconnections are the Western Interconnection, comprising the western parts of both the
U.S. and Canada (approximately the area to the west of the Rocky Mountains), the Eastern Interconnection,
comprising the eastern parts of both the U.S. and Canada (except those part of eastern Canada that are in the Quebec
Interconnection), and the Texas Interconnection (which encompasses the portion of the Texas electricity system
commonly known as the Electric Reliability Council of Texas (ERCOT)). See map of all NERC interconnections at
https://www.nerc.com/AboutNERC/keyplayers/PublishingImages/NERC%20Interconnections.pdf.
30 For example, PJM Interconnection, LLC.
31 For example, Los Angeles Department of Water and Power, Florida Power and Light.
2-13
-------
lines running from the electricity generating sources to substations or from substations to
residences and businesses.
Over the last few decades, several jurisdictions in the U.S. began restructuring the power
industry to separate transmission and distribution from generation, ownership, and operation.
Historically, vertically integrated utilities established much of the existing transmission
infrastructure. However, as parts of the country have restructured the industry, transmission
infrastructure has also been developed by transmission utilities, electric cooperatives, and
merchant transmission companies, among others. Distribution, also historically developed by
vertically integrated utilities, is now often managed by a number of utilities that purchase and
sell electricity, but do not generate it. Electricity restructuring has focused primarily on efforts to
reorganize the industry to encourage competition in the generation segment of the industry,
including ensuring open access of generation to the transmission and distribution services needed
to deliver power to consumers. In many states, such efforts have also included separating
generation assets from transmission and distribution assets to form distinct economic entities.
Transmission and distribution remain price-regulated throughout the country based on the cost of
service.
2.3 Sales, Expenses, and Prices
Electric generating sources provide electricity for ultimate commercial, industrial, and
residential customers. Each of the three major ultimate categories consume roughly a quarter to a
third of the total electricity produced (see Table 2-6).32 Some of these uses are highly variable,
such as heating and air conditioning in residential and commercial buildings, while others are
relatively constant, such as industrial processes that operate 24 hours a day. The distribution
between the end use categories changed very little between 2010 and 2022.
32 Transportation (primarily urban and regional electrical trains) is a fourth ultimate customer category which
accounts less than one percent of electricity consumption.
2-14
-------
Table 2-6 Total U.S. Electric Power Industry Retail Sales, 2010-22 and 2014-22 (billion
kWh)
2010
2022
Sales/Direct , ,
tt /t.-..- Share oi Total
Use (Billion _ , TT
, , End Use
kWh)
Sales/Direct , ,
tt /t.-..- Share oi Total
Use (Billion _ , XT
, , End Use
kWh)
Residential
Commercial
Sales
Industrial
Transportation
1,446 37%
1,330 34%
971 25%
8 0%
1,509 37%
1,391 34%
1,020 25%
7 0%
Total
3,755 97%
3,927 97%
Direct Use
132
140
Total End Use
3,887
4,067
2015
2022
Sales/Direct , ,
tt /t.-..- Share oi Total
Use (Billion , TT
¦ «7i-\ End Use
kWh)
Sales/Direct , ,
tt /t.-..- Share oi Total
Use (Billion , TT
¦ «7i-\ End Use
kWh)
Residential
Commercial
Sales
Industrial
Transportation
1,404 36%
1,361 35%
987 25%
8 0%
1,509 37%
1,391 34%
1,020 25%
7 0%
Total
3,759 96%
3,927 97%
Direct Use
141
140
Total End Use
3,900
4,067
Source: Table 2.2, EIA Electric Power Annual, 2022 (October 19, 2023, release)
Notes: Retail sales are not equal to net generation (Table 2-2) because net generation includes net imported
electricity and loss of electricity that occurs through transmission and distribution, along with data collection frame
differences and non-sampling error. Direct Use represents commercial and industrial facility use of onsite net
electricity generation; electricity sales or transfers to adjacent or co-located facilities; and barter transactions.
2.3.1 Electricity Prices
Electricity prices vary substantially across the U.S., differing both between the ultimate
customer categories and by state and region of the country. Electricity prices are typically
highest for residential and commercial customers because of the relatively high costs of
distributing electricity to individual homes and commercial establishments. The higher prices for
residential and commercial customers are the result of the extensive distribution network
reaching to virtually every building in every part of the country and the fact that generating
stations are increasingly located relatively far from population centers, increasing transmission
costs. Industrial customers generally pay the lowest average prices, reflecting both their
proximity to generating stations and the fact that industrial customers receive electricity at higher
2-15
-------
voltages (which makes transmission more efficient and less expensive). Industrial customers
frequently pay variable prices for electricity, varying by the season and time of day, while
residential and commercial prices have historically been less variable. Overall, industrial
customer prices are usually considerably closer to the wholesale marginal cost of generating
electricity than residential and commercial prices.
On a state-by-state basis, all retail electricity prices vary considerably. In 2022, the
national average retail electricity price (all sectors) was 12.4 cents/kWh, with a range from 8.2
cents (Wyoming) to 39.72 cents (Hawaii).33
The real year prices for 2010 through 2022 are shown in Figure 2-6. Average national
retail electricity prices decreased between 2010 and 2022 by 4 percent in real terms (2022
dollars), and 2 percent between 2015-22.34 The amount of decrease differed for the three major
end use categories (residential, commercial, and industrial). National average commercial prices
decreased the most (4 percent), and industrial prices decreased the least (1 percent) between
2015-21.
18.0
w
2 16.0
£ 6.0
o 4.0
jjj 2.0
LU
0.0
2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021
Residential Commercial Industrial — — — Total
Figure 2-6 Real National Average Electricity Prices (including taxes) for Three Major
End-Use Categories
Source: EIA. Electric Power Annual 2022 and 2021, Table 2.4.
33 EIA State Electricity Profiles with Data for 2022 {http://www.eia.gov/electricity/state/).
34 All prices in this section are estimated as real 2022 prices adjusted using the GDP implicit price deflator unless
otherwise indicated.
2-16
-------
2.3.2 Prices of Fossil Fuel Used for Generating Electricity
Another important factor in the changes in electricity prices are the changes in delivered
fuel prices35 for the three major fossil fuels used in electricity generation: coal, natural gas, and
petroleum products. Relative to real prices in 2015, the national average real price (in 2022
dollars) of coal delivered to EGUs in 2022 had decreased by 12 percent, while the real price of
natural gas increased by 84 percent. The real price of delivered petroleum products also
increased by 102 percent, and petroleum products declined as an EGU fuel (in 2022 petroleum
products generated 1 percent of electricity). The combined real delivered price of all fossil fuels
(weighted by heat input) in 2022 increased by 62 percent over 2015 prices. Figure 2-7 shows the
relative changes in real price of all three fossil fuels between 2010 and 2022.
-80%
Coal Petroleum Natural Gas
-60%
aj
HD
TO
c
a>
-20%
-40%
0%
2010
2022
Figure 2-7 Relative Real Prices of Fossil Fuels for Electricity Generation; Change in
National Average Real Price per MMBtu Delivered to EGU
Source: EI A. Electric Power Annual 2022, Table 7.1.
35 Fuel prices in this section are all presented in terms of price per MMBtu to make the prices comparable.
2-17
-------
2.3.3 Changes in Electricity Intensity of the U.S. Economy from 2010 to 2021
An important aspect of the changes in electricity generation (i.e., electricity demand)
between 2010 and 2022 is that while total net generation increased by 4 percent over that period,
the demand growth for generation was lower than both the population growth (8 percent) and
real GDP growth (30 percent). Figure 2-8 shows the growth of electricity generation, population,
and real GDP during this period.
35%
Real GDP Generation Population
Figure 2-8 Relative Growth of Electricity Generation, Population and Real GDP Since
2010
Sources: Generation: U.S. EIA Electric Power Annual 2022. Population: U.S. Census. Real GDP: U.S. Bureau of
Economic Analysis
Because demand for electricity generation grew more slowly than both the population
and GDP, the relative electric intensity of the U.S. economy improved (i.e., less electricity used
per person and per real dollar of output) during 2010 to 2022. On a per capita basis, real GDP per
capita grew by 20 percent between 2010 and 2022. At the same time, electricity generation per
capita decreased by 3 percent. The combined effect of these two changes improved the overall
electricity generation efficiency in the U.S. market economy. Electricity generation per dollar of
real GDP decreased 20 percent. These relative changes are shown in Figure 2-9.
2-18
-------
Figure 2-9 Relative Change of Real GDP, Population and Electricity Generation
Intensity Since 2010
Sources: Generation: U.S. EIA Electric Power Annual 2021 and 2020. Population: U.S. Census. Real GDP: 2022
Economic Report of the President, Table B-3.
2-19
-------
COSTS, EMISSIONS, AND ENERGY IMPACTS
3.1 Introduction
This section presents the compliance cost, emissions, and energy impact analysis
performed for the MATS RTR. EPA used the Integrated Planning Model (IPM), developed by
ICF Consulting, to conduct its analysis. IPM is a dynamic linear programming model that can be
used to examine air pollution control policies for SO2, NOx, Hg, HC1, PM, and other air
pollutants throughout the U.S. for the entire power system. Documentation for EPA's Power
Sector Modeling Platform 2023 using IPM (hereafter IPM Documentation) can be found at
https://www.epa.gov/power-sector-modeling and is available in the docket for this action.
3.2 EPA's Power Sector Modeling Platform 2023 using IPM
IPM is a state-of-the-art, peer-reviewed, dynamic linear programming model that can be
used to project power sector behavior under future business-as-usual conditions and to examine
prospective air pollution control policies throughout the contiguous U.S. for the entire electric
power system. For this RIA, EPA used IPM to project likely future electricity market conditions
with and without this rulemaking.
IPM, developed by ICF, is a multi-regional, dynamic, deterministic linear programming
model of the contiguous U.S. electric power sector. It provides estimates of least cost capacity
expansion, electricity dispatch, and emissions control strategies while meeting energy demand
and environmental, transmission, dispatch, and reliability constraints. IPM's least-cost dispatch
solution is designed to ensure generation resource adequacy, either by using existing resources or
through the construction of new resources. IPM addresses reliable delivery of generation
resources for the delivery of electricity between the 78 IPM regions, based on current and
planned transmission capacity, by setting limits to the ability to transfer power between regions
using the bulk power transmission system. Notably, the model includes cost and performance
estimates for state-of-the-art air pollution control technologies with respect to Hg, fPM, and
other HAP controls.
EPA has used IPM for almost three decades to better understand power sector behavior
under future business-as-usual conditions and to evaluate the economic and emissions impacts of
prospective environmental policies. The model is designed to reflect electricity markets as
3-1
-------
accurately as possible. EPA uses the best available information from utilities, industry experts,
gas and coal market experts, financial institutions, and government statistics as the basis for the
detailed power sector modeling in IPM. The model documentation provides additional
information on the assumptions discussed here as well as all other model assumptions and
inputs.36
The model incorporates a detailed representation of the fossil-fuel supply system that is
used to estimate equilibrium fuel prices. The model uses natural gas fuel supply curves and
regional gas delivery costs (basis differentials) to simulate the fuel price associated with a given
level of gas consumption within the system. These inputs are derived using ICF's Gas Market
Model (GMM), a supply/demand equilibrium model of the North American gas market.37
IPM also endogenously models the partial equilibrium of coal supply and EGU coal
demand levels throughout the contiguous U.S., taking into account assumed non-power sector
demand and imports/exports. IPM reflects 36 coal supply regions, 14 coal grades, and the coal
transport network, which consists of over four thousand linkages representing rail, barge, and
truck and conveyer linkages. The coal supply curves in IPM were developed during a thorough
bottom-up, mine-by-mine approach that depicts the coal choices and associated supply costs that
power plants would face if selecting that coal over the modeling time horizon. The IPM
documentation outlines the methods and data used to quantify the economically recoverable coal
reserves, characterize their cost, and build the 36 coal regions' supply curves.38
To estimate the annualized costs of additional capital investments in the power sector,
EPA uses a conventional and widely accepted approach that applies a capital recovery factor
(CRF) multiplier to capital investments and adds that to the annual incremental operating
expenses. The CRF is derived from estimates of the power sector's cost of capital (i.e., private
discount rate), the amount of insurance coverage required, local property taxes, and the life of
capital.39 It is important to note that there is no single CRF factor applied in the model; rather, the
36 Detailed information and documentation of EPA's Baseline run using EPA's Power Sector Modeling Platform
2023 using IPM, including all the underlying assumptions, data sources, and architecture parameters can be found
on EPA's website at: https://www.epa.gov/power-sector-modeling.
37 See Chapter 8 of EPA's IPM Documentation, available at: https://www.epa.gov/power-sector-modeling.
38 See Chapter 7 EPA's IPM Documentation, available at: https://www.epa.gov/power-sector-modeling.
39 See Chapter 10 of EPA's IPM Documentation, available at: https://www.epa.gov/power-sector-modeling.
3-2
-------
CRF varies across technologies, book life of the capital investments, and regions in the model in
order to better simulate power sector decision-making.
EPA has used IPM extensively over the past three decades to analyze options for
reducing power sector emissions. Previously, the model has been used to estimate the costs,
emission changes, and power sector impacts in the RIAs for the Clean Air Interstate Rule (U.S.
EPA, 2005), the Cross-State Air Pollution Rule (U.S. EPA, 201 la), the Mercury and Air Toxics
Standards (U.S. EPA, 201 lb), the Clean Power Plan for Existing Power Plants (U.S. EPA,
2015b), the Cross-State Air Pollution Update Rule (U.S. EPA, 2016), the Repeal of the Clean
Power Plan, and the Emission Guidelines for Greenhouse Gas Emissions from Existing Electric
Utility Generating Units (U.S. EPA, 2019), the Revised Cross-State Air Pollution Update Rule
(U.S. EPA, 2021), and the Good Neighbor Plan (2023b).
EPA has also used IPM to estimate the air pollution reductions and power sector impacts
of water and waste regulations affecting EGUs, including contributing to RIAs for the Cooling
Water Intakes (316(b)) Rule (U.S. EPA, 2014a), the Disposal of Coal Combustion Residuals
from Electric Utilities rule (U.S. EPA, 2015c), the Steam Electric Effluent Limitation Guidelines
(U.S. EPA, 2015a), and the Steam Electric Reconsideration Rule (U.S. EPA, 2020).
The model and EPA's input assumptions undergo periodic formal peer review. The
rulemaking process also provides opportunity for expert review and comment by a variety of
stakeholders, including owners and operators of capacity in the electricity sector that is
represented by the model, public interest groups, and other developers of U.S. electricity sector
models. The feedback that the Agency receives provides a highly detailed review of key input
assumptions, model representation, and modeling results. IPM has received extensive review by
energy and environmental modeling experts in a variety of contexts. For example, in September
2019, U.S. EPA commissioned a peer review40 of EPA's v6 Reference Case using the Integrated
Planning Model (IPM). Additionally, and in the late 1990s, the Science Advisory Board
reviewed IPM as part of the CAA Amendments Section 812 prospective studies41 that are
periodically conducted. The Agency has also used the model in a number of comparative
modeling exercises sponsored by Stanford University's Energy Modeling Forum over the past 20
40 See Response and Peer Review Report EPA Reference Case Version 6 Using IPM, available at:
https://www.epa.gov/power-sector-modeling/ipm-peer-reviews.
41 http://www2.epa.gov/clean-air-act-overview/benefits-ancl-costs-clean-air-act.
3-3
-------
years. IPM has also been employed by states (e.g., for the Regional Greenhouse Gas Initiative,
the Western Regional Air Partnership, Ozone Transport Assessment Group), other Federal and
state agencies, environmental groups, and industry.
3.3 Baseline
The modeled "baseline" for any regulatory impact analysis is a business-as-usual
scenario that represents expected behavior in the electricity sector under market and regulatory
conditions in the absence of a regulatory action. As such, the baseline run represents an element
of the baseline for this RIA.42 EPA frequently updates the baseline modeling to reflect the latest
available electricity demand forecasts from the U.S. EIA as well as expected costs and
availability of new and existing generating resources, fuels, emission control technologies, and
regulatory requirements.
For our analysis of the MATS RTR rule, EPA used EPA's Power Sector Modeling
Platform 2023 using IPM to provide power sector emissions projections for air quality modeling,
as well as a companion updated database of EGU units (the National Electricity Energy Data
System or NEEDS for IPM 202343) that is used in EPA's modeling applications of IPM. The
baseline for this final rule includes the Good Neighbor Plan (Final GNP), the Revised CSAPR
Update, CSAPR Update, and CSAPR, as well as MATS. The baseline run also includes the 2015
Effluent Limitation Guidelines (ELG) and the 2015 Coal Combustion Residuals (CCR), and the
recently finalized 2020 ELG and CCR rules.44
This version of the model, which is used as the baseline for this RIA, also includes recent
updates to state and federal legislation affecting the power sector, including Public Law 117-169,
136 Stat. 1818 (August 16, 2022), commonly known as the Inflation Reduction Act of 2022 (the
IRA). The IPM Documentation includes a summary of all legislation reflected in this version of
the model as well as a description of how that legislation is implemented in the model.
42 As described in Chapter 5 of EPA's Guidelines for Preparing Economic Analyses, the baseline "should
incorporate assumptions about exogenous changes in the economy that may affect relevant benefits and costs (e.g.,
changes in demographics, economic activity, consumer preferences, and technology), industry compliance rates,
other regulations promulgated by EPA or other government entities, and behavioral responses to the proposed rule
by firms and the public." (U.S. EPA, 2014b).
43 https://www.epa.gov/power-sector-modeling/national-electric-energy-data-system-needs.
44 For a full list of modeled policy parameters, please see: https://www.epa.gov/power-sector-modeling.
3-4
-------
Under the baseline, the impacts of the IRA result in an acceleration of the ongoing shift
towards lower emitting generation and declining generation share for fossil-fuel fired generation.
A range of studies have outlined how reliability continues to be maintained under high variable
renewable penetration scenarios. U.S. EPA (2023a) summarized results from fourteen multi-
sector and power sector models under the IRA in 2030 and 2035. Across the models, wind and
solar resources provide 22 to 54 percent of generation (with median of 45 percent) in 2030 and
21 to 80 percent (with median of 50 percent) in 2035. The North American Renewable
Integration Study (Brinkman et al., 2021) showed how the U.S. could accommodate between 70
to 79 percent of wind and solar generation by 2050. The Solar Futures Study (DOE, 2021)
illustrated power systems with upwards of 80 percent of renewable energy by 2050. Finally, Cole
et al. (2021) demonstrates a 100 percent renewable power system for the contiguous U.S.
The inclusion of the final GNP and other regulatory actions (including federal, state, and
local actions) in the base case is necessary in order to reflect the level of controls that are likely
to be in place in response to other requirements apart from the scenarios analyzed in this section.
This base case will provide meaningful projections of how the power sector will respond to the
cumulative regulatory requirements for air emissions in totality, while isolating the incremental
impacts of MATS RTR relative to a base case with other air emission reduction requirements
separate from this final action.
The analysis of power sector cost and impacts presented in this section is based on a
single policy run compared to the baseline run. The difference between the two runs represents
the incremental impacts projected solely as a result of compliance with the final MATS RTR.
3.4 Regulatory Options Analyzed
For this RIA, EPA analyzed the regulatory options summarized in the table below, which
are described in more detail in Section 1.3.1. The remainder of this section discusses the
approach used for estimating the costs and/or emissions impacts of each provision of this final
rule.
3-5
-------
Table 3-1 Summary of Final Regulatory Options Examined in this RIA
Regulatory Options Examined in this RIA
Provision
Less Stringent
Final Rule
FPM Standard (Surrogate
Standard for Non-Hg HAP
Metals)
Retain existing fPM standard of
0.030 lb/MMBtu
Revised fPM standard of 0.010
lb/MMBtu
Hg Standard
Retain Hg standard for lignite-fired
EGUs of 4.0 lb/TBtu
Revised Hg standard for lignite-
fired EGUs of 1.2 lb/TBtu
Continuous Emissions
Monitoring Systems (PM CEMS)
Require installation of PM CEMS
to demonstrate compliance
Require installation of PM CEMS
to demonstrate compliance
Startup Definition
Remove startup definition #2
Remove startup definition #2
As explained in Section 1.3.1, both the final rule and less stringent options described in
Table 3-1 have not been changed from the proposed and less stringent options examined in the
RIA for the proposal of this action. The proposal RIA included a more stringent regulatory
option that projected the impacts of lowering the fPM standard to 0.006 lb/MMBtu, while
holding the other three proposed amendments unchanged from the proposed option. EPA
solicited comment on this more stringent fPM standard in the preamble of the proposed rule. As
explained in section V.A.4. of the preamble of the final rule, EPA determined not to pursue a
more stringent standard for fPM emissions, such as a limit of 0.006 lb/MMBtu. After considering
comments to the proposed rule and after conducting additional analysis, EPA determined that a
lower fPM standard would not be compatible with PM CEMS due to measurement uncertainty.
As a result, this RIA does not examine a more stringent option than the suite of requirements that
constitute the final rule; the final rule represents the most stringent suite of regulatory options
available under the technology review.
The revisions to the fPM standard and the Hg standard are modeled endogenously within
IPM. For the fPM standard, emissions controls and associated costs are modeled based on
information available in the memorandum titled "2024 Update to the 2023 Proposed Technology
Review for the Coal- and Oil-Fired EGU Source Category," which is available in the docket.
This memorandum summarizes the fPM emissions rate for each existing EGU. Based on the
emissions rates detailed in this memorandum, EPA assumed various levels of O&M, ESP
3-6
-------
upgrades, upgrades to existing fabric filters, or new fabric filter installations to comply with each
of the finalized standards in the modeling. Those assumptions are detailed in Table 3-2.
Table 3-2 PM Control Technology Modeling Assumptions"
PM
Control Strategy
Cost (in 2019 dollars)
fPM Reduction
Operation &
Maintenance (O&M)
$100,000/year
Unit-specific
Minor
ESP Upgrades
$20/kW
20%
Typical
ESP Upgrades
$40/kW
40%
ESP Rebuild
$80/kW
55%
(0.0051b/MMBlu floor)
Upgrade Existing FF Bags
Unit-specific, approximately $15K
- $500K annual O&M
50%
(0.002 lb/MMBtu floor)
New Fabric Filter
(6.0 A/C Ratio)
Unit-specific,
$150-360/kW*
90%
(0.002 lb/MMBtu floor)
a Capital costs are expressed here in terms of $/kW. O&M costs are expressed here on an annual basis.
* https://www.epa.gov/system/files/documents/2021-09/attachment_5-
7_pm_control_cost_development_methodology.pdf
The cost and reductions associated with control of Hg emissions at lignite-fired EGUs are
also modeled endogenously and reflect the assumption that each of these EGUs replace standard
powdered activated carbon (PAC) sorbent with halogenated PAC sorbent.
While more detail on the costs associated with the PM CEMS requirement and the
change in the startup definition is presented in Section 3.5.2, we note here that these costs were
estimated exogenously without the use of the model that provides the bulk of the cost analysis
for this RIA. As a result, the results of the power sector modeling do not include costs associated
with these provisions, but the costs associated with requiring PM CEMS and the change in the
startup definition are included in the total cost projections for the rule for each of the regulatory
options analyzed in this RIA. As the incremental costs of requiring PM CEMS are small relative
to the ongoing costs of operations, we do not think the endogenous incorporation of these costs
would change any projected results in a meaningful way.
3-7
-------
3.5 Power Sector Impacts
3.5.1 Emissions
As indicated previously, this RIA presents emissions reductions estimates in years 2028,
2030, and 2035 based on IPM projections.45 Table 3-3 presents the estimated impact on power
sector emissions resulting from compliance with the final rule in the contiguous U.S. The
quantified emission estimates presented in the RIA include changes in pollutants directly covered
by this rule, such as Hg and non-Hg HAP metals, and changes in other pollutants emitted from
the power sector as a result of the compliance actions projected under this final rule. The model
projections capture the emissions changes associated with implementation of HAP mitigation
measures at affected sources as well as the resulting effects on dispatch as the relative operating
costs for some affected units have changed. The projections indicate that the final rule results in
reductions in emissions of Hg in all run years, of 16 percent, 17 percent, and 18 percent in 2028,
2030, and 2035, respectively, as well as reductions in PM2.5 and PM10 emissions in all run years.
45 Note that baseline mercury emissions projections are higher than proposal due to a revision in final baseline
modeling to better reflect current ACI performance at existing lignite-fired units.
3-8
-------
Table 3-3 EGU Emissions and Projected Emissions Changes for the Baseline and the
Final Rule for 2028, 2030, and 2035"
Total Emissions
Year
Baseline
Final Rule
Change from
Baseline
% Change
under Final
Rule
2028
6,129
5,129
-999.1
-16.3%
Hg (lbs.)
2030
5,863
4,850
-1,013
-17.3%
2035
4,962
4,055
-907.0
-18.3%
2028
70.5
69.7
-0.77
-1.09%
PM2.5 (thousand tons)
2030
66.3
65.8
-0.53
-0.79%
2035
50.7
50.2
-0.47
-0.93%
2028
79.5
77.4
-2.07
-2.60%
PM10 (thousand tons)
2030
74.5
73.1
-1.33
-1.79%
2035
56.0
54.8
-1.18
-2.11%
2028
454.3
454.0
-0.290
-0.06%
SO2 (thousand tons)
2030
333.5
333.5
0.025
0.01%
2035
239.9
239.9
-0.040
-0.02%
Ozone-season NOx
(thousand tons)
2028
2030
2035
189.0
174.99
116.99
188.8
175.4
119.1
-0.165
0.488
2.282282
-0.09%
0.28%
1.95%
Annual NOx (thousand
tons)
2028
2030
2035
460.55
392.88
253.44
460.3
392.7
253.5
-0.283
-0.022
0.066
-0.06%
-0.01%
0.03%
2028
2.474
2.474
0.000
0.01%
HC1 (thousand tons)
2030
2.184
2.184
0.000
0.01%
2035
1.484
1.485
0.001
0.06%
2028
1,158.8
1,158.7
-0.0655
-0.01%
CO2 (million metric tons)
2030
1,098.3
1,098.3
0.0361
0.00%
2035
724.2
724.1
-0.099
-0.01%
a This analysis is limited to the geographically contiguous lower 48 states. Values are independently rounded and
may not sum.
We also estimate that the final rule will reduce at least seven tons of non-Hg HAP metals in
2028, five tons of non-Hg HAP metals in 2030, and four tons of non-Hg HAP metals in 2035.
These reductions are composed of reductions in emissions of antimony, arsenic, beryllium,
3-9
-------
cadmium, chromium, cobalt, lead, manganese, nickel, and selenium.46 Table 3-4 summarizes the
total emissions reductions projected over the 2028 to 2037 analysis period.
Table 3-4 Cumulative Projected Emissions Reductions for the Final Rule, 2028 to
2037a'b
Pollutant
Emissions Reductions
Hg (pounds)
9,500
PM2.5 (tons)
5,400
CO2 (thousand tons)
650
SO2 (tons)
770
NOx (tons)
220
Non-Hg HAP metals (tons)
49
a Values rounded to two significant figures.
b Estimated reductions from model year 2028 are applied to 2028 and 2029, those from model year 2030 are applied
to 2031 and 2032, and those from model year 2035 are applied to 2032 through 2037. These values are summed to
generate total reduction figures.
Importantly, the continuous monitoring of fPM required in this rule will likely induce
additional emissions reductions that we are unable to quantify. Continuous measurements of
emissions accounts for changes to processes and fuels, fluctuations in load, operations of
pollution controls, and equipment malfunctions. By measuring emissions across all operations,
power plant operators and regulators can use the data to ensure controls are operating properly
and to assess continuous compliance with relevant standards. Because CEMS enable power plant
operators to quickly identify and correct problems with pollution control devices, it is possible
that fPM emissions could be lower than they otherwise would have been for up to three
months—or up to three years if testing less frequently under the LEE program— at a time. This
potential reduction in fPM and non-Hg HAP metals emission resulting from the information
provided by continuous monitoring coupled with corrective actions by plant operators could be
sizeable over the existing coal-fired fleet and is not quantified in this rulemaking.
As we are finalizing the removal of paragraph (2) of the definition of "startup," the time
period for engaging fPM or non-Hg HAP metal controls after non-clean fuel use, as well as for
full operation of fPM or non-Hg HAP metal controls, is expected to be reduced when
46 The estimates on non-mercury HAP metals reductions were obtained my multiplying the ratio of non-mercury
HAP metals to fPM by estimates of PMio reductions under the rule, as we do not have estimates of fPM reductions
using IPM, only PMio. The ratios of non-mercury HAP metals to fPM were based on analysis of 2010 MATS
Information Collection Request (ICR) data. As there may be substantially more fPM than PMio reduced by the
control techniques projected to be used under this rule, these estimates of non-mercury HAP metals reductions are
likely underestimates. More detail on the estimated reduction in non-mercury HAP metals can be found in the
docketed memorandum Estimating Non-Hg HAP Metals Reductions for the 2024 Technology Review for the Coal-
Fired EGUSource Category.
3-10
-------
transitioning to paragraph (1). The reduced time period for engaging controls therefore increases
the duration in which pollution controls are employed and lowers emissions.
To the extent that the CEMS requirement and removal of the second definition of startup
leads to actions that may otherwise not occur absent the amendments to those provisions in this
final rule, there may be emissions impacts we are unable to estimate.
3.5.2 Compliance Costs
3.5.2.1 Power Sector Costs
The power industry's "compliance costs" are represented in this analysis as the change in
electric power generation costs between the baseline and policy scenarios and are presented in
Table 3-5. In other words, these costs are an estimate of the increased power industry
expenditures required to implement the final rule requirements. The total compliance costs,
presented in Section 3.5.2.4, are estimated for this RIA as the sum of two components. The first
component, estimated using the modeling discussed above, is presented below in Table 3-5. This
component constitutes the majority of the incremental costs for the final. The second component,
the costs of the final rule PM CEMS requirement, is discussed in Section 3.5.2.2.
EPA projects that the annual incremental compliance cost of the final rule is $110
million, $110 million, and $93 million (2019 dollars) in 2028, 2030, and 2035, respectively. The
annual incremental cost is the projected additional cost of complying with the final rule in the
year analyzed and includes the amortized cost of capital investment and any applicable costs of
operating additional pollution controls, investments in new generating sources, shifts between or
amongst various fuels, and other actions associated with compliance. This projected cost does
not include the compliance calculated outside of IPM modeling, namely the compliance costs
related to PM CEMS. See Section 3.5.2.2 for further details on these costs. EPA believes that the
cost assumptions used for this RIA reflect, as closely as possible, the best information available
to the Agency today. See Section 3.5.4 for a discussion of projected capacity changes and
Section 3.6 for a discussion of the uncertainty regarding necessary pollution controls.
3-11
-------
Table 3-5 Power Sector Annualized Compliance Cost Estimates under the Final Rule in
2028, 2030, and 2035 (millions of 2019 dollars)
Analysis Year
Final Rule
2028
110
2030
110
2035
93
Note: Values have been rounded to two significant figures. As explained in Section 3.4, the incremental costs of
requiring PM CEMS are small relative to the ongoing costs of operation, so the less stringent regulatory alternative
in this RIA was not modeled using IPM. As a result, power sector impacts are not estimated for the less stringent
regulatory option, but the costs associated with requiring PM CEMS (Table 3-6) are included in the total cost across
regulatory options (Table 3-7).
3.5.2.2 PM CEMS Costs
In addition to revising the PM emission standard for existing coal-fired EGUs, EPA is
revising the requirements for demonstrating compliance with the PM emission standard for coal-
and oil-fired EGUs. The final PM standard renders the current limit for the LEE program moot
since it is lower than the current PM LEE limit. Therefore, EPA is removing PM from the LEE
program. Currently, EGUs that are not LEE units can demonstrate compliance with the fPM
standard either by conducting performance testing quarterly, use of PM continuous parameter
monitoring systems (CPMS) or using PM CEMS.
After considering updated information on the costs for performance testing compared to
the cost of PM CEMS and capabilities of PM CEMS measurement abilities, as well as the
benefits of using PM CEMS, which include increased transparency, compliance assurance, and
accelerated identification of anomalous emissions, EPA is finalizing the requirement that all
coal-fired EGUs and oil-fired EGUs demonstrate compliance with the PM emission standard by
using PM CEMS.
The revision of PM limits alters the composition and duration of testing runs in facilities
that use either compliance testing methodology. Estimated costs for quarterly fPM testing and
PM CEMS are provided in the "Revised Estimated Non-Beta Gauge PM CEMS and Filterable
PM Testing Costs" memorandum, available in the docket. The annualized costs for units
currently employing EPA Method 5 quarterly testing are estimated at about $60,000.47 EPA
calibrated its cost estimates for PM CEMS in response to observed installations, manufacturer
input, public comment, and engineering analyses. These calibrations include an assumed
47 EGUs receiving contractual or quantity discounts from performance test provides may incur lower costs.
3-12
-------
replacement lifespan of 15 years and an interest rate of 7 percent to approximate the prevailing
bank prime rate. For the portion of EGUs that employ PM CEMS, we estimate the annualized
costs to be about $72,000.
To produce an inventory of total units which would require the installation of PM CEMS
under the final rule as well as the incremental costs of the requirement, EPA began with an
inventory of all existing coal-fired EGUs with capacity great enough to be regulated by MATS.
That inventory was then filtered to remove EGUs with planned retirements or coal to gas
conversions prior to 2028 from analysis of both the baseline and final rule. Within that remaining
inventory of 314 EGUs, we used recent compliance data to determine that 120 units have
installed PM CEMS, while 177 units use quarterly testing and do not have existing PM CEMS
installations. The remaining 17 units (for which fPM compliance data were not available) are
assumed to use quarterly testing and not have existing PM CEMS installations.
Table 3-6 Incremental Cost of Final Continuous Emissions Monitoring (PM CEMS)
Requirement
Compliance
Approach in
Baseline
Units
(no.)
Baseline
Cost (per
year per
unit)
Total
Baseline
Costs (per
year)
Final Rule
(per year per
unit)
Final Rule
Costs (per
year)
Incremental
Costs (per
year)
Quarterly Testing
190
$60,000
$12,000,000
$72,000
$14,000,000
$2,300,000
PM CEMS
120
$72,000
$8,700,000
$72,000
$8,700,000
$0
Total
320
—
$20,000,000
—
$23,000,000
$2,300,000
Note: Values rounded to two significant figures. Rows may not appear to add correctly due to rounding.
As detailed in Table 3-6, relative to the baseline scenario, revised PM CEMS cost
estimates in the final rule leads to an estimated incremental cost of about $12,000 per year per
unit for EGUs currently employing quarterly testing. The final rule results in costs of about $2.3
million per year in total.
3.5.2.3 Startup Definition Costs
EPA is finalizing the removal of one of the two options for defining the startup period for
EGUs. The first option defines startup as either the first-ever firing of fuel in a boiler for the
purpose of producing electricity, or the firing of fuel in a boiler after a shutdown event for any
purpose. Startup ends when any of the steam from the boiler is used to generate electricity for
sale over the grid or for any other purpose (including on-site use). In the second option, startup is
3-13
-------
defined as the period in which operation of an EGU is initiated for any purpose. Startup begins
with either the firing of any fuel in an EGU for the purpose of producing electricity or useful
thermal energy (such as heat or steam) for industrial, commercial, heating, or cooling purposes
(other than the first-ever firing of fuel in a boiler following construction of the boiler) or for any
other purpose after a shutdown event. Startup ends four hours after the EGU generates electricity
that is sold or used for any other purpose (including on-site use), or four hours after the EGU
makes useful thermal energy (such as heat or steam) for industrial, commercial, heating, or
cooling purposes, whichever is earlier. This second option, referred to as paragraph (2) of the
definition of "startup," required clean fuel use to the maximum extent possible, operation of PM
control devices within one hour of introduction of primary fuel {i.e., coal, residual oil, or solid
oil-derived fuel) to the EGU, collection and submission of records of clean fuel use and
emissions control device capabilities and operation, as well as adherence to applicable numerical
standards within four hours of the generation of electricity or thermal energy for use either on
site or for sale over the grid {i.e., the end of startup) and to continue to maximize clean fuel use
throughout that period.
According to EPA analysis, owners or operators of coal- and oil-fired EGUs that
generated over 98 percent of electricity in 2022 have made the requisite adjustments, whether
through greater clean fuel capacity, better tuned equipment, better trained staff, a more efficient
and/or better design structure, or a combination of factors, to be able to meet the requirements of
paragraph (1) of the startup definition. This ability points out an improvement in operation that
all EGUs should be able to meet at little to no additional expenditure since the additional
recordkeeping and reporting provisions associated with the work practice standards of paragraph
(2) of the startup definition were more expensive than the requirements of paragraph (1) of the
definition. As a result, this RIA does not incorporate any additional costs of this finalized
provision.
3.5.2.4 Total Compliance Costs
The estimates of the total compliance costs are presented in Table 3-7. The total costs are
composed of the change in electric power generation costs between the baseline and policy
scenarios as presented in Table 3-5 and the incremental cost of the final PM CEMS requirement
as detailed in Table 3-6. There are no anticipated costs associated with this rule prior to 2028.
3-14
-------
3-15
-------
Table 3-7 Stream of Projected Compliance Costs for the Final Rule and Less Stringent
Regulatory Alternative (millions of 2019 dollars)"
Regulatory Alternative
Year
Final Ruleb
Less Stringent
2028 (applied to 2028 and 2029)b
110
2.3
2030 (applied to 2030 and 203 l)b
120
2.3
2035 (applied to 2032 to 2037)b
95
2.3
2% Discount Rate
PV
860
19
EAV
96
2.3
3% Discount Rate
PV
790
18
EAV
92
2.1
7% Discount Rate
PV
560
13
EAV
80
1.8
a Values rounded to two significant figures. PV and EAV discounted to 2023.
b IPM run years apply to particular calendar years as reported in the table. The run year information as applied to
individual calendar years is thus used to calculate PV and EAVs. Values rounded to two significant figures.
3.5.3 Projected Compliance Actions for Emissions Reductions
Electric generating units subject to the Hg and fPM emission limits in this final rule will
likely use various Hg and PM control strategies to comply. This section summarizes the
projected compliance actions related to each of these emissions limits.
The 2028 baseline includes approximately 5 GW of operational minemouth EGU
capacity designed to burn low rank virgin coal. All of this capacity is currently equipped with
Activated Carbon Injection (ACI) technology, and operation of this technology is reflected in the
baseline. Each of these EGUs projected to consume lignite is assigned an additional variable
operating cost that is consistent with achieving a 1.2 lb/MMBtu limit. Under the final rule, this
additional cost does not result in incremental retirements for these units, nor does it result in a
significant change to the projected generation level for these units.
The baseline also includes 11.6 GW of operational coal capacity that, based on the
analysis documented in the EPA docketed memorandum titled "2024 Update to the 2023
Proposed Technology Review for the Coal- and Oil-Fired EGU Source Category," EPA assumes
would either need to improve existing PM controls or install new PM controls to comply with the
3-16
-------
final rule in 2028. The various PM control upgrades that EPA assumes would be necessary to
achieve the emissions limits analyzed are summarized in Table 3-8.
Table 3-8 Projected PM Control Strategies under the Final Rule in 2028 (GW)
PM Control Strategy
Projected Actions and Retrofits
under the Final Rule
Additional O&M
3.7
Minor ESP Upgrades
0.7
Typical ESP Upgrades
2.0
ESP Rebuild
2.4
FF Bag Upgrade
1.3
New Fabric Filter
1.5
Total
11.6
Except for one facility (Colstrip, located in Montana), all of the 11.6 GW of operational
coal capacity that EPA assumes would need to take some compliance action to meet the final
standards are currently operating existing ESPs and/or fabric filters. All of that capacity is
projected to install the controls summarized in Table 3-8 and remain operational in 2028.
3.5.4 Generating Capacity
In this section, we discuss the projected changes in capacity by fuel type, building on and
adding greater context to the information presented in the previous section. We first look at total
capacity by fuel type, then retirements by fuel type, and finally new capacity builds by fuel type
for the 2028, 2030, and 2035 run years.
Table 3-9 shows the total net projected capacity by fuel type for the baseline and the final
rule for 2028, 2030, and 2035. Here, we see the net effects of projected retirements (Table 3-10)
and new capacity builds (see Table 3-11). There are no significant incremental changes in
capacity projected in response to the final rule for any given fuel type.
3-17
-------
Table 3-9 2028, 2030, and 2035 Projected U.S. Capacity by Fuel Type for the Baseline
and the Final Rule
Total Generation Capacity (GW)
Baseline
Final Rule
Change under Final Rule
GW %
2028
Coal
105.8
105.8
0.0
0.0%
Natural Gas
471.0
471.0
0.0
0.0%
Oil/Gas Steam
62.6
62.6
0.0
0.0%
Non-Hydro RE
394.1
394.1
0.0
0.0%
Hydro
102.4
102.4
0.0
0.0%
Energy Storage
46.7
46.7
0.0
0.0%
Nuclear
93.6
93.6
0.0
0.0%
Other
6.5
6.5
0.0
0.0%
Total
1,282.7
1,282.7
0.0
0.0%
2030
Coal
85.0
85.0
0.0
0.0%
Natural Gas
478.6
478.6
0.0
0.0%
Oil/Gas Steam
64.3
64.3
0.0
0.0%
Non-Hydro RE
440.2
440.2
0.0
0.0%
Hydro
103.7
103.7
0.0
0.0%
Energy Storage
58.6
58.6
0.0
0.0%
Nuclear
90.9
90.9
0.0
0.0%
Other
6.5
6.5
0.0
0.0%
Total
1,327.7
1,327.7
0.0
0.0%
2035
Coal
51.6
51.6
0.0
0.0%
Natural Gas
476.0
476.0
0.0
0.0%
Oil/Gas Steam
55.3
55.3
0.0
0.0%
Non-Hydro RE
698.5
698.5
0.0
0.0%
Hydro
107.3
107.3
0.0
0.0%
Energy Storage
113.6
113.6
0.0
0.0%
Nuclear
83.7
83.7
0.0
0.0%
Other
6.5
6.5
0.0
0.0%
Total
1,592.4
1,592.4
0.0
0.0%
Note: In this table, "Non-Hydro RE" includes biomass, geothermal, landfill gas, solar, and wind.
3-18
-------
Table 3-10 shows the total capacity projected to retire by fuel type for the baseline and
the final rule in all run years. The final rule is not projected to result in changes to projected
retirements.
Table 3-10 2028, 2030, and 2035 Projected U.S. Retirements by Fuel Type for the
Baseline and the Final Rule
Projected Retirements (GW)
„ „• . t» . % Change under Final
Baseline Final Rule ® .
2028
Coal
37.8
37.8
0.0%
Natural Gas
1.3
1.3
0.0%
Oil/Gas Steam
12.4
12.4
0.0%
Non-Hydro RE
2.9
2.9
0.0%
Hydro
0.1
0.1
0.0%
Nuclear
0.0
0.0
0.0%
Other
0.0
0.0
0.0%
Total
54.4
54.4
0.0%
2030
Coal
56.7
56.6
0.0%
Natural Gas
1.7
1.7
0.0%
Oil/Gas Steam
12.4
12.4
0.0%
Non-Hydro RE
2.9
2.9
0.0%
Hydro
0.1
0.1
0.0%
Nuclear
2.7
2.7
0.0%
Other
0.0
0.0
0.0%
Total
76.5
76.5
0.0%
2035
Coal
83.7
83.7
0.0%
Natural Gas
4.3
4.3
0.0%
Oil/Gas Steam
22.7
22.7
0.0%
Non-Hydro RE
3.0
3.0
0.0%
Hydro
0.1
0.1
0.0%
Nuclear
9.9
9.9
0.0%
Other
0.1
0.1
0.0%
Total
123.7
123.7
0.0%
Note: In this table, "Non-Hydro RE" includes biomass, geothermal, landfill gas, solar, and wind.
3-19
-------
Finally, Table 3-11 shows the projected U.S. new capacity builds by fuel type for the
baseline and the final rule in all run years. For the final rule, the incremental changes in projected
new capacity for any given fuel type are negligible.
Table 3-11 2028, 2030, and 2035 Projected U.S. New Capacity Builds by Fuel Type for
the Baseline and the Final Rule
New Capacity (GW)
Baseline
Final Rule
% Change under Final
Rule
2028
Coal
0.0
0.0
0.0%
Natural Gas
26.2
26.2
0.0%
Energy Storage
3.2
3.2
0.2%
Non-Hydro RE
44.8
44.8
0.0%
Hydro
0.0
0.0
0.0%
Nuclear
0.0
0.0
0.0%
Other
0.0
0.0
0.0%
Total
74.3
74.3
0.0%
2030
Coal
0.0
0.0
0.0%
Natural Gas
34.3
34.3
0.0%
Energy Storage
15.2
15.2
0.0%
Non-Hydro RE
90.8
90.8
0.0%
Hydro
1.3
1.3
0.0%
Nuclear
0.0
0.0
0.0%
Other
0.0
0.0
0.0%
Total
141.5
141.6
0.0%
2035
Coal
0.0
0.0
0.0%
Natural Gas
34.2
34.2
0.0%
Energy Storage
70.2
70.2
0.1%
Non-Hydro RE
349.4
349.4
0.0%
Hydro
4.9
4.9
0.0%
Nuclear
0.0
0.0
0.0%
Other
0.0
0.0
0.0%
Total
458.6
458.6
0.0%
Note: In this table, "Non-Hydro RE" includes biomass, geothermal, landfill gas, solar, and wind.
3.5.5 Generation Mix
In this section, we discuss the projected changes in generation mix for 2028, 2030, and
2035 for the final rule. Table 3-12 presents the projected generation and percentage changes in
3-20
-------
national generation mix by fuel type for run years 2028, 2030, and 2035. These generation mix
estimates reflect limited changes in energy generation as a result of the final rule in any run year.
Estimated changes in coal and natural gas use under the final rule are examined further in
Section 3.5.6.
Table 3-12 2028, 2030, and 2035 Projected U.S. Generation by Fuel Type for the
Baseline and the Final Rule
Generation Mix (TWh)
Incremental Change under Final Rule
Baseline
Final Rule
TWh
%
2028
Coal
472
472
-0.1
0.0%
Natural Gas
1,652
1,652
0.1
0.0%
Oil/Gas Steam
26
26
0.0
0.0%
Non-Hydro RE
1,141
1,141
0.0
0.0%
Hydro
293
293
0.0
0.0%
Energy Storage
53
53
0.0
0.1%
Nuclear
751
751
0.0
0.0%
Other
31
31
0.0
0.0%
Total
4,418
4,418
0.0
0.0%
2030
Coal
410
410
0.0
0.0%
Natural Gas
1,670
1,670
0.0
0.0%
Oil/Gas Steam
25
25
0.0
0.0%
Non-Hydro RE
1,329
1,329
0.0
0.0%
Hydro
298
298
0.0
0.0%
Energy Storage
69
69
0.0
0.0%
Nuclear
729
729
0.0
0.0%
Other
31
31
0.0
0.0%
Total
4,560
4,560
0.0
0.0%
2035
Coal
236
236
-0.1
0.0%
Natural Gas
1,344
1,344
0.0
0.0%
Oil/Gas Steam
8
8
0.0
-0.4%
Non-Hydro RE
2,229
2,229
0.0
0.0%
Hydro
319
319
0.0
0.0%
Energy Storage
148
148
0.1
0.1%
Nuclear
667
667
0.0
0.0%
Other
31
31
0.0
0.0%
Total
4,981
4,981
0.0
0.0%
Note: In this table, "Non-Hydro RE" includes biomass, geothermal, landfill gas, solar, and wind.
3-21
-------
3.5.6 Coal and Natural Gas Use for the Electric Power Sector
In this section we discuss the estimated changes in coal use and natural gas use in 2028,
2030, and 2035. Table 3-13 and Table 3-14 present percentage changes in national coal usage by
EGUs by coal supply region and coal rank, respectively. These fuel use estimates show small
changes in national coal use in the final rule relative to the baseline in all run years. Additionally,
the final rule is not projected to result in significant coal switching between supply regions or
coal rank.
Table 3-13 2028, 2030, and 2035 Projected U.S. Power Sector Coal Use by Coal Supply
Region for the Baseline and the Final Rule
Million Tons
Region
Year
Baseline
Final Rule
% Change under
Final Rule
Appalachia
39.8
39.8
0.1%
Interior
37.8
37.8
-0.1%
Waste Coal
2028
7.3
7.3
0.0%
West
166.1
166.0
-0.1%
Total
250.9
250.8
0.0%
Appalachia
38.8
38.8
0.0%
Interior
35.1
35.1
0.0%
Waste Coal
2030
7.1
7.1
0.0%
West
141.5
141.5
0.0%
Total
222.5
222.5
0.0%
Appalachia
31.8
31.9
0.1%
Interior
19.4
19.4
-0.1%
Waste Coal
2035
6.8
6.8
0.0%
West
89.0
89.1
0.1%
Total 147.1 147.2 0.0%
3-22
-------
Table 3-14 2028, 2030, and 2035 Projected U.S. Power Sector Coal Use by Rank for the
Baseline and the Final Rule
Million Tons
% Change
Rank
Year
Baseline
Final Rule
under Final
Rule
Bituminous
72.1
72.1
0.00%
Subbituminous
145.1
145.1
0.00%
2028
Lignite
32.5
32.3
-0.60%
Total
249.6
249.5
0.00%
Bituminous
62.8
62.8
0.00%
Subbituminous
125.8
125.8
0.00%
2030
Lignite
29.3
29.3
0.00%
Total
218
218
0.00%
Bituminous
42.4
42.4
0.00%
Subbituminous
2035
74.1
74.2
0.10%
Lignite
24.5
24.5
0.00%
Total
140.9
141
0.00%
Table 3-15 presents the projected changes in national natural gas usage by EGUs in the
2028, 2030, and 2035 run years. These fuel use estimates reflect negligible changes in projected
gas generation in 2028, 2030, and 2035.
Table 3-15 2028, 2030, and 2035 Projected U.S. Power Sector Natural Gas Use for the
Baseline and the Final Rule
Year
Baseline
Trillion Cubic Feet
Final Rule
% Change
under Final Rule
2028
11.6
11.6
0.0%
2030
11.7
11.7
0.0%
2035
9.3
9.3
0.0%
3.5.7 Fuel Price, Market, and Infrastructure
The projected impacts of the final rule on coal and natural gas prices are presented below
in Table 3-16 and Table 3-17, respectively. As with the projected impact of the final rule on fuel
use, there is no significant change projected for minemouth and delivered coal prices due to the
final rule.
3-23
-------
Table 3-16 2028, 2030, and 2035 Projected Minemouth and Power Sector Delivered Coal
Price (2019 dollars) for the Baseline and the Final Rule
$/MMBtu
Year
Baseline
Final Rule
% Change under
Final Rule
Minemouth
Delivered
2028
0.98
1.54
0.98
1.54
0.0%
0.0%
Minemouth
Delivered
2030
1.02
1.56
1.02
1.56
0.0%
0.0%
Minemouth
Delivered
2035
1.07
1.55
1.07
1.55
0.0%
0.0%
Consistent with the projection of no significant change in natural gas use under the final
rule, Henry Hub and power sector delivered natural gas prices are not projected to significantly
change under the final rule over the period analyzed. Table 3-17 summarizes the projected
impacts on Henry Hub and delivered natural gas prices in 2028, 2030, and 2035.
Table 3-17 2028, 2030, and 2035 Projected Henry Hub and Power Sector Delivered
Natural Gas Price (2019 dollars) for the Baseline and the Final Rule
$/MMBtu
Year
Baseline
Final Rule
% Change under
Final Rule
Henry Hub
Delivered
2028
2.78
2.84
2.78
2.84
0.0%
0.0%
Henry Hub
Delivered
2030
2.89
2.95
2.89
2.95
0.0%
0.0%
Henry Hub
Delivered
2035
2.87
2.88
2.87
2.88
0.0%
0.0%
3.5.8 Retail Electricity Prices
EPA estimated the change in the retail price of electricity (2019 dollars) using the Retail
Price Model (RPM).48 The RPM was developed by ICF for EPA and uses the IPM estimates of
changes in the cost of generating electricity to estimate the changes in average retail electricity
prices. The prices are average prices over consumer classes (i.e., consumer, commercial, and
industrial) and regions, weighted by the amount of electricity used by each class and in each
region. The RPM combines the IPM annual cost estimates in each of the 64 IPM regions with
48 See documentation available at: https://www.epa.gov/airmarkets/retail-price-model.
3-24
-------
EIA electricity market data for each of the 25 electricity supply regions (shown in Figure 3-1) in
the electricity market module of the National Energy Modeling System (NEMS).49
Table 3-18, Table 3-19, and Table 3-20 present the projected percentage changes in the
retail price of electricity for the regulatory control alternatives in 2028, 2030, and 2035,
respectively. Consistent with other projected impacts presented above, the projected impacts on
average retail electricity prices at both the national and regional level are projected to be small in
all run years.
49 See documentation available at:
https://www.eia.gov/outlooks/aeo/nems/documentation/electricity/pdf/EMM_2022.pdf.
3-25
-------
Table 3-18 Projected Average Retail Electricity Price by Region for the Baseline and
under the Final Rule, 2028
All Sectors
2028 Average Retail Electricity Price
(2019 mills/kWh)
Region
Baseline
Final Rule
% Change
under Final Rule
TRE
73.4
73.4
0.0%
FRCC
96.4
96.4
0.0%
MISW
92.3
92.3
0.0%
MISC
87.9
88.0
0.2%
MISE
95.2
95.2
0.0%
MISS
81.3
81.3
0.0%
ISNE
141.8
141.8
0.0%
NYCW
208.4
208.4
0.0%
NYUP
121.5
121.5
0.0%
PJME
116.9
116.9
0.0%
PJMW
90.4
90.4
0.0%
PJMC
72.4
72.4
0.0%
PJMD
70.8
70.8
0.0%
SRCA
94.7
94.7
0.0%
SRSE
96.7
96.7
0.0%
SRCE
71.6
71.6
0.0%
SPPS
75.3
75.3
0.0%
SPPC
98.5
98.4
0.0%
SPPN
64.1
64.1
0.0%
SRSG
101.3
101.3
0.0%
CANO
138.7
138.7
0.0%
CASO
170.5
170.5
0.0%
NWPP
75.0
75.4
0.5%
RMRG
96.4
96.4
0.0%
BASN
96.8
96.8
0.0%
National
97.1
97.1
0.0%
3-26
-------
Table 3-19 Projected Average Retail Electricity Price by Region for the Baseline and
under the Final Rule, 2030
All Sectors
2030 Average Retail Electricity Price
(2019 mills/kWh)
Region
Baseline
Final Rule
% Change
under Final Rule
TRE
73.3
73.3
0.0%
FRCC
97.6
97.6
0.0%
MISW
93.2
93.2
0.0%
MISC
91.3
91.5
0.2%
MISE
109.4
109.4
0.0%
MISS
85.7
85.7
0.0%
ISNE
156.6
156.6
0.0%
NYCW
210.3
210.3
0.0%
NYUP
125.7
125.7
0.0%
PJME
109.9
109.9
0.0%
PJMW
97.3
97.3
0.0%
PJMC
89.3
89.3
0.0%
PJMD
76.5
76.5
0.0%
SRCA
92.1
92.2
0.0%
SRSE
94.7
94.7
0.0%
SRCE
70.7
70.7
0.0%
SPPS
77.7
77.8
0.0%
SPPC
97.3
97.3
0.0%
SPPN
65.1
65.1
0.0%
SRSG
101.7
101.6
0.0%
CANO
142.9
142.9
0.0%
CASO
173.8
173.9
0.0%
NWPP
81.6
81.7
0.1%
RMRG
100.7
100.7
0.0%
BASN
96.3
96.3
0.0%
National
99.6
99.6
0.0%
3-27
-------
Table 3-20 Projected Average Retail Electricity Price by Region for the Baseline and
under the Final Rule, 2035
All Sectors
2035 Average Retail Electricity Price
(2019 mills/kWh)
Region
Baseline
Final Rule
% Change
under Final Rule
TRE
78.4
78.4
0.0%
FRCC
91.9
91.9
0.0%
MISW
84.5
84.5
0.0%
MISC
81.5
81.5
0.1%
MISE
95.7
95.7
0.0%
MISS
79.2
79.2
0.0%
ISNE
156.1
155.8
-0.2%
NYCW
208.9
208.9
0.0%
NYUP
124.6
124.6
0.0%
PJME
108.5
108.5
0.0%
PJMW
91.8
91.8
0.0%
PJMC
75.1
75.1
0.0%
PJMD
71.4
71.4
0.0%
SRCA
89.4
89.4
0.0%
SRSE
90.1
90.1
0.0%
SRCE
67.1
67.1
0.0%
SPPS
69.5
69.5
0.0%
SPPC
80.4
80.4
0.0%
SPPN
63.0
63.0
0.0%
SRSG
103.4
103.4
0.0%
CANO
139.5
139.5
0.0%
CASO
172.8
172.8
0.0%
NWPP
78.5
78.9
0.4%
RMRG
93.4
93.4
0.0%
BASN
96.9
97.0
0.0%
National
95.9
95.9
0.0%
3-28
-------
(23)'
NWPP
M(SW,
19
SPPN
MISE]
NYUP
¦21 ¦
[CANO]
25
BASN
24]
R W R G
STlraPf
M L#PJMD '
Figure 3-1 Electricity Market Module Regions
Source: EI A (http:/mww. eia.gov/forecasts/aeo/pdf/nercjnap.pdf)
3.6 Limitations of Analysis and Key Areas of Uncertainty
EPA's power sector modeling is based on expert judgment of various input assumptions
for variables whose outcomes are uncertain. As a general matter, the Agency reviews the best
available information from engineering studies of air pollution controls and new capacity
construction costs to support a reasonable modeling framework for analyzing the cost, emission
changes, and other impacts of regulatory actions for EGUs. The annualized cost of the final rule,
as quantified here, is EPA's best assessment of the cost of implementing the rule on the power
sector.
The IPM-projected annualized cost estimates of private compliance costs provided in this
analysis are meant to show the increase in production (generating) costs to the power sector in
response to the finalized requirements. To estimate these annualized costs, as discussed earlier,
EPA uses a conventional and widely accepted approach that applies a capital recovery factor
(CRF) multiplier to capital investments and adds that to the annual incremental operating
expenses to calculate annual costs. The CRF is derived from estimates of the cost of capital
3-29
-------
(private discount rate), the amount of insurance coverage required, local property taxes, and the
life of capital. The private compliance costs presented earlier are EPA's best estimate of the
direct private compliance costs of the rule.
In addition, there are several key areas of uncertainty related to the electric power sector
that are worth noting, including:
• Electricity demand: The analysis includes an assumption for future electricity demand.
To the extent electricity demand is higher and lower, it may increase/decrease the
projected future composition of the fleet.
• Natural gas supply and demand: To the extent natural gas supply and delivered prices
are higher or lower, it would influence the use of natural gas for electricity generation and
overall competitiveness of other EGUs (e.g., coal and nuclear units).
• Longer-term planning by utilities: Many utilities have announced long-term clean
energy and/or climate commitments, with a phasing out of large amounts of coal capacity
by 2030 and continuing through 2050. These announcements are not necessarily reflected
in the baseline and may alter the amount of coal capacity projected in the baseline that
would be covered under this rule.
• FPM emissions and control: As discussed above, the baseline fPM emissions rates for
each unit are based on the analysis documented in the memorandum titled "2024 Update
to the 2023 Proposed Technology Review for the Coal- and Oil-Fired EGU Source
Category." For those EGUs with rates greater than the final limit, EPA assumes that
control technology summarized in Section 3.4 would be necessary to remain operational.
While the baseline emissions rate for each EGU and the cost and performance
assumption for each PM control technology are the best available to EPA at this time, it
is possible that some EGUs may be able to achieve the revised fPM emissions limits with
less costly control technology (e.g., an ESP upgrade instead of a fabric filter installation).
It is also possible that EPA's cost assumptions reflect higher technology costs than might
be incurred by EGUs.
These are key uncertainties that may affect the overall composition of electric power
generation fleet and/or compliance with the finalized emissions limits and could thus have an
effect on the estimated costs and impacts of this action. While it is important to recognize these
key areas of uncertainty, they do not change EPA's overall confidence in the projected impacts
of the final rule presented in this section. EPA continues to monitor industry developments and
makes appropriate updates to the modeling platforms in order to reflect the best and most current
data available.
3-30
-------
Estimated impacts of the Revised 2023 and Later Model Year Light-Duty Vehicle GHG
Emissions Standards are captured in the baseline,50 while estimated impacts of the Proposed
Rule: Model Years 2027 and Later Light-Duty and Medium-Duty Vehicle Emissions Standards
are not captured in the baseline.51 The latter rule (in its proposal) is projected to increase the total
demand for electricity by 0.4 percent in 2030 and 3.4 percent in 2040 relative to the baseline
electricity demand projections assumed in this analysis. Estimated impacts of the 2023 Final
Standards of Performance for New, Reconstructed, and Modified Sources and Emissions
Guidelines for Existing Sources: Oil and Natural Gas Sector Climate Review are also not
included in this analysis. The RIA for oil and natural gas sector rule projected small increases in
the price of natural gas as result of the requirements (U.S. EPA, 2023c). All else equal, inclusion
of these two programs would likely result in a modest increase in the fPM reductions and total
cost of compliance for this rule. While we might see less retired capacity in the baseline due to
higher electricity demand, and thus more PM controls under the RTR, the magnitude of the
potential incremental impacts would likely be very small.
3.7 References
Brinkman, G., Bain, D., Buster, G., Draxl, C., Das, P., Ho, J., . . . Zhang, J. (2021). The North
American Renewable Integration Study (NARIS): A U.S. Perspective. Retrieved from
United States: https://www.osti.gov/biblio/1804701
Cole, W. J., Greer, D., Denholm, P., Frazier, A. W., Machen, S., Mai, T., . . . Baldwin, S. F.
(2021). Quantifying the challenge of reaching a 100% renewable energy power system
for the United States. Joule, 5(7), 1732-1748. doi:10.1016/j.joule.2021.05.011
DOE. (2021). The Solar Futures Study. Retrieved from United States:
https ://www. osti. gov/biblio/1820105
U.S. EPA. (2005). Regulatory Impact Analysis for the Final Clean Air Interstate Rule. Research
Triangle Park, NC: U.S. Environmental Protection Agency, Office of Air Quality
Planning and Standards, Health and Environmental Impact Division.
https://www.epa.gOv/sites/default/files/2020-07/documents/transport_ria_final-clean-air-
interstate-rule_2005-03 .pdf
U.S. EPA. (201 la). 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. Research Triangle Park, NC: U.S. Environmental Protection
50 86 FR 43726. The RIA for this rule available at: https.V/nepis.epa. sov/Exe/ZvPDF. cei?Dockev=P 1012QNB. pdf.
51 88 FR 29184.
3-31
-------
Agency, Office of Air Quality Planning and Standards, Health and Environmental Impact
Division, https://www3.epa.gov/ttn/ecas/docs/ria/transport_ria_final-csapr_2011-06.pdf
U.S. EPA. (201 lb). Regulatory Impact Analysis for the Final Mercury and Air Toxics Standards.
(EPA-452/R-11-011). Research Triangle Park, NC: U.S. Environmental Protection
Agency, Office of Air Quality Planning and Standards, Health and Environmental Impact
Division. http://www.epa.gov/ttn/ecas/regdata/RIAs/matsriafinal.pdf
U.S. EPA. (2014a). Economic Analysis for the Final Section 316(b) Existing Facilities Rule.
(EPA-821-R-14-001). Washington DC: U.S. Environmental Protection Agency.
https://www.epa.gov/sites/default/files/2015-05/documents/cooling-water_phase-
4_economics_2014 .pdf
U.S. EPA. (2014b). Guidelines for Preparing Economic Analyses. (EPA 240-R-10-001).
Washington DC: U.S. Environmental Protection Agency, Office of Policy, National
Center for Environmental Economics, https://www.epa.gov/environmental-
economics/guidelines-preparing-economic-analyses
U.S. EPA. (2015a). Benefit and Cost Analysis for the Effluent Limitations Guidelines and
Standards for the Steam Electric Power Generating Point Source Category. (EPA-821-R-
15-005). Washington DC: U.S. Environmental Protection Agency.
https://www.epa.gOv/sites/default/files/2015-10/documents/steam-electric_benefit-cost-
analy sis_09-29-2015 .pdf
U.S. EPA. (2015b). Regulatory Impact Analysis for the Clean Power Plan Final Rule. (EPA-
452/R-15-003). Research Triangle Park, NC: U.S. Environmental Protection Agency,
Office of Air Quality Planning and Standards, Health and Environmental Impact
Division, https://www.epa.gov/sites/default/files/2020-07/documents/utilities_ria_final-
clean-power-plan-existing-units_2015-08.pdf
U.S. EPA. (2015c). Regulatory Impact Analysis: EPA's 2015 RCRA Final Rule Regulating Coal
Combustion Residual (CCR) Landfills and Surface Impoundments At Coal-Fired Electric
Utility Power Plants. (EPA-821-R-20-003). Washington DC: U.S. Environmental
Protection Agency. https://www.regulations.gov/document/EPA-HQ-RCRA-2009-0640-
12034
U.S. EPA. (2016). Regulatory Impact Analysis of the Cross-State Air Pollution Rule (CSAPR)
Update for the 2008 National Ambient Air Quality Standards for Ground-Level Ozone.
(EPA-452/R-16-004). Research Triangle Park, NC: U.S. Environmental Protection
Agency, Office of Air Quality Planning and Standards, Health and Environmental Impact
Division, https://www.epa.gov/sites/default/files/2020-07/documents/transport_ria_final-
csapr-update_2016-09.pdf
U.S. EPA. (2019). 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. (EPA-452/R-19-003). Research Triangle Park, NC: U.S.
Environmental Protection Agency, Office of Air Quality Planning and Standards, Health
3-32
-------
and Environmental Impact Division, https://www.epa.gov/sites/production/files/2019-
06/documents/utilities_ria_final_cpp_repeal_and_ace_2019-06.pdf
U.S. EPA. (2020). Benefit and Cost Analysis for Revisions to the Effluent Limitations
Guidelines and Standards for the Steam Electric Power Generating Point Source
Category. (EPA-821-R-20-003). Washington DC: U.S. Environmental Protection
Agency, https://www.epa.gov/sites/default/files/2020-
08/documents/steam_electric_elg_2020_final_reconsideration_rule_benefit_and_cost_an
alysis.pdf
U.S. EPA. (2021). Regulatory Impact Analysis for the Final Revised Cross-State Air Pollution
Rule (CSAPR) Update for the 2008 Ozone NAAQS. (EPA-452/R-21-002). Research
Triangle Park, NC: U.S. Environmental Protection Agency, Office of Air Quality
Planning and Standards, Health and Environmental Impact Division.
https://www.epa.gov/sites/default/files/2021-
03/documents/revi sed_csapr_update_ria_final.pdf
U.S. EPA. (2023a). Electricity Sector Emissions Impacts of the Inflation Reduction Act:
Assessment of projected C02 emission reductions from changes in electricity generation
and use. (EPA 430-R-23-004). Retrieved from
https://www.epa.gov/system/files/documents/2023-
09/Electricity _Emissions_Impacts_Inflation_Reduction_Act_Report_EPA-FINAL.pdf
U.S. EPA. (2023b). Regulatory Impact Analysis for the Final Federal Good Neighbor Plan
Addressing Regional Ozone Transport for the 2015 Ozone National Ambient Air Quality
Standards. (EPA-452/R-23-001). Research Triangle Park, NC: U.S. Environmental
Protection Agency, Office of Air Quality Planning and Standards, Health and
Environmental Impact Division, https://www.epa.gov/system/files/documents/2023-
03/SAN%208670%20Federal%20Good%20Neighbor%20Plan%2020230315%20RIA_Fi
nal.pdf
U.S. EPA. (2023c). Regulatory Impact Analysis of the Standards of Performance for New,
Reconstructed, and Modified Sources and Emissions Guidelines for Existing Sources: Oil
and Natural Gas Sector Climate Review. (EPA-452/R-23-013). Research Triangle Park,
NC: U.S. Environmental Protection Agency, Office of Air Quality Planning and
Standards, Health and Environmental Impact Division.
https://www.epa.gov/system/files/documents/2023-12/eol2866_oil-and-gas-nsps-eg-
climate-review-2060-avl6-ria-20231130.pdf
3-33
-------
BENEFITS ANALYSIS
4.1 Introduction
This rule is projected to reduce emissions of Hg and non-Hg HAP metals, fine particulate
matter (PM2.5), sulfur dioxide (SO2), nitrogen oxides (NOx), and carbon dioxide (CO2)
nationally. The projected reductions in Hg are expected to reduce the bioconcentration of MeHg
in fish. Subsistence fishing is associated with vulnerable populations, including minorities and
those of low socioeconomic status. Further reductions in Hg emissions should reduce fish
concentrations and exposure to HAP particularly for the subsistence fisher sub-population. The
projected reductions in HAP emissions should help EPA maintain an ample margin of safety by
reducing exposure to MeHg and carcinogenic HAP metals.
Regarding the potential health and ecological benefits of the rule from projected HAP
reductions, we note that these are discussed only qualitatively and not quantitatively. Exposure to
the HAP emitted by the source category, depending on the exposure duration and level of
exposure, is associated with a variety of adverse health effects. These adverse health effects may
include chronic health disorders (e.g., irritation of the lung, skin, and mucus membranes;
decreased pulmonary function, pneumonia, or lung damage; detrimental effects on the central
nervous system; cardiovascular disease; damage to the kidneys; and alimentary effects such as
nausea and vomiting), adverse neurodevelopmental impacts, and increased risk of cancer. See 76
FR 25003-25005 for a fuller discussion of the health effects associated with HAP.
The analysis of the overall EGU sector completed for EPA's review of the 2020
appropriate and necessary finding (2023 Final A&N Review) identified significant reductions in
cardiovascular and neuro-developmental effects from exposure to MeHg (88 FR 13956).
However, the amount of Hg reduction projected under this rule is a fraction of the Hg estimates
used in the 2023 Final A&N Review. Overall, the uncertainty associated with modeling potential
benefits of Hg reduction for fish consumers would be sufficiently large as to compromise the
utility of those benefit estimates—though importantly, such uncertainty does not decrease our
confidence that reductions in emissions should result in reduced exposures of HAP to the general
population, including MeHg exposures to subsistence fishers located near these facilities.
Further, estimated risks from exposure to non-Hg HAP metals were not expected to exceed
4-1
-------
acceptable levels, although we note that these emissions reductions should result in decreased
exposure to HAP for individuals living near these facilities.
ReducingPM2.5 and SO2 emissions is expected to reduce ground-level PM2.5
concentrations. Reducing NOx emissions is expected to reduce both ground-level ozone and
PM2.5 concentrations. Below we present the estimated number and economic value of these
avoided PM2.5 and ozone-attributable premature deaths and illnesses. We also present the
estimated monetized climate and health benefits associated with emission reductions projected
under the final rule.
In addition to reporting results, this section details the methods used to estimate the
benefits to human health of reducing concentrations of PM2.5 and ozone resulting from the
projected emissions reductions. This analysis uses methods for determining air quality changes
that have been used in the RIAs from multiple previous proposed and final rules (U.S. EPA,
2019b, 2020a, 2020b, 2021a, 2022c), including the RIA for the proposal of this rule (U.S. EPA,
2023b). The approach involves two major steps: (1) developing spatial fields of air quality across
the U.S. for a baseline scenario and the final rule for 2028, 2030, and 2035 using nationwide
photochemical modeling and related analyses (see Air Quality Modeling Appendix, Appendix A,
for more details); and (2) using these spatial fields in BenMAP-CE to quantify the benefits under
the final rule and each year as compared to the baseline in that year.52 See Section 4.3.3 for more
detail on BenMAP-CE. When estimating the value of improved air quality over a multi-year time
horizon, the analysis applies population growth and income growth projections for each future
year through 2037 and estimates of baseline mortality incidence rates at five-year increments.
Additionally, elevated concentrations of GHGs in the atmosphere have been warming the
planet, leading to changes in the Earth's climate including changes in the frequency and intensity
of heat waves, precipitation, and extreme weather events, rising seas, and retreating snow and
ice. The well-documented atmospheric changes due to anthropogenic GHG emissions are
changing the climate at a pace and in a way that threatens human health, society, and the natural
environment. There will likely be important climate benefits associated with the CO2 emissions
52 Note we do not perform air quality analysis on the less stringent regulatory option because it has no quantified
emissions reductions associated with the finalized requirements for CEMS and the removal of startup definition
number two.
4-2
-------
reductions expected from this rule. In this RIA, we monetize climate benefits from reducing
emissions of CO2 using estimates of the SC-CO2.
EPA is unable to quantify and monetize the potential benefits of requiring facilities to
utilize CEMS rather than continuing to allow the use of quarterly testing, but the requirement has
been considered qualitatively. Relative to periodic testing practices, continuous monitoring of
fPM will result in increased transparency, as well as potential emissions reductions from
identifying problems more rapidly. Hence, the final rule may induce further reductions of fPM
and non-Hg HAP metals than we project in this RIA, and these reductions would likely lead to
additional health benefits. However, due to data and methodological challenges, EPA is unable
to quantify these potential additional reductions. The continuous monitoring of fPM required in
this rule is also likely to provide several additional important benefits to the public which are not
quantified in this rule, including greater certainty, accuracy, transparency, and granularity in fPM
emissions information than exists today. Additionally, to the extent that the removal of the
second definition of startup leads to actions that may otherwise not occur absent this final rule,
there may be beneficial impacts we are unable to estimate. Though the rule is likely to also yield
positive benefits associated with reducing pollutants other than Hg, non-Hg HAP metals, PM2.5,
ozone, and CO2, time, resource, and data limitations prevented us from quantifying and
estimating the economic value of those reductions. Specifically, in this RIA EPA does not
monetize health benefits of reducing direct exposure to NO2 and SO2 nor ecosystem effects and
visibility impairment associated with changes in air quality. We qualitatively discuss these
unquantified impacts in this section of the RIA.
4.2 Hazardous Air Pollutant Benefits
This final rule is projected to reduce emissions of Hg and non-Hg HAP metals.
Specifically, projected reductions in Hg are expected to help reduce exposure to MeHg for sub-
populations that rely on subsistence fishing. In addition, projected emissions reductions should
also reduce exposure to non-Hg HAP metals including carcinogens such as nickel, arsenic, and
hexavalent chromium, for residents located in the vicinity of these facilities.
4-3
-------
4.2.1 Hg
Hg is a persistent, bioaccumulative toxic metal that is emitted from power plants in three
forms: gaseous elemental Hg (HgO), oxidized Hg compounds (Hg+2), and particle-bound Hg
(HgP). Elemental Hg does not quickly deposit or chemically react in the atmosphere, resulting in
residence times that are long enough to contribute to global scale deposition. Oxidized Hg and
HgP deposit quickly from the atmosphere impacting local and regional areas in proximity to
sources. MeHg is formed by microbial action in the top layers of sediment and soils, after Hg has
precipitated from the air and deposited into waterbodies or land. Once formed, MeHg is taken up
by aquatic organisms and bioaccumulates up the aquatic food web. Larger predatory fish may
have MeHg concentrations many times that of the concentrations in the freshwater body in which
they live (ATSDR, 2022). MeHg can adversely impact ecosystems and wildlife.
Human exposure to MeHg is known to have several adverse neurodevelopmental
impacts, such as IQ loss measured by performance on neurobehavioral tests, particularly on tests
of attention, fine motor-function, language, and visual spatial ability. In addition, evidence in
humans and animals suggests that MeHg can have adverse effects on both the developing and the
adult cardiovascular system, including fatal and non-fatal ischemic heart disease (IHD). Further,
nephrotoxicity, immunotoxicity, reproductive effects (impaired fertility), and developmental
effects have been observed with MeHg exposure in animal studies (ATSDR, 2022). MeHg has
some genotoxic activity and is capable of causing chromosomal damage in a number of
experimental systems. EPA has classified MeHg as a "possible" human carcinogen (U.S. EPA,
2001).
The projected reductions in Hg under this final rule are expected to reduce the
bioconcentration of MeHg in fish due to Hg emissions from MATS-affected sources. Risk from
near-field deposition of Hg to subsistence fishers has previously been evaluated, using a site-
specific assessment of a lake near three lignite-fired facilities (U.S. EPA, 2020d). The results
suggest that MeHg exposure to subsistence fishers from lignite-fired units is below the current
RfD for MeHg neurodevelopmental toxicity or IQ loss, with an estimated hazard quotient (HQ)
of 0.06. In general, EPA believes that exposures at or below the RfD are unlikely to be
associated with appreciable risk of deleterious effects.
4-4
-------
Regarding the potential magnitude of human health risk reductions and benefits
associated with this rule, we make the following observations. All of the exposure results
generated as part of the 2020 Residual Risk analysis were below the presumptive acceptable
cancer risk threshold and noncancer health-based thresholds. While these results suggest that the
residual risks from HAP exposure are low, we do recognize that this regulation should still
reduce exposure to HAP.
Regarding potential benefits of the rule to the general population of fish consumers, while
we note that the analysis of the overall EGU sector completed for the 2023 Final A&N Review
did identify significant reductions in cardiovascular and neuro-developmental effects, given the
substantially smaller Hg reduction associated with this rule (approximately 900 to 1000 pounds
per year under the final rule compared to the approximately 29 tons of Hg evaluated in the 2023
Final A&N Review), overall uncertainty associated with modeling potential benefits for the
broader population of fish consumers would be sufficiently large as to compromise the utility of
those benefit estimates.
Despite the lack of quantifiable risks from Hg emissions, reductions would be expected to
have some impact (reduction) on the overall MeHg burden in fish for waterbodies near covered
facilities. In the appropriate and necessary determination, EPA illustrated that the burden of Hg
exposure is not equally distributed across the population and that some subpopulations bore
disproportionate risks associated with exposure to emissions from U.S. EGUs. High levels of fish
consumption observed with subsistence fishing were associated with vulnerable populations,
including minorities and those with low socioeconomic status (SES). Reductions in Hg
emissions should reduce MeHg exposure and body burden for subsistence fishers.
U.S. EGU Hg emissions can lead to increased deposition of Hg to nearby waterbodies.
Deposition of Hg to waterbodies can also have an impact on ecosystems and wildlife. Hg
contamination is present in all environmental media with aquatic systems being particularly
impacted due to bioaccumulation. Bioaccumulation refers to the net uptake of a contaminant
from all possible pathways and includes the accumulation that may occur by direct exposure to
contaminated media as well as uptake from food. Atmospheric Hg enters freshwater ecosystems
by direct deposition and through runoff from terrestrial watersheds. Once Hg deposits, it may be
converted to organic MeHg mediated primarily by sulfate-reducing bacteria. Methylation is
4-5
-------
enhanced in anaerobic and acidic environments, greatly increasing Hg toxicity and potential to
bioaccumulate in aquatic foodwebs (Munthe et al. 2007). The highest levels of MeHg
accumulation are most often measured in fish eating (piscivorous) animals and those which prey
on other fish eaters. In laboratory studies, adverse effects from exposure to MeHg in wildlife
have been observed in fish, mink, otters, and several avian species at exposure levels as low as
0.25 micrograms of MeHg per gram of body weight (U.S. EPA, 1997). The risk of Hg exposure
may also extend to insectivorous terrestrial species such as songbirds, bats, spiders, and
amphibians that receive Hg deposition or from aquatic systems near the forest areas they inhabit
(Bergeron et al., 2010a, 2010b; Cristol et al., 2008; Rimmer et al., 2005; Wada et al., 2009;
Wada et al., 2010)
The projected emissions reductions of Hg are expected to lower deposition of Hg into
ecosystems and reduce U.S. EGU attributable bioaccumulation of MeHg in wildlife, particularly
for areas closer to the effected units subject to near-field deposition. Because Hg emissions from
U.S. EGUs can both become deposited in or bioaccumulate in organisms living in foreign and
international waters, reduction of Hg emissions from U.S. EGUs could lead to some benefits
internationally as well. EPA is currently unable to quantify or monetize such effects.
4.2.2 Non-Hg HAP Metal
U.S. EGUs are the largest source of selenium emissions and a major source of non-Hg
HAP metals emissions including arsenic, chromium, cobalt, and nickel. Additionally, U.S. EGUs
emit beryllium, cadmium, lead, and manganese. These emissions include HAP metals that are
persistent and bioaccumulate (arsenic, cadmium, and lead) and others have cancer-causing
potential (beryllium, cadmium, chromium, cobalt, lead, and nickel). PM controls are expected to
reduce HAP metals emissions and therefore reduce exposure to HAP metals for the general
population including those living near these facilities.
Exposure to these HAP metals, depending on exposure duration and levels of exposures,
is associated with a variety of adverse health effects. These adverse health effects may include
chronic health disorders (e.g., irritation of the lung, skin, and mucus membranes; decreased
pulmonary function, pneumonia, or lung damage; detrimental effects on the central nervous
system; damage to the kidneys; and alimentary effects such as nausea and vomiting). As of 2023,
three of the key HAP metals or their compounds emitted by EGUs (arsenic, chromium as
4-6
-------
hexavalent chromium, and nickel as nickel refinery dust and nickel subsulfide) are classified as
carcinogenic to humans. Specifically, hexavalent chromium is carcinogenic to humans by the
inhalation of exposure. Two other key HAP emitted by EGUs (cadmium and selenium as
selenium sulfide) are classified as probable human carcinogens.
U.S. EGU source category emissions of non-Hg HAP are not expected to exceed 1 in a
million for inhalation cancer risk for those facilities impacted by the control requirements in the
final rule. Further, cancer risk was determined to fall within the acceptable range for
multipathway exposure to the persistent and bioaccumulative non-Hg HAP metals, such as
arsenic, cadmium, and lead.53 However, the projected emissions reductions should reduce levels
of exposure to carcinogenic HAP in communities near the impacted facilities.
EPA also evaluated the potential for noncancer risks from exposure to non-Hg HAP
metals in 2020. To address the risk from chronic inhalation exposure to multiple pollutants, we
aggregated the health risks associated with pollutants that affect the same target organ. Further,
we examined the potential for adverse health effects from acute inhalation exposure to individual
pollutants. Lastly, we also examined the potential for health impacts stemming from multiple
pathways of exposure for arsenic, cadmium, and lead. The estimated risks were not expected to
exceed current health thresholds for adverse effects (U.S. EPA, 2020d). Therefore, we are unable
to identify or quantify noncancer benefits from the projected non-Hg HAP metals emission
reductions, although we do note that emissions reductions associated with this rule should further
reduce exposure to these non-Hg HAP metals in communities near these facilities.
In the subsequent sections, we describe the health effects associated with the main non-
Hg HAP metals of concern: antimony (Section 4.2.2.1), arsenic (Section 4.2.2.2), beryllium
(Section 4.2.2.3), cadmium (Section 4.2.2.4), chromium (Section 4.2.2.5), cobalt (Section
4.2.2.6), lead (Section 4.2.2.7), manganese (Section 4.2.2.8), nickel (Section 4.2.2.9), and
selenium (Section 4.2.2.10). This final rule is projected to reduce at least four to seven tons of
non-Hg HAP metals emissions per year. With the data available, it was not possible to estimate
the change in emissions of each individual HAP.
53 https://www.regulations.gov/document/EPA-HQ-OAR-2018-0794-0014.
4-7
-------
4.2.2.1 Antimony
Antimony (Sb), a naturally occurring element, is released into the environment by
incinerators and coal-burning power plants and is considered toxic through the oral, inhalation
and dermal routes. The respiratory tract is most sensitive to the effects of inhaled Sb. Acute
(short-term) inhalation exposure to Sb results in effects including respiratory irritation,
pulmonary inflammation, increases in lung macrophages and impaired lung clearance. Acute
high-level inhalation exposure to Sb has been associated with degeneration in heart and EKG
alterations (ATSDR, 2019). Chronic (long-term) inhalation exposure to Sb has been associated
with interstitial fibrosis and lung neoplasms. EPA has not assessed Sb for carcinogenicity under
the IRIS program (U.S. EPA, 1987a)
4.2.2.2 Arsenic
Arsenic (As), a naturally occurring element, is found throughout the environment, and is
considered toxic through the oral, inhalation and dermal routes. Acute (short-term) high-level
inhalation exposure to as dust or fumes has resulted in gastrointestinal effects (nausea, diarrhea,
abdominal pain, and gastrointestinal hemorrhage); central and peripheral nervous system
disorders have occurred in workers acutely exposed to inorganic As. Chronic (long-term)
inhalation exposure to inorganic as in humans is associated with irritation of the skin and mucous
membranes. Chronic inhalation can also lead to conjunctivitis, irritation of the throat and
respiratory tract, and perforation of the nasal septum (ATSDR, 2007). Chronic oral exposure has
resulted in gastrointestinal effects, anemia, peripheral neuropathy, skin lesions,
hyperpigmentation, and liver or kidney damage in humans. Inorganic As exposure in humans, by
the inhalation route, has been shown to be strongly associated with lung cancer, while ingestion
of inorganic as in humans has been linked to a form of skin cancer and also to bladder, liver, and
lung cancer. EPA has classified inorganic arsenic as a Group A, human carcinogen (U.S. EPA,
1995a).
4.2.2.3 Beryllium
The major sources of beryllium emissions are from the combustion of fossil fuels like
coal and fuel oil. Acute exposure to beryllium compounds can lead to skin irritation, dermatitis,
upper and lower airway inflammation, and pulmonary edema (Jakubowski and Palczynski,
2007). Inhalation of beryllium compounds can lead to the storage of the compound in the lung
4-8
-------
tissue and cause a specific lung disease called chronic beryllium disease (CBD) which starts with
beryllium sensitization (Seidler et al., 2012). Common symptoms of CBD include fatigue,
coughing, weight loss, and fevers. Research has shown that beryllium exposure causes cancer in
rats and monkeys, and while some research shows a relationship with cancer in humans, it is not
definitive. Beryllium is considered to be a Group B1 probable human carcinogen by EPA (U.S.
EPA, 1998a).
4.2.2.4 Cadmium
The main sources of cadmium in air are the burning of fossil fuels and the incineration of
municipal waste. Acute inhalation in humans causes adverse effects in the lung, such as
pulmonary irritation. Chronic inhalation in humans can result in a build-up of cadmium in the
kidney, and if sufficiently high, may result in kidney disease. Animal studies indicate that
cadmium may cause adverse developmental effects, including reduced body weight, skeletal
malformation, and altered behavior and learning (ATSDR, 2012a). Lung cancer has been found
in some studies of workers exposed to Cd in the air and studies of rats that inhaled cadmium.
EPA has classified cadmium as a probable human carcinogen (Group Bl) (U.S. EPA, 1987b).
4.2.2.5 Chromium
Chromium (Cr) may be emitted in two forms, trivalent Cr (Cr+3) or hexavalent Cr
(Cr+6). The respiratory tract is the major target organ for Cr+6 toxicity, for acute and chronic
inhalation exposures. Shortness of breath, coughing, and wheezing have been reported from
acute exposure to Cr+6, while perforations and ulcerations of the septum, bronchitis, decreased
pulmonary function, pneumonia, and other respiratory effects have been noted from chronic
exposures. Animal studies have reported adverse reproductive effects from exposure to Cr+6.
Human and animal studies have clearly established the carcinogenic potential of Cr+6 by the
inhalation route, resulting in an increased risk of lung cancer (ATSDR, 2012b). EPA has
classified Cr+6 as a Group A, human carcinogen (U.S. EPA, 1998c). Trivalent Cr is less toxic
than Cr+6. The respiratory tract is also the major target organ for Cr+3 toxicity, similar to Cr+6.
EPA has not classified Cr+3 with respect to carcinogenicity (U.S. EPA, 1998b).
4.2.2.6 Cobalt
Cobalt (Co) and cobalt compounds are naturally occuring and possess physiochemical
properties like iron and nickel. The primary anthropogenic sources of Co in the environment are
4-9
-------
from the burning of fossil fuels, mining and smelting of Co ores, and processing of cobalt-
containing alloys. Exposure to Co in the general population occurs through inhalation of ambient
air or ingestion of food and drinking water. The respiratory tract is most sensitive to the effects
of inhaled Co. Acute (short-term) inhalation exposure to Co results in pulmonary irritation and
edema. Chronic (long-term) inhalation exposure to Co results in decreased lung function,
inflammation, and lesions cobalt (ATSDR, 2023a). EPA has not yet assessed Co for
carcinogenicity under the IRIS program (U.S. EPA, 2008).
4.2.2.7 Lead
Lead is found naturally in ore deposits. A major source of lead in the U.S. environment
has historically been from combustion of leaded gasoline, which was phased out of use after
1973. Other sources of lead have included mining and smelting of ore; manufacture of and use of
lead-containing products (e.g., lead-based paints, pigments, and glazes; electrical shielding;
plumbing; storage batteries; solder; and welding fluxes); manufacture and application of lead-
containing pesticides; combustion of coal and oil; and waste incineration. Lead is associated with
toxic effects in every organ system including adverse renal, cardiovascular, hematological,
reproductive, and developmental effects. However, the major target for lead toxicity is the
nervous system, both in adults and children. Long-term exposure of adults to lead at work has
resulted in decreased performance in some tests that measure functions of the nervous system.
Lead exposure may also cause weakness in fingers, wrists, or ankles. Lead exposure also causes
small increases in blood pressure, particularly in middle-aged and older people and may also
cause anemia. Children are more sensitive to the health effects of lead than adults. No safe blood
lead level in children has been determined. At lower levels of exposure, lead can affect a child's
mental and physical growth. Fetuses exposed to lead in the womb may be born prematurely and
have lower weights at birth. Exposure in the womb, in infancy, or in early childhood also may
slow mental development and cause lower intelligence later in childhood. There is evidence that
these effects may persist beyond childhood (ATSDR, 2023b). EPA has determined that lead is a
probable human carcinogen (Group 2B) (U.S. EPA, 1988).
4.2.2.8 Manganese
Manganese (Mn) is a naturally occuring metal found in rock and used in steel production
or as an additive in gasoline. Chronic exposure to high levels of Mn by inhalation in humans
4-10
-------
results primarily in central nervous system effects. Visual reaction time, hand steadiness, and
eye-hand coordination were affected in chronically-exposed workers. Manganism, characterized
by feelings of weakness and lethargy, tremors, a masklike face, and psychological disturbances,
may result from chronic exposure to higher levels. Impotence and loss of libido have been noted
in male workers afflicted with Manganism attributed to inhalation exposures. High levels of
exposure have been associated with lung irritation and reproductive effects. In animals, nervous
system and reproductive effects have been observed (ATSDR, 2012c). EPA has classified Mn in
Group D, not classifiable as to carcinogenicity in humans (U.S. EPA, 1995b).
4.2.2.9 Nickel
Nickel (Ni) is found in ambient air as a result of releases from oil and coal combustion,
nickel metal refining, sewage sludge incineration, manufacturing facilities, and other sources.
Respiratory effects have been reported in humans from inhalation exposure to nickel. Acute
exposure to nickel carbonyl has been associated with reports of pulmonary fibrosis and renal
edema in both animals and humans. Chronic inhalation of nickel in workers can cause chronic
bronchitis and reduced lung function (ATSDR, 2005, 2023b). Human and animal studies have
reported an increased risk of lung and nasal cancers from exposure to nickle refinery dusts and
nickel subsulfide. EPA has classified nickel subsulfide and nickel refinery dusts as human
carcinogens and nickel carbonyl as a probable human carcinogen (U.S. EPA, 1987c, 1987d,
1987e).
4.2.2.10 Selenium
Selenium has many uses including in the electronics industry; the glass industry; in
pigments used in plastics, paints, enamels, inks, and rubber; as a catalyst in the preparation of
pharmaceuticals; and in special trades. Dizziness, fatigue, and irritation of mucous membranes
have been reported in people exposed to high levels of selenium in the air in the workplace. High
amounts of selenium have been associated with adverse reproductive effects in animal studies.
However, the relevance of the effects observed in rats and monkeys to humans is not known
(ATSDR, 2003). One selenium compound, selenium sulfide, is carcinogenic in animals exposed
orally. EPA has classified elemental Se as a Group D2, not classifiable as to human
carcinogenicity, and selenium sulfide as a Group B2, probable human carcinogen (U.S. EPA,
1991).
4-11
-------
4.2.3 Additional HAP Benefits
As discussed in detail in the 2023 Final A&N Review, it is challenging to quantify the
full range of benefits of HAP reductions. But that does not mean that these benefits are small,
insignificant, or nonexistent. In the 2011 MATS RIA (U.S. EPA, 2011), EPA discussed the
potential for non-monetizable benefits from effects on fish, birds, and mammals, in part
represented through the commercial and recreational fishing economy. A report submitted to
EPA in comments concluded that recreational and commercial fishing are substantial
contributors to regional U.S. economies with dollar values in the tens of billions (IEc, 2019). At
this scale of economic activity, even small shifts in consumer behavior prompted by further HAP
reductions can result in substantial economic impacts.
As another example of the potential value of these emissions reductions, EPA received
numerous comments in the public comment periods of past EGU HAP regulations highlighting
that benefits of Hg reductions to tribal health, subsistence, fishing rights, and cultural identity,
while not easily quantified or monetized, are nonetheless important to consider. Finally, EPA
also qualitatively considers impacts on ecosystem services, which are generally defined as the
economic benefits that individuals and organizations obtain from ecosystems. The monetization
of endpoints like ecosystem services, tribal culture, and the activity related to fishing remains
challenging. While EPA is not able to monetize the impacts of reduced HAP exposures projected
for this rule, we note the importance of the contributions of further reductions of HAP emissions
to the sustainability of these important economic and cultural values.
4.3 Criteria Pollutant Benefits
The benefits analysis presented in this section applies methods consistent with those
employed most recently in the RIA for the proposed PM National Ambient Air Quality
Standards (NAAQS). EPA's approach for selecting PM2.5 and ozone-related health endpoints to
quantify and monetize is summarized below and we refer readers to the referenced Health
Benefits TSD for a full description of our methods (U.S. EPA, 2023a).
Estimating the health benefits of reductions in PM2.5 and ozone exposure begins with
estimating the change in exposure for each individual and then estimating the change in each
individual's risks for those health outcomes affected by exposure. The benefit of the reduction in
4-12
-------
each health risk is based on the exposed individual's willingness to pay (WTP) for the risk
change, assuming that each outcome is independent of one another. The greater the magnitude of
the risk reduction from a given change in concentration, the greater the individual's WTP, all
else equal. The social benefit of the change in health risks equals the sum of the individual WTP
estimates across all of the affected individuals residing in the U.S.54
We conduct this analysis 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)
developing spatial fields of air quality for the baseline and final rule (2) selecting air pollution
health endpoints to quantify; (3) calculating counts of air pollution effects using a health impact
function; (4) specifying the health impact function with concentration-response parameters
drawn from the epidemiological literature to calculate the economic value of the health impacts.
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.
As structured, the final rule would affect the distribution of ozone and PM2.5
concentrations in much of the U.S. This RIA estimates avoided ozone- and PIVfo.s-related health
impacts that are distinct from those reported in the RIAs for both ozone and PM NAAQS (U.S.
EPA, 2015, 2022d) The ozone and PM NAAQS RIAs illustrate, but do not predict, the benefits
and costs of strategies that states may choose to enact when implementing a revised NAAQS;
these costs and benefits are illustrative and cannot be added to the costs and benefits of policies
that prescribe specific emission control measures. This RIA estimates the benefits (and costs) of
specific emissions control measures. The benefit estimates are based on these modeled changes
in PM2.5 and summer season average ozone concentrations.
54 This RIA also reports the change in the sum of the risk, or the change in the total incidence, of a health outcome
across the population. If the benefit per unit of risk is invariant across individuals, the total expected change in the
incidence of the health outcome across the population can be multiplied by the benefit per unit of risk to estimate the
social benefit of the total expected change in the incidence of the health outcome.
4-13
-------
4.3.1 Air Quality Modeling Methodology
The final rule influences the level of pollutants emitted in the atmosphere that adversely
affect human health, including directly emitted PM2.5, as well as SO2 and NOx, which are both
precursors to ambient PM2.5. NOx emissions are also a precursor to ambient ground-level ozone.
EPA used air quality modeling to estimate changes in ozone and PM2.5 concentrations that may
occur as a result of the final rule relative to the baseline.
As described in the Air Quality Modeling Appendix (Appendix A), gridded spatial fields
of ozone and PM2.5 concentrations representing the baseline and final rule were derived from
CAMx source apportionment modeling in combination with NOx, SO2, and primary PM2.5 EGU
emissions obtained from the outputs of the IPM runs described in Section 3 of this RIA. While
the air quality modeling includes all inventoried pollution sources in the contiguous U.S.,
contributions from all sources other than EGUs are held constant at projected 2026 levels in this
analysis, and the only changes quantified between the baseline and the final rule are those
associated with the projected impacts of this final rule on EGU emissions. EPA prepared gridded
spatial fields of air quality for the baseline and the final rule for two health-impact metrics:
annual mean PM2.5 and April through September seasonal average eight-hour daily maximum
(MDA8) ozone (AS-M03). These ozone and PM2.5 gridded spatial fields cover all locations in
the contiguous U.S. and were used as inputs to BenMAP-CE which, in turn, was used to quantify
the benefits from this rule.
The basic methodology for determining air quality changes is the same as that used in the
RIAs from multiple previous rules (U.S. EPA, 2019b, 2020a, 2020b, 2021a, 2022c). The Air
Quality Modeling Appendix (Appendix A) provides additional details on the air quality
modeling and the methodologies EPA used to develop gridded spatial fields of summertime
ozone and annual PM2.5 concentrations. The appendix also provides figures showing the
geographical distribution of air quality changes.
4.3.2 Selecting Air Pollution Health Endpoints to Quantify
The methods used in this RIA incorporate evidence reported in the most recent completed
PM Integrated Science Assessment (PM ISA) and Ozone Integrated Science Assessments
(Ozone ISA) and accounts for recommendations from the Science Advisory Board (U.S. EPA,
2022e). When updating each health endpoint EPA considered: (1) the extent to which there
4-14
-------
exists a causal relationship between that pollutant and the adverse effect; (2) whether suitable
epidemiologic studies exist to support quantifying health impacts; (3) and whether robust
economic approaches are available for estimating the value of the impact of reducing human
exposure to the pollutant. Our approach for updating the endpoints and to identify suitable
epidemiologic studies, baseline incidence rates, population demographics, and valuation
estimates is summarized below. Detailed descriptions of these updates are available in the Health
Benefits TSD, which is in the docket for this rulemaking. The Health Benefits TSD describes the
Agency's approach for quantifying the number and value of estimated air pollution-related
impacts. Updates since the publication of the Health Benefits TSD are described below. In this
document the reader can find the rationale for selecting health endpoints to quantify; the
demographic, health and economic data used; modeling assumptions; and our techniques for
quantifying uncertainty.55
55 The analysis was completed using BenMAP-CE version 1.5.8, which is a variant of the current publicly available
version. We also include new estimates of the cost of asthma onset and stroke beyond those described in the Health
Benefits TSD.
4-15
-------
Table 4-1
Health Effects of PM2.5, Ambient Ozone,
and Climate Effects
Category
Effect
Effect
Quantified
Effect
Monetized
More
Information
Premature mortality
from exposure to
Adult premature mortality based on cohort study
estimates and expert elicitation estimates (age 65-99
or age 30-99)
~
~
PMISA
PM2.5
Infant mortality (age <1)
~
~
PMISA
Heart attacks (age >18)
~
~i
PMISA
Hospital admissions—cardiovascular (ages 65-99)
~
~
PMISA
Emergency department visits— cardiovascular (age
0-99)
~
~
PMISA
Hospital admissions—respiratory (ages 0-18 and 65-
99)
Emergency room visits—respiratory (all ages)
Cardiac arrest (ages 0-99; excludes initial hospital
and/or emergency department visits)
Stroke (ages 65-99)
Asthma onset (ages 0-17)
¦/
7
~
~
~
~
~
~i
71
~
PMISA
PMISA
PMISA
PMISA
PMISA
Asthma symptoms/exacerbation (6-17)
~
~
PMISA
Nonfatal morbidity
from exposure to
Lung cancer (ages 30-99)
Allergic rhinitis (hay fever) symptoms (ages 3-17)
Lost work days (age 18-65)
~
~
~
~
~
~
PMISA
PMISA
PMISA
PM2.5
Minor restricted-activity days (age 18-65)
¦/
~
PMISA
Hospital admissions—Alzheimer's disease (ages 65-
99)
¦/
~
PMISA
Hospital admissions—Parkinson's disease (ages 65-
99)
~
~
PMISA
Other cardiovascular effects (e.g., other ages)
—
—
PMISA2
Other respiratory effects (e.g., pulmonary function,
non-asthma ER visits, non-bronchitis chronic
diseases, other ages, and populations)
—
—
PMISA2
Other nervous system effects (e.g., autism, cognitive
decline, dementia)
—
—
PMISA2
Metabolic effects (e.g., diabetes)
—
—
PMISA2
Reproductive and developmental effects (e.g., low
birth weight, pre-term births, etc.)
—
—
PMISA2
Cancer, mutagenicity, and genotoxicity effects
—
—
PMISA2
Mortality from
Premature respiratory mortality based on short-term
study estimates (0-99)
¦/
~
Ozone ISA
exposure to ozone
Premature respiratory mortality based on long-term
study estimates (age 30-99)
~
~
Ozone ISA
Hospital admissions—respiratory (ages 0-99)
~
~
Ozone ISA
Emergency department visits—respiratory (ages 0-
99)
Asthma onset (0-17)
Asthma symptoms/exacerbation (asthmatics age 2-
ill
Allergic rhinitis (hay fever) symptoms (ages 3-17)
Minor restricted-activity days (age 18-65)
~
~
Ozone ISA
Nonfatal morbidity
from exposure to
ozone
~
¦/
7
7
~
~
~
~
Ozone ISA
Ozone ISA
Ozone ISA
Ozone ISA
School absence days (age 5-17)
~
~
Ozone ISA
Decreased outdoor worker productivity (age 18-65)
—
—
Ozone ISA2
Metabolic effects (e.g., diabetes)
—
—
Ozone ISA2
Other respiratory effects (e.g., premature aging of
lungs)
—
—
Ozone ISA2
4-16
-------
Table 4-1
Health Effects of PM2.5, Ambient Ozone, and Climate Effects
Category
Effect
Effect Effect More
Quantified Monetized Information
Reproductive and developmental effects
Climate impacts from carbon dioxide (CO2)
Cardiovascular and nervous svstem effects
— Ozone ISA2
— Ozone ISA2
S Section 4.4
Climate
effects
Other climate impacts (e.g., ozone, black carbon,
aerosols, other impacts)
IPCC,
Ozone ISA,
PMISA
1 Valuation estimate excludes initial hospital and/or emergency department visits.
2 Not quantified due to data availability limitations and/or because current evidence is only suggestive of causality.
4.3.3 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 and summer season
average ozone concentrations for the years 2030, 2035, and 2040 using health impact functions
(Sacks et al., 2020). 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.
BenMAP quantifies counts of attributable effects using health impact functions, which
combine information regarding the: concentration-response relationship between air quality
changes and the risk of a given adverse outcome; population exposed to the air quality change;
baseline rate of death or disease in that population; and air pollution concentration to which the
population is exposed.
The following provides an example of a health impact function, in this case for PM2.5
mortality risk. We estimate counts of PIVfo.s-related total deaths (y;y) during each year i among
adults aged 18 and older (a) in each county j in the contiguous U.S. (where j = 1,...,/ and ./is
the total number of counties) as:
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, fi is the risk coefficient for total mortality for adults associated
4-17
-------
with annual average PM2.5 exposure, Q is the annual mean PM2.5 concentration in county j in
year 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.56
The BenMAP-CE tool is pre-loaded with projected population from the Woods & Poole
company; cause-specific and age-stratified death rates from the Centers for Disease Control and
Prevention, projected to future years; recent-year baseline rates of hospital admissions,
emergency department visits and other morbidity outcomes from the Healthcare Cost and
Utilization Program and other sources; concentration-response parameters from the published
epidemiologic literature cited in the ISAs for fine particles and ground-level ozone; and cost of
illness or WTPWTP economic unit values for each endpoint. Consistent with advice received
from the U.S. EPA Science Advisory Board, EPA will substitute the existing Woods & Poole
population projections with those that are not proprietary (U.S. EPA Science Advisory Board,
2024).
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.
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 analyses.
Thus, similar to work by Kiinzli et al. (2000) and co-authors and other, more recent health
impact analyses, our estimates are based on the best available methods of benefits transfer.
Benefits transfer is the science and art of adapting primary research from similar contexts to
obtain the most accurate measure of benefits for the environmental quality change under
analysis. Adjustments are made for the level of environmental quality change, the socio-
demographic and economic characteristics of the affected population, and other factors to
improve the accuracy and robustness of benefits estimates.
56 In this illustrative example, the air quality is resolved at the county level. For this RIA, we simulate air quality
concentrations at a 12 km grid cell resolution The BenMAP-CE tool assigns the rates of baseline death and disease
stored at the county level to the 12 km grid cells using an area-weighted algorithm. This approach is described in
greater detail in the appendices to the BenMAP-CE user manual.
4-18
-------
4.3.4 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
$1,000, then the WTP for an avoided statistical premature mortality amounts to $10 million
($1,000/0.0001 change in risk). Hence, this value is population-normalized, as it accounts for the
size of the population and the percentage of that population experiencing the 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.
4.3.5 Benefits Analysis Data Inputs
In Figure 4-1, we summarize the key data inputs to the health impact and economic
valuation estimates, which were calculated using BenMAP-CE tool version 1.5.1. (Sacks et al.,
2020). In the sections below we summarize the data sources for each of these inputs, including
4-19
-------
demographic projections, incidence and prevalence rates, effect coefficients, and economic
valuation.
Figure 4-1 Data Inputs and Outputs for the BenMAP-CE Tool
4.3.5.1 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. (2015). The Woods & Poole
database contains county-level projections of population by age, sex, and race 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 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:
4-20
-------
• 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, 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.
4.3.5.2 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 percent. A baseline incidence rate, necessary to convert this
relative change into a number of cases, is the estimate of the 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.
The Health Benefits TSD (see Table 12) 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 endpoints, whereas county-level data are available for premature
mortality. Whenever possible, the national rates used are national averages, because these data
4-21
-------
are most applicable to a national assessment of benefits. When quantifying some endpoints, we
were unable to identify a suitable administrative database supplying baseline rates of the event of
interest; in these cases, we selected an incidence rate reported within the study supplying the risk
estimate.
We projected mortality rates such that future mortality rates are consistent with our
projections of population growth. 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). Further information regarding this procedure may be found in
the Health Benefits TSD and the appendices to the BenMAP user manual (U.S. EPA, 2022a).
The baseline incidence rates for hospital admissions and emergency department visits
reflect the revised rates first applied in the Revised Cross-State Air Pollution Rule Update (U.S.
EPA, 2021a). In addition, we revised the baseline incidence rates for acute myocardial infarction.
These revised rates are more recent than the rates they replace and more accurately represent the
rates at which populations of different ages, and in different locations, visit the hospital and
emergency department for air pollution-related illnesses. Lastly, these rates reflect unscheduled
hospital admissions only, which represents a conservative assumption that most air pollution-
related visits are likely to be unscheduled. If air pollution-related hospital admissions are
scheduled, this assumption would underestimate these benefits.
4.3.5.3 Effect Coefficients
Our approach for selecting and parametrizing effect coefficients for the benefits analysis
is described fully in the Health Benefits TSD. Because of the substantial economic value
associated with estimated counts of PIvfc.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 Health Benefits TSD.
A substantial body of published scientific literature documents the association between
PM2.5 concentrations and the risk of premature death (U.S. EPA, 2019a, 2022e). This body of
literature reflects thousands of epidemiology, toxicology, and clinical studies. The PM ISA,
completed as part of this review of the fPM standards and reviewed by the Clean Air Scientific
Advisory Committee (CASAC) (U.S. EPA Science Advisory Board, 2022) concluded that there
4-22
-------
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. The size of the mortality effect estimates from
epidemiologic studies, the serious nature of the effect itself, and the high monetary value
ascribed to prolonging life make mortality risk reduction 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 Health Benefits TSD
accompanying this RIA that is generally consistent with previous RIAs. Briefly, clinically
significant epidemiologic studies of health endpoints for which ISAs report strong evidence are
evaluated using established minimum and preferred criteria for identifying studies and hazard
ratios best characterizing risk. Following this systematic approach led to the identification of
three studies best characterizing the risk of premature death associated with long-term exposure
to PM2.5 in the U.S. (Pope et al., 2019; Turner et al., 2016; X Wu et al., 2020). The 2019 PM ISA
(U.S. EPA, 2019a), the 2022 Supplement to the PM ISA (U.S. EPA, 2022e), and the 2022 PM
Policy Assessment (U.S. EPA, 2022b) also identified these three studies as providing key
evidence of the association between long-term PM2.5 exposure and mortality. These studies used
data from three U.S. cohorts: (1) an analysis of Medicare beneficiaries (Medicare); (2) the
American Cancer Society (ACS); and (3) the National Health Interview Survey (NHIS). 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 three identified studies and hazard ratios
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 percent 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 km grid resolution using a hybrid ensemble-based prediction model that
combined three machine learning models and relied on satellite data, land-use information,
4-23
-------
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 percent confidence interval (1.058-
1.074) associated with a change in annual mean PM2.5 exposure of 10.0 |ig/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 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 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 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 percent confidence interval (1.08-1.15) associated with a change in annual mean PM2.5
exposure of 10.0 |ig/m3 (Pope et al., 2019, Table 2, Subcohort). This study exhibits two key
4-24
-------
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.
EPA has historically used estimated Hazard Ratios from extended analyses of the ACS
cohort to estimate PM-related risk of premature death Krewski (Krewski et al., 2009; Pope et al.,
2002; Pope et al., 1995). A more recent ACS analysis, Turner et al. (2016):
• extended the follow-up period of the ACS CSP-II to 22 years (1982-2004),
• evaluated 669,046 participants over 12,662,562 person-years of follow up and 237,201
observed deaths, and
applied a more advanced exposure estimation approach than had previously been used
when analyzing the ACS cohort, combining the geostatistical Bayesian Maximum Entropy
framework with national-level land use regression models.
The total mortality hazard ratio best estimating risk from these ACS cohort studies was
based on a random-effects Cox proportional hazard model incorporating multiple individual and
ecological covariates (relative risk =1.06, 95 percent confidence intervals 1.04-1.08 per 10
|ig/m3 increase in PM2.5) from Turner et al. (2016). The relative risk estimate is identical to a risk
estimate drawn from earlier ACS analysis of all-cause long-term exposure PM2.5-attributable
mortality (Krewski et al., 2009). However, as the ACS hazard ratio is quite similar to the
Medicare estimate of (1.066, 1.058-1.074), especially when considering the broader age range
(greater than 29 versus greater than 64), only Wu et al. (2020) and Pope et al. (2019) are
included in the main benefits assessments, with Wu et al. (2020) representing results from both
the Medicare and ACS cohorts.
4.3.6 Quantifying Cases of Ozone-Attributable Premature Death
Mortality risk reductions account for the majority of monetized ozone-related and PM2.5-
related benefits. For this reason, this subsection and the following provide a brief background of
the scientific assessments that underly the quantification of these mortality risks and identifies
the risk studies used to quantify them in this RIA, for ozone and PM2.5 respectively. As noted
above, U.S. EPA (2023a) describes fully the Agency's approach for quantifying the number and
value of ozone and PM2.5 air pollution-related impacts, including additional discussion of how
4-25
-------
the Agency selected the risk studies used to quantify them in this RIA. The Health Benefits TSD
also includes additional discussion of the assessments that support quantification of these
mortality risk than provide here.
In 2008, the National Academies of Science issued a series of recommendations to EPA
regarding the procedure for quantifying and valuing ozone-related mortality due to short-term
exposures (National Research Council, 2008). Chief among these was that"... short-term
exposure to ambient ozone is likely to contribute to premature deaths" and the committee
recommended that "ozone-related mortality be included in future estimates of the health benefits
of reducing ozone exposures..." The NAS also recommended that".. .the greatest emphasis be
placed on the multicity and [National Mortality and Morbidity Air Pollution Studies
(NMMAPS)] ... studies without exclusion of the meta-analyses" (National Research Council,
2008). Prior to the 2015 Ozone NAAQS RIA, the Agency estimated ozone-attributable
premature deaths using an NMMAPS-based analysis of total mortality (Bell et al., 2004), two
multi-city studies of cardiopulmonary and total mortality (Huang et al., 2005; Schwartz, 2005),
and effect estimates from three meta-analyses of non-accidental mortality (Bell et al., 2005; Ito et
al., 2005; Levy et al., 2005). Beginning with the 2015 Ozone NAAQS RIA, the Agency began
quantifying ozone-attributable premature deaths using two newer multi-city studies of non-
accidental mortality (R. L. Smith et al., 2009; Zanobetti and Schwartz, 2008) and one long-term
cohort study of respiratory mortality (Jerrett et al. 2009).
EPA quantifies and monetizes effects the Integrated Science Assessment (ISA) identifies
as having either a causal or likely-to-be-causal relationship with the pollutant. Relative to the
2015 ISA, the 2020 ISA for Ozone reclassified the casual relationship between short-term ozone
exposure and total mortality, changing it from "likely to be causal" to "suggestive of, but not
sufficient to infer, a causal relationship." The 2020 Ozone ISA separately classified short-term
ozone exposure and respiratory outcomes as being "causal" and long-term exposure as being
"likely to be causal." When determining whether there existed a causal relationship between
short- or long-term ozone exposure and respiratory effects, EPA evaluated the evidence for both
morbidity and mortality effects. The ISA identified evidence in the epidemiologic literature of an
association between ozone exposure and respiratory mortality, finding that the evidence was not
entirely consistent and there remained uncertainties in the evidence base.
4-26
-------
EPA continues to quantify premature respiratory mortality attributable to both short- and
long-term exposure to ozone because doing so is consistent with: (1) the evaluation of causality
noted above; and (2) EPA's approach for selecting and quantifying endpoints described in the
Technical Support Document (TSD) "Estimating PM2.5- and Ozone-Attributable Health
Benefits," which was recently reviewed by the U.S. EPA Science Advisory Board (U.S. EPA,
2023; U.S. EPA-SAB 2024).
We estimate counts of ozone-attributable respiratory death from short-term exposures a
pooled risk estimate calculated using parameters from Zanobetti and Schwartz (2008) and
Katsouyanni et al. (2009). Consistent with the RIA for the Final Revised CSAPR Update (U.S.
EPA, 2021a), we use two estimates of ozone-attributable respiratory deaths from short-term
exposures are estimated using the risk estimate parameters from Zanobetti and Schwartz (2008)
and Katsouyanni et al. (2009). Ozone-attributable respiratory deaths from long-term exposures
are estimated using Turner et al. (2016). Due to time and resource limitations, we were unable to
reflect the warm season defined by Zanobetti and Schwartz (2008) as June-August. Instead, we
apply this risk estimate to our standard warm season of May-September.
4.3.7 Quantifying Cases of PM2 5-A ttributable Premature Death
When quantifying PM-attributable cases of adult mortality, we use the effect coefficients
from two epidemiology studies examining two large population cohorts: the American Cancer
Society cohort (Turner et al., 2016) and the Medicare cohort (Di et al., 2017). The 2019 PM ISA
indicates that the ACS and Medicare cohorts provide strong evidence of an association between
long-term PM2.5 exposure and premature mortality with support from additional cohort studies.
There are distinct attributes of both the ACS and Medicare cohort studies that make them well-
suited to being used in a PM benefits assessment and so here we present PM2.5 related effects
derived using relative risk estimates from both cohorts.
The PM ISA, which was reviewed by the Clean Air Scientific Advisory Committee of
EPA's Science Advisory Board (U.S. EPA Science Advisory Board, 2022), concluded that there
is a causal relationship between mortality and both long-term and short-term exposure to PM2.5
based on the entire body of scientific evidence. The PM ISA also concluded that the scientific
literature supports the use of a no-threshold log-linear model to portray the PM-mortality
concentration-response relationship while recognizing potential uncertainty about the exact
4-27
-------
shape of the concentration-response relationship. The 2019 PM ISA, which informed the setting
of the 2020 PM NAAQS, reviewed available studies that examined the potential for a
population-level threshold to exist in the concentration-response relationship. Based on such
studies, the ISA concluded that the evidence supports the use of a "no-threshold" model and that
"little evidence was observed to suggest that a threshold exists" (U.S. EPA, 2009a). Consistent
with this evidence, the Agency historically has estimated health impacts above and below the
prevailing NAAQS (U.S. EPA, 2019b, 2021a, 2022c).
4.3.8 Characterizing Uncertainty in the Estimated Benefits
Like other complex analyses using estimated parameters and inputs from numerous
models, there are sources of uncertainty. The Health Benefits TSD details our approach to
characterizing uncertainty in both quantitative and qualitative terms (U.S. EPA, 2023a). The
Health Benefits 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 Health Benefits 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 Health Benefits TSD:
1. For adult all-cause mortality:
4-28
-------
a. The distributions of air quality concentrations experienced by the original
cohort population (Health Benefits TSD Section 7.1.2.1);
b. Methods of estimating and assigning exposures in epidemiologic studies
(Health Benefits TSD Section 7.1.2.2);
c. Confounding by ozone (Health Benefits TSD Section 7.1.2.3); and
d. The statistical technique used to generate hazard ratios in the epidemiologic
study (Health Benefits TSD Section 7.1.2.4).
2. Plausible alternative risk estimates for asthma onset in children (TSD Section 7.1.3),
cardiovascular hospital admissions (Health Benefits TSD Section 7.1.4,), and
respiratory hospital admissions (Health Benefits TSD Section 7.1.5);
3. Effect modification of PM2.5-attributable health effects in at-risk populations (Health
Benefits 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 Health Benefits TSD, respectively. Qualitative
discussions of various sources of uncertainty can be found in Section 7.5 of the Health Benefits
TSD.
4.3.8.1 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 estimated using 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 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
4-29
-------
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.
4.3.8.2 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 Health Benefits TSD.
Key assumptions underlying the estimates for premature mortality, which account for
over 98 percent 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 Science Advisory Board, 2022).
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 Science Advisory Board, 2022).
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
board (U.S. EPA Science Advisory Board, 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.
4.3.9 Estimated Number and Economic Value of Health Benefits
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 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
4-30
-------
recommended by the Guidelines for Preparing Economic Analyses (U.S. EPA, 2014). Below we
report the estimated number of reduced premature deaths and illnesses in each year relative to
the baseline along with the 95 percent confidence interval (Table 4-2 or ozone-related health
impacts and Table 4-3 for PIVfo.s-related impacts). The number of reduced estimated deaths and
illnesses from the final are calculated from the sum of individual reduced mortality and illness
risk across the population.
Table 4-2 Estimated Avoided Ozone-Related Premature Respiratory Mortalities and
Illnesses for the Final Rule for 2028, 2030, and 2035 (95 percent confidence interval)"
2028
2030
20358
Avoided premature respiratory mortalities
Long-
term
exposure
Turner et al. (2016)b
0.37
0.019
-0.07
(0.26 to 0.48)
(0.013 to 0.025)
(-0.091 to -0.049)
Short-
term
exposure
Katsouyanni et al.
(2009)bc and Zanobetti et
al. (2008)° pooled
0.017
(0.0068 to 0.027)
0.0009
(0.0004 to 0.0014)
-0.0032
(-0.005 to-0.0013)
Morbidity effects
Long-
term
exposure
Asthma onsetd
2.3
(2 to 2.6)
0.25
(0.22 to 0.29)
-0.9
(-1.0 to-0.78)
Allergic rhinitis
symptomsf
14
(7.1 to 20)
1.5
(0.79 to 2.2)
-5.1
(-7.4 to -2.7)
Hospital admissions—
0.055
0.0041
-0.0098
respiratory0
(-0.014 to 0.12)
(-0.0011 to 0.009)
(-0.022 to 0.0026)
Short-
term
ED visits—respiratory6
0.62
(0.17 to 1.31)
0.58
(0.016 to 0.12)
-0.14
(-0.3 to -0.039)
Asthma symptoms
440
(-54 to 920)
48
(-5.9 to 100)
-160
(-340 to 20)
exposure
Minor restricted-activity
190
21
-64
days0,6
(76 to 300)
(8.2 to 32)
(-100 to -26)
School absence days
160
(-22 to 330)
17
(-2.5 to 37)
-58
(-120 to 8.2)
a Values rounded to two significant figures.
b Applied risk estimate derived from April-September exposures to estimates of ozone across the May-September
warm season.
0 Converted ozone risk estimate metric from MDA1 to MDA8.
d Applied risk estimate derived from June-August exposures to estimates of ozone across the May-September warm
season.
e Applied risk estimate derived from full year exposures to estimates of ozone across the May-September warm
season.
f Converted ozone risk estimate metric from DA24 to MDA8.
g In 2035, the IPM model projects a small projected increase in NOX emissions results from very small, modeled
changes in fossil dispatch and coal use relative to the baseline. As shown in Figure 8-8, while there are small
predicted ozone decreases from the final rule compared to the baseline evident in North Dakota in 2028 and
Montana in 2035, there are also small predicted ozone increases evident near the border of Arizona and New Mexico
in 2035. These small increases result in the very small negative health impacts presented in this table.
4-31
-------
Table 4-3 Estimated Avoided PlVh.s-Related Premature Mortalities and Illnesses for the
Final Rule in 2028, 2030, and 2035 (95 percent confidence interval)
Avoided Mortality
2028
2030
2035
(Pope et al., 2019) (adult
mortality ages 18-99 years)
7.2
(5.2 to 9.2)
2.7
(1.9 to 3.4)
1.7
(1.2 to 2.1)
(X. Wu et al., 2020) (adult
mortality ages 65-99 years)
3.4
(3 to 3.8)
1.3
(1.1 to 1.4)
0.84
(0.74 to 0.94)
(Woodruff et al., 2008) (infant
mortality)
0.0087
(-0.0055 to 0.022)
0.0026
(-0.0016 to 0.0066)
0.0013
(-0.00083 to 0.0034)
Avoided Morbidity
2028
2030
2035
Hospital admissions—
cardiovascular (age >18)
0.5
(0.37 to 0.64)
0.19
(0.13 to 0.24)
0.12
(0.084 to 0.15)
Hospital admissions—respiratory
0.73
(0.25 to 1.2)
0.23
(0.076 to 0.37)
0.12
(0.038 to 0.20)
ED visits-cardiovascular
1.1
(-0.4 to 2.5)
0.37
(-0.14 to 0.87)
0.23
(-0.088 to 0.53)
ED visits—respiratory
2
(0.4 to 4.3)
0.72
(0.14 to 1.5)
0.41
(0.081 to 0.86)
Acute Myocardial Infarction
0.12
(0.07 to 0.17)
0.042
(0.024 to 0.059)
0.025
(0.015 to 0.036)
Cardiac arrest
0.053
(-0.022 to 0.12)
0.019
(-0.0076 to 0.043)
0.011
(-0.0045 to 0.25)
Hospital admissions-
Alzheimer's Disease
2
(1.5 to 2.5)
0.6
(0.44 to 0.74)
0.33
(0.24 to 0.41)
Hospital admissions-
Parkinson's Disease
0.23
(0.12 to 0.34)
0.087
(0.044 to 0.13)
0.054
(0.027 to 0.08)
Stroke
0.21
(0.0055 to 0.36)
0.077
(0.02 to 0.13)
0.047
(0.012 to 0.081)
Lung cancer
0.24
(0.072 to 0.4)
0.087
(0.026 to 0.15)
0.055
(0.017 to 0.092)
Hay Fever/Rhinitis
52
(13 to 91)
17
(4.2 to 30)
9.7
(2.3 to 17)
Asthma Onset
8.1
(7.8 to 8.4)
2.7
(2.5 to 2.8)
1.4
(1.4 to 1.5)
Asthma symptoms - Albuterol
use
1,500
(-743 to 3,700)
510
(-250 to 1,200)
290
(-140 to 690)
Lost work days
390
(330 to 450)
130
(110 to 150)
73
(62 to 84)
Minor restricted-activity days
2,300
(1,900 to 2,700)
780
(640 to 930)
430
(350 to 510)
Note: Values rounded to two significant figures.
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 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 U.S. EPA (2014). Table 4-4 reports the estimated economic value of
avoided premature deaths and illness in each year relative to the baseline along with the 95
4-32
-------
percent confidence interval. Table 4-5 through Table 4-7 presents the stream of health benefits
from 2028 through 2037 for the final rule using the monetized sums of long-term ozone and
PM2.5 mortality and morbidity impacts discounted at 2, 3, and 7 percent, respectively.57 Note the
benefits of the less stringent regulatory alternative are described qualitatively. As a result, there
are no quantified benefits associated with this regulatory option.
57 EPA continues to refine its approach for estimating and reporting PM-related effects at lower concentrations. The
Agency acknowledges the additional uncertainty associated with effects estimated at these lower levels and seeks to
develop quantitative approaches for reflecting this uncertainty in the estimated PM benefits.
4-33
-------
Table 4-4 Estimated Discounted Economic Value of Avoided Ozone and PM2.5-
Attributable Premature Mortality and Illness for the Final Rule 2028, 2030, and 2035 (95
perceni
confidence interval; millions of 2C
19 dollars)a'b'c
Disc.
Rate
Pollutant
2028
2030
2035
2%
Ozone
Benefits
$1.3 and $5.2
$0.13 and $0.34
-$1.2 and -$0.48
PM25
Benefits
$41 and $82
$15 and $30
$10 and $19
Ozone
plus PM2 5
Benefits
$42 and $87
$15 and $30
$9.50 and $18
3%
Ozone
Benefits
$0.71 $4
($0.34 and ($0.66 to
to $1.3) $11)
$0,066 $0.26
($0.36 to and ($0,053
$0.11) to $0.63)
$-0.96 $-0.24
($-2.3 to and ($-0.38 to
$-0.19) -$0.13)
PM25
Benefits
$38 $78
($5 to and ($8.4 to
$97) $210)
$14 $29
($1.8 to and ($3.1 to
$37) $76)
$9.5 $19
($1.1 to and ($1.9 to
$24) $49)
Ozone
plus PM2 5
Benefits
$39 $82
($5.3 to and ($9.1 to
$98) $220)
$14 $29
($2.4 to and ($3.2 to
37) $77)
$9.3 $18
($0.72 and ($-0.4 to
to $24) $49)
7%
Ozone
Benefits
$0.53 $3.8
($0.18 and ($0.48 to
to $1.1) $9.9)
$0,047
($0t019 and ($0,034
$0,084) t0 $0 55:)
$-0.17 $-0.81
($-0.3 to and ($-2 to $-
$-0,068) 0.13)
PM25
Benefits
$34 $70
($4.1 to and ($7.2 to
$86) $180)
$13 $26
($1.5 to and ($2.6 to
$33) $69)
$8.5 $17
($0.95 and ($1.7 to
to $22) $44)
Ozone
plus PM2 5
Benefits
$35 $7
($4.3 to and ($7.7 to
$87) $190)
$13 $26
($1.5 to and ($2.6 to
$33) $70)
$8.3 $16
($0.65 and ($-0.3 to
to $22) $44)
a Values rounded to two significant figures. The two benefits estimates are separated by the word "and" to signify
that they are two separate estimates. The estimates do not represent lower- and upper-bound estimates and should
not be summed.
b We estimated changes in NOx for the ozone season and changes in PM2 5 and PM2 5 precursors in 2028, 2030, and
2035.
°EPA is unable to provide confidence intervals for 2 percent-based estimates currently.
d Sum of ozone mortality estimated using the pooled short-term ozone exposure risk estimate and the Wu et al.
(2020) long-term PM2 5 exposure mortality risk estimate.
e Sum of the Turner et al. (2016) long-term ozone exposure risk estimate and the Pope et al. (2019) long-term PM2 5
exposure mortality risk estimate.
4-34
-------
Table 4-5 Stream of Estimated Human Health Benefits from 2028 through 2037:
Monetized Benefits Quantified as Sum of Long-Term Ozone Mortality and Long-Term
PM2.5 Mortality (discounted at 2 percent to 2023; millions of 2019 dollars)"
Year
Under the
Final Rule
2028b
$38
and
$79
2029
$38
and
$79
2030b
$13
and
$27
2031
$14
and
$27
2032
$7.4
and
$14
2033
2034
$7.5
$7J
and
and
$14
$14
203 5b
$7.6
and
$14
2036
$7.6
and
$14
2037
$7.6
and
$14
PV
EAV
$150
$17
and
and
$300
$33
a Benefits for all other years were extrapolated from years with model-based air quality estimates. Benefits
calculated as value of avoided: PM2 5-attributable deaths quantified using a concentration-response relationship from
Wu et al. (2020) and Pope et al. (2019); Ozone-attributable deaths quantified using a concentration-response
relationship from the Turner et al. (2017); and PM2 5 and ozone-related morbidity effects. The two benefits estimates
are separated by the word "and" to signify that they are two separate estimates. The estimates do not represent
lower- and upper-bound estimates and should not be summed.
b Analysis year in which air quality models were run.
4-35
-------
Table 4-6 Stream of Estimated Human Health Benefits from 2028 through 2037:
Monetized Benefits Quantified as Sum of Long-Term Ozone Mortality and Long-Term
PM2.5 Mortality (discounted at 3 percent to 2023; millions of 2019 dollars)"
Year
Under the
Final Rule
2028b
$34
and
$71
2029
$33
and
$71
2030b
$12
and
$24
2031
$12
and
$24
2032
$6.6
and
$13
2033
$6.6
and
$13
2034
$6.5
and
$12
2035b
$6.5
and
$12
2036
$6.5
and
$12
2037
$6.4
and
$12
PV
$130
and
$260
EAV
$15
and
$31
a Benefits for all other years were extrapolated from years with model-based air quality estimates. Benefits
calculated as value of avoided: PM2 5-attributable deaths quantified using a concentration-response relationship from
Wu et al. (2020) and Pope et al. (2019); Ozone-attributable deaths quantified using a concentration-response
relationship from the Turner et al. (2017); and PM2 5 and ozone-related morbidity effects. The two benefits estimates
are separated by the word "and" to signify that they are two separate estimates. The estimates do not represent
lower- and upper-bound estimates and should not be summed.
b Analysis year in which air quality models were run.
4-36
-------
Table 4-7 Stream of Estimated Human Health Benefits from 2028 through 2037:
Monetized Benefits Quantified as Sum of Long-Term Ozone Mortality and Long-Term
PM2.5 Mortality (discounted at 7 percent to 2023; millions of 2019 dollars)"
Under the
Year
Final Rule
2028b
$25
and
$52
2029
$24
and
$50
2030b
$8.0
and
$16
2031
$7.7
and
$16
2032
$4.2
and
$8.0
2033
$4.0
and
$7.7
2034
$3.9
and
$7.3
203 5b
$3.7
and
$7.0
2036
$3.5
and
$6.7
2037
$3.4
and
$6.4
PV
$86
and
$180
EAV
$12
and
$25
a Benefits for all other years were extrapolated from years with model-based air quality estimates. Benefits
calculated as value of avoided: PM2 5-attributable deaths quantified using a concentration-response relationship from
Wu et al. (2020) and Pope et al. (2019); Ozone-attributable deaths quantified using a concentration-response
relationship from the Turner et al. (2017); and PM2 5 and ozone-related morbidity effects. The two benefits estimates
are separated by the word "and" to signify that they are two separate estimates. The estimates do not represent
lower- and upper-bound estimates and should not be summed.
b Analysis year in which air quality models were run.
This analysis uses several recent improvements in health endpoint valuation. School loss
days now account for lost human capital formation, as was discussed in the Health Benefits TSD
which was reviewed by the EPA Scientific Advisory Board's Review of BenMAP and Benefits
Methods. We include new estimates of the cost asthma onset and stroke beyond those described
in the Health Benefits TSD.
The new valuation estimate for school loss days is described in the Health Benefits TSD
in Section 5.3.8. We include two costs of school loss days: caregiver costs and loss of learning.
We calculate each separately and then sum. Caregiver costs are valued at their employers'
average cost for employed caregivers. For unemployed caregivers, the opportunity cost of their
time is calculated as the average take-home pay. The loss of learning is calculated based on the
impact of absences on learning multiplied by the impact of school learning on adult earnings.
The loss of learning estimate is currently only available for middle and high school students. The
two costs are summed.
4-37
-------
The caregiver costs assume that an adult caregiver stays home with the child and loses
any wage income they would have earned that day. For working caregivers, we follow EPA
guidance and value their time at the average wage including fringe benefits and overhead costs.
The average daily wage in 2021 was $195 (2015 dollars, assumed to be the average weekly wage
divided by 5),58 which yields an average daily labor cost of $340 for employed parents after
applying average multipliers of 1.46 for fringe benefits and 1.2 for overhead. For nonworking
caregivers, we assume that the opportunity cost of time is the average after-tax earnings. We
estimate the income tax rate for a median household to be 7 percent, yielding net earnings of
$195 multiplied by 0.93 or $181 (2015 dollars). The income tax rate of 7 percent is the
percentage difference in median post-tax income and median income from Tables A1 and CI in
Shrider et al. (2021).
The probability that a parent is working is measured with the employment population
ratio among people with their own children under 18 and is 77.2 percent.59 Combining the cost of
working and nonworking caregivers yields a caregiver cost of $305 per school loss day.
To measure the loss of learning, we update the Liu et al. (2021) estimate. Liu et al. (2021)
estimated the impact of a school absence on learnings as measured by an end-of-course test
score. We multiply by an estimate of the impact of learning as measured by end-of-course test
scores on adult income from Chetty et al. (2014). This approach yields an estimated learning loss
of $2,842 per school absence (discounted at 2 percent), $2,230 per school absence (discounted at
3 percent) and $975 per school absence (discounted at 7 percent).
We updated the Chetty et al. (2014) estimate to use 2010 income and to estimate lifetime
incomes discounted at 3 percent and 7 percent. Liu et al. (2021) estimate that a school absence
leads to a $1,200 reduction in lifetime earnings, based on the Chetty et al. (2014) estimate that
lifetime earnings are $522,000 (2010 dollars). We use 2010 ACS data from IPUMS to calculate
expected lifetime earnings of $1,137,732 (discounting at 2 percent), $892,579 (discounting at 3
percent) and $390,393 (discounting at 7 percent). We then multiply the Liu et al. (2021) estimate
of $1,200 by ($1,137,732 divided by $522,000) and ($892,579 divided by $522,000) and
58 U.S, Bureau of Labor Statistics (2022), series Employment, Hours, and Earnings from the Current Employment
Statistics (Series ID CES0500000011).
59 US Bureau of Labor Statistics Employment Characteristics of Families, 2021, Table 5.
4-38
-------
($390,393 divided by $522,000) and convert from 2010 dollars to 2015 dollars based on the
Consumer Price Index for All Urban Consumers.
We use caregiver costs for preschool and elementary school children and the sum of
caregiver costs and loss of learning for middle school and high school students. We calculate that
31 percent of children under 18 are middle school and high school ages 13-18, assuming each
bin is distributed equally, so the combined average effect is $1,186 ($305 plus $2,842 multiplied
by 0.31) with 2 percent discounting, $1,000 ($305 plus $2,230 multiplied by 0.31) with 3 percent
discounting, and $610 ($305 plus $975 multiplied by 0.31) with 7 percent discounting in 2015
dollars (U.S. Census Bureau, 2010).60
We include a new estimate of the cost of illness of asthma onset based on Maniloff and
Fann (2023). These estimates are $181,249 with a 2 percent discount rate, $146,370 with a 3
percent discount rate, and $76,629 with a 7 percent discount rate (2015 dollars). We also include
a new estimate of the cost of illness of stroke onset based on Maniloff and Fann (2023).These
estimates are $158,763 with a 2 percent discount rate, $150,675 with a 3 percent discount rate,
and $123,984 with a 7 percent discount rate (2015 dollars).
4.3.10 Additional Unquantified Benefits
Data, time, and resource limitations prevented EPA from quantifying the estimated health
impacts or monetizing estimated benefits associated with direct exposure to NO2 and SO2,
independent of the role NO2 and SO2 play as precursors to PM2.5 and ozone, ecosystem effects,
and visibility impairment due to the absence of air quality modeling data for these pollutants in
this analysis. While all health benefits and welfare benefits were not able to be quantified, it does
not imply that there are not additional benefits associated with reductions in exposures to ozone,
PM2.5, NO2 or SO2. Criteria pollutants from U.S. EGUs can also be transported downwind into
foreign countries, in particular Canada and Mexico. Therefore, reduced criteria pollutants from
U.S. EGUs can lead to public health and welfare benefits in foreign countries. EPA is currently
unable to quantify or monetize these effects.
The EPA is also unable to quantify and monetize the incremental potential benefits of
requiring facilities to utilize CEMS rather than continuing to allow the use of quarterly testing,
60 U.S. Census Bureau, Age and Sex Composition in the United States: 2010, Table 1,
https://www.census.gOv/data/tables/2010/demo/age-and-sex/2010-age-sex-composition.html.
4-39
-------
but the requirement has been considered qualitatively. The continuous monitoring of fPM
required in this rule is also likely to provide several additional benefits to the public which are
not quantified in this rule, including greater certainty, accuracy, transparency, and granularity in
fPM emissions information than exists today.
Table 4-8 Additional Unquantified Benefit Categories
Category
Effect
Effect
Quantified
Effect
Monetized
More
Information
Improved Human
Health
Asthma hospital admissions
—
—
NO2 ISA1
Chronic lung disease hospital admissions
—
—
NO2 ISA1
Reduced incidence of
morbidity from exposure
Respiratory emergency department visits
—
—
NO2 ISA1
Asthma exacerbation
—
—
NO2 ISA1
to NO2
Acute respiratory symptoms
—
—
NO2 ISA1
Premature mortality
—
—
NO2 ISA1-2-3
Other respiratory effects (e.g., airway
hyperresponsiveness and inflammation, lung
function, other ages, and populations)
—
—
NO2 ISA2-3
Improved Environment
Reduced visibility
impairment
Visibility in Class 1 areas
Visibility in residential areas
—
—
PM ISA1
PM ISA1
Reduced effects on
materials
Household soiling
Materials damage (e.g., corrosion, increased
wear)
—
—
PMISA1-2
PM ISA2
Reduced effects from
PM deposition (metals
and organics)
Effects on individual organisms and ecosystems
—
—
PMISA2
Visible foliar injury on vegetation
—
—
Ozone ISA1
Reduced vegetation growth and reproduction
—
—
Ozone ISA1
Reduced vegetation and
ecosystem effects from
exposure to ozone
Yield and quality of commercial forest products
and crops
Damage to urban ornamental plants
Carbon sequestration in terrestrial ecosystems
—
—
Ozone ISA1
Ozone ISA2
Ozone ISA1
Recreational demand associated with forest
aesthetics
—
—
Ozone ISA2
Other non-use effects
Ozone ISA2
4-40
-------
Table 4-8 Additional Unquantified Benefit Categories
Category
Effect
Effect
Quantified
Effect
Monetized
More
Information
Ecosystem functions (e.g., water cycling,
biogeochemical cycles, net primary productivity,
leaf-gas exchange, community composition)
—
—
Ozone ISA2
Recreational fishing
—
—
NOx SOx
ISA1
Tree mortality and decline
—
—
NOx SOx
ISA2
Reduced effects from
acid deposition
Commercial fishing and forestry effects
—
—
NOx SOx
ISA2
Recreational demand in terrestrial and aquatic
ecosystems
—
—
NOx SOx
ISA2
Other non-use effects
NOx SOx
ISA2
Ecosystem functions (e.g., biogeochemical
cycles)
—
—
NOx SOx
ISA2
Species composition and biodiversity in
terrestrial and estuarine ecosystems
—
—
NOx SOx
ISA2
Coastal eutrophication
—
—
NOx SOx
ISA2
Reduced effects from
nutrient enrichment from
deposition.
Recreational demand in terrestrial and estuarine
ecosystems
Other non-use effects
—
—
NOx SOx
ISA2
NOx SOx
ISA2
Ecosystem functions (e.g., biogeochemical
cycles, fire regulation)
—
—
NOx SOx
ISA2
Reduced vegetation
effects from ambient
exposure to SO2 and NOx
Injury to vegetation from SO2 exposure
—
—
NOx SOx
ISA2
Injury to vegetation from NOx exposure
—
—
NOx SOx
ISA2
1 We assess these benefits qualitatively due to data and resource limitations for this RIA.
2 We assess these benefits qualitatively because we do not have sufficient confidence in available data or methods.
3 We assess these benefits qualitatively because current evidence is only suggestive of causality or there are other significant
concerns over the strength of the association.
4.3.10.1 NO2 Health Benefits
In addition to being a precursor to PM2.5 and ozone, NOx emissions are also linked to a
variety of adverse health effects associated with direct exposure. We were unable to estimate the
health benefits associated with reduced NO2 exposure in this analysis. Following a
comprehensive review of health evidence from epidemiologic and laboratory studies, the ISA for
Oxides of Nitrogen —Health Criteria (NOx ISA) concluded that there is a likely causal
4-41
-------
relationship between respiratory health effects and short-term exposure to NO2 (U.S. EPA,
2016). 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. The NOx ISA also 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 NOx 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.
4.3.10.2 SO2 Health Benefits
In addition to being a precursor to PM2.5, SO2 emissions are also linked to a variety of
adverse health effects associated with direct exposure. We were unable to estimate the health
benefits associated with reduced SO2 in this analysis. Therefore, this analysis only quantifies and
monetizes the PM2.5 benefits associated with the reductions in SO2 emissions. Following an
extensive evaluation of health evidence from epidemiologic and laboratory studies, the ISA for
Oxides of Sulfur—Health Criteria (SO2 ISA) concluded that there is a causal relationship
between respiratory health effects and short-term exposure to SO2 sulfur (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 at concentrations between 20 and 100 parts per
billion (ppb), 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": 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. 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 mortality,
there was a lack of robustness of the observed associations to adjustment for other pollutants.
4-42
-------
4.3.10.3 Ozone Welfare Benefits
Exposure to ozone has been associated with a wide array of vegetation and ecosystem
effects in the published literature ecological (U.S. EPA, 2020c). Sensitivity to ozone is highly
variable across species, with over 65 plant species identified as "ozone-sensitive," many of
which occur in state and national parks and forests. These effects include those that damage or
impair the intended use of the plant or ecosystem. Such effects can include reduced growth
and/or biomass production in sensitive plant species, including forest trees, reduced yield and
quality of crops, visible foliar injury, species composition shift, and changes in ecosystems and
associated ecosystem services. See Section F of the Ozone Transport Policy Analysis Proposed
Rule TSD for a summary of an assessment of risk of ozone-related growth impacts on selected
forest tree species (U.S. EPA, 2022f).
4.3.10.4 NO2 and SO2 Welfare Benefits
As described in the IS As for Oxides of Nitrogen, Oxides of Sulfur and Particulate Matter
Ecological Criteria (U.S. EPA, 2020c), NOx and SO2 emissions also contribute to a variety of
adverse welfare effects, including those associated with acidic deposition, visibility impairment,
and nutrient enrichment. Deposition of nitrogen and sulfur causes acidification, which can cause
a loss of biodiversity of fishes, zooplankton, and macro invertebrates in aquatic ecosystems, as
well as a decline in sensitive tree species, such as red spruce (Picea rubens) and sugar maple
(Acer saccharum) in terrestrial ecosystems. In the northeastern U.S., the surface waters affected
by acidification are a source of food for some recreational and subsistence fishermen and for
other consumers and support several cultural services, including aesthetic and educational
services and recreational fishing. Biological effects of acidification in terrestrial ecosystems are
generally linked to aluminum toxicity, which can cause reduced root growth, restricting the
ability of the plant to take up water and nutrients. These direct effects can, in turn, increase the
sensitivity of these plants to stresses, such as droughts, cold temperatures, insect pests, and
disease, leading to increased mortality of canopy trees. Terrestrial acidification affects several
important ecological services, including declines in habitat for threatened and endangered
species (cultural), declines in forest aesthetics (cultural), declines in forest productivity
(provisioning), and increases in forest soil erosion and reductions in water retention (cultural and
regulating).
4-43
-------
Deposition of nitrogen is also associated with aquatic and terrestrial nutrient enrichment.
In estuarine waters, excess nutrient enrichment can lead to eutrophication. Eutrophication of
estuaries can disrupt an important source of food production, particularly fish and shellfish
production, and a variety of cultural ecosystem services, including water-based recreational and
aesthetic services. Terrestrial nutrient enrichment is associated with changes in the types and
number of species and biodiversity in terrestrial systems. Excessive nitrogen deposition upsets
the balance between native and nonnative plants, changing the ability of an area to support
biodiversity. When the composition of species changes, then fire frequency and intensity can
also change, as nonnative grasses fuel more frequent and more intense wildfires.
4.3.10.5 Visibility Impairment Benefits
Reducing secondary formation of 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. Visibility has direct significance to people's
enjoyment of daily activities and their overall sense of wellbeing. Good visibility 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, 2009b). Previous analyses such as U.S. EPA (2012)
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 associated with this rule would be likely to have a
significant impact on visibility in urban areas or Class I areas.
Reductions in emissions of NO2 will improve the level of visibility throughout the U.S.
because these gases (and the particles of nitrate and sulfate formed from these gases) impair
visibility by scattering and absorbing light (U.S. EPA, 2009b). Visibility is also referred to as
visual air quality (VAQ), and it directly affects people's enjoyment of a variety of daily activities
(U.S. EPA, 2009b). Good visibility increases quality of life where individuals live and work, and
where they travel for recreational activities, including sites of unique public value, such as the
Great Smoky Mountains National Park (U.S. EPA, 2009b).
4-44
-------
4.4 Climate Benefits
EPA estimates the climate benefits of CO2 emissions reductions expected from the final
rule using estimates of the social cost of carbon (SC-CO2) that reflect recent advances in the
scientific literature on climate change and its economic impacts and incorporate
recommendations made by the National Academies of Science, Engineering, and Medicine
(National Academies, 2017). EPA published and used these estimates in the RIA for the
December 2023 final oil and natural gas sector rulemaking, "Standards of Performance for New,
Reconstructed, and Modified Sources and Emissions Guidelines for Existing Sources: Oil and
Natural Gas Sector Climate Review" (US EPA 2023c). EPA solicited public comment on the
methodology and use of these estimates in the RIA for the Agency's December 2022 oil and
natural gas sector supplemental proposal and has conducted an external peer review of these
estimates, as described further below.61
The SC-CO2 is the monetary value of the net harm to society associated with a marginal
increase in CO2 emissions in a given year, or the net benefit of avoiding that increase. In
principle, SC-CO2 includes the value of all climate change impacts (both negative and positive),
including (but not limited to) changes in net agricultural productivity, human health effects,
property damage from increased flood risk and natural disasters, disruption of energy systems,
risk of conflict, environmental migration, and the value of ecosystem services. The SC-CO2,
therefore, reflects the societal value of reducing emissions of CO2 by one metric ton and is the
theoretically appropriate value to use in conducting benefit-cost analyses of policies that affect
CO2 emissions. In practice, data and modeling limitations restrain the ability of SC-CO2
estimates to include all physical, ecological, and economic impacts of climate change, implicitly
assigning a value of zero to the omitted climate damages. The estimates are, therefore, a partial
accounting of climate change impacts and likely underestimate the marginal benefits of
abatement.
Since 2008, EPA has used estimates of the social cost of various GHGs (i.e., SC-CO2,
SC-CH4, and SC-N2O), collectively referred to as the "social cost of greenhouse gases" (SC-
GHG), in analyses of actions that affect GHG emissions. The values used by EPA from 2009 to
61 See https://www.epa.gov/environmental-economics/scghg for a copy of the final report and other related materials.
4-45
-------
2016, and since 2021 — including in the proposal for this rulemaking — have been consistent
with those developed and recommended by the Interagency Working Group (IWG) on the SC-
GHG; and the values used from 2017 to 2020 were consistent with those required by E.O. 13783,
which disbanded the IWG. During 2015-2017, the National Academies conducted a
comprehensive review of the SC-CO2 and issued a final report in 2017 recommending specific
criteria for future updates to the SC-CO2 estimates, a modeling framework to satisfy the
specified criteria, and both near-term updates and longer-term research needs pertaining to
various components of the estimation process. The IWG was reconstituted in 2021 and E.O.
13990 directed it to develop a comprehensive update of its SC-GHG estimates, recommendations
regarding areas of decision-making to which SC-GHG should be applied, and a standardized
review and updating process to ensure that the recommended estimates continue to be based on
the best available economics and science going forward.
EPA is a member of the IWG and is participating in the IWG's work under E.O. 13990.
As noted in previous EPA RIAs, while that process continues, EPA is continuously reviewing
developments in the scientific literature on the SC-GHG, including more robust methodologies
for estimating damages from emissions, and looking for opportunities to further improve SC-
GHG estimation.62 In the December 2022 oil and natural gas sector supplemental proposal RIA,
the Agency included a sensitivity analysis of the climate benefits of the supplemental proposal
using a new set of SC-GHG estimates that incorporates recent research addressing
recommendations of the National Academies (National Academies, 2017) in addition to using
the interim SC-GHG estimates presented in the Technical Support Document: Social Cost of
Carbon, Methane, and Nitrous Oxide Interim Estimates under Executive Order 13990 (IWG,
2021) that the IWG recommended for use until updated estimates that address the National
Academies' recommendations are available.
EPA solicited public comment on the sensitivity analysis and the accompanying draft
technical report, External Review Draft of Report on the Social Cost of Greenhouse Gases:
Estimates Incorporating Recent Scientific Advances, which explains the methodology underlying
62 EPA strives to base its analyses on the best available science and economics, consistent with its responsibilities,
for example, under the Information Quality Act.
4-46
-------
the new set of estimates, in the December 2022 oil and natural gas supplemental proposal RIA.
The response to comments document can be found in the docket for that action.63
To ensure that the methodological updates adopted in the technical report are consistent
with economic theory and reflect the latest science, EPA also initiated an external peer review
panel to conduct a high-quality review of the technical report, completed in May 2023. The peer
reviewers commended the agency on its development of the draft update, calling it a much-
needed improvement in estimating the SC-GHG and a significant step toward addressing the
National Academies' recommendations with defensible modeling choices based on current
science. The peer reviewers provided numerous recommendations for refining the presentation
and for future modeling improvements, especially with respect to climate change impacts and
associated damages that are not currently included in the analysis. Additional discussion of
omitted impacts and other updates have been incorporated in the technical report to address peer
reviewer recommendations. Complete information about the external peer review, including the
peer reviewer selection process, the final report with individual recommendations from peer
reviewers, and EPA's response to each recommendation is available on EPA's website.64
The remainder of this section provides an overview of the methodological updates
incorporated into the SC-GHG estimates used in this final RIA. A more detailed explanation of
each input and the modeling process is provided in the final technical report, Report on the
Social Cost of Greenhouse Gases: Estimates Incorporating Recent Scientific Advances.65
Appendix B presents the projected benefits of the final rule using the interim SC-GHG (IWG,
2021) estimates used in the proposal RIA for comparison purposes.
The steps necessary to estimate the SC-GHG with a climate change integrated assessment
model (IAM) can generally be grouped into four modules: socioeconomics and emissions,
climate, damages, and discounting. The emissions trajectories from the socioeconomic module
are used to project future temperatures in the climate module. The damage module then
translates the temperature and other climate endpoints (along with the projections of
socioeconomic variables) into physical impacts and associated monetized economic damages,
where the damages are calculated as the amount of money the individuals experiencing the
63 https://www.regulations.gov/docket/EPA-HQ-OAR-2021-0317.
64 https://www. epa.gov/environmental-economics/scghg-tsd-peer-review.
65 See https://www.epa.gov/environmental-economics/scghg for a copy of the final report and other related materials.
4-47
-------
climate change impacts would be willing to pay to avoid them. To calculate the marginal effect
of emissions, i.e., the SC-GHG in year "f," the entire model is run twice - first as a baseline and
second with an additional pulse of emissions in year "t" After recalculating the temperature
effects and damages expected in all years beyond "f resulting from the adjusted path of
emissions, the losses are discounted to a present value in the discounting module. Many sources
of uncertainty in the estimation process are incorporated using Monte Carlo techniques by taking
draws from probability distributions that reflect the uncertainty in parameters.
The SC-GHG estimates used by EPA and many other federal agencies since 2009 have
relied on an ensemble of three widely used IAMs: Dynamic Integrated Climate and Economy
(DICE) (Nordhaus, 2010); Climate Framework for Uncertainty, Negotiation, and Distribution
(FUND) (Anthoff and Tol, 2013a, 2013b); and Policy Analysis of the Greenhouse Gas Effect
(PAGE) (Hope, 2013). In 2010, the IWG harmonized key inputs across the IAMs, but all other
model features were left unchanged, relying on the model developers' best estimates and
judgments. That is, the representation of climate dynamics and damage functions included in the
default version of each IAM as used in the published literature was retained.
The SC-GHG estimates in this RIA no longer rely on the three IAMs (i.e., DICE, FUND,
and PAGE) used in previous SC-GHG estimates. As explained previously, EPA uses a modular
approach to estimate the SC-GHG, consistent with the National Academies' near-term
recommendations. That is, the methodology underlying each component, or module, of the SC-
GHG estimation process is developed by drawing on the latest research and expertise from the
scientific disciplines relevant to that component. Under this approach, each step in the SC-GHG
estimation improves consistency with the current state of scientific knowledge, enhances
transparency, and allows for more explicit representation of uncertainty.
The socioeconomic and emissions module relies on a new set of probabilistic projections
for population, income, and GHG emissions developed under the Resources for the Future (RFF)
Social Cost of Carbon Initiative (Rennert, Prest, et al., 2022). These socioeconomic projections
(hereinafter collectively referred to as the RFF-SPs) are an internally consistent set of
probabilistic projections of population, GDP, and GHG emissions (CO2, CH4, and N2O) to 2300.
Based on a review of available sources of long-run projections necessary for damage
calculations, the RFF-SPs stand out as being most consistent with the National Academies'
4-48
-------
recommendations. Consistent with the National Academies' recommendation, the RFF-SPs were
developed using a mix of statistical and expert elicitation techniques to capture uncertainty in a
single probabilistic approach, taking into account the likelihood of future emissions mitigation
policies and technological developments, and provide the level of disaggregation necessary for
damage calculations. Unlike other sources of projections, they provide inputs for estimation out
to 2300 without further extrapolation assumptions. Conditional on the modeling conducted for
the SC-GHG estimates, this time horizon is far enough in the future to capture the majority of
discounted climate damages. Including damages beyond 2300 would increase the estimates of
the SC-GHG. As discussed in U.S. EPA (2023c), the use of the RFF-SPs allows for capturing
economic growth uncertainty within the discounting module.
The climate module relies on the Finite Amplitude Impulse Response (FaIR) model
(IPCC, 2021b; Millar et al., 2017; C. J. Smith et al., 2018), a widely used Earth system model
which captures the relationships between GHG emissions, atmospheric GHG concentrations, and
global mean surface temperature. The FaIR model was originally developed by Richard Millar,
Zeb Nicholls, and Myles Allen at Oxford University, as a modification of the approach used in
IPCC AR5 to assess the GWP and GTP (Global Temperature Potential) of different gases. It is
open source, widely used (e.g., IPCC (2018, 2021a)) and was highlighted by the National
Academies (2017) as a model that satisfies their recommendations for a near-term update of the
climate module in SC-GHG estimation. Specifically, it translates GHG emissions into mean
surface temperature response and represents the current understanding of the climate and GHG
cycle systems and associated uncertainties within a probabilistic framework. The SC-GHG
estimates used in this RIA rely on FaIR version 1.6.2 as used by the IPCC (2021a). It provides,
with high confidence, an accurate representation of the latest scientific consensus on the
relationship between global emissions and global mean surface temperature and offers a code
base that is fully transparent and available online. The uncertainty capabilities in FaIR 1.6.2 have
been calibrated to the most recent assessment of the IPCC (which importantly narrowed the
range of likely climate sensitivities relative to prior assessments). See U.S. EPA (2023c) for
more details.
4-49
-------
The socioeconomic projections and outputs of the climate module are inputs into the
damage module to estimate monetized future damages from climate change.66 The National
Academies' recommendations for the damage module, scientific literature on climate damages,
updates to models that have been developed since 2010, as well as the public comments received
on individual EPA rulemakings and the IWG's February 2021 TSD, have all helped to identify
available sources of improved damage functions. The IWG (e.g., IWG 2010, 2016a, 2021), the
National Academies (2017), comprehensive studies (e.g., Rose et al. (2014)), and public
comments have all recognized that the damages functions underlying the IWG SC-GHG
estimates used since 2013 (taken from DICE 2010 (Nordhaus, 2010); FUND 3.8 (Anthoff and
Tol, 2013a, 2013b); and PAGE 2009 (Hope, 2013)) do not include all the important physical,
ecological, and economic impacts of climate change. The climate change literature and the
science underlying the economic damage functions have evolved, and DICE 2010, FUND 3.8,
and PAGE 2009 now lag behind the most recent research. The IWG (e.g., IWG (2010, 2016a,
2021)), the National Academies (2017), comprehensive studies (e.g., Rose et al. (2014)), and
public comments have all recognized that the damages functions underlying the IWG SC-GHG
estimates used since 2013 (taken from DICE 2010 (Nordhaus, 2010); FUND 3.8 (Anthoff and
Tol, 2013a, 2013b); and PAGE 2009 (Hope, 2013)) do not include all of the important physical,
ecological, and economic impacts of climate change. The climate change literature and the
science underlying the economic damage functions have evolved, and DICE 2010, FUND 3.8,
and PAGE 2009 now lag behind the most recent research.
The challenges involved with updating damage functions have been widely recognized.
Functional forms and calibrations are constrained by the available literature and need to
extrapolate beyond warming levels or locations studied in that literature. Research and public
resources focused on understanding how these physical changes translate into economic impacts
have been significantly less than the resources focused on modeling and improving our
understanding of climate system dynamics and the physical impacts from climate change
66 In addition to temperature change, two of the three damage modules used in the SC-GHG estimation require
global mean sea level (GMSL) projections as an input to estimate coastal damages. Those two damage modules use
different models for generating estimates of GMSL. Both are based off reduced complexity models that can use the
FaIR temperature outputs as inputs to the model and generate projections of GMSL accounting for the contributions
of thermal expansion and glacial and ice sheet melting based on recent scientific research. Absent clear evidence on
a preferred model, the SC-GHG estimates presented in this RIA retain both methods used by the damage module
developers. See U.S. EPA (2023c) for more details.
4-50
-------
(Auffhammer, 2018). Even so, there has been a large increase in research on climate impacts and
damages in the time since DICE 2010, FUND 3.8, and PAGE 2009 were published. Along with
this growth, there continues to be wide variation in methodologies and scope of studies, such that
care is required when synthesizing the current understanding of impacts or damages. Based on a
review of available studies and approaches to damage function estimation, EPA uses three
separate damage functions to form the damage module: (1) a subnational-scale, sectoral damage
function (based on the Data-driven Spatial Climate Impact Model (DSCIM) developed by the
Climate Impact Lab (Carleton et al., 2022; Climate Impact Lab (CIL), 2023; Rode et al., 2021);
(2) a country-scale, sectoral damage function (based on the Greenhouse Gas Impact Value
Estimator (GIVE) model developed under RFF's Social Cost of Carbon Initiative (Rennert,
Errickson, et al., 2022); and (3) a meta-analysis-based damage function (based on Howard and
Sterner (2017)).
The damage functions in DSCIM and GIVE represent substantial improvements relative
to the damage functions underlying the SC-GHG estimates used by EPA to date and reflect the
forefront of scientific understanding about how temperature change and SLR lead to monetized
net (market and nonmarket) damages for several categories of climate impacts. The models'
spatially explicit and impact-specific modeling of relevant processes allow for improved
understanding and transparency about mechanisms through which climate impacts are occurring
and how each damage component contributes to the overall results, consistent with the National
Academies' recommendations. DSCIM addresses common criticisms related to the damage
functions underlying current SC-GHG estimates (e.g., Pindyck (2017)) by developing multi-
sector, empirically grounded damage functions. The damage functions in the GIVE model offer a
direct implementation of the National Academies' near-term recommendation to develop
updated sectoral damage functions that are based on recently published work and reflective of
the current state of knowledge about damages in each sector. Specifically, the National
Academies noted that "[t]he literature on agriculture, mortality, coastal damages, and energy
demand provide immediate opportunities to update the [models]" (p. 199 in National Academies
(2017)), which are the four damage categories currently in GIVE. A limitation of both models is
that the sectoral coverage is still limited, and even the categories that are represented are
incomplete. Neither DSCIM nor GIVE yet accommodate estimation of several categories of
temperature driven climate impacts (e.g., morbidity, conflict, migration, biodiversity loss) and
4-51
-------
only represent a limited subset of damages from changes in precipitation. For example, while
precipitation is considered in the agriculture sectors in both DSCIM and GIVE, neither model
takes into account impacts of flooding, changes in rainfall from tropical storms, and other
precipitation related impacts. As another example, the coastal damage estimates in both models
do not fully reflect the consequences of SLR-driven salt-water intrusion and erosion, or SLR
damages to coastal tourism and recreation. Other missing elements are damages that result from
other physical impacts (e.g., ocean acidification, non-temperature-related mortality such as
diarrheal disease and malaria) and the many feedbacks and interactions across sectors and
regions that can lead to additional damages.67 See U.S. EPA (2023c) for more discussion of
omitted damage categories and other modeling limitations. DSCIM and GIVE do account for the
most commonly cited benefits associated with CO2 emissions and climate change - CO2 crop
fertilization and declines in cold related mortality. As such, while the GIVE- and DSCIM-based
results provide state-of-the-science assessments of key climate change impacts, they remain
partial estimates of future climate damages resulting from incremental changes in CO2, CH4, and
N2O.68
Finally, given the still relatively narrow sectoral scope of the recently developed DSCIM
and GIVE models, the damage module includes a third damage function that reflects a synthesis
of the state of knowledge in other published climate damages literature. Studies that employ
meta-analytic techniques69 offer a tractable and straightforward way to combine the results of
multiple studies into a single damage function that represents the body of evidence on climate
damages that pre-date CIL and RFF's research initiatives. The first use of meta-analysis to
combine multiple climate damage studies was done by Tol (2009) and included 14 studies. The
studies in Tol (2009) served as the basis for the global damage function in DICE starting in
version 2013R (Nordhaus, 2014). The damage function in the most recent published version of
67 The one exception is that the agricultural damage function in DSCIM and GIVE reflects the ways that trade can
help mitigate damages arising from crop yield impacts.
68 One advantage of the modular approach used by these models is that future research on new or alternative damage
functions can be incorporated in a relatively straightforward way. DSCIM and GIVE developers have work
underway on other impact categories that may be ready for consideration in future updates (e.g., morbidity and
biodiversity loss).
69 Meta-analysis is a statistical method of pooling data and/or results from a set of comparable studies of a problem.
Pooling in this way provides a larger sample size for evaluation and allows for a stronger conclusion than can be
provided by any single study. Meta-analysis yields a quantitative summary of the combined results and current state
of the literature.
4-52
-------
DICE, DICE 2016, is from an updated meta-analysis based on a rereview of existing damage
studies and included 26 studies published over 1994-2013 (Nordhaus and Moffat, 2017). Howard
and Sterner (2017) provide a more recent published peer-reviewed meta-analysis of existing
damage studies (published through 2016) and account for additional features of the underlying
studies. This study addresses differences in measurement across studies by adjusting estimates
such that the data are relative to the same base period. They also eliminate double counting by
removing duplicative estimates. Howard and Sterner's final sample is drawn from 20 studies that
were published through 2015. Howard and Sterner (2017) present results under several
specifications and show that the estimates are somewhat sensitive to defensible alternative
modeling choices. As discussed in detail in U.S. EPA (2023c), the damage module underlying
the SC-GHG estimates in this RIA includes the damage function specification (that excludes
duplicate studies) from Howard and Sterner (2017) that leads to the lowest SC-GHG estimates,
all else equal.
The discounting module discounts the stream of future net climate damages to its present
value in the year when the additional unit of emissions was released. Given the long-time
horizon over which the damages are expected to occur, the discount rate has a large influence on
the present value of future damages. Consistent with the findings of National Academies (2017),
the economic literature, OMB Circular A-4's guidance for regulatory analysis, and IWG
recommendations to date (IWG, 2010, 2013, 2016a, 2016b, 2021), EPA continues to conclude
that the consumption rate of interest is the theoretically appropriate discount rate to discount the
future benefits of reducing GHG emissions and that discount rate uncertainty should be
accounted for in selecting future discount rates in this intergenerational context. OMB's Circular
A-4 points out that "the analytically preferred method of handling temporal differences between
benefits and costs is to adjust all the benefits and costs to reflect their value in equivalent units of
consumption and to discount them at the rate consumers and savers would normally use in
discounting future consumption benefits" (OMB, 2003).70 The damage module described above
calculates future net damages in terms of reduced consumption (or monetary consumption
equivalents), and so an application of this guidance is to use the consumption discount rate to
70 Similarly, OMB's Circular A-4 (2023) points out that "The analytically preferred method of handling temporal
differences between benefits and costs is to adjust all the benefits and costs to reflect their value in equivalent units
of consumption before discounting them."
4-53
-------
calculate the SC-GHG. Thus, EPA concludes that the use of the social rate of return on capital (7
percent under the 2003 OMB Circular A-4 guidance), which does not reflect the consumption
rate, to discount damages estimated in terms of reduced consumption would inappropriately
underestimate the impacts of climate change for the purposes of estimating the SC-GHG.71
For the SC-GHG estimates used in this RIA, EPA relies on a dynamic discounting
approach that more fully captures the role of uncertainty in the discount rate in a manner
consistent with the other modules. Based on a review of the literature and data on consumption
discount rates, the public comments received on individual EPA rulemakings, and the February
2021 TSD, and the National Academies (2017) recommendations for updating the discounting
module, the SC-GHG estimates rely on discount rates that reflect more recent data on the
consumption interest rate and uncertainty in future rates. Specifically, rather than using a
constant discount rate, the evolution of the discount rate over time is defined following the latest
empirical evidence on interest rate uncertainty and using a framework originally developed by
Ramsey (1928) that connects economic growth and interest rates. The Ramsey approach
explicitly reflects (1) preferences for utility in one period relative to utility in a later period and
(2) the value of additional consumption as income changes. The dynamic discount rates used to
develop the SC-GHG estimates applied in this RIA have been calibrated following the Newell et
al. (2022) approach, as applied in Rennert, Errickson, et al. (2022); Rennert, Prest, et al. (2022).
This approach uses the Ramsey (1928) discounting formula in which the parameters are
calibrated such that (1) the decline in the certainty-equivalent discount rate matches the latest
empirical evidence on interest rate uncertainty estimated by Bauer and Rudebusch (2020, 2023)
and (2) the average of the certainty-equivalent discount rate over the first decade matches a near-
term consumption rate of interest. Uncertainty in the starting rate is addressed by using three
near-term target rates (1.5, 2.0, and 2.5 percent) based on multiple lines of evidence on observed
market interest rates.
The resulting dynamic discount rate provides a notable improvement over the constant
discount rate framework used for SC-GHG estimation in previous EPA RIAs. Specifically, it
provides internal consistency within the modeling and a more complete accounting of
71 See also the discussion of the inappropriateness of discounting consumption-equivalent measures of benefits and
costs using a rate of return on capital in Circular A-4 (OMB, 2003).
4-54
-------
uncertainty consistent with economic theory (Arrow et al., 2013; Cropper et al., 2014) and the
National Academies' (2017) recommendation to employ a more structural, Ramsey-like
approach to discounting that explicitly recognizes the relationship between economic growth and
discounting uncertainty. This approach is also consistent with the National Academies (2017)
recommendation to use three sets of Ramsey parameters that reflect a range of near-term
certainty-equivalent discount rates and are consistent with theory and empirical evidence on
consumption rate uncertainty. Finally, the value of aversion to risk associated with net damages
from GHG emissions is explicitly incorporated into the modeling framework following the
economic literature. See U.S. EPA (2023c) for a more detailed discussion of the entire
discounting module and methodology used to value risk aversion in the SC-GHG estimates.
Taken together, the methodologies adopted in this SC-GHG estimation process allow for
a more holistic treatment of uncertainty than past estimates used by EPA. The updates
incorporate a quantitative consideration of uncertainty into all modules and use a Monte Carlo
approach that captures the compounding uncertainties across modules. The estimation process
generates nine separate distributions of discounted marginal damages per metric ton - the
product of using three damage modules and three near-term target discount rates - for each gas
in each emissions year. These distributions have long right tails reflecting the extensive evidence
in the scientific and economic literature that shows the potential for lower-probability but higher-
impact outcomes from climate change, which would be particularly harmful to society. The
uncertainty grows over the modeled time horizon. Therefore, under cases with a lower near-term
target discount rate - that give relatively more weight to impacts in the future - the distribution
of results is wider. To produce a range of estimates that reflects the uncertainty in the estimation
exercise while also providing a manageable number of estimates for policy analysis, EPA
combines the multiple lines of evidence on damage modules by averaging the results across the
three damage module specifications. The full results generated from the updated methodology
for methane and other GHGs (SC-CO2, SC-CH4, and SC-N2O) for emissions years 2020 through
2080 are provided in U.S. EPA (2023c).
Table 4-9 summarizes the resulting averaged certainty-equivalent SC-CO2 estimates
under each near-term discount rate that are used to estimate the climate benefits of the CO2
emission reductions expected from the final rule. These estimates are reported in 2019 dollars
but are otherwise identical to those presented in U.S. EPA (2023c). The SC-CO2 increase over
4-55
-------
time within the models — i.e., the societal harm from one metric ton emitted in 2030 is higher
than the harm caused by one metric ton emitted in 2027 — because future emissions produce
larger incremental damages as physical and economic systems become more stressed in response
to greater climatic change, and because GDP is growing over time and many damage categories
are modeled as proportional to GDP.
Table 4-9 Estimates of the Social Cost of CO2 Values, 2028-2037 (2019 dollars per
Metric Tonne CO2) a
Near-term Ramsey Discount Rate
Emission Year
2.5%
2%
1.5%
2028
140
220
370
2029
140
220
380
2030
140
230
380
2031
150
230
380
2032
150
230
390
2033
150
240
390
2034
150
240
400
2035
160
240
400
2036
160
250
410
2037
160
250
410
a Source: U.S. EPA (2023c). Note: These SC-CO2 values are identical to those reported in the technical report U.S
EPA (2023c) adjusted for inflation to 2019 dollars using the annual GDP Implicit Price Deflator values in the U.S.
Bureau of Economic Analysis' (BEA) NIPA Table 1.1.9 (U.S. BEA, 2021). The values are stated in $/metric ton
CO2 and vary depending on the year of CO2 emissions. This table displays the values rounded to two significant
figures. The annual unrounded values used in the calculations in this RIA are available in Appendix A.4 of U.S.
EPA (2023c) and at: www.epa.gov/environmental-economics/scghg.
The methodological updates described above represent a major step forward in bringing
SC-GHG estimation closer to the frontier of climate science and economics and address many of
the National Academies' (2017) near-term recommendations. Nevertheless, the resulting SC-
CO2 estimates presented in Table 4-9, still have several limitations, as would be expected for any
modeling exercise that covers such a broad scope of scientific and economic issues across a
complex global landscape. There are still many categories of climate impacts and associated
damages that are only partially or not reflected yet in these estimates and sources of uncertainty
that have not been fully characterized due to data and modeling limitations. For example, the
modeling omits most of the consequences of changes in precipitation, damages from extreme
weather events, the potential for nongradual damages from passing critical thresholds (e.g.,
tipping elements) in natural or socioeconomic systems, and non-climate mediated effects of
4-56
-------
GHG emissions. Importantly, the updated SC-GHG methodology does not yet reflect
interactions and feedback effects within, and across, Earth and human systems. For example, it
does not explicitly reflect potential interactions among damage categories, such as those
stemming from the interdependencies of energy, water, and land use. These, and other,
interactions and feedbacks were highlighted by the National Academies as an important area of
future research for longer-term enhancements in the SC-GHG estimation framework.
Table 4-10 presents the estimated annual, undiscounted climate benefits of the estimated
changes in CO2 emissions the final rule, using the SC-CO2 estimates presented in Table 4-9, for
the stream of years beginning in 2028 through 2037. Also shown are the present value (PV) of
monetized climate benefits discounted back to 2023 and equivalent annualized values (EAV)
associated with each of the three SC-CO2 values. To calculate the present and annualized values
of climate benefits in Table 4-10, EPA uses the same discount rate as the near-term target
Ramsey rate used to discount the climate benefits from future CO2 reductions.72 That is, future
climate benefits estimated with the SC-CO2 at the near-term 2.5 percent, 2 percent, and 1.5
percent Ramsey rate are discounted to the base year of the analysis using a constant 2.5, 2, and
1.5 percent rate, respectively. Note the less stringent regulatory alternative only has unquantified
benefits associated with the finalized requirements for PM CEMS. As a result, there are no
quantified benefits associated with this regulatory option.
72 As discussed in U.S. EPA (2023c), the error associated with using a constant discount rate rather than the
certainty-equivalent rate path to calculate the present value of a future stream of monetized climate benefits is small
for analyses with moderate time frames (e.g., 30 years or less). EPA (2023c) also provides an illustration of the
amount that climate benefits from reductions in future emissions will be underestimated by using a constant discount
rate relative to the more complicated certainty-equivalent rate path.
4-57
-------
Table 4-10 Stream of Projected Climate Benefits under the Final Rule from 2028
through 2037 (discounted to 2023, millions of 2019 dollars)"
Near-term Ramsey Discount Rate
Emission Year
2.5%
2%
1.5%
2028b
7.9
13
22
2029
7.9
13
22
2030b
-4.3
-7.1
-12
2031
-4.3
-7.1
-12
2032
12
19
34
2033
12
19
33
2034
12
19
33
203 5b
11
19
33
2036
11
19
33
2037
11
19
33
PV and EAV
PV
76
130
220
EAV
8.7
14
24
a Climate benefits are based on changes (reductions) in C02 emissions and are calculated using updated estimates of
the SC-C02 from U.S. EPA (2023c).
b IPM run years.
Unlike many environmental problems where the causes and impacts are distributed more
locally, GHG emissions are a global externality making climate change a true global challenge.
GHG emissions contribute to damages around the world regardless of where they are emitted.
Because of the distinctive global nature of climate change, in the RIA for this final rule EPA
centers attention on a global measure of climate benefits from GHG reductions.
Consistent with all IWG recommended SC-GHG estimates to date, the SC-GHG values
presented in Table 4-9 provide a global measure of monetized damages from CO2, and Table
4-10and Table 4-1 lpresent the monetized global climate benefits of the CO2 emission reductions
expected from the final rule. This approach is the same as that taken in EPA regulatory analyses
from 2009 through 2016 and since 2021. It is also consistent with guidance in OMB Circular A-4
(OMB 2003, 2023) that recommends reporting of important international effects.73 EPA also
73 The 2003 version of OMB Circular A-4 states when a regulation is likely to have international effects, "these
effects should be reported"; while OMB recommends that international effects be reported separately, the guidance
also explains that "[d]ifferent regulations may call for different emphases in the analysis, depending on the nature
and complexity of the regulatory issues." (OMB 2003). The 2023 update to Circular A-4 states that "In certain
contexts, it may be particularly appropriate to include effects experienced by noncitizens residing abroad in your
primary analysis. Such contexts include, for example, when:
4-58
-------
notes that EPA's cost estimates in RIAs, including the cost estimates contained in this RIA,
regularly do not differentiate between the share of compliance costs expected to accrue to U.S.
firms versus foreign interests, such as to foreign investors in regulated entities.74 A global
perspective on climate effects is therefore consistent with the approach EPA takes on costs.
There are many reasons, as summarized in this section —and as articulated by OMB and in IWG
assessments (IWG, 2010, 2013, 2016a, 2016b, 2021), the 2015 Response to Comments (IWG,
2015), in detail in U.S. EPA (2023c), in Appendix A of the Response to Comments document for
the December 2023 final oil and natural gas sector rulemaking — why EPA focuses on the
global value of climate change impacts when analyzing policies that affect GHG emissions.
International cooperation and reciprocity are essential to successfully addressing climate
change, as the global nature of GHGs means that a ton of GHGs emitted in any other country
harms those in the U.S. just as much as a ton emitted within the territorial U.S. Assessing the
benefits of U.S. GHG mitigation activities requires consideration of how those actions may
affect mitigation activities by other countries, as those international mitigation actions will
provide a benefit to U.S. citizens and residents by mitigating climate impacts that affect U.S.
citizens and residents. This is a classic public goods problem because each country's reductions
benefit everyone else, and no country can be excluded from enjoying the benefits of other
countries' reductions. The only way to achieve an efficient allocation of resources for emissions
reduction on a global basis — and so benefit the U.S. and its citizens and residents — is for all
• assessing effects on noncitizens residing abroad provides a useful proxy for effects on U.S. citizens and residents
that are difficult to otherwise estimate;
• assessing effects on noncitizens residing abroad provides a useful proxy for effects on U.S. national interests that
are not otherwise fully captured by effects experienced by particular U.S. citizens and residents (e.g., national
security interests, diplomatic interests, etc.);
• regulating an externality on the basis of its global effects supports a cooperative international approach to the
regulation of the externality by potentially inducing other countries to follow suit or maintain existing efforts; or
• international or domestic legal obligations require or support a global calculation of regulatory effects" (OMB
2023. Due to the global nature of the climate change problem, the OMB recommendations of appropriate contexts
for considering international effects are relevant to the CO2 emission reductions expected from the final rule. For
example, as discussed in this RIA, a global focus in evaluating the climate impacts of changes in CO2 emissions
supports a cooperative international approach to GHG mitigation by potentially inducing other countries to follow
suit or maintain existing efforts, and the global SC-CO2 estimates better capture effects on U.S. citizens and
residents and U.S. national interests that are difficult to estimate and not otherwise fully captured.
74 For example, in the RIA for the 2018 Proposed Reconsideration of the Oil and Natural Gas Sector Emission
Standards for New, Reconstructed, and Modified Sources, EPA acknowledged that some portion of regulatory costs
will likely "accru[e] to entities outside U.S. borders" through foreign ownership, employment, or consumption (EPA
2018, p. 3-13). In general, a significant share of U.S. corporate debt and equities are foreign-owned, including in the
oil and gas industry.
4-59
-------
countries to base their policies on global estimates of damages. A wide range of scientific and
economic experts have emphasized the issue of international cooperation and reciprocity as
support for assessing global damages of GHG emission in domestic policy analysis. Using a
global estimate of damages in U.S. analyses of regulatory actions allows the U.S. to continue to
actively encourage other nations, including emerging major economies, to also assess global
climate damages of their policies and to take steps to reduce emissions. For example, many
countries and international institutions have already explicitly adapted the global SC-GHG
estimates used by EPA in their domestic analyses (e.g., Canada, Israel) or developed their own
estimates of global damages (e.g., Germany), and recently, there has been renewed interest by
other countries to update their estimates since the draft release of the updated SC-GHG estimates
presented in the December 2022 oil and natural gas sector supplemental proposal RIA.75 Several
recent studies have empirically examined the evidence on international GHG mitigation
reciprocity, through both policy diffusion and technology diffusion effects. See U.S. EPA
(2023c) for more discussion.
For all of these reasons, EPA believes that a global metric is appropriate for assessing the
climate benefits of avoided GHG emissions in this final RIA. In addition, as emphasized in the
National Academies (2017) recommendations, "[i]t is important to consider what constitutes a
domestic impact in the case of a global pollutant that could have international implications that
impact the United States." The global nature of GHG pollution and its impacts means that U.S.
interests are affected by climate change impacts through a multitude of pathways and these need
to be considered when evaluating the benefits of GHG mitigation to U.S. citizens and residents.
The increasing interconnectedness of global economy and populations means that impacts
occuring outside of U.S. borders can have significant impacts on U.S. interests. Examples of
affected interests include direct effects on U.S. citizens and assets located abroad, international
trade, and tourism, and spillover pathways such as economic and political destabilization and
global migration that can lead to adverse impacts on U.S. national security, public health, and
75 In April 2023, the government of Canada announced the publication of an interim update to their SC-GHG
guidance, recommending SC-GHG estimates identical to EPA's updated estimates presented in the December 2022
Supplemental Proposal RIA. The Canadian interim guidance will be used across all Canadian federal departments
and agencies, with the values expected to be finalized by the end of the year.
https://www.canada.ca/en/environment-climate-change/services/climate-change/science-research-data/social-cost-
ghg.html.
4-60
-------
humanitarian concerns. Those impacts point to the global nature of the climate change problem
and are better captured within global measures of the social cost of GHGs.
In the case of these global pollutants, for the reasons articulated in this section, the
assessment of global net damages of GHG emissions allows EPA to fully disclose and
contextualize the net climate benefits of CO2 emission reductions expected from this final rule.
EPA disagrees with public comments received on the December 2022 oil and natural gas sector
supplemental proposal that suggested that EPA can or should use a metric focused on benefits
resulting solely from changes in climate impacts occuring within U.S. borders. The global
models used in the SC-GHG modeling described above do not lend themselves to be
disaggregated in a way that could provide sufficiently robust information about the distribution
of the rule's climate benefits to citizens and residents of particular countries, or population
groups across the globe and within the U.S. Two of the models used to inform the damage
module, the GIVE and DSCIM models, have spatial resolution that allows for some geographic
disaggregation of future climate impacts across the world. This permits the calculation of a
partial GIVE and DSCIM-based SC-GHG measuring the damages from four or five climate
impact categories projected to physically occur within the U.S., respectively, subject to caveats.
As discussed at length in U.S. EPA (2023c), these damage modules are only a partial accounting
and do not capture all of the pathways through which climate change affects public health and
welfare. For example, this modeling omits most of the consequences of changes in precipitation,
damages from extreme weather events (e.g., wildfires), the potential for nongradual damages
from passing critical thresholds (e.g., tipping elements) in natural or socioeconomic systems, and
non-climate mediated effects of GHG emissions other than CO2 fertilization (e.g., tropospheric
ozone formation due to CH4 emissions). Thus, they only cover a subset of potential climate
change impacts. Furthermore, as discussed at length in U.S. EPA (EPA, 2023f), the damage
modules do not capture spillover or indirect effects whereby climate impacts in one country or
region can affect the welfare of residents in other countries or regions—such as how economic
and health conditions across countries will impact U.S. business, investments, and travel abroad.
Additional modeling efforts can and have shed further light on some omitted damage
categories. For example, the Framework for Evaluating Damages and Impacts (FrEDI) is an
open-source modeling framework developed by EPA to facilitate the characterization of net
annual climate change impacts in numerous impact categories within the contiguous U.S. and
4-61
-------
monetize the associated distribution of modeled damages (Sarofim et al., 2021; U.S. EPA,
2021b)).76 The additional impact categories included in FrEDI reflect the availability of U.S.-
specific data and research on climate change effects. As discussed in U.S. EPA (2023c), results
from FrEDI show that annual damages resulting from climate change impacts within the
contiguous U.S. (CONUS) (i.e., excluding Hawaii, Alaska, and U.S. territories) and for impact
categories not represented in GIVE and DSCIM are expected to be substantial. For example,
FrEDI estimates a partial SC-CO2 of $36/mtC02 for damages physically occurring within
CONUS for 2030 emissions (under a 2 percent near-term Ramsey discount rate), compared to a
GIVE and DSCIM-based U.S.-specific SC-CO2 of $16/mtC02 and $14/mtC02, respectively, for
2030 emissions (2019 dollars).
While the FrEDI results help to illustrate how monetized damages physically occurring
within CONUS increase as more impacts are reflected in the modeling framework, they are still
subject to many of the same limitations associated with the DSCIM and GIVE damage modules,
including the omission or partial modeling of important damage categories.77'78 Finally, none of
these modeling efforts-GIVE, DSCIM, and FrEDI-reflect non-climate mediated effects of GHG
emissions experienced by U.S. populations (other than CO2 fertilization effects on agriculture).
Taken together, applying the U.S.-specific partial SC-GHG estimates derived from the
multiple lines of evidence described above to the GHG emissions reduction expected under the
final rule would yield substantial benefits. For example, the present value of the climate benefits
of the final rule over the 2028 to 2037 period as measured by FrEDI from climate change
76 The FrEDI framework and Technical Documentation have been subject to a public review comment period and an
independent external peer review, following guidance in the EPA Peer-Review Handbook for Influential Scientific
Information (ISI). Information on the FrEDI peer-review is available at the EPA Science Inventory (EPA Science
Inventory, 2021).
77 Another method that has produced estimates of the effect of climate change on U.S.-specific outcomes uses a top-
down approach to estimate aggregate damage functions. Published research using this approach include total-
economy empirical studies that econometrically estimate the relationship between GDP and a climate variable,
usually temperature. As discussed in U.S. EPA (2023c), the modeling framework used in the existing published
studies using this approach differ in important ways from the inputs underlying the SC-GHG estimates described
above (e.g., discounting, risk aversion, and scenario uncertainty). Hence, we do not consider this line of evidence in
the analysis for this RIA. Updating the framework of total-economy empirical damage functions to be consistent
with the methods described in this RIA and U.S. EPA (2023c) would require new analysis. Finally, because total-
economy empirical studies estimate market impacts, they do not include any non-market impacts of climate change
(e.g., heat related mortality) and therefore are also only a partial estimate. EPA will continue to review
developments in the literature and explore ways to better inform the public of the full range of GHG impacts.
78 FrEDI estimates a partial SC-CO2 of $33/mtC02 for damages physically occurring within CONUS for 2030
emissions (under a 2 percent near-term Ramsey discount rate) (Hartin et al., 2023), compared to a GIVE and
DSCIM-based U.S.-specific SC-CO2 of $14/mtC02 and $12/mtC02, respectively, for 2030 emissions (2019 USD).
4-62
-------
impacts in CONUS are estimated to be $19 million under a 2 percent near-term Ramsey discount
rate.79 However, the numerous explicitly omitted damage categories and other modeling
limitations discussed above and throughout U.S. EPA (2023c) make it likely that these estimates
underestimate the benefits to U.S. citizens and residents of the GHG reductions from the final
rule; the limitations in developing a U.S.-specific estimate that accurately captures direct and
spillover effects on U.S. citizens and residents further demonstrates that it is more appropriate to
use a global measure of climate benefits from GHG reductions. EPA will continue to review
developments in the literature, including more robust methodologies for estimating the
magnitude of the various damages to U.S. populations from climate impacts and reciprocal
international mitigation activities, and explore ways to better inform the public of the full range
of GHG impacts.
4.5 Total Benefits
Table 4-11 presents the total health and climate benefits80 for the final rule. Note that
while we do not project emissions reductions under the less stringent option, we do expect there
to be benefits from the CEMS requirement. However, since we are unable to quantify these
benefits, for simplicity, we omit results for the less stringent option in this section.
79DCIM and GIVE use global damage functions. Damage functions based on only U.S.-data and research, but not
for other parts of the world, were not included in those models. FrEDI does make use of some of this U.S.-specific
data and research and as a result has a broader coverage of climate impact categories.
80 Monetized climate benefits are discounted using a 2 percent discount rate, consistent with EPA's updated
estimates of the SC-CO2. OMB has long recognized that climate effects should be discounted only at appropriate
consumption-based discount rates. Because the SC-CO2 estimates reflect net climate change damages in terms of
reduced consumption (or monetary consumption equivalents), the use of the social rate of return on capital (7
percent under OMB Circular A-4 (2003)) to discount damages estimated in terms of reduced consumption would
inappropriately underestimate the impacts of climate change for the purposes of estimating the SC-CO2. See Section
4 for more discussion.
4-63
-------
Table 4-11 Stream of Monetized Benefits under the Final Rule from 2028 through 2037
(discounted to 2023, millions of 2019 dollars)"
Values Calculated using 2%
Discount Rate
Values Calculated using 3%
Discount Rate
Values Calculated using 7%
Discount Rate
Year
Health
Benefitsb
Climate
Benefits0'"1
Total
Health
Benefits
Climate
Benefits
(discounted
at 2%)c'd
Total
Health
Benefits
Climate
Benefits
(discounted
at 2%)c'd
Total
2028
79
13
92
71
13
84
52
13
66
2029
79
13
92
71
13
84
50
13
63
2030
27
-7.1
20
24
-7.1
17
16
-7.1
9.1
2031
27
-7.1
20
24
-7.1
16
16
-7.1
8.4
2032
14
19
33
13
19
32
8.0
19
27
2033
14
19
34
13
19
32
7.7
19
27
2034
14
19
34
12
19
32
7.3
19
27
2035
14
19
33
12
19
31
7.0
19
26
2036
14
19
33
12
19
31
6.7
19
26
2037
14
19
33
12
19
31
6.4
19
25
PV
300
130
420
260
130
390
180
130
300
EAV
33
14
47
31
14
45
25
14
39
Non-Monetized Benefits®
Benefits from reductions of about 900 to 1000 pounds of Hg annually
Benefits from reductions about 4 to 7 tons of non-Hg HAP metals annually
Benefits from the increased transparency, compliance assurance, and accelerated identification of anomalous
emission anticipated from requiring PM CEMS
a Values have been rounded to two significant figures. Rows may not appear to add correctly due to rounding.
b Monetized air quality-related benefits include those related to public health associated with reductions in PM2 5 and
ozone concentrations. The estimated value of the air quality-related health benefits included here are the larger of
the two estimates presented in Table 4-5, Table 4-6, and Table 4-7.
0 Monetized climate benefits are based on reductions in CO2 emissions and are calculated using three different
estimates of the social cost of carbon dioxide (SC-CO2) (under 1.5 percent, 2 percent, and 2.5 percent near-term
Ramsey discount rates). For the presentational purposes of this table, we show the climate benefits associated with
the SC-CO2 at the 2 percent near-term Ramsey discount rate. Please see Table 4-10 for the full range of monetized
climate benefit estimates.
d The small increases and decreases in climate and health benefits and related EJ impacts result from very small
changes in fossil dispatch and coal use relative to the baseline. For context, the projected increase in CO2 emission
of less than 40,000 tons in 2030 is roughly one percent of the emissions of a mid-size coal plant operating at
availability (about 4 million tons).
e The list of non-monetized benefits does not include all potential non-monetized benefits. See Table 4-8 for a more
complete list.
4-64
-------
4.6 References
Anthoff, D., & Tol, R. S. J. (2013a). Erratum to: The uncertainty about the social cost of carbon:
A decomposition analysis using FUND (vol 117, pg 515, 2013). Climatic Change,
121(2), 413-413. doi:10.1007/s 10584-013-0959-1
Anthoff, D., & Tol, R. S. J. (2013b). The uncertainty about the social cost of carbon: A
decomposition analysis using FUND. Climatic Change, 117(3), 515-530.
doi: 10.1007/s 10584-013 -0706-7
Arrow, K., Cropper, M., Gollier, C., Groom, B., Heal, G., Newell, R., . . . Weitzman, M. (2013).
Determining Benefits and Costs for Future Generations. Science, 341(6144), 349-350.
doi:doi: 10.1126/science. 1235665
ATSDR. (2003). Toxicological Profile for Selenium. Retrieved from Atlanta, GA:
https://wwwn.cdc.gov/TSP/substances/ToxSubstance.aspx?toxid=28
ATSDR. (2005). Toxicological Profile for Nickel. Retrieved from Atlanta, GA:
https://wwwn.cdc.gov/TSP/substances/ToxSubstance.aspx?toxid=23
ATSDR. (2007). Toxicological Profile for Arsenic (Update). Retrieved from Atlanta, GA:
https://wwwn.cdc.gov/TSP/substances/ToxSubstance.aspx?toxid=3
ATSDR. (2012a). Toxicological Profile for Cadmium. Retrieved from Atlanta, GA:
https://wwwn.cdc.gov/TSP/substances/ToxSubstance.aspx?toxid=15
ATSDR. (2012b). Toxicological Profile for Chromium. Retrieved from Atlanta, GA:
https://wwwn.cdc.gov/TSP/substances/ToxSubstance.aspx?toxid=17
ATSDR. (2012c). Toxicological Profile for Manganese. Retrieved from Atlanta, GA:
https://wwwn.cdc.gov/TSP/substances/ToxSubstance.aspx?toxid=23
ATSDR. (2019). Toxicological Profile for Antimony. Retrieved from Atlanta, GA:
https://wwwn.cdc.gov/TSP/ToxProfiles/ToxProfiles.aspx?id=332&tid=58
ATSDR. (2022). Toxicological Profile for Mercury (Draft for Public Comment). (CAS#: 7439-
97-6). U.S. Center for Desease Control.
https://wwwn.cdc.gov/TSP/ToxProfiles/ToxProfiles.aspx?id=115&tid=24
ATSDR. (2023a). Toxicological Profile for Cobalt (Draft). Retrieved from Atlanta, GA:
https://wwwn.cdc.gov/TSP/ToxProfiles/ToxProfiles.aspx?id=373&tid=64
ATSDR. (2023b). Toxicological Profile for Nickel: Draft for Public Comment. Retrieved from
Atlanta, GA: https://www.atsdr.cdc.gov/ToxProfiles/tpl5.pdf
Auffhammer, M. (2018). Quantifying Economic Damages from Climate Change. Journal of
Economic Perspectives, 32(4), 33-52. doi:10.1257/jep.32.4.33
4-65
-------
Bauer, M. D., & Rudebusch, G. D. (2020). Interest Rates under Falling Stars. American
Economic Review, 110(5), 1316-1354. doi:10.1257/aer.20171822
Bauer, M. D., & Rudebusch, G. D. (2023). The Rising Cost of Climate Change: Evidence from
the Bond Market. The Review of Economics and Statistics, 105(5), 1255-1270.
doi: 10.1162/rest_a_01109
Bell, M. L., Dominici, F., & Samet, J. M. (2005). A meta-analysis of time-series studies of ozone
and mortality with comparison to the national morbidity, mortality, and air pollution
study. Epidemiology, 16(4), 436-445. doi:10.1097/01.ede.0000165817.40152.85
Bell, M. L., McDermott, A., Zeger, S. L., Samet, J. M., & Dominici, F. (2004). Ozone and short-
term mortality in 95 US urban communities, 1987-2000. JAMA, 292(19), 2372-2378.
doi: 10.1001/jama.292.19.2372
Bergeron, C. M., Bodinof, C. M., Unrine, J. M., & Hopkins, W. A. (2010a). Bioaccumulation
and maternal transfer of mercury and selenium in amphibians. Environmental Toxicology
and Chemistry, 29(4), 989-997. doi:https://doi.org/10.1002/etc,125
Bergeron, C. M., Bodinof, C. M., Unrine, J. M., & Hopkins, W. A. (2010b). Mercury
accumulation along a contamination gradient and nondestructive indices of
bioaccumulation in amphibians. Environmental Toxicology and Chemistry, 29(4), 980-
988. doi:https://doi.org/10.1002/etc,121
Carleton, T., Jina, A., Delgado, M., Greenstone, M., Houser, T., Hsiang, S., . . . Zhang, A. T.
(2022). Valuing the Global Mortality Consequences of Climate Change Accounting for
Adaptation Costs and Benefits*. The Quarterly Journal of Economics, 137(4), 2037-
2105. doi:10.1093/qje/qjac020
Chetty, R., Friedman, J. N., & Rockoff, J. E. (2014). Measuring the Impacts of Teachers II:
Teacher Value-Added and Student Outcomes in Adulthood. American Economic
Review, 104(9), 2633-2679. doi: 10.1257/aer. 104.9.2633
Climate Impact Lab (CIL). (2023). Data-driven Spatial Climate Impact Model User Manual,
Version 092023-EPA. https://impactlab.org/wp-
content/uploads/2023/10/DSCIM_UserManual_Version092023-EPA.pdf
Cristol, D. A., Brasso, R. L., Condon, A. M., Fovargue, R. E., Friedman, S. L., Hallinger, K. K., .
. . White, A. E. (2008). The Movement of Aquatic Mercury Through Terrestrial Food
Webs. Science, 320(5874), 335-335. doi:doi:10.1126/science,1154082
Cropper, M. L., Freeman, M. C., Groom, B., & Pizer, W. A. (2014). Declining Discount Rates.
American Economic Review, 104(5), 538-543. doi:10.1257/aer,104.5.538
Di, Q., Wang, Y., Zanobetti, A., Wang, Y., Koutrakis, P., Choirat, C., . . . Schwartz, J. D. (2017).
Air pollution and mortality in the Medicare population. New England Journal of
Medicine, 376(26), 2513-2522. doi: 10.1056/NEJMoal702747
4-66
-------
EPA. (2023f). Supplementary Material for the Regulatory Impact Analysis for the Final
Rulemaking: Standards of Performance for New, Reconstructed, and Modified Sources
and Emissions Guidelines for Existing Sources: Oil and Natural Gas Sector Climate
Review. Retrieved from Washington, DC:
EPA Science Inventory. (2021). Technical Documentation on The Framework for Evaluating
Damages and Impacts (FrEDI). Retrieved from
https ://cfpub. epa.gov/ si/si_public_record_report. cfm? dirEntryId=3 51316&Lab=0 AP& si
mplesearch=0&showcriteria=2&sortby=pubDate&searchall=fredi&timstype=&datebegin
publi shedpresented=02/14/2021
Hartin, C., McDuffie, E. E., Noiva, K., Sarofim, M., Parthum, B., Martinich, J., . . . Fawcett, A.
(2023). Advancing the estimation of future climate impacts within the United States.
EGUsphere, 2023, 1-32. doi:10.5194/egusphere-2023-114
Hollmann, F., Mulder, T., & Kalian, J. J. W., 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.
Hope, C. (2013). Critical issues for the calculation of the social cost of C02: why the estimates
from PAGE09 are higher than those from PAGE2002. Climatic Change, 117(3), 531-543.
doi: 10.1007/s 105 84-012-063 3 -z
Howard, P. H., & Sterner, T. (2017). Few and Not So Far Between: A Meta-analysis of Climate
Damage Estimates. Environmental and Resource Economics, 68(1), 197-225.
doi: 10.1007/s 10640-017-0166-z
Huang, Y., Dominici, F., & Bell, M. L. (2005). Bayesian hierarchical distributed lag models for
summer ozone exposure and cardio-respiratory mortality. Environmetrics, 16(5), 547-
562. doi: 10.1002/env.721
IEc. (2019). The Economic Benefits of the Mercury and Air Toxics Standards (MATS) Rule to
the Commercial and Recreational Fishery Sectors of Northeast and Midwest States (Final
Report, April 17, 2019) Retrieved from Cambridge, MA: Accessible at:
https://www.regulations.gOv/comment/EPA-HQ-OAR-2018-0794-1175
IPCC. (2018). 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 (V. Masson-Delmotte,
P. Zhai, H.-O. Portner, D. Roberts, J. Skea, P. R. Shukla, A. Pirani, W. Moufouma-Okia,
C. Pean, R. Pidcock, S. Connors, J. B. R. Matthews, Y. Chen, X. Zhou, M. I. Gomis, E.
Lonnoy, T. Maycock, a. M. Tignor, & T. Waterfield Eds.).
IPCC. (2021a). The Earth's Energy Budget, Climate Feedbacks, and Climate Sensitivity:
Contribution of Working Group I to the Sixth Assessment Report of the
Intergovernmental Panel on Climate Change. In V. Masson-Delmotte, P. Zhai, A. Pirani,
S.L. Connors, C. Pean, S. Berger, N. Caud, Y. Chen, L. Goldfarb, M.I. Gomis, M.
4-67
-------
Huang, K. Leitzell, E. Lonnoy, J.B.R. Matthews, T.K. Maycock, T. Waterfield, O.
Yelektji, R. Yu, & B. Zhou (Eds.), Climate Change 2021: The Physical Science Basis
(pp. 923-1054). Cambridge, United Kingdom and New York, NY, USA: Cambridge
University Press.
IPCC. (2021b). Short-Lived Climate Forcers. In Climate Change 2021: The Physical Science
Basis. Cambridge, UK; New York, NY: IPCC
Ito, K., De Leon, S. F., & Lippmann, M. (2005). Associations between ozone and daily
mortality: analysis and meta-analysis. Epidemiology, 16(4), 446-457.
doi:10.1097/01. ede.0000165821.90114.7f
IWG. (2010). Technical Support Document: Social Cost of Carbon for Regulatory Impact
Analysis under Executive Order 12866. Washington DC: U.S. Government, Interagency
Working Group (IWG) on Social Cost of Carbon.
https://www.epa.gov/sites/default/files/2016-12/documents/scc_tsd_2010.pdf
IWG. (2013). Technical Support Document: Technical Update of the Social Cost of Carbon for
Regulatory Impact Analysis Under Executive Order 12866.
IWG. (2015). Response to Comments: Social Cost of Carbon for Regulatory Impact Analysis
Under Executive Order 12866. Washington DC.
https://obamawhitehouse.archives.gov/sites/default/files/omb/inforeg/scc-response-to-
comments-final-july-2015.pdf
IWG. (2016a). Addendum to Technical Support Document on Social Cost of Carbon for
Regulatory Impact Analysis under Executive Order 12866: Application of the
Methodology to Estimate the Social Cost of Methane and the Social Cost of Nitrous
Oxide. Washington DC: U.S. Government, Interagency Working Group (IWG) on Social
Cost of Greenhouse Gases, https://www.epa.gov/sites/default/files/2016-
12/documents/addendum_to_sc-ghg_tsd_august_2016.pdf
IWG. (2016b). Technical Support Document: Technical Update of the Social Cost of Carbon for
Regulatory Impact Analysis Under Executive Order 12866. Washington DC: U.S.
Government, Interagency Working Group (IWG) on Social Cost of Greenhouse Gases.
https://www.epa.gov/sites/default/files/2016-12/documents/sc_co2_tsd_august_2016.pdf
IWG. (2021). Technical Support Document: Social Cost of Carbon, Methane, and Nitrous Oxide
Interim Estimates under Executive Order 13990. Washington DC: U.S. Government,
Interagency Working Group (IWG) on Social Cost of Greenhouse Gases.
https://www.whitehouse.gov/wp-
content/uploads/2021/02/TechnicalSupportDocument_SocialCostofCarbonMethaneNitro
usOxide.pdf?source=email
Jakubowski, M., & Palczynski, C. (2007). CHAPTER 21 - Beryllium. In G. F. Nordberg, B. A.
Fowler, M. Nordberg, & L. T. Friberg (Eds.), Handbook on the Toxicology of Metals
(Third Edition) (pp. 415-431). Burlington: Academic Press.
4-68
-------
Katsouyanni, K., Samet, J. M., Anderson, H. R., Atkinson, R., Le Tertre, A., Medina, S., . . .
Committee, H. E. I. H. R. (2009). Air pollution and health: a European and North
American approach (APHENA). Res Rep Health Eff Inst(142), 5-90. Retrieved from
https://www.ncbi.nlm.nih.gov/pubmed/20073322
Krewski, D., Jerrett, M., Burnett, R. T., Ma, R., Hughes, E., Shi, Y., . . . Calle, E. E. (2009).
Extended follow-up and spatial analysis of the American Cancer Society study linking
particulate air pollution and mortality: Health Effects Institute Boston, MA.
Kiinzli, N., Kaiser, R., Medina, S., Studnicka, M., Chanel, O., Filliger, P., . . . Quenel, P. (2000).
Public-health impact of outdoor and traffic-related air pollution: a European assessment.
The Lancet, 356(9232), 795-801.
Levy, J. I., Chemerynski, S. M., & Sarnat, J. A. (2005). Ozone exposure and mortality: an
empiric bayes metaregression analysis. Epidemiology, 16(4), 458-468.
doi:10.1097/01.ede.0000165820.08301.b3
Liu, J., Lee, M., & Gershenson, S. (2021). The short- and long-run impacts of secondary school
absences. Journal of Public Economics, 199, 104441.
doi: http s: //doi. org /10.1016/j. j pub eco.2021.104441
Maniloff, P., & Fann, N. (2023). Estimates of the cost of illness of myocardial infarction, asthma
and stroke. Applied Economics, 1-11. doi: 10.1080/00036846.2023.2257935
Millar, R. J., Nicholls, Z. R., Friedlingstein, P., & Allen, M. R. (2017). A modified impulse-
response representation of the global near-surface air temperature and atmospheric
concentration response to carbon dioxide emissions. Atmos. Chem. Phys., 17(11), 7213-
7228. doi: 10.5194/acp-17-7213 -2017
National Academies. (2017). Valuing Climate Damages: Updating Estimation of the Social Cost
of Carbon Dioxide. Washington DC: The National Academies Press.
National Research Council. (2008). Estimating Mortality Risk Reduction and Economic Benefits
from Controlling Ozone Air Pollution. In. Washington (DC): National Academies Press
(US).
Newell, R. G., Pizer, W. A., & Prest, B. C. (2022). A Discounting Rule for the Social Cost of
Carbon. Journal of the Association of Environmental and Resource Economists, 9(5),
1017-1046. doi: 10.1086/718145
Nordhaus, W. D. (2010). Economic aspects of global warming in a post-Copenhagen
environment. Proceedings of the National Academy of Sciences of the United States of
America, 107(26), 11721-11726. doi: 10.1073/pnas. 1005985107
Nordhaus, W. D. (2014). Estimates of the Social Cost of Carbon: Concepts and Results from the
DICE-2013R Model and Alternative Approaches. Journal of the Association of
Environmental and Resource Economists, 1(1/2), 273-312. doi: 10.1086/676035
4-69
-------
Nordhaus, W. D., & Moffat, A. (2017). A survey of global impacts of climate change:
replication, survey methods, and a statistical analysis. National Bureau of Economic
Research, Working Paper 23646. Retrieved from http://www.nber.org/papers/w23646
OMB. (2003). Circular A-4: Regulatory Analysis. Washington DC.
https://www.whitehouse.gov/wp-
content/uploads/legacy_drupal_files/omb/circulars/A4/a-4.pdf
Pindyck, R. S. (2017). Comments on Proposed Rule and Regulatory Impact Analysis on the
Delay and Suspension of Certain Requirements for Waster Prevention and Resource
Conservation. Comment submitted on Nov. 6, 2017.
https://downloads.regulations.gov/EPA-HQ-OAR-2018-0283-6184/attachment_6.pdf
Pope, C. A., Burnett, R. T., Thun, M. J., Calle, E. E., Krewski, D., Ito, K., & Thurston, G. D.
(2002). Lung cancer, cardiopulmonary mortality, and long-term exposure to fine
particulate air pollution. JAMA, 287(9), 1132-1141.
Pope, C. A., Lefler, J. S., Ezzati, M., Higbee, J. D., Marshall, J. D., Kim, S.-Y., . . . Robinson, A.
L. (2019). Mortality risk and fine particulate air pollution in a large, representative cohort
of US adults. Environmental Health Perspectives, 127(7), 077007.
Pope, C. A., Thun, M. J., Namboodiri, M. M., Dockery, D. W., Evans, J. S., Speizer, F. E., &
Heath, C. W. (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.
Ramsey, F. P. (1928). A Mathematical Theory of Saving. The Economic Journal, 38(152), 543-
559. doi: 10.2307/2224098
Rennert, K., Errickson, F., Prest, B. C., Rennels, L., Newell, R. G., Pizer, W., . . . Anthoff, D.
(2022). Comprehensive Evidence Implies a Higher Social Cost of C02. Nature,
610(7933), 687-692. doi:10.1038/s41586-022-05224-9
Rennert, K., Prest, B. C., Pizer, W. A., Newell, R. G., Anthoff, D., Kingdon, C., . . . Errickson, F.
(2022). The social cost of carbon: Advances in long-term probabilistic projections of
population, GDP, emissions, and discount rates. Brookings Papers on Economic Activity,
Fall 2021, 223-305.
Rimmer, C. C., McFarland, K. P., Evers, D. C., Miller, E. K., Aubry, Y., Busby, D., & Taylor, R.
J. (2005). Mercury Concentrations in Bicknell's Thrush and Other Insectivorous
Passerines in Montane Forests of Northeastern North America. Ecotoxicology, 14(1),
223-240. doi: 10.1007/s 10646-004-6270-1
Rode, A., Carleton, T., Delgado, M., Greenstone, M., Houser, T., Hsiang, S., . . . Yuan, J. (2021).
Estimating a social cost of carbon for global energy consumption. Nature, 598(7880),
308-314. doi: 10.1038/s41586-021-03883-8
4-70
-------
Rose, S., Turner, D., Blanford, G., Bistline, J., de la Chesnaye, F., & Wilson, T. (2014).
Understanding the Social Cost of Carbon: A Technical Assessment. EPRI Technical
Update Report. Palo Alto, CA.
Sacks, J. D., Fann, N., Gumy, S., Kim, I., Ruggeri, G., & Mudu, P. (2020). Quantifying the
Public Health Benefits of Reducing Air Pollution: Critically Assessing the Features and
Capabilities of WHO's AirQ+ and U.S. EPA's Environmental Benefits Mapping and
Analysis Program—Community Edition (BenMAP—CE). Atmosphere, 11(5), 516.
Retrieved from https://www.mdpi.eom/2073-4433/ll/5/516
Sarofim, M. C., Martinich, J., Neumann, J. E., Willwerth, J., Kerrich, Z., Kolian, M., . . . Hartin,
C. (2021). A Temperature Binning Approach for Multi-sector Climate Impact Analysis.
Climatic Change, 165(1), 22. doi:10.1007/sl0584-021-03048-6
Schwartz, J. (2005). How sensitive is the association between ozone and daily deaths to control
for temperature? Am J Respir Crit Care Med, 171(6), 627-631.
doi: 10.1164/rccm.200407-9330C
Seidler, A., Euler, U., Muller-Quernheim, J., Gaede, K. I., Latza, U., Groneberg, D., & Letzel, S.
(2012). Systematic review: progression of beryllium sensitization to chronic beryllium
disease. Occupational Medicine, 62(7), 506-513. doi:10.1093/occmed/kqs069
Shrider, E. A., Kollar, M., Chen, F., & Semega, J. (2021). Current Population Reports, P60-273,
Income and Poverty in the United States: 2020,. Washington, DC: U.S. Government
Publishing Office.
https://www.census.gov/content/dam/Census/library/publications/2021/demo/p60-
273.pdf
Smith, C. J., Forster, P. M., Allen, M., Leach, N., Millar, R. J., Passerello, G. A., & Regayre, L.
A. (2018). FAIR vl.3: a simple emissions-based impulse response and carbon cycle
model. Geosci. Model Dev., 11(6), 2273-2297. doi:10.5194/gmd-ll-2273-2018
Smith, R. L., Xu, B., & Switzer, P. (2009). Reassessing the relationship between ozone and
short-term mortality in US urban communities. Inhalation toxicology, 21(sup2), 37-61.
Tol, R. (2009). An analysis of mitigation as a response to climate change. Copenhagen
Consensus on Climate. Copenhagen Consensus Center.
https://copenhagenconsensus.eom/sites/default/files/ap_mitigation_tol_v_3.0.pdf
Turner, M. C., Jerrett, M., Pope, A., Ill, Krewski, D., Gapstur, S. M., Diver, W. R., . . . Burnett,
R. T. (2016). Long-term ozone exposure and mortality in a large prospective study.
American Journal of Respiratory and Critical Care Medicine, 193(10), 1134-1142.
doi:10.1164/rccm.201508-1633OC
U.S. EPA. (1987a). Antimony. Retrieved from
https://iris.epa.gov/ChemicalLanding/&substance_nmbr=6
4-71
-------
U.S. EPA. (1987b). Cadmium. Retrieved from
https://iris.epa.gov/ChemicalLanding/&substance_nmbr=141
U.S. EPA. (1987c). Nickel Carbonyl. Retrieved from
https://iris.epa.gov/ChemicalLanding/&substance_nmbr=274
U.S. EPA. (1987d). Nickel Refinery Dust. Retrieved from
https://iris.epa.gov/ChemicalLanding/&substance_nmbr=272
U.S. EPA. (1987e). Nickel Subsulfide. Retrieved from
https://iris.epa.gov/ChemicalLanding/&substance_nmbr=273
U.S. EPA. (1988). Lead. Retrieved from https://iris.epa.gov/static/pdfs/0277_summary.pdf
U.S. EPA. (1991). Selenium sulfide; CASRN 7446-34-6. Retrieved from
https://iris.epa.gov/static/pdfs/0458_summary.pdf
U.S. EPA. (1995a). Arsenic, inorganic; CASRN 7440-38-2 Retrieved from
https://iris.epa.gov/static/pdfs/0278_summary.pdf
U.S. EPA. (1995b). Manganese. Retrieved from
https://iris.epa.gov/ChemicalLanding/&substance_nmbr=373
U.S. EPA. (1997). Mercury Study Report to Congress (EPA-452/R-97-003). Retrieved from
United States:
U.S. EPA. (1998a). Beryllium and compounds. Retrieved from
https://iris.epa.gov/ChemicalLanding/&substance_nmbr=12
U.S. EPA. (1998b). Toxicological Review of Hexavalent Chromium. Washington, DC: U.S.
Environmental Protection Agency, https://iris.epa.gov/static/pdfs/0144tr.pdf
U.S. EPA. (1998c). Toxicological Review of Trivalent Chromium. Washington, DC: U.S.
Environmental Protection Agency, https://iris.epa.gov/static/pdfs/0028tr.pdf
U.S. EPA. (2001). IRIS Summary for Methylmercury (MeHg). (CASRN 22967-92-6). U.S.
Environmental Protection Agency.
https://iris.epa.gov/ChemicalLanding/&substance_nmbr=73
U.S. EPA. (2008). Provisional Peer-Reviewed Toxicity Values for Cobalt (EPA/690/R-08/008F).
Retrieved from https://cfpub.epa.gov/ncea/pprtv/recordisplay.cfm?deid=338894
U.S. EPA. (2009a). Integrated Science Assessment for Particulate Matter (Final Report). (EPA-
600/R-08-139F). Research Triangle Park, NC: Office of Research and Development,
National Center for Environmental Assessment.
https://cfpub.epa.gov/ncea/risk/recordisplay.cfm?deid=216546
4-72
-------
U.S. EPA. (2009b). Integrated Science Assessment for Particulate Matter (second external
review draft). (EPA-600/R-08-139B). Research Triangle Park, NC: Office of Research
and Development, National Center for Environmental Assessment,
http: //cfpub. epa. gov/ncea/ cfm / recordi spl ay. cfm? dei d=210586
U.S. EPA. (2011). Regulatory Impact Analysis for the Final Mercury and Air Toxics Standards.
(EPA-452/R-11-011). Research Triangle Park, NC: U.S. Environmental Protection
Agency, Office of Air Quality Planning and Standards, Health and Environmental Impact
Division. http://www.epa.gov/ttn/ecas/regdata/RIAs/matsriafinal.pdf
U.S. EPA. (2012). Regulatory Impact Analysis for the Final Revisions to the National Ambient
Air Quality Standards for Particulate Matter. (EPA-452/R-12-005). Research Triangle
Park, NC: U.S. Environmental Protection Agency, Office of Air Quality Planning and
Standards, Health and Environmental Impact Division.
https://www3.epa.gov/ttnecasl/regdata/RIAs/finalria.pdf
U.S. EPA. (2014). Guidelines for Preparing Economic Analyses. (EPA 240-R-10-001).
Washington DC: U.S. Environmental Protection Agency, Office of Policy, National
Center for Environmental Economics, https://www.epa.gov/environmental-
economics/guidelines-preparing-economic-analyses
U.S. EPA. (2015). Regulatory Impact Analysis of the Final Revisions to the National Ambient
Air Quality Standards for Ground-Level Ozone. (EPA-452/R-15-007). Research Triangle
Park, NC: Office of Air Quality Planning and Standards, Health and Environmental
Impacts Division, https://www3.epa.gov/ttnecasl/docs/20151001ria.pdf
U.S. EPA. (2016). Integrated Science Assessment for Oxides of Nitrogen - Health Criteria.
(EPA/600/R-15/068). Research Triangle Park, NC: Office of Research and Development,
National Center for Environmental Assessment.
https://cfpub.epa.gOv/ncea/i sa/recordisplay.cfm?deid=310879
U.S. EPA. (2017). Integrated Science Assessment for Sulfur Oxides - Health Criteria.
(EPA/600/R-17/451). Research Triangle Park, NC: Office of Research and Development,
National Center for Environmental Assessment.
https://cfpub.epa.gOv/ncea/i sa/recordisplay.cfm?deid=338596
U.S. EPA. (2019a). Integrated Science Assessment (ISA) for Particulate Matter (Final Report).
(EPA/600/R-19/188). Research Triangle Park, NC: U.S. Environmental Protection
Agency, Office of Research and Development, Center for Public Health and
Environmental Assessment, https://www.epa.gov/naaqs/particulate-matter-pm-standards-
integrated-science-assessments-current-review
U.S. EPA. (2019b). 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. (EPA-452/R-19-003). Research Triangle Park, NC: U.S.
Environmental Protection Agency, Office of Air Quality Planning and Standards, Health
and Environmental Impact Division, https://www.epa.gov/sites/production/files/2019-
06/documents/utilities_ria_final_cpp_repeal_and_ace_2019-06.pdf
4-73
-------
U.S. EPA. (2020a). Analysis of Potential Costs and Benefits for the National Emission Standards
for Hazardous Air Pollutants: Coal- and Oil-Fired Electric Utility Steam Generating
Units - Subcategory of Certain Existing Electric Utility Steam Generating Units Firing
Eastern Bituminous Coal Refuse for Emissions of Acid Gas Hazardous Air Pollutants.
Research Triangle Park, NC: U.S. Environmental Protection Agency, Office of Air
Quality Planning and Standards, Health and Environmental Impact Division.
https://www.epa.gov/sites/default/files/2020-04/documents/mats_coal_refuse_cost-
benefit_memo.pdf
U.S. EPA. (2020b). Benefit and Cost Analysis for Revisions to the Effluent Limitations
Guidelines and Standards for the Steam Electric Power Generating Point Source
Category. (EPA-821-R-20-003). Washington DC: U.S. Environmental Protection
Agency, https://www.epa.gov/sites/default/files/2020-
08/documents/steam_electric_elg_2020_final_reconsideration_rule_benefit_and_cost_an
alysis.pdf
U.S. EPA. (2020c). Integrated Science Assessment (ISA) for Oxides of Nitrogen, Oxides of
Sulfur and Particulate Matter Ecological Criteria. (EPA/600/R-20/278). Washington DC:
U.S. Environmental Protection Agency.
https://cfpub.epa.gOv/ncea/i sa/recordisplay.cfm?deid=349473
U.S. EPA. (2020d). Residual Risk Assessment for the Coal- and Oil-Fired EGU Source Category
in Support of the 2020 Risk and Technology Review Final Rule. Research Triangle Park,
NC: U.S. Environmental Protection Agency, Office of Air Quality Planning and
Standards. https://www.regulations.gov/document/EPA-HQ-OAR-2018-0794-4553
U.S. EPA. (2021a). Regulatory Impact Analysis for the Final Revised Cross-State Air Pollution
Rule (CSAPR) Update for the 2008 Ozone NAAQS. (EPA-452/R-21-002). Research
Triangle Park, NC: U.S. Environmental Protection Agency, Office of Air Quality
Planning and Standards, Health and Environmental Impact Division.
https://www.epa.gov/sites/default/files/2021-
03/documents/revi sed_csapr_update_ria_final.pdf
U.S. EPA. (2021b). Technical Documentation on the Framework for Evaluating Damages and
Impacts (FrEDI). (EPA 430-R-21-004). https://www.epa.gov/cira/fredi
U.S. EPA. (2022a). BenMAP-CE User Manual and Appendices. Research Triangle Park, NC:
Office of Air Quality Planning and Standards, https://www.epa.gov/benmap/benmap-ce-
manual-and-appendices
U.S. EPA. (2022b). Policy Assessment for the Reconsideration of the National Ambient Air
Quality Standards for Particulate Matter. Office of Air Quality Planning and Standards.
(EPA-452/R-22-004). Research Triangle Park, NC: U.S. EPA, OAQPS, HEID.
https://www.epa.gov/system/files/documents/2022-
05/Final%20Policy%20Assessment%20for%20the%20Reconsideration%20of%20the%2
0PM%20NAAQS_May2022_0.pdf
4-74
-------
U.S. EPA. (2022c). Regulatory Impact Analysis for Proposed Federal Implementation Plan
Addressing Regional Ozone Transport for the 2015 Ozone National Ambient Air Quality
Standard. (EPA-452/D-22-001). Research Triangle Park, NC: U.S. Environmental
Protection Agency, Office of Air Quality Planning and Standards, Health and
Environmental Impact Division, https://www.epa.gov/system/files/documents/2022-
03/transport_ria_proposal_fip_2015_ozone_naaqs_2022-02.pdf
U.S. EPA. (2022d). Regulatory Impact Analysis for the Proposed Reconsideration of the
National Ambient Air Quality Standards for Particulate Matter. (EPA-452/P-22-001).
Research Triangle Park, NC: U.S. Environmental Protection Agency, Office of Air
Quality Planning and Standards, Health and Environmental Impact Division.
https://www.epa.gOv/system/files/documents/2023-01/naaqs-pm_ria_proposed_2022-
12.pdf
U.S. EPA. (2022e). Supplement to the 2019 Integrated Science Assessment for Particulate
Matter (Final Report). (EPA/600/R-22/028). Research Triangle Park, NC: U.S.
Environmental Protection Agency, Office of Research and Development, Center for
Public Health and Environmental Assessment.
https://cfpub.epa.gOv/ncea/i sa/recordisplay.cfm?deid=354490
U.S. EPA. (2022f). Technical Support Document (TSD) for the Proposed Federal
Implementation Plan Addressing Regional Ozone Transport for the 2015 Ozone National
Ambient Air Quality Standard: Ozone Transport Policy Analysis Proposed Rule TSD.
U.S. Environmental Protection Agency, Office of Air and Radiation.
https://www.epa.gOv/system/files/documents/2022-03/ozone-transport-policy-analysis-
proposed-rule-tsd.pdf
U.S. EPA. (2023a). Estimating PM2.5- and Ozone-Attributable Health Benefits. Research
Triangle Park, NC: U.S. Environmental Protection Agency, Office of Air Quality
Planning and Standards, Health and Environmental Impact Division.
https://www.epa.gov/system/files/documents/2023-01/Estimating%20PM2.5-
%20and%200zone-Attributable%20Health%20Benefits%20TSD_0.pdf
U.S. EPA. (2023b). Regulatory Impact Analysis for the Proposed National Emission Standards
for Hazardous Air Pollutants: Coal- and Oil-Fired Electric Utility Steam Generating
Units Review of the Residual Risk and Technology Review. (EPA-452/R-23-002).
Research Triangle Park, NC: U.S. Environmental Protection Agency, Office of Air
Quality Planning and Standards, Health and Environmental Impact Division.
https://www.epa.gov/system/files/documents/2023-
04/MATS%20RTR%20Proposal%20RIA%20Formatted.pdf
U.S. EPA. (2023c). Supplementary Material for the Regulatory Impact Analysis for the Final
Rulemaking, "Standards of Performance for New, Reconstructed, and Modified Sources
and Emissions Guidelines for Existing Sources: Oil and Natural Gas Sector Climate
Review": EPA Report on the Social Cost of Greenhouse Gases: Estimates Incorporating
Recent Scientific Advances. Washington, DC: U.S. EPA
4-75
-------
U.S. EPA Science Advisory Board. (2004). Advisory on Plans for Health Effects Analysis in the
Analytical Plan for EPA's Second Prospective Analysis - Benefits and Costs of the Clean
Air Act, 1990-2020 Advisory by the Health Effects Subcommittee of the Advisory
Council on Clean Air Compliance Analysis. (EPA-SAB-COUNCIL-ADV-4-002).
Washington DC: U.S. Environmental Protection Agency, Office of Air Quality Planning
and Standards, Heath and Environmental Impacts Division.
https://www.epa.gov/naaqs/particulate-matter-pm-standards-policy-assessments-current-
review-0
U.S. EPA Science Advisory Board. (2022). CASAC Review of the EPA's Policy Assessment for
the Reconsideration of the National Ambient Air Quality Standards for Particulate Matter
(External Review Draft - October 2021). (EPA-CASAC-22-002). Research Triangle
Park, NC: U.S. Environmental Protection Agency,.
https://casac.epa.gov/ords/sab/f?p=l 13:12:1342972375271::: 12
U.S. EPA Science Advisory Board. (2024). Review of BenMAP and Benefits Methods. (EPA-
SAB-24-003). U.S. Environmental Protection Agency,.
https://sab.epa.gov/ords/sab/f?p=l 14:0:15383733289761 :APPLICATION_PROCESS=R
EPORT DOC:: : REPORT ID: 1124
Wada, H., Cristol, D. A., McNabb, F. M. A., & Hopkins, W. A. (2009). Suppressed
Adrenocortical Responses and Thyroid Hormone Levels in Birds near a Mercury-
Contaminated River. Environmental Science & Technology, 43(15), 6031-6038.
doi:10.1021/es803707f
Wada, H., Yates, D. E., Evers, D. C., Taylor, R. J., & Hopkins, W. A. (2010). Tissue mercury
concentrations and adrenocortical responses of female big brown bats (Eptesicus fuscus)
near a contaminated river. Ecotoxicology, 19(7), 1277-1284. doi: 10.1007/s 10646-010-
0513-0
Woodruff, T. J., Darrow, L. A., & Parker, J. D. (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. Retrieved from
https://www.woodsandpoole.com/
Wu, X., Braun, D., Schwartz, J., Kioumourtzoglou, M., & Dominici, F. (2020). Evaluating the
impact of long-term exposure to fine particulate matter on mortality among the elderly.
Science advances, 6(29), eaba5692.
Wu, X., Braun, D., Schwartz, J., Kioumourtzoglou, M. A., & Dominici, F. (2020). Evaluating the
impact of long-term exposure to fine particulate matter on mortality among the elderly.
Sci Adv, 6(29), eaba5692. doi:10.1126/sciadv.aba5692
Zanobetti, A., & Schwartz, J. (2008). Mortality displacement in the association of ozone with
mortality: an analysis of 48 cities in the United States. Am J Respir Crit Care Med,
177(2), 184-189. doi:10.1164/rccm.200706-8230C//
4-76
-------
ECONOMIC IMPACTS
5.1 Overview
Economic impact analyses focus on changes in market prices and output levels. If
changes in market prices and output levels in the primary markets are significant enough,
impacts on other markets may also be examined. Both the magnitude of costs needed to comply
with a rule and the distribution of these costs among affected facilities can have a role in
determining how the market will change in response to a rule. This section analyzes the potential
impacts on small entities and the potential labor impacts associated with this rulemaking. For
additional discussion of impacts on fuel use and electricity prices, see Section 3.
5.2 Small Entity Analysis
For the final rule, EPA performed a small entity screening analysis for impacts on all
affected EGUs and non-EGU facilities by comparing compliance costs to historic revenues at the
ultimate parent company level. This is known as the cost-to-revenue or cost-to-sales test, or the
"sales test." The sales test is an impact methodology EPA employs in analyzing entity impacts as
opposed to a "profits test," in which annualized compliance costs are calculated as a share of
profits. The sales test is frequently used because revenues or sales data are commonly available
for entities impacted by EPA regulations, and profits data normally made available are often not
the true profit earned by firms because of accounting and tax considerations. Also, the use of a
sales test for estimating small business impacts for a rulemaking is consistent with guidance
offered by EPA on compliance with the Regulatory Flexibility Act (RFA)81 and is consistent
with guidance published by the U.S. Small Business Administration's (SBA) Office of Advocacy
that suggests that cost as a percentage of total revenues is a metric for evaluating cost increases
on small entities in relation to increases on large entities.82
81 See U.S. EPA. (2006). Final Guidance for EPA Rulewriters: Regulatory Flexibility Act as Amended by the Small
Business and Regulatory Enforcement Fairness Act. Available at: https://www.epa.gov/sites/production/files/2015-
06/documents/guidance-regflexact.pdf.
82 See U.S. SBA Office of Advocacy. (2017). A Guide for Government Agencies: How to Comply with the
Regulatory Flexibility Act. Available at: https://advocacy.sba.gov/2017/08/31/a-guide-for-government-agencies-
how-to-comply-with-the-regulatory-flexibility-act.
5-1
-------
5.2.1 Methodology
This section presents the methodology and results for estimating the impact of the rule on
small EGU entities in the year of compliance, 2028, based on the following endpoints:
• annual economic impacts of the final rule on small entities, and
• ratio of small entity impacts to revenues from electricity generation.
For this analysis, EPA first considered EGUs that are subject to MATS requirements and
for which EPA assumed additional controls would be necessary to meet the requirements of the
finalized rule. We then refined this list of MATS-affected EGUs, complementing the list with
units for which the projected impacts exceeds either of the two criteria below relative to the
baseline:
• Fuel use (BTUs) changes by +/- 1 percent or more
• Generation (GWh) changes by +/- 1 percent or more
Please see Section 3 for more discussion of the power sector modeling.
Based on these criteria, EPA identified a total of 377 potentially affected EGUs
warranting examination in 2028 in this RFA analysis. Next, we determined power plant
ownership information, including the name of associated owning entities, ownership shares, and
each entity's type of ownership. We primarily used data from Hitachi — Power Grids, The
Velocity Suite 12020 ("VS"), supplemented by limited research using publicly available data.
Majority owners of power plants with affected EGUs were categorized as one of the seven
ownership types. These ownership types are:
1. Investor-Owned Utility (IOU): Investor-owned assets (e.g., a marketer, independent
power producer, financial entity) and electric companies owned by stockholders, etc.
2. Cooperative (Co-Op): Non-profit, customer-owned electric companies that generate
and/or distribute electric power.
3. Municipal: A municipal utility, responsible for power supply and distribution in a small
region, such as a city.
4. Sub-division: Political subdivision utility is a county, municipality, school district,
hospital district, or any other political subdivision that is not classified as a municipality
under state law.
5. Private: Similar to an investor-owned utility, however, ownership shares are not openly
traded on the stock markets.
5-2
-------
6. State: Utility owned by the state.
7. Federal: Utility owned by the federal government.
Next, EPA used both the D&B Hoovers online database and the VS database to identify
the ultimate owners of power plant owners identified in the VS database. This was necessary, as
many majority owners of power plants (listed in VS) are themselves owned by other ultimate
parent entities (listed in D&B Hoovers). In these cases, the ultimate parent entity was identified
via D&B Hoovers, whether domestically or internationally owned.
EPA followed SBA size standards to determine which non-government ultimate parent
entities should be considered small entities in this analysis. These SBA size standards are
specific to each industry, each having a threshold level of either employees, revenue, or assets
below which an entity is considered small. SBA guidelines list all industries, along with their
associated North American Industry Classification System (NAICS) code and SBA size
standard. Therefore, it was necessary to identify the specific NAICS code associated with each
ultimate parent entity in order to understand the appropriate size standard to apply. Data from
D&B Hoovers were used to identify the NAICS codes for most of the ultimate parent entities. In
many cases, an entity that is a majority owner of a power plant is itself owned by an ultimate
parent entity with a primary business other than electric power generation. Therefore, it was
necessary to consider SBA entity size guidelines for the range of NAICS codes listed in Table
5-1. This table represents the range of NAICS codes and areas of primary business of ultimate
parent entities that are majority owners of potentially affected EGUs in EPA's IPM base case.
5-3
-------
Table 5-1
SBA Size Standards by NAICS Code
Size Standard Size Standard
(millions of (number of
NAICS Code
NAICS U.S. Industry Title
dollars)
employees)
211120
Crude Petroleum Extraction
1,250
212221
Gold Ore Mining
1,500
221111
Hydroelectric Power Generation
500
221112
Fossil Fuel Electric Power Generation
750
221113
Nuclear Electric Power Generation
750
221114
Solar Electric Power Generation
250
221115
Wind Electric Power Generation
250
221116
Geothermal Electric Power Generation
250
221117
Biomass Electric Power Generation
250
221118
Other Electric Power Generation
250
221121
Electric Bulk Power Transmission and Control
500
221122
Electric Power Distribution
1,000
221210
Natural Gas Distribution
1,000
221310
Water Supply and Irrigation Systems
$41.00
221320
Sewage Treatment Facilities
$35.00
221330
Steam and Air Conditioning Supply
$30.00
311221
Wet Corn Milling
1,250
311224
Soybean and Other Oilseed Processing
1,000
322121
Paper (except Newsprint) Mills
1,250
325611
Soap and Other Detergent Manufacturing
1,000
325920
Explosives Manufacturing
750
331110
Iron and Steel Mills and Ferroalloy Manufacturing
1,500
332313
Plate Work Manufacturing
750
332911
Industrial Valve Manufacturing
750
333611
Turbine and Turbine Generator Set Unit Manufacturing
1,500
333613
Mechanical Power Transmission Equipment Manufacturing
750
423520
Coal and Other Mineral and Ore Merchant Wholesalers
200
423990
Other Miscellaneous Durable Goods Merchant Wholesalers
100
424690
Other Chemical and Allied Products Merchant Wholesalers
175
424720
Petroleum and Petroleum Products Merchant Wholesalers
200
522110
Commercial Banking
$750.00
523210
Securities and Commodity Exchanges
$47.00
523910
Miscellaneous Intermediation
$44.25
523930
Investment Advice
$41.50
524126
Direct Property and Casualty Insurance Carriers
1,500
525910
Open-End Investment Funds
$37.50
525990
Other Financial Vehicles
$40.00
541330
Engineering Services
$22.50
541611
Administrative Management and General Management
$21.50
Consulting Services
541715
Research and Development in the Physical, Engineering, and Life Sciences
1,000
(except Nanotechnology and Biotechnology)
551112
Offices of Other Holding Companies
$45.50
5-4
-------
NAICS Code
NAICS U.S. Industry Title
Size Standard
(millions of
dollars)
Size Standard
(number of
employees)
611310
Colleges, Universities and Professional Schools
$30.50
721110
Hotels (except Casino Hotels) and Motels
$35.00
813910
Business Associations
$13.50
Note: Based on size standards effective at the time EPA conducted this analysis (SBA size standards, effective
December 19, 2022. Available at the following link: https://www.sba.gov/document/support—table-size-standards).
Source: SBA, 2022.
EPA compared the relevant entity size criterion for each ultimate parent entity to the SBA
size standard noted in Table 5-1. We used the following data sources and methodology to
estimate the relevant size criterion values for each ultimate parent entity:
• Employment, Revenue, and Assets: EPA used the D&B Hoovers database as the
primary source for information on ultimate parent entity employee numbers, revenue, and
assets.83 In parallel, EPA also considered estimated revenues from affected EGUs based
on analysis of IPM parsed-file84 estimates for the baseline for 2028. EPA assumed that
the ultimate parent entity revenue was the larger of the two revenue estimates. In limited
instances, supplemental research was also conducted to estimate an ultimate parent
entity's number of employees, revenue, or assets.
• Population: Municipal entities are defined as small if they serve populations of less than
50,000.85 EPA primarily relied on data from the Ventyx database and the U.S. Census
Bureau to inform this determination.
Ultimate parent entities for which the relevant measure is less than the SBA size standard were
identified as small entities and carried forward in this analysis.
In the projected results for 2028, EPA identified 377 potentially affected EGUs, owned
by 104 entities. Of these, EPA identified 45 potentially affected EGUs owned by 24 small
entities included in the power sector baseline.
83 Estimates of sales were used in lieu of revenue estimates when revenue data were unavailable.
84 IPM output files report aggregated results for "model" plants (i.e., aggregates of generating units with similar
operating characteristics). Parsed files approximate the IPM results at the generating unit level.
85 The Regulatory Flexibility Act defines a small government jurisdiction as the government of a city, county,
town, township, village, school district, or special district with a population of less than 50,000
(5 U.S.C. section 601(5)). For the purposes of the RFA, States and tribal governments are not considered small
governments. EPA's Final Guidance for EPA Rulewriters: Regulatory Flexibility Act is located here:
https://www.epa.gov/sites/default/files/2015-06/documents/guidance-regflexact.pdf.
5-5
-------
The chosen compliance strategy will be primarily a function of the unit's marginal
control costs and its position relative to the marginal control costs of other units. To attempt to
account for each potential control strategy, EPA estimates compliance costs as follows:
Ccompliance A Coperating+Retrofit A CFuel + A R
where C represents a component of cost as labeled and A R represents the change in revenues,
calculated as the difference in value of electricity generation between the baseline case and the
rule in in 2028.
Realistically, compliance choices and market conditions can combine such that an entity
may actually experience a reduction in any of the individual components of cost. Under the rule,
some units will forgo some level of electricity generation (and thus revenues) to comply, and this
impact will be lessened on these entities by the projected increase in electricity prices under the
rule. On the other hand, those units increasing generation levels will see an increase in electricity
revenues and as a result, lower net compliance costs. If entities are able to increase revenue more
than an increase in fuel cost and other operating costs, ultimately, they will have negative net
compliance costs (or increased profit). Overall, small entities are not projected to install
relatively costly emissions control retrofits but may choose to do so in some instances. Because
this analysis evaluates the total costs along each of the compliance strategies laid out above for
each entity, it inevitably captures gains such as those described. As a result, what we describe as
cost is actually a measure of the net economic impact of the rule on small entities.
For this analysis, EPA used IPM-parsed output to estimate costs based on the parameters
above, at the unit level. These impacts were then summed for each small entity, adjusting for
ownership share. Net impact estimates were based on the following: operating and retrofit costs,
sale or purchase of allowances, and the change in fuel costs or electricity generation revenues
under the finalized MATS requirements relative to the base case. These individual components
of compliance costs were estimated as follows:
1. Operating and retrofit costs (A Coperating+Retrofit): EPA projected which compliance
option would be selected by each EGU in 2028 and applied the appropriate cost to this
choice (for details, please see Section 3 of this RIA). For 2028, IPM projected retrofit
costs were also included in the calculation.
5-6
-------
2. Fuel costs (A CFuei): The change in fuel expenditures under the final requirements was
estimated by taking the difference in projected fuel expenditures between the IPM
estimates under the final requirements and the baseline.
3. Value of electricity generated (A CFuei): To estimate the value of electricity generated,
the projected level of electricity generation is multiplied by the regional-adjusted retail
electricity price ($/MWh) estimate, for all entities except those categorized as private in
Ventyx. See Section 3 for a discussion of the Retail Price Model, which was used to
estimate the change in the retail price of electricity. For private entities, EPA used the
wholesale electricity price instead of the retail electricity price because most of the
private entities are independent power producers (IPP). IPPs sell their electricity to
wholesale purchasers and do not own transmission facilities. Thus, their revenue was
estimated with wholesale electricity prices.
5.2.2 Results
As indicated above, the use of a sales test for estimating small business impacts for a
rulemaking is consistent with guidance offered by EPA on compliance with the RFA and is
consistent with guidance published by the SBA's Office of Advocacy that suggests that cost as a
percentage of total revenues is a metric for evaluating cost increases on small entities in relation
to increases on large entities. EPA assessed the economic and financial impacts of the rule using
the ratio of compliance costs to the value of revenues from electricity generation, focusing in
particular on entities for which this measure is greater than 1 percent.
The projected impacts, including compliance costs, of the rule on small entities are
summarized in Table 5-2. All costs are presented in 2019 dollars. We projected the annual net
compliance cost to small entities to be approximately $2.0 million in 2028. Relative to the
baseline, the rule is projected to generate compliance cost reductions greater than 1 percent of
baseline revenue for one of the 24 small entities directly impacted, and compliance cost increases
greater than 1 percent are projected for two. The remaining 23 entities are not projected to
experience compliance cost changes of more than 1 percent. Of the 24 entities considered in this
analysis, two are holding units projected to experience compliance cost increases greater than 1
percent of generation revenue at a facility level as well as at a parent holding company level.
5-7
-------
Table 5-2 Pro jected Impacts of Final Rule on Small Entities in 2028
EGU
Ownership Type
Number of Potentially
Affected Entities
Total Net Compliance
Cost (millions 2019
dollars)
Number of Small
Entities with
Compliance Costs >1%
of Generation Revenues
Subdivision
1
-0.029
0
Investor Owned
3
-0.056
0
Private
7
-0.059
0
Co-op
13
2.1
1
Total
24
2.0
1
5.2.3 Conclusion
Making a determination that there is not a significant economic impact on a substantial
number of small entities (often referred to as a "SISNOSE") requires an assessment of whether
an estimated economic impact is significant and whether that impact affects a substantial number
of small entities. EPA identified 104 potentially affected EGU entities in the projection year of
2028. Of these, EPA identified 24 small entities affected by the rule, and of these, three small
entities may experience costs of greater than 1 percent of revenues. Based on this analysis, for
this rule overall we conclude that the estimated costs for the final rule will not have a significant
economic impact on a substantial number of small entities.
5.3 Labor Impacts
This section discusses potential employment impacts of this regulation. 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. There are significant challenges when trying to evaluate
the employment effects of an environmental regulation due to a wide variety of other economic
changes that can affect employment, including the impact of the coronavirus pandemic on labor
markets and the state of the macroeconomy generally. Considering these challenges, we look to
the economics literature to provide a constructive framework and empirical evidence. To
simplify, we focus on impacts on labor demand related to compliance behavior. Environmental
regulation may also affect labor supply through changes in worker health and productivity (Zivin
and Neidell, 2018).
5-8
-------
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 their demand at all (Berman and Bui, 2001; Deschenes, 2018;
Morgenstern et al., 2002). 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 et al. (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. Changes in
employment may also occur in different sectors related to the regulated industry, both upstream
and downstream, or in sectors producing substitute or complimentary products. Employment
impacts in related sectors are often difficult to measure. Consequently, we focus our labor
impacts analysis primarily on the directly regulated facilities and other EGUs and related fuel
markets.
This section discusses and projects potential employment impacts for the utility power,
coal and natural gas production sectors that may result from the final rule. EPA has a long
history of analyzing the potential impacts of air pollution regulations on changes in the amount
of labor needed in the power generation sector and directly related sectors. The analysis
conducted for this RIA builds upon the approaches used in the past and takes advantage of newly
available data to improve the assumptions and methodology.86
The results presented in this section are based on a methodology that estimates the impact
on employment based on the differences in projections between two modeling scenarios: the
baseline scenario, and a scenario that represents the implementation of the rule. The estimated
employment difference between these scenarios can be interpreted as the incremental effect of
the rule on employment in this sector. As discussed in Section 3, there is uncertainty related to
the future baseline projections. Because the incremental employment estimates presented in this
section are based on projections discussed in Section 3, it is important to highlight the relevance
86 For a detailed overview of this methodology, including all underlying assumptions, see the U.S. EPA
Methodology for Power Sector-Specific Employment Analysis, available in the docket.
5-9
-------
of the Section 3 uncertainty discussion to the analysis presented in this section. Note that there is
also uncertainty related to the employment factors applied in this analysis, particularly factors
informing job-years related to relatively new technologies, such as energy storage, on which
there is limited data to base assumptions.
Like previous analyses, this analysis represents an evaluation of "first-order employment
impacts" using a partial equilibrium modeling approach. It includes some of the potential ripple
effects of these impacts on the broader economy. These ripple effects include the secondary job
impacts in both upstream and downstream sectors. The analysis includes impacts on upstream
sectors including coal, natural gas, and uranium. However, the approach does not analyze
impacts on other fuel sectors, nor does it analyze potential impacts related to transmission or
distribution. This approach excludes the economy-wide employment effects of changes to energy
markets (such as higher or lower forecasted electricity prices). This approach also excludes labor
impacts that are sometimes reflected in a benefits analysis for an environmental policy, such as
increased productivity from a healthier workforce and reduced absenteeism due to fewer sick
days of employees and dependent family members (e.g., children).
5.3.1 Overview of Methodology
The methodology includes the following two general approaches, based on the available
data. The first approach uses detailed employment data that are available for several types of
generation technologies in the 2020 U.S. Energy and Employment Report (USEER).87 For
employment related to other electric power sector generating and pollution control technologies,
the second approach uses information available in the U.S. Economic Census.
Detailed employment inventory data are available regarding recent employment related to
coal, hydro, natural gas, geothermal, wind, and solar generation technologies as well as battery
storage. The data enables the creation of technology-specific factors that can be applied to model
projections of capacity (reported in MW) and generation (reported in megawatt-hours, or MWh)
to estimate impacts on employment. Since employment data are only available in aggregate by
fuel type, it is necessary to disaggregate by labor type to differentiate between types of jobs or
tasks for categories of workers. For example, some types of employment remain constant
87 https://www.usenergyjobs.org/
5-10
-------
throughout the year and are largely a function of the size of a generator, e.g., fixed operation and
maintenance activities, while others are variable and are related to the amount of electricity
produced by the generator, e.g., variable operation and maintenance activities.
The approach can be summarized in three basic steps:
• Quantify the total number of employees by fuel type in a given year;
• Estimate total fixed operating & maintenance (FOM), variable operating &
maintenance (VOM), and capital expenditures by fuel type in that year; and
• Disaggregate total employees into three expenditure-based groups and develop factors
for each group (FTE/MWh, FTE/MW-year, FTE/MW new capacity).
Where detailed employment data are unavailable, it is possible to estimate labor impacts
using labor intensity ratios. These factors provide a relationship between employment and
economic output and are used to estimate employment impacts related to construction and
operation of pollution control retrofits, as well as some types of electric generation technologies.
For a detailed overview of this methodology, including all underlying assumptions and
the types of employment represented by this analysis, see the U.S. EPA Methodology for Power
Sector-Specific Employment Analysis, available in the docket.
5.3.2 Overview of Power Sector Employment
In this section we focus on employment related to electric power generation, as well as
coal and natural gas extraction because these are the segments of the power sector that are most
relevant to the projected impacts of the rule. Other segments not discussed here include other
fuels, energy efficiency, and transmission, distribution, and storage. The statistics presented here
are based on the 2020 USEER, which reports data from 2019.88
In 2019, the electric power generation sector employed nearly 900,000 people. Relative
to 2018, this sector grew by over 2 percent, despite job losses related to nuclear and coal
generation. These losses were offset by increases in employment related to other generating
technologies, including natural gas, solar, and wind. The largest component of total 2019
88 While 2020 data are available in the 2021 version of this report, this section of the RIA utilizes 2019 data because
this year does not reflect any short-term trends related to the COVID-19 pandemic. The annual report is available at:
https://www .usenergyjobs. org/.
5-11
-------
employment in this sector is construction (33 percent). Other components of the electric power
generation workforce include utility workers (20 percent), professional and business service
employees (20 percent), manufacturing (13 percent), wholesale trade (8 percent), and other (5
percent). In 2019, jobs related to solar and wind generation represent 31 percent and 14 percent
of total jobs, respectively, and jobs related to coal generation represent 10 percent of total
employment.
In addition to generation-related employment, we also look at employment related to coal
and natural gas use in the electric power sector. In 2019, the coal industry employed about
75,000 workers. Mining and extraction jobs represent the vast majority of total coal-related
employment in 2019 (74 percent). The natural gas fuel sector employed about 276,000
employees in 2019. About 60 percent of those jobs were related to mining and extraction.
5.3.3 Projected Sectoral Employment Changes due to the Final Rule
Electric generating units subject to the Hg and fPM emission limits in this rule will likely
use various Hg and PM control strategies to comply. EPA estimates that 11.6 GW of operational
coal capacity would either need to improve existing PM controls or install new PM controls to
comply with the final rule in 2028. The various PM control upgrades that EPA assumes would be
necessary to achieve with the emissions limits analyzed are summarized in Table 3-8.
Based on these power sector modeling projections, we estimate an increase in
construction-related job-years related to the installation of new pollution controls under the rule,
as well as the construction of new generating capacity. In 2028, we estimate an increase of
approximately 1,600 construction-related job-years related to the construction of new pollution
controls or control upgrades and an increase of approximately 200 job-years related to the
construction of new capacity. In 2030, we estimate a small decrease in construction job-years for
new pollution controls and new capacity, followed by an increase of 500 construction job-years
for new capacity in 2035. Construction-related job-year changes are one-time impacts, occurring
during each year of the multi-year periods during which construction of new capacity is
completed. Construction-related figures in Table 5-3 represent a point estimate of incremental
changes in construction jobs for each year (for a three-year construction projection, this table
presents one-third of the total jobs for that project).
5-12
-------
Table 5-3 Projected Changes in Labor Utilization: Construction-Related (Number of
Job-Years of Employment in a Single Year)
2028
2030
2035
New Pollution Controls
1,600
<100
<100
New Capacity
200
<100
500
Notes: "<100" denotes an increase or decrease of fewer than 100 job-years. A large share of the construction-related
job years is attributable to construction of energy storage, a relatively new technology on which there is limited data
to base labor assumptions.
We also estimate changes in the number of job-years related to recurring non-
construction employment. Recurring employment changes are job-years associated with annual
recurring jobs including operating and maintenance activities and fuel extraction jobs. Newly
built generating capacity creates a recurring stream of positive job-years, while retiring
generating capacity, as well as avoided capacity builds, create a stream of negative job-years.
Consistent with the small projected changes in generation over 2028 through 2035, this rule is
expected to result in small impacts in recurring non-construction jobs. Table 5-4 provides
detailed estimates of recurring non-construction employment changes.
Table 5-4 Projected Changes in Labor Utilization: Recurring Non-Construction
(Number of Job-Years of Employment in a Single Year)
2028
2030
2035
Pollution Controls
<100
<100
<100
Existing Capacity
<100
<100
<100
New Capacity
<100
<100
<100
Fuels (Coal, Natural Gas, Uranium)
<100
<100
<100
Coal
<100
<100
<100
Natural Gas
<100
<100
<100
Uranium
<100
<100
<100
Note: "<100" denotes an increase or decrease of fewer than 100 job-years; Numbers may not sum due to rounding.
5.3.4 Conclusions
Generally, there are significant challenges when trying to evaluate the employment
effects due to an environmental regulation from employment effects due to a wide variety of
other economic changes, including the impact of the coronavirus pandemic on labor markets and
the state of the macroeconomy generally. For EGUs, this rule may result in a sizable near-term
increase in construction-related jobs related to the installation of new pollution controls, and any
changes in recurring non-construction employment are expected to be small.
5-13
-------
5.4 References
Berman, E., & Bui, L. T. M. (2001). Environmental regulation and labor demand: evidence from
the South Coast Air Basin. Journal of Public Economics, 79(2), 265-295.
doi:https://doi.org/10.1016/S0047-2727(99)00101-2
Curtis, E. M. (2018). Who Loses under Cap-and-Trade Programs? The Labor Market Effects of
the NOx Budget Trading Program. The Review of Economics and Statistics, 100(1), 151-
166. doi: 10.1162/REST_a_00680
Curtis, E. M. (2020). Reevaluating the ozone nonattainment standards: Evidence from the 2004
expansion. Journal of Environmental Economics and Management, 99, 102261.
doi: 10.1016/j.jeem.2019.102261
Deschenes, O. (2018). Environmental regulations and labor markets. IZA World of Labor, 22.
doi:10.15185/izawol.22.v2
Ferris, A. E., Shadbegian, R., & Wolverton, A. (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 Economists, 1(4), 521-553.
doi: 10.1086/679301
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. doi: 10.1086/342808
Morgenstern, R. D., Pizer, W. A., & Shih, J.-S. (2002). Jobs Versus the Environment: An
Industry-Level Perspective. Journal of Environmental Economics and Management,
43(3), 412-436. doi:https://doi.org/10.1006/jeem.2001.1191
Zivin, J. G., & Neidell, M. (2018). Air pollution's hidden impacts. Science, 359(6371), 39-40.
doi:doi: 10.1126/science.aap7711
5-14
-------
ENVIRONMENTAL JUSTICE IMPACTS
6.1 Introduction
E.O. 12898 directs 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 section.
Additionally, E.O. 13985 was signed to advance racial equity and support underserved
communities through Federal government actions (86 FR 7009, January 20, 2021). Most
recently, E.O. 14096 (88 FR 25251, April 26, 2023) strengthens the directives for achieving
environmental justice that are set out in E.O. 12898. EPA defines EJ as the just treatment and
meaningful involvement of all people regardless of race, color, national origin, Tribal affiliation,
disability, or income with respect to the development, implementation, and enforcement of
environmental laws, regulations, and policies. EPA further defines the term just 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,
governmental, and commercial operations or programs and policies."89 Meaningful involvement
means that: (1) potentially affected populations have an appropriate opportunity to participate in
decisions about a proposed 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.90 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
89 See, e.g., "Environmental Justice." EPA.gov, U.S. Environmental Protection Agency, 4 Mar. 2021,
https://www.epa.gov/environmentaljustice.
90 See https://www.epa.gov/environmentaljustice/technical-guidance-assessing-environmental-justice-regulatory-
analysis.
6-1
-------
groups of concern for both the baseline and regulatory options, using the best available
information (both quantitative and qualitative) to inform the decision-maker and the public.
The Presidential Memorandum on Modernizing Regulatory Review (86 FR 7223;
January 20, 2021) calls for procedures to "take into account the distributional consequences of
regulations, including as part of a quantitative or qualitative analysis of the costs and benefits of
regulations, to ensure that regulatory initiatives appropriately benefit, and do not inappropriately
burden disadvantaged, vulnerable, or marginalized communities." Under E.O. 13563, federal
agencies may consider equity, human dignity, fairness, and distributional considerations, where
appropriate and permitted by law. For purposes of analyzing regulatory impacts, EPA relies upon
its June 2016 "Technical Guidance for Assessing Environmental Justice in Regulatory
Analysis,"91 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. The Technical Guidance states that a
regulatory action may involve potential EJ concerns if it could: (1) create new disproportionate
impacts; (2) exacerbate existing disproportionate impacts; or (3) present opportunities to address
existing disproportionate impacts through the action under development.
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,
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.
91 See https://www. epa.gov/environmentaljustice/technical-guidance-assessing-environmental-justice-regulatory-
analysis.
6-2
-------
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.
6.2 Analyzing E J Impacts in this Final Rule
In addition to the benefits assessment (see Section 4), EPA considers potential EJ
concerns associated with this final rulemaking. A potential EJ concern is defined as "the actual
or potential lack of fair treatment or meaningful involvement of communities with EJ concerns in
the development, implementation and enforcement of environmental laws, regulations and
policies."92 For analytical purposes, this concept refers more specifically to "disproportionate
impacts on communities with EJ concerns that may exist prior to or that may be created by the
final regulatory action." Although EJ concerns for each rulemaking are unique and should be
considered on a case-by-case basis, EPA's EJ Technical Guidance 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, EPA developed an analytical approach that considers the
purpose and specifics of the rulemaking, as well as the nature of known and potential exposures
across various demographic groups. While the final rule targets HAP emissions, other local air
pollutants emissions may also be reduced, such as NOx and SO2. NOx and SO2 emissions can
lead to localized exposures that may be associated with health effects in nearby populations at
sufficiently high concentrations and certain populations may be at increased risk of exposure-
related health effects, such as people with asthma.
92 See https://www. epa.gov/environmentaljustice/technical-guidance-assessing-environmental-justice-regulatory-
analysis.
6-3
-------
As HAP exposure results generated as part of the 2020 Residual Risk analysis were
below both the presumptive acceptable cancer risk threshold and the noncancer health
benchmarks, and this final regulation should further reduce exposure to HAP, there are no
'disproportionate and adverse effects' of potential EJ concern. Therefore, we did not perform a
quantitative EJ assessment of HAP risk. In addition, technical limitations prevented analysis of
NOx and SO2 emission reductions. While HAP, NO2, and SO2 exposures and concentrations
were not directly evaluated as part of this EJ assessment, due to the potential for reductions in
these and other environmental stressors nearby affected sources, EPA qualitatively discussed EJ
impacts of HAP (Section 6.3) and conducted a proximity analysis to evaluate the potential EJ
implications of changes in localized exposures (Section 6.4).93
As this final rule is also expected to reduce ambient PM2.5 and ozone concentrations,
EPA conducted a quantitative analysis of modeled changes in PM2.5 and ozone concentrations
across the continental U.S. resulting from the control strategies projected to occur under the rule,
characterizing aggregated and distributional exposures both prior to and following
implementation of the final regulatory option in 2028, 2030, and 2035 (Section 6.5 and 6.7). It is
important to note that due to the small magnitude of underlying emissions changes, and the
corresponding small magnitude of the ozone and PM2.5 concentration changes, the rule is
expected to have only a small impact on the distribution of exposures across each demographic
group. As the final rule is also focused on climate impacts resulting from emissions reductions
directly targeted in this rulemaking, EPA qualitatively discussed climate impacts in Section 6.6.
Unique limitations and uncertainties are specific to each type of analysis, which are
described prior to presentation of analytic results in the subsections below.
6.3 Qualitative Assessment of HAP Impacts
As required by section 112(n)(l)(A) of the CAA, EPA has determined that it is
appropriate and necessary to regulate HAP emissions from coal- and oil-fired EGUs. This
determination was driven by the significant public health risks and harms posed by prior levels
of EGU emissions as evaluated against the availability and costs of emissions controls that could
be employed to reduce this harmful pollution. As part of the appropriate and necessary
93 The 2016 NOx ISA and 2017 SOx ISA identified people with asthma, children, and older adults as being at
increased risk of NO2- and SCh-related health effects and the 2017 SOx ISA.
6-4
-------
determination, the Administrator specifically considered the impacts of EGU HAP emissions on
different populations and concluded that certain parts of the U.S. population may be especially
vulnerable to Hg emissions based on their characteristics or circumstances. In some cases, the
enhanced vulnerability relates to life stage (e.g., fetuses, infants, young children). In other cases,
the enhanced vulnerability can be ascribed to the communities in which the population lives. In
this second category, the greater sensitivity to HAP emissions can be attributed to poorer levels
of overall health (e.g., higher rates of cardiovascular disease, nutritional deficiencies) or to
dietary practices which are more common in some low-income communities of color (e.g.,
subsistence fishers). The net effect is that certain sub-populations may be especially vulnerable
to EGU HAP emissions and that these emissions are a potential EJ concern.
Of the HAP potentially impacted by this final rulemaking, Hg is a persistent and
bioaccumulative toxic metal that can be readily transported and deposited to soil and aquatic
environments where it is transformed by microbial action into MeHg.94 Consumption of fish is
the primary pathway for human exposure to MeHg. MeHg bioaccumulates in the aquatic food
web eventually resulting in highly concentrated levels of MeHg within larger fish.95 A NAS
Study reviewed the effects of MeHg on human health and concluded that it is highly toxic to
multiple human and animal organ systems. Of particular concern is chronic prenatal exposure via
maternal consumption of foods containing MeHg. Elevated exposure has been associated with
developmental neurotoxicity and manifests as poor performance on neurobehavioral tests,
particularly on tests of attention, fine motor function, language, verbal memory, and visual-
spatial ability. Because the impacts of the neurodevelopmental effects of MeHg are greatest
during periods of rapid brain development, developing fetuses, infants, and young children are
particularly vulnerable. In particular, children born to populations with high fish consumption
(e.g., people consuming fish as a dietary staple) or impaired nutritional status may be especially
susceptible to adverse neurodevelopmental outcomes. As part of the 2023 Final A&N Review,
EPA evaluated how the neurodevelopmental and cardiovascular risks varied across populations.
That analysis completed in support of the appropriate and necessary determination (addressing
the EGU sector collectively) suggested that subsistence fisher populations that are racially,
94 U.S. EPA. 1997. Mercury Study Report to Congress. EPA-452/R-97-003 December 1997.
95 National Research Council (NAS). 2000. Toxicological Effects of MeHg. Committee on the Toxicological Effects
of MeHg, Board on Environmental Studies and Toxicology, National Research Council.
6-5
-------
culturally, geographically, and/or income-differentiated could experience elevated exposures
relative to not only the general population but also the population of subsistence fishers
generally. As noted in Section 4 of this document, while previous EPA assessments have shown
that current modeled exposures are well below the RfD, we conclude that further reductions in
Hg emissions from lignite-fired EGUs covered in this final action should further reduce
exposures for the subsistence fisher sub-population. However, as we do not expect appreciable
adverse health effects as a result of HAP emissions from this source category, we have not
conducted quantitative or qualitative analyses to assess specific Hg-related impacts of this action
for EJ communities of potential concern or how those impacts differ from U.S. population-wide
effects.
6.4 Demographic Proximity Analyses of Existing Facilities
Demographic proximity analyses allow one to assess the potentially vulnerable
populations residing near affected facilities as a proxy for exposure and the potential for adverse
health impacts that may occur at a local scale due to economic activity at a given location
including noise, odors, traffic, and emissions such as NO2 and SO2 covered under this EPA
action and not modeled elsewhere in this RIA.
Although baseline proximity analyses are presented here, several important caveats
should be noted. Emissions are expected to both decrease and increase from the rulemaking in
the three modeled future years, so communities near affected facilities could experience either
improvements or worsening in air quality from directly emitted pollutants. It should also be
noted that facilities may vary widely in terms of the impacts they already pose to nearby
populations. In addition, proximity to affected facilities does not capture variation in baseline
exposure across communities, nor does it indicate that any exposures or impacts will occur and
should not be interpreted as a direct measure of exposure or impact. These points limit the
usefulness of proximity analyses when attempting to answer questions from EPA's EJ Technical
Guidance.
Demographic proximity analyses were performed for all plants with at least one coal-
fired unit greater than 25 MW without retirement or gas conversion plans before 2029 affected
by this final rulemaking. Due to the distinct regulatory requirements, the following subsets of
affected facilities were separately evaluated:
6-6
-------
• Coal plants with units potentially impacted by the final Hg standard revision (12
facilities): Comparison of the percentage of various populations (race/ethnicity, age,
education, poverty status, income, and linguistic isolation) living near the facilities to
average national levels.
• Coal plants with units potentially impacted by the final fPM standard revision (21
facilities): Comparison of the percentage of various populations (race/ethnicity, age,
education, poverty status, income, and linguistic isolation) living near the facilities to
average national levels.
The current analysis identified all census blocks with centroids within a 10-km radius of
the latitude/longitude location of each facility, and then linked each block with census-based
demographic data.96 The total population within a specific radius around each facility is the sum
of the population for every census block within that specified radius, based on each block's
population provided by the 2020 decennial Census.97 Statistics on race, ethnicity, age, education
level, poverty status and linguistic isolation were obtained from the Census' American
Community Survey (ACS) 5-year averages for 2016-2020. These data are provided at the block
group level. For the purposes of this analysis, the demographic characteristics of a given block
group - that is, the percentage of people in different races/ethnicities, the percentage without a
high school diploma, the percentage that are below the poverty level, the percentage that are
below two times the poverty level, and the percentage that are linguistically isolated - are
presumed to also describe each census block located within that block group.
In addition to facility-specific demographics, the demographic composition of the total
population within the specified radius (e.g., 10 km) for all facilities was also computed (e.g., all
EGUs potentially impacted by the Hg standard revision). In calculating the total populations, to
avoid double-counting, each census block population was only counted once. That is, if a census
block was located within the selected radius (i.e., 10 km) for multiple facilities, the population of
that census block was only counted once in the total population. Finally, this analysis compares
the demographics at each specified radius (i.e., 10 km) to the demographic composition of the
nationwide population.
96 The 10-km distance was determined to be the shortest radius around these units that captured a large enough
population to avoid excessive demographic uncertainty.
97 The location of the Census block centroid is used to determine if the entire population of the Census block is
assumed to be within the specified radius. It is unknown how sensitive these results may be to different methods of
population estimation, such as aerial apportionment.
6-7
-------
Table 6-lFor the population living within 10 km of lignite-fired coal plants potentially
impacted by the Hg standard, the percentage of the population that is American Indian and
Alaska Native Tribes is above the national average (0.9 percent versus 0.6 percent), and the
percentage of the population that is Hispanic/Latino or Other/Multiracial is below the
corresponding national averages. The percentage of the population that is Black, below the
poverty level and below two times the poverty level is similar to the national averages. Finally,
the percentage of the population that is in linguistic isolation is below the national average.
The population living within 10 km of the units potentially impacted by the PM standard
is 86 percent White. The percentage of the population that is below two times the poverty level is
above the national average (32 percent versus 29 percent). The percentage of the population in
the other demographic categories is near or below the national averages.
6-8
-------
Table 6-1 Proximity Demographic Assessment Results Within 10 km of Coal-Fired
Units Greater than 25 MW Without Retirement or Gas Conversion Plans Before 2029
Affected by this Rulemaking a,b
Population within 10 km
Demographic Group
Nationwide Average for
Comparison
Coal plants potentially
impacted by Hg
standard
Coal plants potentially
impacted by fPM
standard
Total Population
329,824,950
17,790
233,575
Number of Facilities
-
12
28
Race and Ethnicity by Percent
White
60%
79%
86%
Black
12%
12%
7%
American Indian and
Alaska Native Tribes
0.60%
0.9%
0.3%
Hispanic or Latino2
19%
5%
5%
Other and Multiracial
9%
2%
3%
Income by Percent
Below Poverty Level
13%
12%
14%
Below 2x Poverty Level
29%
28%
32%
Education by Percent
>25 and w/o a HS
Diploma
12%
13%
12%
Linguistically Isolated by Percent
Linguistically Isolated
5%
2%
1%
a The nationwide population count and all demographic percentages are based on the Census' 2016-2020 American
Community Survey five-year block group averages and include Puerto Rico. Demographic percentages based on
different averages may differ. The total population counts are based on the 2020 Decennial Census block
populations.
b To avoid double counting, the "Hispanic or Latino" category is treated as a distinct demographic category for these
analyses. A person is identified as one of five racial/ethnic categories above: White, Black, American Indian and
Alaska Native Tribes, Other and Multiracial, or Hispanic/Latino. A person who identifies as Hispanic or Latino is
counted as Hispanic/Latino for this analysis, regardless of what race this person may have also identified as in the
Census. Includes white and nonwhite.
6.5 EJ PM2.5 and Ozone Exposure Impacts
This EJ air pollutant exposure98 analysis aims to evaluate the potential for EJ concerns
related to PM2.5 and ozone exposures" among potentially vulnerable populations. To assess EJ
ozone and PM2.5 exposure impacts, we focus on the first and third of the three EJ questions from
98 The term exposure is used here to describe estimated PM2 5 and ozone concentrations and not individual dosage.
99 Air quality surfaces used to estimate exposures are based on 12-km grids. Additional information on air quality
modeling can be found in the air quality modeling information section.
6-9
-------
EPA's 2016 EJ Technical Guidance,100 which ask if there are potential EJ concerns associated
with stressors affected by the regulatory action for population groups of concern in the baseline
and if those potential EJ concerns in the baseline are exacerbated, unchanged, or mitigated under
the regulatory options being considered.101
To address these questions with respect to the PM2.5 and ozone exposures, 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 impacts. Specifically, as 1) this final
rule affects EGUs across the U.S., which typically have tall stacks that result in emissions from
these sources being dispersed over large distances, and 2) both ozone and PM2.5 can undergo
long-range transport, it is appropriate to conduct an EJ assessment of the contiguous U.S. Given
the availability of modeled PM2.5 and ozone air quality surfaces under the baseline and final
regulatory option, we conduct an analysis of changes in PM2.5 and ozone concentrations resulting
from the emission changes projected under the final rule as compared to the baseline scenario,
characterizing average and distributional exposures the analysis years 2028, 2030, and 2035.
However, several important caveats of this analysis are as follows:
• The baseline scenarios for 2028, 2030, and 2035 represent EGU emissions expected in
2028, 2030, and 2035 respectively, but emissions from all other sources are projected to
the year 2026. The 2028, 2030, and 2035 baselines therefore do not capture any
anticipated changes in ambient ozone and PM2.5 between 2026 and 2028, 2030, or 2035
that would occur due to emissions changes from sources other than EGUs.
• Modeling of post-policy air quality concentration changes are based on state-level
emission data paired with facility-level baseline 2026 emissions that were available in the
summer 2021 version of IPM. While the baseline spatial patterns represent ozone and
PM2.5 concentrations associated with the facility level emissions described above, the
post-policy air quality surfaces will capture expected ozone and PM2.5 changes that result
100 U.S. Environmental Protection Agency (EPA), 2015. Guidance on Considering Environmental Justice During the
Development of Regulatory Actions, https://www.epa.gov/sites/default/files/2015-06/documents/considering-ej-in-
rulemaking-guide-final.pdf.
101 EJ question 2 asks if there are potential EJ concerns (i.e., disproportionate burdens across population groups)
associated with environmental stressors affected by the regulatory action for population groups of concern for the
regulatory options under consideration We use the results from questions 1 and 3 to gain insight into the answer to
EJ question 2 in the summary (Section 6.7), for several reasons. Importantly, the total magnitude of differential
exposure burdens with respect to ozone and PM2 5 among population groups at the national scale has been fairly
consistent pre- and post-policy implementation across recent rulemakings. As such, differences in nationally
aggregated exposure burden averages between population groups before and after the rulemaking tend to be very
similar. Therefore, as disparities in pre- and post-policy burden results appear virtually indistinguishable, the
difference attributable to the rulemaking can be more easily observed when viewing the change in exposure impacts,
and as we had limited available time and resources, we chose to provide quantitative results on the pre-policy
baseline and policy-specific impacts only, which related to EJ questions 1 and 3.
6-10
-------
from state-to-state emissions changes but will not capture heterogenous changes in
emissions from multiple facilities within a single state.
• Air quality simulation input information are at a 12-km grid resolution and population
information is either at the Census tract- or county-level, potentially masking impacts at
geographic scales more highly resolved than the input information.
• The two specific air pollutant metrics evaluated in this assessment, warm season
maximum daily eight-hour ozone average concentrations and average annual PM2.5
concentrations, are focused on longer-term exposures that have been linked to adverse
health effects. This assessment does not evaluate disparities in other potentially health-
relevant metrics, such as shorter-term exposures to ozone and PM2.5.
• PM2.5 EJ impacts were limited to exposures, and do not extend to health effects, given
additional uncertainties associated with estimating health effects stratified by
demographic population and the ability to predict differential PIvfc.s-attributable EJ health
impacts.
Population variables considered in this EJ exposure assessment include race, ethnicity,
educational attainment, employment status, health insurance status, life expectancy, linguistic
isolation, poverty status, redlined areas, tribal land, age, and sex (Table 6-2).102,103,104,105 Note that
these variables are different than the proximity analysis because criteria pollutants have
nationwide impacts rather than the localized impacts that are investigated for HAP in the
proximity analysis. There are also fewer demographic uncertainties at a national scale which
allows us to use an expanded set of variables for a nationwide analysis.
102 Population projections stratified by race/ethnicity, age, and sex are based on economic forecasting models
developed by Woods and Poole (2015). The Woods and Poole database contains county-level projections of
population by age, sex, and race out to 2050, relative to a baseline using the 2010 Census data. Population
projections for each county are determined simultaneously with every other county in the U.S to consider patterns of
economic growth and migration. County-level estimates of population percentages within the poverty status and
educational attainment groups were derived from 2015-2019 5-year average ACS estimates. Additional information
can be found in Appendix J of the BenMAP-CE User's Manual (https://www.epa.gov/benmap/benmap-ce-manual-
and-appendices).
103 The Tribal Land variable was also added in response to recent Executive Orders that have emphasized the need
for more detailed analysis on the impacts on American Indians. The Tribal Lands variable focuses specifically on
populations who live on Tribal lands in addition to quantifying those whose race is American Indian but may or may
not live on Tribal lands.
104 EPA acknowledges the recent comments about cumulative risk assessment and is currently in the process of
developing cumulative risk assessment methods for our quantitative environmental justice analyses. In the interim,
this rulemaking utilizes the "life expectancy" and "redlining" variables as a proxy to identify communities with
higher or lower exposure to cumulative risks. EPA continues to improve its methodology based on its framework for
a Cumulative Risk Assessment as well as guidance from multiple Executive Orders and intend to assess cumulative
risk more accurately in future rulemakings.
105 An additional population variable that is not included in this analysis is persons with disability. Persons with
disability is a new environmental justice metric listed inE.O. 14096 (88 FR 25251, April 26, 2023), and EPA is
currently developing analytical techniques/tools to evaluate its impact on our environmental analyses.
6-11
-------
The demographic groups and processing methodology for each dataset are described
below. County-level datasets were generated for 3,109 counties in the contiguous U.S.
Table 6-2 Demographic Populations Included in the PM2.5 and Ozone EJ Exposure
Analyses
Demographic
Groups
Ages
Spatial Scale of
Population Data
Race
Asian; American Indian; Black; White
0-99
Census tract
Ethnicity
Hispanic; Non-Hispanic
0-99
Census tract
Educational
Attainment
High school degree or more; No high school degree
25-99
Census tract
Employment
Status
Employed; Unemployed; Not in the labor force
0-99
County
Health Insurance
Insured; Uninsured
0-64
County
Linguistic
Isolation
Speaks English "very well" or better; Speaks English less
than "very well" OR
Speaks English "well" or better; Speaks English less than
"well"
0-99
Census tract
Poverty Status
Above the poverty line; Below the poverty line OR
Above 2x the poverty line; Below 2x the poverty line
0-99
Census tract
Redlined Areas
HOLC3 Grades A-C; HOLC Grade D; Not graded by
HOLC
0-99
Census tract
Life Expectancy
Top 75%; Bottom 25%
0-99
Census tract
Tribal Land
Tribal land; Not Tribal land
0-99
Census tract
Children
0-17
Age
Adults
Older Adults
18-64
65-99
Census tract
Sex
Female; Male
0-99
Census tract
"Home Owners' Loan Corporation (HOLC)
6.5.1 Populations Predicted to Experience PM2.5 and Ozone Air Quality Changes
While EPA projects the final rule will lead to both decreases and increases in emissions
in different regions, the magnitude of the air pollution exposure changes from the final rule is
quite small across the three future years analyzed. For all three future years evaluated, there were
no discernable PM2.5 or ozone concentration changes out to the hundredths digit, reiterating the
small magnitude of national average PM2.5 or ozone changes (Figure 6-1 and Figure 6-2).
6.5.2 PM2.5 EJ Exposure Analysis
We evaluated the potential for EJ concerns among potentially vulnerable populations
resulting from exposure to PM2.5 under the baseline and final regulatory option in this rule. This
was done by characterizing the projected distribution of PM2.5 exposures both prior to and
following implementation of the final rule in 2028, 2030, and 2035.
6-12
-------
As this analysis is based on the same PM2.5 spatial fields as the benefits assessment (see
Appendix A for a discussion of the spatial fields), it is subject to similar types of uncertainty (see
Section 4.3.8 for a discussion of the uncertainty). A particularly germane limitation for this
analysis is that the magnitude of the expected concentration changes is quite small, likely making
uncertainties associated with the various input data more relevant.
6.5.2.1 National Aggregated Results
National average baseline PM2.5 concentrations in micrograms per cubic meter (|ig/m3) in
2028, 2030, and 2035 are shown in the colored column labeled "baseline" in the Figure 6-1 heat
map. Concentrations in the "baseline" columns represent the total estimated PM2.5 exposure
burden averaged over the 12-month calendar year and are colored to visualize differences more
easily in average concentrations (lighter blue coloring representing smaller average
concentrations and darker blue coloring representing larger average concentrations). Average
national disparities observed in the baseline of this rule are similar to those described by recent
rules (e.g., the Final PMNAAQS), that is, populations with national average PM2.5
concentrations higher than the reference population ordered from most to least difference were:
residents of HOLC Grade D (i.e., redlined) census tracts, linguistically isolated, residents of
HOLC Grade A-C (i.e., not redlined) census tracts, Hispanic individuals, Asian individuals,
those without a high school diploma, Black individuals, below the poverty line, the unemployed,
and the uninsured. Average national disparities observed in the baseline of this rule are generally
consistent across the three future years and similar to those described by recent rules (e.g., the
Final PMNAAQS).
For all three future years evaluated, there were no discernable PM2.5 changes under the
final regulatory option for any population analyzed when showing concentrations out to the
hundredths digit, reiterating the small magnitude of national average PM2.5 changes.
The national-level assessment of PM2.5 before and after implementation of this final
rulemaking suggests that while EJ exposure disparities are present in the pre-policy scenario, EJ
exposure concerns are not likely created or exacerbated by the rule for the population groups
evaluated, due to the small magnitude of the PM2.5 concentration reductions. It is also important
to note that at the national-level the PM2.5 concentrations before and after implementation for all
6-13
-------
three future years evaluated the concentrations for each demographic group are below the
recently revised standard of 9 ug/m' . ""
2028
2030
2035
Group
Population
Baseline
Absolute
Reductions
Baseline
Absolute
Reductions
Baseline
Absolute
Reductions
Reference
Reference (0-99)
7.16
0.00
7.11
0.00
7.08
0.00
American Indian (0-99)
6.69
0.00
6.66
0.00
6.64
0.00
Race
Asian (0-99)
7.73
0.00
7.67
0.00
7.62
0.00
Black (0-99)
7.41
0.00
7.35
0.00
7.29
0.00
White (0-99)
7.07
0.00
7.02
0.00
7.00
0.00
Ethnicity
Hispanic (0-99)
7.94
0.00
7.90
0.00
7.85
0.00
Non-Hispanic (0-99)
6.94
0.00
6.89
0.00
6.85
0.00
Educational
Less educated (>24; no HS)
7.49
0.00
7.44
0.00
7.43
0.00
Attainment
More educated (>24: HS or more)
7.06
0.00
7.01
0.00
6.99
0.00
Employment
Status
Employed (0-99)
7.15
0.00
7.10
0.00
7.07
0.00
Not in the labor force (0-99)
Unemployed (0-99)
7.16
7.31
0.00
0.00
7.11
7.26
0.00
0.00
7.08
7.24
0.00
0.00
Insurance
Insured (0-64)
7.20
0.00
7.15
0.00
7.12
0.00
Status
Uninsured (0-64)
7.27
0.00
7.23
0.00
7.20
0.00
Linquistic
English < well (0-99)
S.09
0.00
8.05
0.00
8.04
0.00
Isolation
English well or better (0-99)
7.11
0.00
7.06
0.00
7.04
0.00
Life
Expectancy
Bottom 25% life expectancy (0-99)
7.20
0.00
7.13
0.00
7.10
0.00
Life expectancy data unavailable (0-99)
Top 75% life expectancy (0-99)
7.11
7.15
0.00
0.00
7.07
7.10
0.00
0.00
7.04
7.08
0.00
0.00
Poverty
Poverty line (0-99)
7.12
0.00
7.08
0.00
7.05
0.00
Redlined
Areas
HOLC Grade D (0-99)
HOLC Grades A-C (0-99)
8.20
7.95
0.00
0.00
8,15
7.90
0.00
0.00
8.12
7.86
0.00
0.00
Not Graded by HOLC (0-99)
6.99
0.00
6.94
0.00
6.92
0.00
Tribal Land
NotTribal land (0-99)
7.16
0.00
7.11
0.00
7.09
0.00
Designation
Tribal land (0-99)
6.63
0.00
6.58
0.00
6.53
0.00
Adults (IS-64)
7.20
0.00
7.15
0.00
7.13
0.00
Ages
Children (0-17)
7.22
0.00
7.17
0.00
7.14
0.00
Older Adults (65-99)
6.94
0.00
6.90
0.00
6.89
0.00
Sex
Females (0-99)
7.17
0.00
7.12
o.oo
7.09
0.00
Males (0-99)
7.14
0.00
7.10
0.00
7.07
0.00
Figure 6-1 Heat Map of the National Average PM2.5 Concentrations in the Baseline and
Reductions in Concentrations Due to the Final Regulatory Option Across Demographic
Groups in 2028, 2030, and 2035 (jig/m3)
6.5.2.2 State Aggregated Results
We also assess PM2.5 concentration reductions by state and demographic population in
2028, 2030, and 2035 for the 48 states in the contiguous U.S, for the final rule
See https'/Zwrnv.epa.gov/systern/file.sMocuments/2024-02/pm-naaqs-fmal-frn-pre-pubIication.pdf
6-14
-------
The magnitude of state-level PM2.5 concentration changes under the final regulatory
option is not discernable out to the hundredths digit, reiterating the small magnitude of state-
level average PM2.5 changes. The small magnitude of differential PM2.5 exposure impacts
expected by the final rule is not likely to exacerbate or mitigate EJ concerns within individual
states.
6.5.2.3 Distributional Results
We also assess the cumulative proportion of each population exposed to ascending levels
of PM2.5 concentration changes across the contiguous U.S. Results allow evaluation of what
percentage of each subpopulation (e.g., Hispanics) in the contiguous U.S. experience what
change in PM2.5 concentrations compared to what percentage of the overall reference group (i.e.,
the total population of contiguous U.S.) experiences similar concentration changes from EGU
emission changes under the final regulatory option in 2028, 2030, and 2035.
This distributional EJ analysis is also subject to additional uncertainties related to more
highly resolved input parameters and additional assumptions. For example, this analysis does not
account for potential difference in underlying susceptibility, vulnerability, or risk factors across
populations to PM2.5 exposure. Nor could we include information about differences in other
factors that could affect the likelihood of adverse impacts (e.g., exercise patterns) across groups.
Therefore, this analysis should not be used to assert that there are meaningful differences in
PM2.5 exposure impacts associated with either the baseline or the rule across population groups.
As the baseline scenario is similar to that described by other RIAs, we focus on the PM2.5
changes due to this final rulemaking. Distributions of 12-km gridded PM2.5 concentration
changes from EGU control strategies of affected facilities analyzed for the years 2028, 2030, and
2035 were evaluated.
The vast majority of PM2.5 concentration changes for each population distribution round
to 0.00 |ig/m3 under the final regulatory option for all three future years analyzed. Therefore,
there are no discernable differences in impacts in the distributional analyses of PM2.5
concentration changes under the final regulatory option, which provides additional evidence that
the final rule is not likely to exacerbate or mitigate EJ PM2.5 exposure concerns for population
groups evaluated.
6-15
-------
6.5.3 Ozone EJ Exposure Analysis
To evaluate the potential for EJ concerns among potentially vulnerable populations
resulting from exposure to ozone under the baseline and final rule, we characterize the projected
distribution of ozone exposures both prior to and following implementation of the final rule in
2028, 2030, and 2035.
As this analysis is based on the same ozone spatial fields as the benefits assessment (see
Appendix A for a discussion of the spatial fields), it is subject to similar types of uncertainty (see
Section 4.3.8 for a discussion of the uncertainty). In addition to the small magnitude of
differential ozone concentration changes associated with this final rulemaking when comparing
across demographic populations, a particularly germane limitation is that ozone, being a
secondary pollutant, is the byproduct of complex atmospheric chemistry such that direct linkages
cannot be made between specific affected facilities and downwind ozone concentration changes
based on available air quality modeling.
Ozone concentration and exposure metrics can take many forms, although only a small
number are commonly used. The analysis presented here is based on the average April-
September warm season maximum daily eight-hour average ozone concentrations (AS-M03),
consistent with the health impact functions used in the benefits assessment (Section 4). As
developing spatial fields is time and resource intensive, the same spatial fields used for the
benefits analysis were also used for the ozone exposure analysis performed here to assess EJ
impacts.
The construct of the AS-M03 ozone metric used for this analysis should be kept in mind
when attempting to relate the results presented here to the ozone NAAQS and when interpreting
the confidence in the association between exposures and health effects. Specifically, the seasonal
average ozone metric used in this analysis is not constructed in a way that directly relates to
NAAQS design values, which are based on daily maximum eight-hour concentrations.107 Thus,
AS-M03 values reflecting seasonal average concentrations well below the level of the NAAQS
at a particular location do not necessarily indicate that the location does not experience any daily
107 Level of 70 ppb with an annual fourth-highest daily maximum eight-hour concentration, averaged over three
years.
6-16
-------
(eight-hour) exceedances of the ozone NAAQS. Relatedly, EPA is confident that reducing the
highest ambient ozone concentrations will result in substantial improvements in public health,
including reducing the risk of ozone-associated mortality. However, the Agency is less certain
about the public health implications of changes in relatively low ambient ozone concentrations.
Most health studies rely on a metric such as the warm-season average ozone concentration; as a
result, EPA typically utilizes air quality inputs such as the AS-M03 spatial fields in the benefits
assessment, and we judge them also to be the best available air quality inputs for this EJ ozone
exposure assessment.
6.5.3.1 National Aggregated Results
National average baseline ozone concentrations in ppb in 2028, 2030, and 2035 are
shown in the colored column labeled "baseline" in the heat map (Figure 6-2). Concentrations in
the "baseline" columns represent the total estimated daily eight-hour maximum ozone exposure
burden averaged over the six-month April-September ozone season and are colored to visualize
differences more easily in average concentrations, with lighter green coloring representing
smaller average concentrations and darker green coloring representing larger average
concentrations. Populations with national average ozone concentrations higher than the reference
population ordered from most to least difference were: American Indian individuals, Hispanic
individuals, those who are linguistically isolated, residents of Tribal Lands, Asian individuals,
residents of HOLC Grades A-C (i.e., not redlined) census tracts, those without a high school
diploma, the unemployed, populations with higher life expectancy or with life expectancy data
unavailable, children, residents of HOLC Grade D (i.e., redlined) census tracts, and the insured.
Average national disparities observed in the baseline of this rule are fairly consistent across the
three future years and similar to those described by recent rules (e.g., the RIA for the Final
GNP).
For all three future years evaluated, there were no discernable ozone changes under the
final rule for any population analyzed when showing concentrations out to the hundredths digit,
reiterating the small magnitude of national average ozone changes.
The national-level assessment of ozone burden concentrations in the baseline and ozone
exposure changes due to the final rule suggests that while EJ exposure disparities are present in
the pre-policy scenario, EJ exposure concerns are not likely created or exacerbated by the rule
6-17
-------
for the population groups evaluated, due to the small magnitude of the ozone concentration
changes. Note that while we were able to compare the annual average PM2.5 concentrations to the
newly revised NAAQS, the estimated ozone impacts in terms of annual average change are
difficult to compare to the ozone NAAQS as the annual fourth-highest daily maximum 8-hour
concentration.
2028
2030
2035
Group
Population
Baseline
Absolute
Reductions
Baseline
Absolute
Reductions
Baseline
Absolute
Reductions
Reference
Reference (0-99)
40.25
0.00
40.21
0.00
40.01
0.00
American Indian (0-99)
42.61
0.00
42.57
0.00
42.41
0.00
Race
Asian (0-99)
41.61
0.00
41.54
0.00
41.27
0.00
Black (0-99)
33.86
0.00
33.81
0.00
33.56
0.00
White (0-99)
40.35
0.00
40.31
o.oo
40.12
0.00
Ethnicity
Hispanic (0-99)
42.50
0.00
42.44
0.00
42.20
0.00
Non-Hispanic (0-99)
39.64
0.00
39.58
0.00
39.34
0.00
Educational
Less educated (>24; no HS)
40.75
0.00
40.72
0.00
40.55
0.00
Attainment
More educated (>24: HS or more)
40.07
0.00
40.03
0.00
39.83
0.00
Employment
Status
Employed (0-99)
Not in the labor force (0-99)
Unemployed (0-99)
40.26
40.23
40.70
0.00
0.00
0.00
40.21
40.19
40.66
0.00
0.00
0.00
40.01
39.99
40.48
0.00
0.00
0.00
Insurance
Insured (0-64)
40.40
0.00
40.36
0.00
40.16
0.00
Status
Uninsured (0-64)
39.97
0.00
39.93
0.00
39.71
0.00
Linquistic
English < well (0-99)
41.86
0.00
41.82
0.00
41.64
0.00
Isolation
English well or better (0-99)
40.18
0.00
40.13
0.00
39.93
0.00
Life
Expectancy
Bottom 25% life expectancy (0-99)
39.11
0.00
39.07
0.00
38.84
0.00
Life expectancy data unavailable (0-99)
Top 75% life expectancy (0-99)
40.56
40.54
0.00
0.00
40.51
40.50
0.00
0.00
40.32
40.31
0.00
0.00
Poverty
Povertyline (0-99)
40.25
0.00
40.21
0.00
40.01
0.00
Redlined
Areas
HOLC Grade D (0-99)
40.44
0.00
40.40
0.00
40.16
0.00
HOLC Grades A-C (0-99)
41.18
0.00
41.14
0.00
40.89
0.00
Not Graded by HOLC (0-99)
40.11
0.00
40.07
0.00
39.88
0.00
Tribal Land
NotTribal land (0-99)
40.24
0.00
40.20
0.00
40.00
0.00
Designation
Tribal land (0-99)
41.64
0.00
41.58
0.00
41.24
0.00
Adults (18-64)
40.30
0.00
40.27
0.00
40.07
0.00
Ages
Children (0-17)
40.47
0.00
40.43
o.oo
40.23
0.00
Older Adults (65-99)
39.85
0.00
39.81
0.00
39.63
0.00
Sex
Females (0-99)
40.24
0.00
40.20
0.00
40.00
0.00
Males (0-99)
40.27
0.00
40.22
0.00
40.03
0.00
Figure 6-2 Heat Map of the National Average Ozone Concentrations in the Baseline and
Reductions in Concentrations under the Final Rule Across Demographic Groups in 2028,
2030, and 2035 (ppb)
6.5.3.2 State Aggregated Results
We also provide ozone concentration reductions by state and demographic population in
2028, 2030, and 2035 for the 48 states in the contiguous U.S, for the final regulatory option
6-18
-------
(Figure 6-3). In this heat map, dark blue indicates larger ozone reductions, with demographic
groups shown as rows and each state as a column.
The magnitude of state-level ozone concentration changes under the final regulatory
option is very small, with the vast majority of state-level ozone concentrations changes not
discernable out to the hundredths digit. State-level average populations that are projected to
experience reductions in ozone concentrations by up to 0.01 ppb are residents of HOLC Grade D
(i.e., redlined) census tracts and Black individuals in Arkansas (AR), and most population groups
in North Dakota (ND). Only state-level average reductions in ozone concentrations were
observed for populations in 2028. The small magnitude of differential ozone exposure impacts
expected by the final rule is not likely to exacerbate or mitigate EJ concerns within individual
states.
Y8ar p°P"lation ji3S8bQQs;gs9=igaS5llliillgzizl5lzzoooS5^ggg5>&gigg ozone(PPb)
Reference (0-99)
American Indian (0-99)
Asian (0-99)
Black (0-99)
White (0-99)
Hispanic (0-99)
2028 LeSS ec'ucatGC' (>24; no HS)
Unemployed (0-99)
Unisured (0-64)
Bottom 25% life expectancy (0-9
English < well (0-99)
-------
populations expected to experience post-policy ozone exposure changes. Nor could we include
information about differences in other factors that could affect the likelihood of adverse impacts
(e.g., exercise patterns) across groups. Therefore, this analysis should not be used to assert that
there are meaningful differences in ozone exposures impacts in either the baseline or the rule
across population groups.
As the baseline scenario is similar to that described by other RIAs, we focus on the ozone
changes due to this final rulemaking. Distributions of 12-km gridded ozone concentration
changes from EGU control strategies of affected facilities under the final rule were evaluated.
The vast majority of ozone concentration changes round to 0.00 ppb under the final
regulatory option for all three future years analyzed. Therefore, there are no discernable
differences in impacts in the distribution of ozone concentration changes across population
demographics under the final regulatory option. This also provides additional evidence that the
final rule is not likely to exacerbate or mitigate EJ ozone exposure concerns for population
groups evaluated.
6.6 GHG Impacts on Environmental Justice and other Populations of Concern
In the 2009 Endangerment Finding, the Administrator considered how climate change
threatens the health and welfare of the U.S. population. As part of that consideration, she also
considered risks to people of color and low-income individuals and communities, finding that
certain parts of the U.S. population may be especially vulnerable based on their characteristics or
circumstances. These groups include economically and socially disadvantaged communities;
individuals at vulnerable life stages, such as the elderly, the very young, and pregnant or nursing
women; those already in poor health or with comorbidities; persons with disabilities; those
experiencing homelessness, mental illness, or substance abuse; and Indigenous or other
populations dependent on one or limited resources for subsistence due to factors including but
not limited to geography, access, and mobility.
Scientific assessment reports produced over the past decade by the U.S. Global Change
Research Program (USGCRP), the IPCC, the National Research Council, and the National
Academies of Science, Engineering, and Medicine add more evidence that the impacts of climate
change raise potential EJ concerns (IPCC, 2018; National Academies, 2017; National Research
6-20
-------
Council, 2011; Oppenheimer et al., 2014; Porter et al., 2014; Smith et al., 2014; U.S. EPA, 2021;
USGCRP, 2016, 2018). These reports conclude that less-affluent, traditionally marginalized, and
predominantly non-White communities can be especially vulnerable to climate change impacts
because they tend to have limited resources for adaptation, are more dependent on climate-
sensitive resources such as local water and food supplies or have less access to social and
information resources. Some communities of color, specifically populations defined jointly by
ethnic/racial characteristics and geographic location (e.g., African-American, Black, and
Hispanic/Latino communities; individuals who identify as Native American, particularly those
living on tribal lands and Alaska Natives), may be uniquely vulnerable to climate change health
impacts in the U.S., as discussed below. In particular, the 2016 scientific assessment on the
Impacts of Climate Change on Human Health found with high confidence that vulnerabilities are
place- and time-specific, lifestages and ages are linked to immediate and future health impacts,
and social determinants of health are linked to greater extent and severity of climate change-
related health impacts (USGCRP, 2016).
Per the Fourth National Climate Assessment (NCA4), "Climate change affects human
health by altering exposures to heat waves, floods, droughts, and other extreme events; vector-,
food- and waterborne infectious diseases; changes in the quality and safety of air, food, and
water; and stresses to mental health and well-being" (Ebi et al., 2018). Many health conditions
such as cardiopulmonary or respiratory illness and other health impacts are associated with and
exacerbated by an increase in GHGs and climate change outcomes, which is problematic as these
diseases occur at higher rates within vulnerable communities. Importantly, negative public health
outcomes include those that are physical in nature, as well as mental, emotional, social, and
economic.
The scientific assessment literature, including the aforementioned reports, demonstrates
that there are myriad ways in which these populations may be affected at the individual and
community levels. Individuals face differential exposure to criteria pollutants, in part due to the
proximities of highways, trains, factories, and other major sources of pollutant-emitting sources
to less-affluent residential areas. Outdoor workers, such as construction or utility crews and
agricultural laborers, who frequently are comprised of already at-risk groups, are exposed to poor
air quality and extreme temperatures without relief. Furthermore, people in communities with EJ
concerns face greater housing, clean water, and food insecurity and bear disproportionate and
6-21
-------
adverse economic impacts and health burdens associated with climate change effects. They have
less or limited access to healthcare and affordable, adequate health or homeowner insurance
(USGCRP, 2016). Finally, resiliency and adaptation are more difficult for economically
vulnerable communities; these communities have less liquidity, individually and collectively, to
move or to make the types of infrastructure or policy changes to limit or reduce the hazards they
face. They frequently are less able to self-advocate for resources that would otherwise aid in
building resilience and hazard reduction and mitigation.
The assessment literature cited in EPA's 2009 and 2016 Endangerment and Cause or
Contribute Findings, as well as Impacts of Climate Change on Human Health, also concluded
that certain populations and life stages, including children, are most vulnerable to climate-related
health effects (USGCRP, 2016). The assessment literature produced from 2016 to the present
strengthens these conclusions by providing more detailed findings regarding related
vulnerabilities and the projected impacts youth may experience. These assessments - including
the Fourth National Climate Assessment (USGCRP, 2018) and The Impacts of Climate Change
on Human Health in the United States (USGCRP, 2016) - describe how children's unique
physiological and developmental factors contribute to making them particularly vulnerable to
climate change. Impacts to children are expected from heat waves, air pollution, infectious and
waterborne illnesses, and mental health effects resulting from extreme weather events
(USGCRP, 2016). In addition, children are among those especially susceptible to allergens, as
well as health effects associated with heat waves, storms, and floods. Additional health concerns
may arise in low-income households, especially those with children, if climate change reduces
food availability and increases prices, leading to food insecurity within households. More
generally, these reports note that extreme weather and flooding can cause or exacerbate poor
health outcomes by affecting mental health because of stress; contributing to or worsening
existing conditions, again due to stress or also as a consequence of exposures to water and air
pollutants; or by impacting hospital and emergency services operations (Ebi et al., 2018).
Further, in urban areas in particular, flooding can have significant economic consequences due to
effects on infrastructure, pollutant exposures, and drowning dangers. The ability to withstand and
recover from flooding is dependent in part on the social vulnerability of the affected population
and individuals experiencing an event (National Academy of Sciences, 2019). In addition,
children are among those especially susceptible to allergens, as well as health effects associated
6-22
-------
with heat waves, storms, and floods. Additional health concerns may arise in low-income
households, especially those with children, if climate change reduces food availability and
increases prices, leading to food insecurity within households.
The Impacts of Climate Change on Human Health also found that some communities of
color, low-income groups, people with limited English proficiency, and certain immigrant groups
(especially those who are undocumented) are subject to many factors that contribute to
vulnerability to the health impacts of climate change (USGCRP, 2016). While difficult to isolate
from related socioeconomic factors, race appears to be an important factor in vulnerability to
climate-related stress, with elevated risks for mortality from high temperatures reported for
Black or African American individuals compared to White individuals after controlling for
factors such as air conditioning use. Moreover, people of color are disproportionately more
exposed to air pollution based on where they live, and disproportionately vulnerable due to
higher baseline prevalence of underlying diseases such as asthma. As explained earlier, climate
change can exacerbate local air pollution conditions, so this increase in air pollution is expected
to have disproportionate and adverse effects on these communities. Locations with greater health
threats include urban areas (due to, among other factors, the "heat island" effect where built
infrastructure and lack of green spaces increases local temperatures), areas where airborne
allergens and other air pollutants already occur at higher levels, and communities that have
experienced depleted water supplies or vulnerable energy and transportation infrastructure.
The 2021 EPA report on climate change and social vulnerability examined four socially
vulnerable groups (individuals who are low income, minority, without high school diplomas,
and/or 65 years and older) and their exposure to several different climate impacts (air quality,
coastal flooding, extreme temperatures, and inland flooding) (U.S. EPA, 2021). This report
found that Black and African-American individuals were 40 percent more likely to currently live
in areas with the highest projected increases in mortality rates due to climate-driven changes in
extreme temperatures, and 34 percent more likely to live in areas with the highest projected
increases in childhood asthma diagnoses due to climate-driven changes in particulate air
pollution. The report found that Hispanic and Latino individuals are 43 percent more likely to
live in areas with the highest projected labor hour losses in weather-exposed industries due to
climate-driven warming, and 50 percent more likely to live in coastal areas with the highest
projected increases in traffic delays due to increases in high-tide flooding. The report found that
6-23
-------
American Indian and Alaska Native individuals are 48 percent more likely to live in areas where
the highest percentage of land is projected to be inundated due to sea level rise, and 37 percent
more likely to live in areas with high projected labor hour losses. Asian individuals were found
to be 23 percent more likely to live in coastal areas with projected increases in traffic delays
from high-tide flooding. Persons with low income or no high school diploma are about 25
percent more likely to live in areas with high projected losses of labor hours, and 15 percent
more likely to live in areas with the highest projected increases in asthma due to climate-driven
increases in particulate air pollution, and in areas with high projected inundation due to sea level
rise.
In a more recent 2023 report, Climate Change Impacts on Children's Health and Weil-
Being in the U.S., EPA considered the degree to which children's health and well-being may be
impacted by five climate-related environmental hazards—extreme heat, poor air quality, changes
in seasonality, flooding, and different types of infectious diseases (U.S. EPA, 2023). The report
found that children's academic achievement is projected to be reduced by 4-7 percent per child,
as a result of moderate and higher levels of warming, impacting future income levels. The report
also projects increases in the numbers of annual emergency department visits associated with
asthma, and that the number of new asthma diagnoses increases by 4-11 percent due to climate-
driven increases in air pollution relative to current levels. In addition, more than 1 million
children in coastal regions are projected to be temporarily displaced from their homes annually
due to climate-driven flooding, and infectious disease rates are similarly anticipated to rise, with
the number of new Lyme disease cases in children living in 22 states in the eastern and
midwestern U.S. increasing by approximately 3,000-23,000 per year compared to current levels.
Overall, the report confirmed findings of broader climate science assessments that children are
uniquely vulnerable to climate-related impacts and that in many situations, children in the U.S.
who identify as Black, Indigenous, and People of Color, are limited English-speaking, do not
have health insurance, or live in low-income communities may be disproportionately more
exposed to the most severe adverse impacts of climate change.
Indigenous communities face disproportionate and adverse risks from the impacts of
climate change, particularly those communities impacted by degradation of natural and cultural
resources within established reservation boundaries and threats to traditional subsistence
lifestyles. Indigenous communities whose health, economic well-being, and cultural traditions
6-24
-------
depend upon the natural environment will likely be affected by the degradation of ecosystem
goods and services associated with climate change. The IPCC indicates that losses of customs
and historical knowledge may cause communities to be less resilient or adaptable (Porter et al.,
2014). The NCA4 (2018) noted that while Indigenous peoples are diverse and will be impacted
by the climate changes universal to all Americans, there are several ways in which climate
change uniquely threatens Indigenous peoples' livelihoods and economies (Jantarasami et al.,
2018; USGCRP, 2018). In addition, as noted in the following paragraph, there can be
institutional barriers (including policy-based limitations and restrictions) to their management of
water, land, and other natural resources that could impede adaptive measures.
For example, Indigenous agriculture in the Southwest is already being adversely affected
by changing patterns of flooding, drought, dust storms, and rising temperatures leading to
increased soil erosion, irrigation water demand, and decreased crop quality and herd sizes. The
Confederated Tribes of the Umatilla Indian Reservation in the Northwest have identified climate
risks to salmon, elk, deer, roots, and huckleberry habitat. Housing and sanitary water supply
infrastructure are vulnerable to disruption from extreme precipitation events. Native Americans'
ability to respond to these conditions is impeded by limitations imposed by statutes including the
Dawes Act of 1887 and the Indian Reorganization Act of 1934, which ultimately restrict
Indigenous peoples' autonomy regarding land-management decisions through Federal trusteeship
of certain tribal lands and mandated Federal oversight of these peoples' management decisions.
Additionally, NCA4 noted that Indigenous peoples generally are subjected to institutional racism
effects, such as poor infrastructure, diminished access to quality healthcare, and greater risk of
exposure to pollutants. Consequently, Native Americans often have disproportionately higher
rates of asthma, cardiovascular disease, Alzheimer's disease, diabetes, and obesity. These health
conditions and related effects (disorientation, heightened exposure to PM2.5, etc.) can all
contribute to increased vulnerability to climate-driven extreme heat and air pollution events,
which also may be exacerbated by stressful situations, such as extreme weather events, wildfires,
and other circumstances.
NCA4 and IPCC's Fifth Assessment Report also highlighted several impacts specific to
Alaskan Indigenous Peoples (Porter et al., 2014). Coastal erosion and permafrost thaw will lead
to more coastal erosion, rendering winter travel riskier and exacerbating damage to buildings,
roads, and other infrastructure—impacts on archaeological sites, structures, and objects that will
6-25
-------
lead to a loss of cultural heritage for Alaska's Indigenous people. In terms of food security, the
NCA4 discussed reductions in suitable ice conditions for hunting, warmer temperatures
impairing the use of traditional ice cellars for food storage, and declining shellfish populations
due to warming and acidification. While the NCA4 also noted that climate change provided more
opportunity to hunt from boats later in the fall season or earlier in the spring, the assessment
found that the net impact was an overall decrease in food security.
6.7 Summary
As with all EJ analyses, data limitations make it quite 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. For
example, here we provide qualitative EJ assessment of ozone and PM2.5 concentration changes
from this rule but can only qualitatively discuss EJ impacts of CO2 emission reductions.
Therefore, this analysis is only a partial representation of the distributions of potential impacts.
Additionally, EJ concerns for each rulemaking are unique and should be considered on a case-
by-case basis, so results similar to those presented here should not be assumed for other
rulemakings.
For the rule, we quantitatively evaluate the proximity of affected facilities populations of
potential EJ concern (Section 6.4) and the potential for disproportionate pre- and policy-policy
PM2.5 and ozone exposures across different demographic groups (Section 6.5). As exposure
results generated as part of the 2020 Residual Risk analysis were below both the presumptive
acceptable cancer risk threshold and the noncancer health benchmarks, and this final regulation
should still reduce exposure to HAP, there are no 'disproportionate and adverse effects' of
potential EJ concern. Therefore, we did not perform a quantitative EJ assessment of HAP risk.
Each of these analyses presented depend on mutually exclusive assumptions, was performed to
answer separate questions, and is associated with unique limitations and uncertainties.
Baseline demographic proximity analyses provide information as to whether there may
be potential EJ concerns associated with local environmental stressors such as local NO2 and SO2
emitted from sources affected by the regulatory action, traffic, or noise for certain population
groups of concern in the baseline (Section 6.4). The baseline demographic proximity analyses
examined the demographics of populations living within 10 km of the following sources: lignite-
6-26
-------
fired coal plants with units potentially impacted by the Hg standard revision and coal plants with
units potentially impacted by the fPM standard revision. The proximity demographic analysis
indicates that on average, the population living within 10 km of coal plants potentially impacted
by the fPM standards shas a higher percentage of people living below two times the poverty level
than the national average. In addition, on average the percentage of the Native American
population living within 10 km of lignite-fired coal plants potentially impacted by Hg standard is
higher than the national average. Relating these results to question 1 from Section 6.3, we
conclude that there may be potential EJ concerns associated with directly emitted pollutants that
are affected by the regulatory action (e.g., local NOx or SO2) for certain population groups of
concern in the baseline (question 1). However, as proximity to affected facilities does not capture
variation in baseline exposure across communities, nor does it indicate that any exposures or
impacts will occur, these results should not be interpreted as a direct measure of exposure or
impact.
While the demographic proximity analyses may appear to parallel the baseline analysis of
nationwide ozone and PM2.5 exposures in certain ways, the two should not be directly compared.
The baseline ozone and PM2.5 exposure assessments are in effect an analysis of total burden in
the contiguous U.S., and include various assumptions, such as the implementation of
promulgated regulations. It serves as a starting point for both the estimated ozone and PM2.5
changes due to this final rule as well as a snapshot of air pollution concentrations in the near
future. This final rule is also expected to reduce emissions of direct PM2.5, NOx, and SO2
nationally throughout the year. Because NOx and SO2 are also precursors to secondary formation
of ambient PM2.5 and NOx is a precursor to ozone formation, reducing these emissions would
impact human exposure. Quantitative ozone and PM2.5 exposure analyses can provide insight
into all three EJ questions, so they are performed to evaluate potential disproportionate impacts
of this rulemaking.
The baseline ozone and PM2.5 exposure analyses respond to question 1 from EPA's EJ
Technical Guidance document more directly than the proximity analyses, as they evaluate a form
of the environmental stressor primarily affected by the regulatory action (Section 6.5). Baseline
PM2.5 and ozone exposure analyses show that certain populations, such as residents of redlined
census tracts, those linguistically isolated, Hispanic individuals, Asian individuals, those without
a high school diploma, and the unemployed may experience disproportionately higher ozone and
6-27
-------
PM2.5 exposures as compared to the national average. American Indian individuals, residents of
Tribal Lands, populations with higher life expectancy or with life expectancy data unavailable,
children, and insured populations may also experience disproportionately higher ozone
concentrations than the reference group. Hispanic individuals, Black individuals, those below the
poverty line, and uninsured populations may also experience disproportionately higher PM2.5
concentrations than the reference group. Therefore, there likely are potential EJ concerns
associated with environmental stressors affected by the regulatory action for population groups
of concern in the baseline.
Finally, we evaluate how the post-policy options of this final rulemaking are expected to
differentially impact demographic populations, informing questions 2 and 3 from EPA's EJ
Technical Guidance with regard to ozone and PM2.5 exposure changes. Due to the small
magnitude of the exposure changes across population demographics associated with the
rulemaking relative to the magnitude of the baseline disparities, we infer that baseline disparities
in ozone and PM2.5 concentration burdens are likely to remain after implementation of the final
regulatory option (question 2). Also, due to the very small differences in the magnitude of post-
policy ozone and PM2.5 exposure impacts across demographic populations, we do not find
evidence that potential EJ concerns related to ozone or PM2.5 exposures will be exacerbated or
mitigated in the final regulatory option, compared to the baseline (question 3).
This EJ air quality analysis concludes that there are PM2.5 and ozone exposure disparities
across various populations in the pre-policy baseline scenario (EJ question 1) and infer that these
disparities are likely to persist after promulgation of this final rulemaking (EJ question 2). This
EJ assessment also suggests that this action will neither mitigate nor exacerbate PM2.5 and ozone
exposure disparities across populations of EJ concern analyzed (EJ question 3) at the national
scale.
6.8 References
Ebi, K. L., Hasegawa, T., Hayes, K., Monaghan, A., Paz, S., & Berry, P. (2018). Health risks of
warming of 1.5 °C, 2 °C, and higher, above pre-industrial temperatures. Environmental
Research Letters, 13(6), 063007. doi: 10.1088/1748-9326/aac4bd
IPCC. (2018). 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
6-28
-------
change, sustainable development, and efforts to eradicate poverty (V. Masson-Delmotte,
P. Zhai, H.-O. Portner, D. Roberts, J. Skea, P. R. Shukla, A. Pirani, W. Moufouma-Okia,
C. Pean, R. Pidcock, S. Connors, J. B. R. Matthews, Y. Chen, X. Zhou, M. I. Gomis, E.
Lonnoy, T. Maycock, a. M. Tignor, & T. Waterfield Eds.).
National Academies. (2017). Valuing Climate Damages: Updating Estimation of the Social Cost
of Carbon Dioxide. Washington DC: The National Academies Press.
National Academy of Sciences. (2019). Climate Change and Ecosystems. Washington DC: The
National Academies Press.
National Research Council. (2011). America's Climate Choices. Washington, DC: The National
Academies Press.
Oppenheimer, M., Campos, M., Warren, R., Birkmann, J., Luber, G., O'Neill, B., & Takahashi,
K. (2014). Emergent risks and key vulnerabilities. In C.B. Field, V.R. Barros, D.J.
Dokken, K.J. Mach, M.D. Mastrandrea, T.E. Bilir, M. Chatterjee, K.L. Ebi, Y.O. Estrada,
R.C. Genova, B. Girma, E.S. Kissel, A.N. Levy, S. MacCracken, P.R. Mastrandrea, &
L.L.White (Eds.), Climate Change 2014: Impacts, Adaptation, and Vulnerability. Part A:
Global and Sectoral Aspects. Contribution of Working Group II to the Fifth Assessment
Report of the Intergovernmental Panel on Climate Change (pp. 1039-1099). Cambridge,
United Kingdom and New York, NY: Cambridge University Press.
Porter, J. R., Xie, L., Challinor, A. J., Cochrane, K., Howden, M., Iqbal, M. M., & Lobell, D. B.
(2014). Food security and food production systems. In C.B. Field, V.R. Barros, D.J.
Dokken, K.J. Mach, M.D. Mastrandrea, T.E. Bilir, M. Chatterjee, K.L. Ebi, Y.O. Estrada,
R.C. Genova, B. Girma, E.S. Kissel, A.N. Levy, S. MacCracken, P.R. Mastrandrea, &
L.L.White (Eds.), Climate Change 2014: Impacts, Adaptation, and Vulnerability. Part A:
Global and Sectoral Aspects. Contribution of Working Group II to the Fifth Assessment
Report of the Intergovernmental Panel on Climate Change (pp. 485-533). Cambridge,
United Kingdom and New York, NY: Cambridge University Press.
Smith, K. R., Woodward, A., Campbell-Lendrum, D., Chadee, D. D., Honda, Y., Liu, Q., . . .
Sauerborn, R. (2014). Human Health: Impacts, Adaptation, and Co-Benefits. In C.B.
Field, V.R. Barros, D.J. Dokken, K.J. Mach, M.D. Mastrandrea, T.E. Bilir, M. Chatterjee,
K.L. Ebi, Y.O. Estrada, R.C. Genova, B. Girma, E.S. Kissel, A.N. Levy, S. MacCracken,
P.R. Mastrandrea, & L.L.White (Eds.), Climate Change 2014: Impacts, Adaptation, and
Vulnerability. Part A: Global and Sectoral Aspects. Contribution of Working Group II to
the Fifth Assessment Report of the Intergovernmental Panel on Climate Change (pp. 709-
754). Cambridge, United Kingdom and New York, NY: Cambridge University Press.
U.S. EPA. (2021). Climate Change and Social Vulnerability in the United States: A Focus on Six
Impacts. (EPA 43O-R-21-003). Washington DC.
https://www.epa.gov/system/files/documents/2021-09/climate-vulnerability_september-
2021_508.pdf
6-29
-------
USGCRP. (2016). The Impacts of Climate Change on Human Health in the United States: A
Scientific Assessment. Washington DC: U.S. Global Change Research Program.
http://dx.doi.org/10.7930/J0R49NQX
USGCRP. (2018). Impacts, Risks, and Adaptation in the United States: Fourth National Climate
Assessment, Volume II. Washington DC: U.S. Global Change Research Program.
http://dx.doi.org/10.7930/NCA4.2018
6-30
-------
COMPARISON OF BENEFITS AND COSTS
7.1 Introduction
This section presents the estimates of the projected benefits, costs, and net benefits
associated with the final MATS review relative to baseline MATS requirements. There are
potential benefits and costs that may result from this rule that have not been quantified or
monetized. Due to current data and modeling limitations, quantified and monetized benefits from
reducing Hg and non-Hg HAP metals emissions are not included in the monetized benefits
presented here. We are also unable to quantify the potential benefits from the CEMS
requirement. Due to data and modeling limitations, there are also still many categories of climate
impacts and associated damages that are not reflected yet in the monetized climate benefits from
reducing CO2 emissions. For example, the modeling omits most of the consequences of changes
in precipitation, damages from extreme weather events, the potential for nongradual damages
from passing critical thresholds (e.g., tipping elements) in natural or socioeconomic systems, and
non-climate mediated effects of GHG emissions (e.g., ocean acidification).
The projections indicate that the final rule results in 9,500 pounds of reductions in
emissions of Hg as well as 5,400 tons of reductions in PM2.5 across all run years. The final rule is
projected to also reduce CO2, SO2, and NOx by 650,000 tons, 770 tons, and 220 tons,
respectively, and we estimate that the final rule will reduce at least 49 tons of non-Hg HAP
metals. These reductions are composed of reductions in emissions of antimony, arsenic,
beryllium, cadmium, chromium, cobalt, lead, manganese, nickel, and selenium.108 Table 7-1
summarizes the total emissions reductions projected over the 2028 to 2037 analysis period.
108 The estimates on non-mercury HAP metals reductions were obtained my multiplying the ratio of non-mercury
HAP metals to fPM by estimates of PM10 reductions under the rule, as we do not have estimates of fPM reductions
using IPM, only PM10. The ratios of non-mercury HAP metals to fPM were based on analysis of 2010 MATS
Information Collection Request (ICR) data. As there may be substantially more fPM than PM10 reduced by the
control techniques projected to be used under this rule, these estimates of non-mercury HAP metals reductions are
likely underestimates. More detail on the estimated reduction in non-mercury HAP metals can be found in the
docketed memorandum Estimating Non-Hg HAP Metals Reductions for the 2024 Technology Review for the Coal-
Fired EGUSource Category.
7-1
-------
Table 7-1 Cumulative Projected Emissions Reductions for the Final Rule, 2028 to
2037a'b
Pollutant
Emissions Reductions
Hg (pounds)
PM2.5 (tons)
9,500
5,400
650
770
220
49
C02 (thousand tons)
S02 (tons)
NOx (tons)
Non-Hg HAP metals (tons)
a Values rounded to two significant figures.
b Estimated reductions from model year 2028 are applied to 2028 and 2029, those from model year 2030 are applied
to 2031 and 2032, and those from model year 2035 are applied to 2032 through 2037. These values are summed to
generate total reduction figures.
The compliance costs reported in this RIA are not social costs, although in this analysis
we use compliance costs as a proxy for social costs. We do not account for changes in costs and
benefits due to changes in economic welfare of suppliers to the electricity market or to non-
electricity consumers from those suppliers. Furthermore, costs due to interactions with pre-
existing market distortions outside the electricity sector are omitted.
7.2 Methods
EPA calculated the PV of benefits, costs, and net benefits for the years 2028 through
2037, using 2, 3, and 7 percent end-of-period discount rates from the perspective of 2023. All
dollars are in 2019 dollars. In addition to the final rule, we assess a less stringent alternative to
the final requirements.
This calculation of a PV requires an annual stream of values for each year of the 2028 to
2037 timeframe. EPA used IPM to estimate cost and emission changes for the projection years
2028, 2030, and 2035. The year 2028 approximates the compliance year for the final
requirements. In the IPM modeling for this RIA, the 2028 projection year is representative of
2028 and 2029, the 2030 projection year is representative of 2030 and 2031, and the 2035
projection year is representative of 2032 to 2037. Estimates of costs and emission changes in
other years are determined from the mapping of projection years to the calendar years that they
represent. Consequently, the cost and emission estimates from IPM in each projection year are
applied to the years which it represents.109
109 Projected costs associated with the CEMS requirement are not based on IPM. For information on these cost
estimates, see Section 3.
7-2
-------
Health benefits are based on projection year emission estimates and also account for
year-specific variables that influence the size and distribution of the benefits. These variables
include population growth, income growth, and the baseline rate of death.110 Climate benefits
estimates are based on these projection year emission estimates, and also account for year-
specific SC-CO2 values.
EPA calculated the PV and EAV of costs, benefits, and net benefits over the 2028
through 2037 timeframe for the three regulatory options examined in this RIA. The EAV
represents a flow of constant annual values that, had they occurred in each year from 2028 to
2037, would yield an equivalent present value. The EAV represents the value of a typical cost or
benefit for each year of the analysis, in contrast to the year-specific estimates presented
elsewhere for the snapshot years of 2028, 2030, and 2035.
7.3 Results
We first present net benefit analysis for the three years of detailed analysis, 2028, 2030,
and 2035. Table 7-2, Table 7-3, and Table 7-4 present the estimates of the projected compliance
costs, health benefits, climate benefits, and net benefits projected for the final rule. Table 7-5,
Table 7-6, and Table 7-7 present results for the less stringent regulatory option.
The comparison of benefits and costs in PV and EAV terms for the final rule can be
found in for the final regulatory option. Table 7-9 presents the results for the less stringent
regulatory option. Estimates in the tables are presented as rounded values. Note the less stringent
regulatory option only has unquantified benefits associated with requirements for PM CEMS. As
a result, there are no quantified benefits associated with this regulatory option.
110 As these variables differ by year, the health benefit estimates vary by year, including when different years are
based on the same IPM projection year emission estimate.
7-3
-------
Table 7-2 Pro jected Net Benefits of the Final Rule in 2028 (millions of 2019 dollars)a,b
Final Rule, 2028
Health Benefits0
42
and
87
Climate Benefits'1
14
Total Benefits6
57
and
100
Compliance Costs
110
Net Benefits
-58
and
-13
Non-Monetized Benefits®
Benefits from reductions of about 1000 pounds of Hg
Benefits from reductions of about 7 tons of non-Hg HAP metals
Benefits from the increased transparency, compliance assurance, and accelerated identification of anomalous
emission anticipated from requiring PM CEMS
a We focus results to provide a snapshot of projected benefits and costs in 2028, using the best available information
to approximate social costs and social benefits recognizing uncertainties and limitations in those estimates.
b Values have been rounded to two significant figures. Rows may not appear to add correctly due to rounding.
0 Monetized air quality related benefits include those related to public health associated with reductions in PM2 5 and
ozone concentrations. For the presentational purposes of this table, the projected health benefits reported here are
associated with several point estimates and are presented at a real discount rate of 2 percent. See Table 4-4 for the
full range of monetized health benefit estimates.
d Monetized climate benefits are based on reductions in CO2 emissions and are calculated using three different
estimates of the social cost of carbon dioxide (SC-CO2) (under 1.5 percent, 2 percent, and 2.5 percent near-term
Ramsey discount rates). For the presentational purposes of this table, we show the climate benefits associated with
the SC-CO2 at the 2 percent near-term Ramsey discount rate. See Table 4-10 for the full range of monetized climate
benefit estimates.
e Several categories of benefits remain unmonetized and are thus not directly reflected in the quantified benefit
estimates in the table. Non-monetized benefits include benefits from reductions in Hg and non-Hg HAP metals
emissions and from the increased transparency, compliance assurance, and accelerated identification of anomalous
emission anticipated from requiring CEMS.
7-4
-------
Table 7-3 Pro jected Net Benefits of the Final Rule in 2030 (millions of 2019 dollars)a,b
Final Rule, 2030
Health Benefits0
15
and
31
Climate Benefits'1
-8.2
Total Benefits6
7.3
and
22
Compliance Costs
120
Net Benefits
-110
and
-94
Non-Monetized Benefits®
Benefits from reductions of about 1000 pounds of Hg
Benefits from reductions of about 4 tons of non-Hg HAP metals
Benefits from the increased transparency, compliance assurance, and accelerated identification of anomalous
emission anticipated from requiring PM CEMS
a We focus results to provide a snapshot of projected benefits and costs in in 2030, using the best available
information to approximate social costs and social benefits recognizing uncertainties and limitations in those
estimates.
b Values have been rounded to two significant figures. Rows may not appear to add correctly due to rounding.
0 Monetized air quality related health benefits include those related to public health associated with reductions in
PM2 5 and ozone concentrations. For the presentational purposes of this table, the projected health benefits reported
here are associated with several point estimates and are presented at a real discount rate of 2 percent. See Table 4-4
for the full range of monetized health benefit estimates.
d Monetized climate benefits are based on reductions in C02 emissions and are calculated using three different
estimates of the social cost of methane (SC-CO2) (under 1.5 percent, 2 percent, and 2.5 percent near-term Ramsey
discount rates). For the presentational purposes of this table, we show the climate benefits associated with the SC-
CO2 at the 2 percent near-term Ramsey discount rate. See Table 4-10 for the full range of monetized climate benefit
estimates.
e Several categories of benefits remain unmonetized and are thus not directly reflected in the quantified benefit
estimates in the table. Non-monetized benefits include benefits from reductions in Hg and non-Hg HAP metals
emissions and from the increased transparency, compliance assurance, and accelerated identification of anomalous
emission anticipated from requiring CEMS.
7-5
-------
Table 7-4 Projected Net Benefits of the Final Rule in 2035 (millions of 2019 dollars)a,b
Final Rule, 2035
Health Benefits0
10
and
18
Climate Benefits'1
24
Total Benefits6
34
and
42
Compliance Costs
95
Net Benefits
-61
and
-53
Non-Monetized Benefits®
Benefits from reductions of about 900 pounds of Hg
Benefits from reductions of about 4 tons of non-Hg HAP metals
Benefits from the increased transparency, compliance assurance, and accelerated identification of anomalous
emission anticipated from requiring PM CEMS
a We focus results to provide a snapshot of projected benefits and costs in 2035, using the best available information
to approximate social costs and social benefits recognizing uncertainties and limitations in those estimates.
b Values have been rounded to two significant figures. Rows may not appear to add correctly due to rounding.
0 Monetized air quality related health benefits include those related to public health associated with reductions in
PM2 5 and ozone concentrations. For the presentational purposes of this table, the projected health benefits reported
here are associated with several point estimates and are presented at a real discount rate of 2 percent. See Table 4-4
for the full range of monetized health benefit estimates.
d Monetized climate benefits are based on reductions in C02 emissions and are calculated using three different
estimates of the social cost of carbon dioxide (SC-CO2) (under 1.5 percent, 2 percent, and 2.5 percent near-term
Ramsey discount rates). For the presentational purposes of this table, we show the climate benefits associated with
the SC-CO2 at the 2 percent near-term Ramsey discount rate. See Table 4-10 for the full range of monetized climate
benefit estimates.
e Several categories of benefits remain unmonetized and are thus not directly reflected in the quantified benefit
estimates in the table. Non-monetized benefits include benefits from reductions in Hg and non-Hg HAP metals
emissions and from the increased transparency, compliance assurance, and accelerated identification of anomalous
emission anticipated from requiring CEMS.
7-6
-------
Table 7-5 Projected Monetized Benefits, Costs, and Net Benefits of the Less Stringent
Option in 2028 (millions of 2019 dollars) a,b
Final Rule, 2028
Health Benefits0
0
and
0
Climate Benefits'1
0
Total Benefits6
0
and
0
Compliance Costs
2.3
Net Benefits
-2.3
and
-2.3
Non-Monetized Benefits
Benefits from the increased transparency, compliance assurance, and accelerated identification of anomalous
emission anticipated from requiring PM CEMS
a We focus results to provide a snapshot of projected benefits and costs in 2035, using the best available information
to approximate social costs and social benefits recognizing uncertainties and limitations in those estimates.
b Values have been rounded to two significant figures. Rows may not appear to add correctly due to rounding.
Table 7-6 Projected Monetized Benefits, Costs, and Net Benefits of the Less Stringent
Option in 2030 (millions of 2019 dollars) a,b
Final Rule, 2030
Health Benefits0
0
and
0
Climate Benefits'1
0
Total Benefits0
0
and
0
Compliance Costs
2.3
Net Benefits
-2.3
and
-2.3
Non-Monetized Benefits
Benefits from the increased transparency, compliance assurance, and accelerated identification of anomalous
emission anticipated from requiring PM CEMS
a We focus results to provide a snapshot of projected benefits and costs in 2035, using the best available information
to approximate social costs and social benefits recognizing uncertainties and limitations in those estimates.
b Values have been rounded to two significant figures. Rows may not appear to add correctly due to rounding.
Table 7-7 Projected Monetized Benefits, Costs, and Net Benefits of the Less Stringent
Option in 2035 (millions of 2019 dollars)a,b
Final Rule, 2035
Health Benefits0
0
and
0
Climate Benefits'1
0
Total Benefits0
0
and
0
Compliance Costs
2.3
Net Benefits
-2.3
and
-2.3
Non-Monetized Benefits
Benefits from the increased transparency, compliance assurance, and accelerated identification of anomalous
emission anticipated from requiring PM CEMS
a We focus results to provide a snapshot of projected benefits and costs in 2035, using the best available information
to approximate social costs and social benefits recognizing uncertainties and limitations in those estimates.
b Values have been rounded to two significant figures. Rows may not appear to add correctly due to rounding.
7-7
-------
Table 7-8 Stream of Projected Monetized Benefits, Costs, and Net Benefits of the Final
Rule, 2028 to 2037 (discounted to 2023, millions of 2019 dollars)"
Health
Benefitsb
Climate
Benefits04
Compliance
Costs
Net
Benefits0
Year
2%
3%
7%
2%
2%
3%
7%
2%
3%
7%
2028
79
71
52
13
100
99
82
-12
-15
-16
2029
79
71
50
13
100
96
77
-10
-13
-13
2030
27
24
16
-7.1
100
95
73
-82
-78
-64
2031
27
24
16
-7.1
100
92
68
-80
-76
-60
2032
14
13
8
19
79
73
52
-46
-41
-24
2033
14
13
8
19
78
71
48
-44
-39
-21
2034
14
12
7.3
19
76
69
45
-43
-37
-19
2035
14
12
7.0
19
75
67
42
-41
-35
-16
2036
14
12
6.7
19
73
65
39
-40
-33
-14
2037
14
12.0
6.4
19
72
63
37
-39
-32
-11
Health
Benefitsb
Climate
Benefits04
Compliance
Costs
Net
Benefits0
Discount Rate
2%
3%
7%
2%
2%
3%
7%
2%
3%
7%
PV
300
260
180
130
860
790
560
-440
-400
-260
EAV
33
31
25
14
96
92
80
-49
-47
-41
Non-Monetized Benefits®
Benefits from reductions of about 900 to 1000 pounds of Hg annually
Benefits from reductions of about 4 to 7 tons of non-Hg HAP metals annually
Benefits from the increased transparency, compliance assurance, and accelerated identification of anomalous
emission anticipated from requiring PM CEMS
a Values have been rounded to two significant figures. Rows may not appear to add correctly due to rounding.
b The estimated value of the air quality-related health benefits reported here are from Table 4-5, Table 4-6, and
Table 4-7. Monetized benefits include those related to public health associated with reductions in PM2 5 and ozone
concentrations. For discussions of the uncertainty associated with these health benefits estimates, see Section 4.3.8.
0 Monetized climate benefits are based on reductions in CO2 emissions and are calculated using three different
estimates of the social cost of carbon dioxide (SC-CO2) (under 1.5 percent, 2 percent, and 2.5 percent near-term
Ramsey discount rates). For the presentational purposes of this table, we show the climate benefits associated with
the SC-CO2 at the 2 percent near-term Ramsey discount rate. See Table 4-10 for the full range of monetized climate
benefit estimates.
d The small increases and decreases in climate and health benefits and related EJ impacts result from very small
changes in fossil dispatch and coal use relative to the baseline. For context, the projected increase in CO2 emission
of less than 40,000 tons in 2030 is roughly one percent of the emissions of a mid-size coal plant operating at
availability (about 4 million tons).
e Several categories of benefits remain unmonetized and are thus not reflected in the table.
7-8
-------
Table 7-9 Stream of Projected Monetized Benefits, Costs, and Net Benefits of the Less
Stringent Option, 2028 to 2037 (millions of 2019 dollars, discounted to 2023)"
Health Benefits
Climate
Benefits
Compliance Costs
Net Benefits
Year
2%
3%
7%
2%
2%
3%
7%
2%
3%
7%
2023
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
2024
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
2025
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
2026
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
2027
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
2028
0.0
0.0
0.0
0.0
2.1
2.0
1.7
-2.1
-2.0
-1.7
2029
0.0
0.0
0.0
0.0
2.1
2.0
1.6
-2.1
-2.0
-1.6
2030
0.0
0.0
0.0
0.0
2.0
1.9
1.5
-2.0
-1.9
-1.5
2031
0.0
0.0
0.0
0.0
2.0
1.9
1.4
-2.0
-1.9
-1.4
2032
0.0
0.0
0.0
0.0
2.0
1.8
1.3
-2.0
-1.8
-1.3
2033
0.0
0.0
0.0
0.0
1.9
1.7
1.2
-1.9
-1.7
-1.2
2034
0.0
0.0
0.0
0.0
1.9
1.7
1.1
-1.9
-1.7
-1.1
2035
0.0
0.0
0.0
0.0
1.9
1.6
1.0
-1.9
-1.6
-1.0
2036
0.0
0.0
0.0
0.0
1.8
1.6
1.0
-1.8
-1.6
-1.0
2037
0.0
0.0
0.0
0.0
1.8
1.6
0.9
-1.8
-1.6
-0.9
Health Benefits
Climate
Benefits
Compliance Costs
Net Benefits
Discount Rate
2%
3%
7%
2%
2%
3%
7%
2%
3%
7%
PV
0.0
0.0
0.0
0.0
20
18
13
-20
-18
-13
EAV
0.0
0.0
0.0
0.0
2.2
2.1
1.8
-2.2
-2.1
-1.8
Non-Monetized Benefitsb
Benefits from the increased transparency, compliance assurance, and accelerated identification of anomalous
emission anticipated from requiring PM CEMS
a Values have been rounded to two significant figures. Rows may not appear to add correctly due to rounding.
b Several categories of benefits remain unmonetized and are thus not reflected in the table.
The monetized estimates of benefits presented in this section are underestimated because
important categories of benefits, including benefits from reducing Hg and non-Hg HAP metals
emissions and the increased transparency, compliance assurance, and the potential emissions
reductions from the accelerated identification of anomalous emissions anticipated from requiring
PM CEMS, were not monetized in our analysis. Additionally, to the extent that the removal of
the second definition of startup leads to actions that may otherwise not occur absent this final
rule, there may be benefit and cost impacts we are unable to estimate. As a result, the estimates
of compliance costs used in the net benefits analysis may provide an incomplete characterization
of the true costs of the rule. We nonetheless consider these potential impacts in our evaluation of
the net benefits of the rule.
7-9
-------
7.4 Uncertainties and Limitations
Throughout the RIA, we considered several sources of uncertainty, both quantitatively
and qualitatively, regarding the emissions reductions, benefits, and costs estimated for the final
rule. We summarize the key elements of our discussions of uncertainty below.
Compliance costs: The IPM-projected annualized cost estimates of private compliance
costs provided in this analysis are meant to show the increase in production (generating) costs to
the power sector in response to the finalized requirements. As discussed in more detail in section
3.6, there are several key areas of uncertainty related to the electric power sector that are worth
noting, including assumptions about electricity demand, natural gas supply and demand, longer-
term planning by utilities, and assumptions about the cost and performance of controls. There are
also uncertainties associated with the estimated costs for the CEMS requirement as well as
associated with the potential costs of the removal of the startup definition if these amendments
lead to actions by affected facilities that otherwise would not occur absent the finalized
amendments.
Non-monetized benefits: Several categories of health, welfare, and climate benefits are
not quantified in this RIA. These unquantified benefits are described in detail in Section 4. As
noted above, EPA is unable to quantify and monetize the incremental potential benefits of
requiring facilities to utilize CEMS rather than continuing to allow the use of quarterly testing,
but the requirement has been considered qualitatively.
Monetized PM2.5 and ozone-related benefits: The analysis of monetized PM2.5 and
ozone-related benefits described in Section 4.3 includes many data sources as inputs that are
each subject to uncertainty. Input parameters include projected emissions inventories, projected
compliance methods, air quality data from models (with their associated parameters and inputs),
population data, population estimates, health effect estimates from epidemiology studies,
economic data, and assumptions regarding the future state of the world (i.e., regulations,
technology, and human behavior). When compounded, even small uncertainties can greatly
influence the size of the total quantified benefits. Below are key uncertainties associated with
estimating the number and value of PM2.5 and ozone-related premature deaths. Additional detail
regarding specific uncertainties associated with ozone health benefit estimates can be found in
7-10
-------
the Health Benefits TSD (U. S. EPA, 2023). A discussion of uncertainties and limitations related
to the air quality modeling informing the PM2.5 and ozone-related benefits analysis is presented
in section 8.6
Monetized C02-related climate benefits: EPA considered the uncertainty associated
with the SC-CO2) estimates, which were used to calculate the monetized climate impacts of the
changes in CO2 emissions projected to result from this action. Section 4.4 provides a detailed
discussion of the limitations and uncertainties associated with the SC-CO2 estimates used in this
analysis and describes ways in which the modeling addresses quantified sources of uncertainty.
7.5 References
U. S. EPA. (2023). Air Quality Modeling Technical Support Document for Regulatory Impact
Analysis of the Standards of Performance for New, Reconstructed, and Modified Sources
and Emissions Guidelines for Existing Sources: Oil and Natural Gas Sector Climate
Review. (EPA-454/R-23-007). Research Triangle Park, NC: U.S. Environmental
Protection Agency, Office of Air Quality Planning and Standards
7-11
-------
APPENDIX A: AIR QUALITY MODELING
A.l Introduction
As noted in Section 4, EPA used photochemical modeling to create air quality surfaces111
that were then used in air pollution health benefits calculations of the final rule. The modeling-
based surfaces captured air pollution impacts resulting from changes in NOx, SO2, and direct
PM2.5 emissions from EGUs. This appendix describes the source apportionment modeling and
associated methods used to create air quality surfaces for the baseline scenario and the final rule
scenario in three analytic years: 2028, 2030, and 2035. EPA created air quality surfaces for the
following pollutants and metrics: annual average PM2.5; April-September average of 8-hr daily
maximum (MDA8) ozone (AS-M03).
New ozone and PM source apportionment modeling outputs were created to support
analyses in the RIAs for multiple final EGU rulemaking efforts. The basic methodology for
determining air quality changes is the same as that used in the RIAs from multiple previous rules
(U.S. EPA, 2019, 2020a, 2020b, 2021b, 2022a). EPA calculated baseline and final rule EGU
emissions estimates of NOx and SO2 for all three analysis years using IPM (Section 3 of this
RIA). EPA also used IPM outputs to estimate EGU emissions of PM2.5 based on emission factors
described in U.S. EPA (2021a). This appendix provides additional details on the source
apportionment modeling simulations and the associated analysis used to create ozone and PM2.5
air quality surfaces.
A.2 Air Quality Modeling Simulations
The air quality modeling utilized a 2016-based modeling platform which included
meteorology and base year emissions from 2016 and projected future-year emissions for 2026
for all sectors other than EGUs and 2030 for EGUs. The air quality modeling included
photochemical model simulations for a 2016 base year and a future year representing the
combined 2026/2030 emissions described above to provide hourly concentrations of ozone and
PM2.5 component species nationwide. In addition, source apportionment modeling was
performed for the future year to quantify the contributions to ozone from NOx emissions and to
PM2.5 from NOx, SO2 and directly emitted PM2.5 emissions from EGUs on a state-by-state and
111 "Air quality surfaces" refers to continuous gridded spatial fields using a 12 km grid-cell resolution.
A-l
-------
fuel-type basis. As described below, the modeling results for 2016 and the future year, in
conjunction with EGU emissions data for the baseline and the final rule in 2028, 2030, and 2035
were used to construct the air quality surfaces that reflect the influence of emissions changes
between the baseline and the final rule in each year.
The air quality model simulations (i.e., model runs) were performed using the
Comprehensive Air Quality Model with Extensions (CAMx) version 7.10112 (Ramboll Environ,
2021). The nationwide modeling domain (i.e., the geographic area included in the modeling)
covers all lower 48 states plus adjacent portions of Canada and Mexico using a horizontal grid
resolution of 12 km shown in Figure A-l. CAMx requires a variety of input files that contain
information pertaining to the modeling domain and simulation period. These include gridded,
hourly emissions estimates and meteorological data, and initial and boundary concentrations.
The meteorological data and the initial and boundary concentrations were identical to those
described in U.S. EPA (2023a). Separate emissions inventories were prepared for the 2016 base
year and the projected future year. All other inputs (i.e., meteorological fields, initial
concentrations, ozone column, photolysis rates, and boundary concentrations) were specified for
the 2016 base year model application and remained unchanged for the projection-year model
simulation.
2016 base year emissions are described in detail in U.S. EPA (2023b). 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. Future year emissions from all sources other
than EGUs were based on the 2026 emissions projections described in U.S. EPA (2023b). The
Post-IRA 2022 Reference Case of EPA's Power Sector Platform v6 using Integrated Planning
Model (IPM), which includes the Final GNP, was also reflected. The EGU projected inventory
represents demand growth, fuel resource availability, generating technology cost and
112 This CAMx simulation set the Rscale NH3 dry deposition parameter to 0 which resulted in more realistic model
predictions of PM2 5 nitrate concentrations than using a default Rscale parameter of 1.
A-2
-------
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
provi sions of tax 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, Modi fied, 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.epci.gov/power-
sector-modeling/final-pm-naaqs).
Model predictions of ozone and PM2.5 concentrations were compared against ambient
measurements (U.S. EPA, 2023a, 2024). Ozone and PM2.5 model evaluations showed model
performance that was adequate for applying these model simulations for the purpose of creating
air quality surfaces to estimate ozone and PM2.5 benefits.
i3
/ " ' s
JSC
/
•-1 N
\ - *
r
\ '
— J \
*
...1
s 1 w [ ir
\r*
r1 \ \
\ | \
wv \ j
i luS. /
I f ( v *
' 1 ¦
.JVM*™* • Vv
(.TengNv llIMMOm
Ml JM ^ V
r
V '-A v
£ "V
Figure A-1 Air Quality Modeling Domain
The contributions to ozone and PM2.5 component species (e.g., sulfate, nitrate,
ammonium, elemental carbon (EC), organic aerosol (OA), and crustal material113) from EGU
emissions in individual states and from each EGU-fuel type were modeled using the "source
113 Crustal material refers to elements that are commonly found in the earth's crast such as Aluminum, Calcium,
Iron. Magnesium, Manganese, Potassium, Silicon. Titanium, and the associated oxygen atoms.
A-3
-------
apportionment" tool approach. In general, source apportionment modeling quantifies the air
quality concentrations formed from individual, user-defined groups of emissions sources or
"tags." These source tags are tracked through the transport, dispersion, chemical transformation,
and deposition processes within the model to obtain hourly gridded114 contributions from the
emissions in each individual tag to hourly gridded modeled concentrations. For this RIA we used
the source apportionment contribution data to provide a means to estimate of the effect of
changes in emissions from each group of emissions sources (i.e., each tag) to changes in ozone
and PM2.5 concentrations. Specifically, we applied outputs from source apportionment modeling
for ozone and PM2.5 component species using the future year modeled case to obtain the
contributions from EGUs emissions in each state and fuel-type to ozone and PM2.5 component
species concentrations in each 12 km model grid cell nationwide. Ozone contributions were
modeled using the Anthropogenic Precursor Culpability Assessment (APCA) tool and PM2.5
contributions were modeled using the Particulate Matter Source Apportionment Technology
(PSAT) tool (Ramboll Environ, 2021). The ozone source apportionment modeling was
performed for the period April through September to provide data for developing spatial fields
for the April through September maximum daily eight-hour (MDA8) (i.e., AS-M03) average
ozone concentration exposure metric. The PM2.5 source apportionment modeling was performed
for a full year to provide data for developing annual average PM2.5 spatial fields. Source
apportionment simulations were set-up to separately track ozone and PM2.5 contributions from
coal EGU emissions in each contiguous U.S. state, natural gas EGU emissions in each
contiguous U.S. state, and emissions from all other EGUs aggregated across all contiguous U.S.
states. In cases where projected EGU emissions for a specific tag and pollutant were either 0 or
very small, emissions were combined with nearby states to make multi-state tags. Tables A-l, A-
2, and A-3 provide emissions that were tracked for each source apportionment tag.
114 Hourly contribution information is provided for each grid cell to provide spatial patterns of the contributions from
each tag.
A-4
-------
Table A-l Future-Year Emissions Allocated to Each Modeled Coal EGU State Source
Apportionment Tag
State
Ozone Season
NOx (tons)
Annual NOx
(tons)
Annual SO2
(tons)
Annual PM2.:
(tons)
AL
2,537
5,046
1,929
700
AR4
NA
304
331
51
AZ
1,005
2,536
4,515
609
CA
222
511
99
27
CO
19
269
287
21
CT
0
0
0
0
DC
0
0
0
0
DE
0
0
0
0
FL
1,110
1,401
7,163
277
GA
1,654
2,534
3,247
159
IA
8,354
18,776
9,656
1,203
ID
0
0
0
0
IL
1,639
3,742
6,773
270
IN
4,886
18,146
26,584
2,252
KS1
NA
214
121
NA
KY
3,551
7,333
7,127
560
LA2'4
NA
47
NA
NA
MA
0
0
0
0
MD3
NA
139
272
31
MD + PA3
708
NA
NA
NA
ME
0
0
0
0
MI
1,532
4,071
12,478
380
MN
724
1,549
3,289
94
MO
2,947
23,480
38,989
853
MS4
NA
252
507
23
MT
3,771
8,842
4,056
1,252
NC
266
482
634
35
ND
8,583
19,562
25,398
1,923
NE1
7,817
17,507
43,858
NA
NE + KS1
NA
NA
NA
374
NH
0
0
0
0
NJ
0
0
0
0
NM
1,442
2,757
6,800
1,739
NV
0
1
1
0
NY
0
0
0
0
OH
3,152
10,485
21,721
901
OK4
NA
212
152
21
OR
0
0
0
0
PA3
NA
1,530
4,932
167
RI
0
0
0
0
A-5
-------
sc
807
1,939
3,429
364
SD
418
1,100
1,022
27
TN
259
259
269
32
TX2'4
NA
7,031
NA
NA
TX + LA2
NA
NA
11,607
1,578
TX-reg4
2,698
NA
NA
NA
UT
2,702
4,236
7,625
232
VA
466
1,124
259
445
VT
0
0
0
0
WA
0
0
0
0
WI
866
2,137
838
90
WV
6,824
16,358
17,631
1,753
WY
6,066
13,222
11,754
1,024
1KS and NE emissions grouped into multi-state tag for direct PM2 5
2LA and TX emissions grouped into multi-state tag for S02 and direct PM2 5
3MD and PA emissions grouped into multi-state tag for ozone season NOx
4AR, KS, LA, MS, OK and TX emissions grouped into multi-state tag ("TX-reg") for ozone season NOx
A-6
-------
Table A-2 Future-Year Emissions Allocated to Each Modeled Natural Gas EGU State
Source Apportionment Tag
State
Ozone Season NOx
(tons)
Annual NOx
(tons)
Annual SO2
(tons)
Annual P]
(tons)
AL
2,833
5,132
0
1,979
AR
1,651
2,957
0
632
AZ
1,759
3,146
0
686
CA
1,960
5,773
0
1,964
CO
957
1,825
0
461
CT
461
778
0
160
DC
6
11
0
7
DE
383
502
0
134
FL
7,550
14,372
0
4,996
GA
2,279
4,182
0
1,740
IA
875
1,106
0
327
ID
336
513
0
185
IL
1,624
2,705
0
825
IN
1,180
2,166
0
955
KS
329
621
0
54
KY
980
2,806
0
699
LA
3,771
8,706
0
2,158
MA
482
725
0
244
MD
402
710
0
435
ME
232
273
0
21
MI
6,523
11,372
0
1,508
MN
661
928
0
87
MO
587
875
0
342
MS
1,926
3,860
0
1,140
MT
11
19
0
7
NC
1,803
3,426
0
1,213
ND
25
41
0
3
NE
13
47
0
4
NH
120
136
0
34
NJ
1,024
1,910
0
608
NM
733
1,128
0
131
NV
1,693
2,471
0
648
NY
2,793
5,125
0
1,270
OH
1,838
3,824
0
1,617
OK
1,558
2,448
0
546
OR
5
188
0
87
PA
6,811
12,386
0
3,280
RI
115
153
0
73
SC
1,092
2,090
0
917
A-7
-------
SD
93
105
0
11
TN
464
1,107
0
388
TX
7,652
14,715
0
3,567
UT
1,189
1,779
0
514
VA
1,836
3,409
0
1,087
VT
4
8
0
6
WA
485
1,311
0
464
WI
847
1,447
0
369
WV
109
180
0
50
WY
203
206
0
28
Table A-3 Future-Year Emissions Allocated to the Modeled Other EGU Source
Apportionment Tag
State
Ozone Season NOx
Annual NOx
Annual SO2
Annual PM2.5
(tons)
(tons)
(tons)
(tons)
USa
20,611
48,619
9,631
7,915
aOnly includes US emissions from the contiguous 48 states
Examples of the magnitude and spatial extent of ozone and PM2.5 contributions are
provided in through Figure A-5 for EGUs in California, Georgia, Iowa, and Ohio. These figures
show how the magnitude and the spatial patterns of contributions of EGU emissions to ozone
and PM2.5 component species depend on multiple factors including the magnitude and location of
emissions as well as the atmospheric conditions that influence the formation and transport of
these pollutants. For instance, NOx emissions are a precursor to both ozone and PM2.5 nitrate.
However, ozone and nitrate form under very different types of atmospheric conditions, with
ozone formation occurring in locations with ample sunlight and ambient VOC concentrations
while nitrate formation requires colder and drier conditions and the presence of gas-phase
ammonia. California's complex terrain that tends to trap air and allow pollutant build-up
combined with warm sunny summer and cooler dry winters and sources of both ammonia and
VOCs make its atmosphere conducive to formation of both ozone and nitrate. While the
magnitude of EGU NOx emissions from gas plus coal EGUs is substantially larger in Iowa than
in California (Table A-l and Table A-2) the emissions from California lead to larger maximum
contributions to the formation of those pollutants due to the conducive conditions in that state.
Georgia and Ohio both had substantial NOx emissions. While maximum ozone impacts shown
for Georgia and Ohio EGUs are similar order of magnitude to maximum ozone impacts from
A-8
-------
California EGUs, nitrate impacts are negligible in both Georgia and Ohio due to less conducive
atmospheric conditions for nitrate formation in those locations. California EGU SO2 emissions in
the future year source apportionment modeling are several orders of magnitude smaller than SO2
emissions in Ohio and Georgia (Table A-l) leading to much smaller sulfate contributions from
California EGUs than from Ohio and Georgia EGUs. PM2.5 organic aerosol EGU contributions
in this modeling come from primary PM2.5 emissions rather than secondary atmospheric
formation. Consequently, the impacts of EGU emissions on this pollutant tend to occur closer to
the EGU sources than impacts of secondary pollutants (ozone, nitrate, and sulfate) which have
spatial patterns showing a broader regional impact. These patterns demonstrate how the model
captures important atmospheric processes which impact pollutant formation and transport from
emissions sources. Finally, Figures A-6 and A-7 show EGU ozone and PM2.5 contributions from
all contiguous U.S. EGUs split out by fuel type. The spatial differences between coal EGU,
natural gas EGU, and other EGU contributions reflect the varying location and magnitude of
emissions from each type of EGU.
A-9
-------
»
a) Apr-Sep MDA8 03
•" " b) Annual PM2 5 nitrate
.»
»
LJr—rr^
£jy-o^i M
t\TVnllt^
1 F —™ ,
r—\) J f /
™ / ? 'V L fX\\f
L / m—f \ Cwl J UK
i w / f 1 \ 11 \ ) 1y*'
i / ri —S—VUU^r^f
» L ; : T~—J-1 \ / 1 ! Ls#
r \ f—Y\ y/rf
0«
009
sot
CM
* » V\ Y i^~ri 1 r*»»«•«
M>• 000t»0aft|1.J) Ma. - Hi!*tJ2.UW
I H W » HI B)
J"
c) Annual PM2 5 sulfate
d) Annual PM2 5 OA
OW
-
~~NO^
huii Urf\ r 5*
furfWM
V \i—i J
•» 1'
/ /
°« m 1 I t 1 , /
«• ijl ¦ / i_T—• - - \ n l»s-//
4 ¦ ; M" \ ' K/f
••
-
n
\iJ H/TTy
1 [ 1 | r ¥
" W ' 1 " 7 ------
¦MJ Hmy
™ w| Sk Uo>-v^ i\
. \ M V
IS*
to
1 t-i yu4
M >
Ml'
« U» HI US M
Mn • 0 OM»0 «f* (1.1X M*> • OOyi it
Figure A-2 Maps of California EGU Tag Contributions to a) April-September Seasonal
Average MDA8 Ozone (ppb); b) Annual Average PM2.5 Nitrate (jng/ni3); c) Annual
Average PM2.5 Sulfate (ju,g/m3); d) Annual Average PM2.5 Organic Aerosol (jxg/m3)
A-10
-------
a) Apr-Sep MDA8 03
b) Annual PM2.5 nitrate
-
t*
nr-nv f\
» / / \ L 1 rXhr
/ l—Tl ( tN\ Hi
2U / J f / \ | I \ )
V \ / I I—VX iJ/Tf
\ \H—-J
CO*
OM
1 \ 1 - — i f' _ —- / V.'
00?
om
cos
a ) In i' • • /
H M-ZTTV
Oat
-
V M
VsUvi
S v\
,1 V V
m
80 IB 2® BIS »'
Mil - 0 DQJfD * U.JX MlD • 1177 at
i « us a® as tp
Min-ao«. / J | i j j! A r* \- ~y&
in- rvHi/fi
' \ / 7 — 1 J 1 I ll -rfPU*'
Mr\ / 1—\\ wn
\ \J—1— 1 v^K-%
i 'I' 1 •• i • v
j~ -4 1 \ 4
- Y Hf=; —>P—7 ?
0O' » < \ 1 \ 7 v /
H P-~urfv
"
^-ClnWw
• \ j]J4d
y y
nJ.-M
1 v£ \J
1
« u» » ns an
Mn - B Bono « RU Ma - a aw «
-------
a) Apr-Sep MDA8 03
T\
f
jM
Y\
\ I
y
j
%
w
If
c) Annual PM2.5 sulfate
AA
¦ M*
r
a
/
/
im
1—\ *
*
< -Tr^A
^ \\
\
y
b) Annual PM2.5 nitrate
i
-J;
iA
f
X
V «\
\\
\ 1
V
d) Annual PM2 5 OA
1?
& aM
fr I r \S
/
/
K~
7tw
y
sJt
Figure A-4 Maps of Iowa EGU Tag contributions to a) April-September Seasonal
Average MDA8 Ozone (ppb); b) Annual Average PM2.5 Nitrate (jig/m3); c) Annual Average
PM2.5 Sulfate (jiig/1113); d) Annual Average P.M2.5 Organic Aerosol (jig/m3)
A-12
-------
a) Apr-Sep MDA8 03
[A
*
r
tUJf
im- a o».a atCUM Mv ¦ 2 J5r*j»s.li«>
K
c) Annual PM2.5 sulfate
fe* rM
\f\J3\
U
/
V/
, J?
y
¦UUl w« ' o w M ISOftlW
b) Annual PM2.5 nitrate
'¦/i
f
f
dJu
/
/
i p.
/W? 1 \
y
p
d) Annual PM2 5 OA
rvl
rA/
/-r
/
\ /
Vr
•js
Y\
H) rtU.n Mm • 0131 «I«}.!«»
Figure A-5 Maps of Ohio EGU Tag Contributions to a) April-September Seasonal
Average MDA8 Ozone (ppb); b) Annual Average PM2.5 Nitrate (jig/m3); c) Annual Average
PM2.5 Sulfate (jiig/1113); d) Annual Average P.M2.5 Organic Aerosol (jig/m3)
A-13
-------
a) Apr-Sep MDA8 Ozone contributions b) Apr-Sep MDA8 Ozone contributions c) Apr-Sep MDA8 Ozone contributions
from US Coal EGUs
from US Natural Gas EGUs
from US Other EGUs
V -
f
/
_ 'A,
; i (
,\
rjr
nr—
TXfe*. J}
TjJjfhjjfc
»
L - ]( ] > ter
\ Vy vT
t* V if
1%, \ J —1—
F ^ I I
J><' , ff
J JpH " f
¦
nrf^v
Hffij
\r\ jf
-
\i V
Ay ^
,
Figure A-6 Maps of national EGU Tag Contributions to April-September Seasonal
Average MDA8 Ozone (ppb) by fuel for a) coal EGUs; b) natural gas EGUs; c) all other
EGUs
a) Annual PM2.5 contributions from US b) Annual PM2.5 contributions from US c) Annual PM2 5 contributions from US
Coal EGUs
/
¦i 2
V
•Jf
r
*3
Natural Gas EGUs
ff
Ah\
TTfe, r4/\
~TU tfh £&\
.
L
Vt \ /f 1
H
V '¦¦¦
\^XjT$y
•
-
XT y
Other EGUs
1
iM
v,
y
Figure A-7 Maps of national EGU Tag Contributions Annual Average PM2.5 (jig/m3) by
fuel for a) coal EGUs; b) natural gas EGUs; c) all other EGUs
A.3 Applying Modeling Outputs to Create Spatial Fields
In this section we describe the method for creating spatial fields of AS-M03 and annual
average PM2.5 based on the 2016 and future year modeling. The foundational data include (1)
ozone and speciated PM2.5 concentrations in each model grid cell from the 2016 and the future
A-14
-------
year modeling, (2) ozone and speciated PM2.5 contributions in the future year of EGUs emissions
from each state in each model grid cell,115 (3) future year emissions from EGUs that were input to
the contribution modeling (Table A-l, Table A-2, Table A-3), and (4) the EGU emissions from
IPM for baseline and the final rule scenarios in each analytic year. The method to create spatial
fields applies scaling factors to gridded source apportionment contributions based on emissions
changes between future year projections and the baseline and the final rule options to the
modeled contributions. This method is described in detail below.
Spatial fields of ozone and PIVh.sfor the future year were created based on "fusing"
modeled data with measured concentrations at air quality monitoring locations. To create the
spatial fields for each future emissions scenario, these fused future year model fields are used in
combination with the EGU source apportionment modeling and the EGU emissions for each
scenario and analytic year. Contributions from each state and fuel EGU contribution "tag" were
scaled based on the ratio of emissions in the year/scenario being evaluated to the emissions in the
modeled scenario. Contributions from tags representing sources other than EGUs are held
constant at 2026 levels for each of the scenarios and year. For each scenario and year analyzed,
the scaled contributions from all sources were summed together to create a gridded surface of
total modeled ozone and PM2.5. The process is described in a step-by-step manner below starting
with the methodology for creating AS-M03 spatial fields followed by a description of the steps
for creating annual PM2.5 spatial fields.
Ozone:
1. Create fused spatial fields of future year AS-M03 incorporating information from the air
quality modeling and from ambient measured monitoring data. The enhanced Voronoi
Neighbor Average (eVNA) technique (Ding et al., 2016; Gold et al., 1997; U.S. EPA, 2007)
was applied to ozone model predictions in conjunction with measured data to create
modeled/measured fused surfaces that leverage measured concentrations at air quality
monitor locations and model predictions at locations with no monitoring data.
1.1. The AS-M03 eVNA spatial fields are created for the 2016 base year with EPA's
software package, Software for the Modeled Attainment Test - Community Edition
115 Contributions from EGUs were modeled using projected emissions for the modeled scenario. The resulting
contributions were used to construct spatial fields in 2028, 2030, and 2035.
A-l 5
-------
(SMAT-CE)116 (U.S. EPA, 2022b) using three years of monitoring data (2015-2017) and
the 2016 modeled data.
1.2. The model-predicted spatial fields (i.e., not the eVNA fields) of AS-M03 in 2016 were
paired with the corresponding model-predicted spatial fields in the future year to
calculate the ratio of AS-M03 between 2016 and the future year in each model grid cell.
1.3. To create a gridded future year eVNA surfaces, the spatial fields of 2016/future year
ratios created in step 1.2 were multiplied by the corresponding eVNA spatial fields for
2016 created in step 1.1 to produce an eVNA AS-M03 spatial field for future year using
• eVNAg future is the eVNA concentration of AS-M03 or PM2.5 component species in grid-
cell, g, in the future year
• eVNAg 2016 is the eVNA concentration of AS-M03 or PM2.5 component species in grid-
cell, g, in 2016
• Modelg future is the CAMx modeled concentration of AS-M03 or PM2.5 component
species in grid-cell, g, in the future year
• Modelg 2016 is the CAMx modeled concentration of AS-M03 or PM2.5 component in
grid-cell, g, in 2016
2. Create gridded spatial fields of total EGU AS-M03 contributions for each combination of
scenario and analytic year evaluated.
2.1. Use the EGU ozone season NOx emissions for the 2028 baseline and the corresponding
future year modeled EGU ozone season emissions (Table A-l, Table A-2, and Table A-
3) to calculate the ratio of 2028 baseline emissions to future year modeled emissions for
116 SMAT-CE available for download at https://www.epa.gov/scram/photochemical-modeling-tools.
(Eq-1).
eVNAg,future — (®VN.Ag,2oi6) x
Modelg future
ModeL7ni6
g,2016
Eq-1
A-16
-------
each EGU tag (i.e., an ozone scaling factor calculated for each state-fuel tag).117 These
scaling factors are provided in Table A-, A-5 and A-l 1.
2.2. Calculate adjusted gridded AS-M03 EGU contributions that reflect differences in state-
fuel EGU NOx emissions between the modeled future year and the 2028 baseline by
multiplying the ozone season NOx scaling factors by the corresponding gridded AS-
M03 ozone contributions118 from each state-fuel EGU tag.
2.3. Add together the adjusted AS-M03 contributions for each state-fuel EGU tag to produce
spatial fields of adjusted EGU totals for the 2028 baseline.119
2.4. Repeat steps 2.1 through 2.3 for the 2028 final rule scenario and for the baseline and
final rule scenarios for each additional analytic year. All scaling factors for the baseline
scenario and the regulatory control alternatives are provided in Tables A-4, A-5, and A-
3. Create a gridded spatial field of AS-M03 associated with IPM emissions for the 2028
baseline by combining the EGU AS-M03 contributions from step 2.3 with the corresponding
contributions to AS-M03 from all other sources. Repeat for each of the EGU contributions
created in step 2.4 to create separate gridded spatial fields for the 2028 final rule scenario and
the baseline and final rule scenario for the two other analytic years.
Steps 2 and 3 in combination can be represented by equation 2:
117 State-level tags were tracked for separately for coal EGUs and for natural gas EGUs. All other EGU emissions
were tracked using a single national tag. In addition, preliminary testing of this methodology showed unstable
results when very small magnitudes of emissions were tagged especially when being scaled by large factors. To
mitigate this issue, in cases where state-fuel EGU tags were associated with no or very small emissions, tags were
combined into multi-state regions.
118 The source apportionment modeling provided separate ozone contributions for ozone formed in VOC-limited
chemical regimes (03 V) and ozone formed in NOx-limited chemical regimes (03N). The emissions scaling factors
are multiplied by the corresponding 03N gridded contributions to MD A8 concentrations. Since there are no
predicted changes in VOC emissions in the control scenarios, the 03 V contributions remain unchanged.
119 The contributions from the unaltered 03 V tags are added to the summed adjusted 03N EGU tags.
11.
AS-M03giy — eVNA&future
Eq-2
A-17
-------
• AS-M03g i y is the estimated fused model-obs AS-M03 for grid-cell, "g," scenario, ""i."l2u and
year, "y;"121
• eVNAg future is the future year eVNA future year AS-M03 concentration for grid-cell "g"
calculated using Eq-1;
• Tot is the total modeled AS-M03 for grid-cell "g" from all sources in the future year source
apportionment modeling;
• Cg,bc is the future year AS-M03 modeled contribution from the modeled boundary inflow;
• Cg int is the future year AS-M03 modeled contribution from international emissions within the
modeling domain;
• Cg,bio is the future year AS-M03 modeled contribute/on from biogenic emissions;
• Cg fires is the future year AS-M03 modeled contribution from fires;
• Cg,USanthro is the total future year AS-M03 modeled contribution from U.S. anthropogenic
sources other than EGUs;
• CEGUVOc,g,t's t'10 future year AS-M03 modeled contribution from EGU emissions of VOCs from
tag, "t";
• CEGUNOx,g,t 's future year AS-M03 modeled contribution from EGU emissions of NOx from
tag, "t"; and
• SNOx,t,i,y is the EGU NOx scaling factor for tag, "t," scenario, "i," and year, "y."
PM25
4. Create fused spatial fields of future year annual PM2.5 component species incorporating
information from the air quality modeling and from ambient measured monitoring data. The
eVNA technique was applied to PM2.5 component species model predictions in conjunction
with measured data to create modeled/measured fused surfaces that leverage measured
concentrations at air quality monitor locations and model predictions at locations with no
monitoring data.
120 Scenario "i" can represent either the baseline or the final rule scenario.
121 Year "y" can represent 2028, 2030, or 2035.
A-18
-------
4.1. The quarterly average PM2.5 component species eVNA spatial fields are created for the
2016 base year with EPA's SMAT-CE software package using three years of monitoring
data (2015-2017) and the 2016 modeled data.
4.2. The model-predicted spatial fields (i.e., not the eVNA fields) of quarterly average PM2.5
component species in 2016 were paired with the corresponding future year model-
predicted spatial fields to calculate the ratio of PM2.5 component species between 2016
and the future year in each model grid cell.
4.3. To create a gridded future year eVNA surfaces, the spatial fields of 2016/future year
ratios created in step 4.2 were multiplied by the corresponding eVNA spatial fields for
2016 created in step 4.1 to produce an eVNA annual average PM2.5 component species
spatial field for the future year using Eq-1.
5. Create gridded spatial fields of total EGU speciated PM2.5 contributions for each combination
of scenario and analytic year evaluated.
5.1. Use the EGU annual total NOx, SO2, and PM2.5 emissions for the 2028 baseline scenario
and the corresponding future year modeled EGU NOx, SO2, and PM2.5 emissions from
Table A-l, Table A-2 and Table A-3 to calculate the ratio of 2028 baseline emissions to
future year modeled emissions for each EGU state-fuel contribution tag (i.e., annual
nitrate, sulfate and directly emitted PM2.5 scaling factors calculated for each state-fuel
tag).122 These scaling factors are provided in Table A-6 through Table A-l 1.
5.2. Calculate adjusted gridded annual PM2.5 component species EGU contributions that
reflect differences in state-EGU NOx, SO2, and primary PM2.5 emissions between the
modeled future year and the 2028 baseline by multiplying the annual nitrate, sulfate and
directly emitted PM2.5 scaling factors by the corresponding annual gridded PM2.5
component species contributions from each state-fuel EGU tag.123
122 State-level tags were tracked for separately for coal EGUs and for natural gas EGUs. All other EGU emissions
were tracked using a single national tag. In addition, preliminary testing of this methodology showed unstable
results when very small magnitudes of emissions were tagged especially when being scaled by large factors. To
mitigate this issue, in cases where state-fuel EGU tags were associated with no or very small emissions, tags were
combined into multi-state regions.
123 Scaling factors for components that are formed through chemical reactions in the atmosphere were created as
follows: scaling factors for sulfate were based on relative changes in annual SO2 emissions; scaling factors for
A-19
-------
5.3. Add together the adjusted PM2.5 contributions of for each EGU state tag to produce
spatial fields of adjusted EGU totals for each PM2.5 component species.
5.4. Repeat steps 5.1 through 5.3 for the final rule scenario in 2028 and for the baseline and
the final rule scenario for each additional analytic year. The scaling factors for all PM2.5
component species for the baseline and final rule scenarios are provided in Table A-6
through Table A-l 1
6. Create gridded spatial fields of each PM2.5 component species for the 2028 baseline by
combining the EGU annual PM2.5 component species contributions from step 5.3 with the
corresponding contributions to annual PM2.5 component species from all other sources.
Repeat for each of the EGU contributions created in step 5.4 to create separate gridded
spatial fields for the baseline and final rule scenarios for all other analytic years.
7. Create gridded spatial fields of total PM2.5 mass by combining the component species
surfaces for sulfate, nitrate, organic aerosol, elemental carbon, and crustal material with
ammonium, and particle-bound water. Ammonium and particle-bound water concentrations
are calculated for each scenario based on nitrate and sulfate concentrations along with the
ammonium degree of neutralization in the base year modeling (2016) in accordance with
equations from the SMAT-CE modeling software (U.S. EPA, 2022bfi).
Steps 5 and 6 result in Eq-3 for PM2.5 component species: sulfate, nitrate, organic aerosol,
elemental carbon, and crustal material.
nitrate were based on relative changes in annual NOx emissions. Scaling factors for PM2 5 components that are
emitted directly from the source (OA, EC, crustal) were based on the relative changes in annual primary PM2 5
emissions between the future year modeled emissions and the baseline or the final rule scenario in each year.
PMs,g,i,y = eVNA,
s,g,future
Eq-3
A-20
-------
PMg g i y is the estimated fused model-obs PM component species "s" for grid-cell, "g," scenario,
"i,"124 and year, "y;"125
• eVNAs g future is the future year eVNA PM concentration for component species "s" in grid-cell
"g" calculated using Eq-1;
• ^s,g,Tot is the total modeled PM component species "s" for grid-cell "g" from all sources in the
2026 source apportionment modeling;
• Cs,&bc is the future year PM component species "s" modeled contribution from the modeled
boundary inflow;
• Cs &int is the future year PM component species "s" modeled contribution from international
emissions within the modeling domain;
• Cs,g,bio is the future year PM component species "s" modeled contribution from biogenic
emissions;
• Cs g fires is the future year PM component species "s" modeled contribution from fires;
• Cs g USanthro is the total future year PM component species "s" modeled contribution from U.S.
anthropogenic sources other than EGUs;
• CEGUs,g,t is the future year PM component species "s" modeled contribution from EGU emissions
of NOx, SO2, or primary PM2.5 from tag, "t"; and
• Ss t J y is the EGU scaling factor for component species "s," tag "t," scenario "i," and year "y."
Scaling factors for nitrate are based on annual NOx emissions, scaling factors for sulfate are based
on annual SO2 emissions, scaling factors for primary PM2.5 components are based on primary
PM2.5 emissions.
124 Scenario "i" can represent either baseline or the final rule scenario.
125 Year "y" can represent 2028, 2030, or 2035.
A-21
-------
A.4 Scaling Factors Applied to Source Apportionment Tags
Table A-4 Ozone Seasonal NOx Scaling Factors for Coal EGU Tags in the Baseline and
the Final Rule
Baseline
Final Rule
State Tag
2028
2030
2035
2028
2030
2035
AL
1.20
1.40
1.47
1.20
1.40
1.47
AZ
0.01
1.43
1.13
0.01
1.43
1.17
CA
0.00
0.00
0.00
0.00
0.00
0.00
CO
139.01
1.28
1.98
139.01
1.28
1.98
CT
0.00
0.00
0.00
0.00
0.00
0.00
DC
0.00
0.00
0.00
0.00
0.00
0.00
DE
0.00
0.00
0.00
0.00
0.00
0.00
FL
0.47
1.24
0.10
0.47
1.24
0.10
GA
0.00
0.18
0.00
0.00
0.18
0.00
IA
1.17
1.18
0.77
1.17
1.18
0.77
ID
0.00
0.00
0.00
0.00
0.00
0.00
IL
0.97
0.96
0.81
0.97
0.96
0.81
IN
1.35
0.76
0.19
1.35
0.76
0.19
KY
0.79
0.95
0.97
0.79
0.95
0.98
MA
0.00
0.00
0.00
0.00
0.00
0.00
MDPAa
3.14
3.17
2.58
3.14
3.17
2.58
ME
0.00
0.00
0.00
0.00
0.00
0.00
MI
0.75
0.00
0.00
0.75
0.00
0.00
MN
2.41
2.25
0.00
2.41
2.25
0.00
MO
2.72
1.57
0.67
2.71
1.57
0.67
MT
1.07
1.12
1.11
1.07
1.12
1.09
NC
9.89
6.41
2.86
9.92
6.43
2.86
ND
1.09
1.08
0.25
1.06
1.08
0.25
NE
1.16
1.18
0.73
1.16
1.18
0.74
NH
0.00
0.00
0.00
0.00
0.00
0.00
NJ
0.00
0.00
0.00
0.00
0.00
0.00
NM
0.98
0.98
0.01
0.98
0.98
0.01
NV
0.00
0.00
0.00
0.00
0.00
0.00
NY
0.00
0.00
0.00
0.00
0.00
0.00
OH
0.58
1.07
0.00
0.58
1.07
0.00
OR
0.00
0.00
0.00
0.00
0.00
0.00
RI
0.00
0.00
0.00
0.00
0.00
0.00
SC
0.81
2.22
3.18
0.81
2.22
3.18
SD
0.87
1.33
0.00
0.87
1.33
0.00
TN
3.89
0.01
0.00
3.89
0.01
0.00
TX-regb
4.69
4.26
1.64
4.70
4.26
1.64
UT
1.00
0.06
0.06
1.00
0.06
0.06
A-22
-------
Baseline
Final Rule
State Tag
2028
2030
2035
2028
2030
2035
VA
0.65
0.45
0.00
0.65
0.45
0.00
VT
0.00
0.00
0.00
0.00
0.00
0.00
WA
0.00
0.00
0.00
0.00
0.00
0.00
WI
1.66
2.16
0.36
1.69
2.16
0.36
WV
0.92
1.16
0.92
0.92
1.16
0.91
WY
1.26
1.12
1.12
1.26
1.12
1.12
Note: Emissions from Maryland, Arkansas, Kansas, Louisiana, Oklahoma, and Mississippi are less than 10 tpy in
the original source apportionment modeling. Air quality impacts and emissions from those states were combined
with nearby states.
aMDPA: Maryland and Pennsylvania
bTX-reg: Arkansas, Kansas, Louisiana, Oklahoma, Mississippi, Texas
A-23
-------
Table A-5 Ozone Seasonal NOx Scaling Factors for Gas EGU Tags in the Baseline and
the Final Rule
Baseline
Final Rule
State Tag
2028
2030
2035
2028
2030
2035
AL
0.53
0.61
0.49
0.53
0.61
0.49
AR
0.65
0.68
0.43
0.63
0.68
0.43
AZ
0.69
0.68
0.67
0.69
0.68
0.67
CA
0.92
0.94
0.85
0.92
0.94
0.85
CO
3.26
0.63
0.50
3.26
0.63
0.50
CT
1.04
0.98
0.89
1.04
0.98
0.89
DC
0.86
0.59
0.33
0.86
0.59
0.33
DE
0.79
0.80
0.38
0.79
0.80
0.38
FL
1.08
1.03
1.04
1.08
1.03
1.04
GA
0.58
0.54
0.52
0.58
0.54
0.52
IA
0.53
0.42
0.16
0.53
0.42
0.16
ID
0.60
0.90
0.90
0.59
0.91
0.89
IL
0.69
0.61
0.42
0.68
0.61
0.42
IN
0.75
0.63
0.38
0.75
0.63
0.38
KS
1.38
1.32
0.25
1.38
1.32
0.24
KY
0.87
0.81
0.69
0.86
0.81
0.69
LA
1.04
1.00
0.72
1.04
1.00
0.72
MA
0.60
0.67
0.66
0.60
0.67
0.66
MD
1.51
1.33
1.12
1.51
1.33
1.12
ME
1.16
1.15
0.59
1.16
1.15
0.59
MI
0.68
0.70
0.55
0.68
0.70
0.55
MN
0.92
0.84
0.34
0.92
0.84
0.34
MO
0.59
0.59
0.20
0.58
0.59
0.20
MS
0.64
0.62
0.50
0.64
0.62
0.50
MT
0.95
1.10
0.08
0.95
1.10
0.09
NC
0.77
0.59
0.68
0.77
0.59
0.68
ND
0.85
1.85
0.34
0.82
1.85
0.34
NE
5.91
5.92
0.28
5.91
5.92
0.29
NH
0.67
0.51
0.41
0.67
0.51
0.41
NJ
0.81
0.85
0.61
0.81
0.85
0.62
NM
1.00
0.84
0.77
1.00
0.84
0.77
NV
0.33
0.25
0.19
0.33
0.25
0.19
NY
1.03
0.99
0.65
1.03
0.99
0.64
OH
1.02
0.97
0.84
1.03
0.97
0.84
OK
1.69
1.57
0.48
1.69
1.57
0.47
OR
63.29
0.00
0.00
63.55
0.00
0.00
PA
0.79
0.69
0.34
0.79
0.69
0.34
RI
0.69
0.75
0.71
0.69
0.75
0.71
A-24
-------
Baseline
Final Rule
State Tag
2028
2030
2035
2028
2030
2035
SC
0.93
0.96
0.59
0.93
0.96
0.59
SD
0.59
0.59
0.17
0.59
0.59
0.17
TN
1.12
1.09
1.07
1.12
1.09
1.07
TX
0.99
0.89
0.47
0.99
0.89
0.47
UT
0.50
0.43
0.34
0.50
0.43
0.34
VA
0.89
0.85
0.54
0.88
0.85
0.54
VT
0.00
0.37
3.53
0.00
0.37
3.53
WA
0.08
0.23
0.79
0.08
0.23
0.79
WI
0.74
0.70
0.58
0.74
0.70
0.57
WV
1.19
1.12
0.33
1.19
1.12
0.33
WY
0.01
0.04
0.06
0.01
0.04
0.07
A-25
-------
Table A-6 Nitrate Scaling Factors for Coal EGU Tags in the Baseline and the Final
Rule
Baseline
Final Rule
State Tag
2028
2030
2035
2028
2030
2035
AL
1.33
1.45
1.65
1.33
1.45
1.65
AR
39.93
8.30
3.83
39.92
8.32
3.83
AZ
0.47
0.97
0.59
0.47
0.97
0.61
CA
0.24
0.36
0.16
0.24
0.36
0.16
CO
25.56
0.97
0.37
25.57
0.97
0.37
CT
0.00
0.00
0.00
0.00
0.00
0.00
DC
0.00
0.00
0.00
0.00
0.00
0.00
DE
0.00
0.00
0.00
0.00
0.00
0.00
FL
0.89
1.20
0.26
0.89
1.20
0.26
GA
0.23
0.12
0.00
0.23
0.12
0.00
IA
1.20
1.16
0.68
1.20
1.16
0.68
ID
0.00
0.00
0.00
0.00
0.00
0.00
IL
0.98
0.92
0.62
0.98
0.92
0.62
IN
1.29
0.64
0.11
1.28
0.65
0.11
KS
45.15
46.03
3.08
45.15
46.03
3.08
KY
1.38
1.12
1.15
1.38
1.12
1.16
LA
24.63
16.33
25.37
24.63
16.33
25.37
MA
0.00
0.00
0.00
0.00
0.00
0.00
MD
3.54
3.54
3.54
3.54
3.54
3.54
ME
0.00
0.00
0.00
0.00
0.00
0.00
MI
0.74
0.00
0.00
0.74
0.00
0.00
MN
2.97
2.31
0.00
2.97
2.31
0.00
MO
1.41
1.06
0.43
1.40
1.06
0.43
MS
4.02
3.60
1.06
4.01
3.60
1.06
MT
1.07
1.09
1.08
1.07
1.09
1.07
NC
19.19
11.95
3.66
19.22
11.95
3.67
ND
1.03
1.03
0.25
1.02
1.03
0.25
NE
1.14
1.13
0.61
1.14
1.13
0.62
NH
0.00
0.00
0.00
0.00
0.00
0.00
NJ
0.00
0.00
0.00
0.00
0.00
0.00
NM
0.99
0.99
0.01
0.99
0.99
0.01
NV
0.00
0.00
0.00
0.00
0.00
0.00
NY
0.00
0.00
0.00
0.00
0.00
0.00
OH
0.90
0.94
0.19
0.90
0.94
0.19
OK
12.10
5.08
3.11
12.08
5.07
3.11
OR
0.00
0.00
0.00
0.00
0.00
0.00
PA
3.05
2.94
2.61
3.05
2.94
2.61
RI
0.00
0.00
0.00
0.00
0.00
0.00
A-26
-------
Baseline
Final Rule
State Tag
2028
2030
2035
2028
2030
2035
SC
1.15
1.92
2.98
1.14
1.92
2.98
SD
0.93
1.11
0.00
0.93
1.11
0.00
TN
7.49
1.00
0.00
7.49
1.00
0.00
TX
1.02
1.13
0.87
1.02
1.13
0.87
UT
3.50
0.09
0.09
3.50
0.09
0.09
VA
0.67
0.41
0.12
0.67
0.41
0.12
VT
0.00
0.00
0.00
0.00
0.00
0.00
WA
0.00
0.00
0.00
0.00
0.00
0.00
WI
1.84
2.07
0.38
1.85
2.07
0.38
WV
1.25
1.30
0.97
1.25
1.30
0.97
WY
1.32
1.15
1.14
1.32
1.15
1.14
A-27
-------
Table A-7 Nitrate Scaling Factors for Gas EGU Tags in the Baseline and the Final Rule
Baseline
Final Rule
State Tag
2028
2030
2035
2028
2030
2035
AL
0.59
0.60
0.45
0.59
0.60
0.45
AR
0.56
0.68
0.38
0.55
0.68
0.38
AZ
0.73
0.85
0.83
0.73
0.85
0.83
CA
0.76
0.88
0.97
0.76
0.88
0.97
CO
2.02
0.71
0.72
2.02
0.71
0.72
CT
0.92
0.81
0.66
0.92
0.81
0.66
DC
0.63
0.47
0.26
0.63
0.47
0.26
DE
0.79
0.76
0.33
0.79
0.76
0.33
FL
1.11
1.06
1.01
1.10
1.06
1.01
GA
0.68
0.63
0.54
0.68
0.63
0.54
IA
0.49
0.42
0.13
0.49
0.42
0.13
ID
1.02
1.36
1.39
1.01
1.36
1.38
IL
0.54
0.54
0.29
0.53
0.54
0.29
IN
0.67
0.59
0.34
0.66
0.59
0.34
KS
0.96
0.87
0.20
0.96
0.88
0.20
KY
0.81
0.76
0.46
0.81
0.76
0.46
LA
0.96
0.94
0.61
0.96
0.94
0.61
MA
0.64
0.66
0.54
0.64
0.66
0.54
MD
1.47
1.35
1.05
1.47
1.35
1.05
ME
1.64
1.34
0.63
1.64
1.34
0.63
MI
0.65
0.71
0.43
0.65
0.71
0.43
MN
1.02
0.95
0.36
1.02
0.95
0.36
MO
0.52
0.52
0.19
0.52
0.52
0.19
MS
0.61
0.56
0.36
0.61
0.56
0.36
MT
0.66
0.80
0.05
0.66
0.80
0.06
NC
0.89
0.67
0.72
0.89
0.67
0.72
ND
0.66
1.32
0.26
0.65
1.31
0.26
NE
2.05
1.80
0.13
2.05
1.80
0.13
NH
0.78
0.59
0.44
0.78
0.59
0.44
NJ
0.82
0.83
0.51
0.82
0.83
0.52
NM
0.74
0.66
0.64
0.74
0.66
0.64
NV
0.50
0.39
0.44
0.50
0.39
0.44
NY
0.91
0.89
0.55
0.91
0.89
0.55
OH
1.00
0.98
0.87
1.00
0.98
0.87
OK
1.43
1.20
0.34
1.44
1.20
0.34
OR
5.58
0.96
0.50
5.59
0.96
0.49
PA
0.69
0.61
0.35
0.69
0.61
0.35
RI
0.76
0.76
0.64
0.77
0.76
0.64
SC
0.94
0.96
0.67
0.94
0.96
0.67
SD
0.55
0.55
0.16
0.55
0.55
0.16
A-28
-------
Baseline
Final Rule
State Tag
2028
2030
2035
2028
2030
2035
TN
1.02
0.97
0.79
1.02
0.97
0.80
TX
0.97
0.88
0.42
0.97
0.89
0.42
UT
0.52
0.62
0.56
0.52
0.62
0.56
VA
0.84
0.80
0.43
0.84
0.80
0.43
VT
0.10
0.16
1.53
0.10
0.16
1.53
WA
0.43
0.36
0.72
0.43
0.36
0.72
WI
0.66
0.67
0.45
0.66
0.67
0.44
WV
1.02
0.89
0.22
1.02
0.89
0.22
WY
0.01
0.04
0.06
0.01
0.04
0.06
A-29
-------
Table A-8 Sulfate Scaling Factors for Coal EGU Tags in the Baseline and the Final Rule
Baseline
Final Rule
State Tag
2028
2030
2035
2028
2030
2035
AL
4.96
5.39
7.07
4.96
5.39
7.07
AR
118.10
7.02
4.45
118.07
7.04
4.45
AZ
0.48
1.42
1.16
0.48
1.42
1.16
CA
0.33
0.50
0.26
0.33
0.50
0.26
CO
14.31
0.98
0.20
14.31
0.98
0.20
CT
0.00
0.00
0.00
0.00
0.00
0.00
DC
0.00
0.00
0.00
0.00
0.00
0.00
DE
0.00
0.00
0.00
0.00
0.00
0.00
FL
0.98
1.16
0.50
0.98
1.16
0.50
GA
0.04
0.09
0.00
0.04
0.09
0.00
IA
1.31
1.25
0.78
1.31
1.25
0.78
ID
0.00
0.00
0.00
0.00
0.00
0.00
IL
1.01
0.73
0.48
1.01
0.73
0.48
IN
0.89
0.56
0.12
0.89
0.56
0.12
KS
52.35
51.92
11.39
52.35
51.92
11.39
KY
2.68
2.12
1.88
2.68
2.11
1.88
MA
0.00
0.00
0.00
0.00
0.00
0.00
MD
3.54
3.54
3.54
3.54
3.54
3.54
ME
0.00
0.00
0.00
0.00
0.00
0.00
MI
0.85
0.00
0.00
0.85
0.00
0.00
MN
1.68
1.47
0.00
1.68
1.47
0.00
MO
2.20
1.08
0.71
2.20
1.08
0.71
MS
4.02
3.60
1.06
4.01
3.60
1.06
MT
1.85
2.06
1.92
1.85
2.06
1.86
NC
7.31
5.14
1.88
7.32
5.14
1.88
ND
0.94
1.00
0.94
0.94
1.00
0.94
NE
0.96
0.95
0.58
0.96
0.95
0.58
NH
0.00
0.00
0.00
0.00
0.00
0.00
NJ
0.00
0.00
0.00
0.00
0.00
0.00
NM
1.00
1.00
0.01
1.00
1.00
0.01
NV
0.00
0.00
0.00
0.00
0.00
0.00
NY
0.00
0.00
0.00
0.00
0.00
0.00
OH
0.78
0.61
0.29
0.78
0.60
0.29
OK
37.84
4.77
2.54
37.83
4.77
2.54
OR
0.00
0.00
0.00
0.00
0.00
0.00
PA
4.25
4.06
3.94
4.25
4.06
3.94
RI
0.00
0.00
0.00
0.00
0.00
0.00
SC
0.73
1.22
1.76
0.73
1.22
1.76
SD
1.05
1.27
0.00
1.06
1.27
0.00
TN
20.55
1.57
0.00
20.55
1.57
0.00
A-30
-------
Baseline
Final Rule
State Tag
2028
2030
2035
2028
2030
2035
TXLAa
1.86
2.39
2.25
1.86
2.39
2.25
UT
0.93
0.06
0.06
0.93
0.06
0.06
VA
0.11
0.07
0.02
0.11
0.07
0.02
VT
0.00
0.00
0.00
0.00
0.00
0.00
WA
0.00
0.00
0.00
0.00
0.00
0.00
WI
3.50
3.83
1.15
3.51
3.83
1.14
WV
1.40
1.39
1.08
1.40
1.39
1.08
WY
1.26
0.98
0.97
1.26
0.98
0.97
Note: Emissions from Louisiana are less than 10 tpy in the original source apportionment modeling. Air quality
impacts and emissions from Texas and Louisiana were combined.
aTXLA: Louisiana and Texas
A-31
-------
Table A-9 Primary PM2.5 Scaling Factors for Coal EGU Tags in the Baseline and the
Final Rule
Baseline
Final Rule
State Tag
2028
2030
2035
2028
2030
2035
AL
1.20
1.31
1.43
1.20
1.31
1.43
AR
20.02
7.10
3.14
19.96
7.12
3.14
AZ
0.38
1.17
0.61
0.38
1.17
0.61
CA
0.24
0.36
0.16
0.24
0.36
0.16
CO
13.37
1.19
0.51
13.38
1.19
0.51
CT
0.00
0.00
0.00
0.00
0.00
0.00
DC
0.00
0.00
0.00
0.00
0.00
0.00
DE
0.00
0.00
0.00
0.00
0.00
0.00
FL
1.40
1.84
0.25
1.38
1.82
0.25
GA
0.03
0.06
0.00
0.03
0.06
0.00
IA
1.17
1.14
0.67
1.17
1.14
0.67
ID
0.00
0.00
0.00
0.00
0.00
0.00
IL
1.17
0.95
0.57
1.15
0.95
0.57
IN
1.28
0.60
0.20
1.28
0.60
0.20
KY
1.30
1.19
0.77
1.28
1.17
0.75
MA
0.00
0.00
0.00
0.00
0.00
0.00
MD
3.54
3.54
3.54
3.54
3.54
3.54
ME
0.00
0.00
0.00
0.00
0.00
0.00
MI
0.83
0.00
0.00
0.83
0.00
0.00
MN
3.50
2.70
0.00
3.51
2.70
0.00
MO
3.04
1.33
0.54
2.78
1.33
0.54
MS
4.02
3.60
1.06
3.33
2.99
0.88
MT
0.98
0.98
0.98
0.71
0.71
0.72
NC
21.57
17.32
6.08
21.44
17.30
6.09
ND
0.94
0.98
0.78
0.94
0.98
0.78
NEKS3
3.70
3.68
0.80
3.70
3.68
0.80
NH
0.00
0.00
0.00
0.00
0.00
0.00
NJ
0.00
0.00
0.00
0.00
0.00
0.00
NM
0.98
0.99
0.01
0.98
0.99
0.01
NV
0.00
0.00
0.00
0.00
0.00
0.00
NY
0.00
0.00
0.00
0.00
0.00
0.00
OH
0.83
1.08
0.19
0.83
1.08
0.19
OK
14.75
8.14
8.94
14.74
8.12
8.94
OR
0.00
0.00
0.00
0.00
0.00
0.00
PA
3.12
3.04
2.28
2.98
2.91
2.15
RI
0.00
0.00
0.00
0.00
0.00
0.00
SC
1.03
2.17
3.78
1.03
2.17
3.78
SD
0.93
1.11
0.00
0.93
1.11
0.00
A-32
-------
Baseline
Final Rule
State Tag
2028
2030
2035
2028
2030
2035
TN
16.88
1.00
0.00
16.88
1.00
0.00
TXLAb
1.10
1.30
1.15
1.10
1.30
1.15
UT
2.92
0.06
0.06
2.92
0.06
0.06
VA
0.46
0.29
0.08
0.46
0.29
0.08
VT
0.00
0.00
0.00
0.00
0.00
0.00
WA
0.00
0.00
0.00
0.00
0.00
0.00
WI
2.11
2.36
0.46
2.13
2.36
0.46
WV
1.29
1.45
1.23
1.22
1.38
1.17
WY
1.03
1.10
1.08
1.02
1.09
1.07
Note: Emissions from Louisiana and Kansas are less than 10 tpy in the original source apportionment modeling. Air
quality impacts and emissions from those states were combined with nearby states.
aNEKS: Nebraska and Kansas
bTXLA: Louisiana and Texas
A-33
-------
Table A-10 Primary PM2.5 Scaling Factors for Gas EGU Tags in the Baseline and the
Final Rule
Baseline
Final Rule
State Tag
2028
2030
2035
2028
2030
2035
AL
0.85
0.84
0.71
0.85
0.84
0.71
AR
0.63
0.82
0.43
0.63
0.82
0.43
AZ
0.70
0.85
0.86
0.70
0.85
0.86
CA
0.96
1.06
0.98
0.96
1.06
0.98
CO
1.23
0.74
0.77
1.23
0.74
0.77
CT
0.78
0.67
0.60
0.78
0.67
0.60
DC
0.15
0.13
0.11
0.15
0.13
0.11
DE
0.62
0.64
0.31
0.62
0.64
0.31
FL
0.97
0.98
0.95
0.97
0.98
0.95
GA
0.84
0.81
0.72
0.84
0.81
0.72
IA
0.50
0.48
0.20
0.51
0.47
0.20
ID
1.22
1.65
1.68
1.22
1.65
1.67
IL
0.49
0.55
0.28
0.49
0.55
0.28
IN
0.67
0.67
0.44
0.67
0.67
0.44
KS
1.12
1.02
0.19
1.12
1.02
0.19
KY
0.75
0.72
0.49
0.74
0.72
0.49
LA
0.79
0.80
0.64
0.79
0.80
0.64
MA
0.48
0.46
0.34
0.48
0.46
0.34
MD
1.05
1.08
0.85
1.05
1.09
0.85
ME
1.75
1.44
0.51
1.75
1.44
0.52
MI
0.75
0.87
0.63
0.75
0.87
0.63
MN
0.57
0.52
0.21
0.57
0.52
0.21
MO
0.30
0.33
0.10
0.30
0.33
0.10
MS
0.88
0.84
0.51
0.88
0.85
0.51
MT
0.17
0.21
0.03
0.17
0.21
0.04
NC
0.87
0.70
0.76
0.87
0.69
0.76
ND
0.47
0.92
0.19
0.46
0.91
0.19
NE
2.17
2.04
0.27
2.17
2.04
0.28
NH
0.59
0.43
0.31
0.59
0.43
0.31
NJ
0.82
0.84
0.52
0.82
0.84
0.52
NM
0.52
0.52
0.89
0.52
0.52
0.89
NV
0.72
0.84
0.83
0.72
0.84
0.83
NY
0.86
0.85
0.59
0.86
0.85
0.59
OH
0.95
0.95
0.89
0.95
0.95
0.89
OK
1.00
0.79
0.22
1.01
0.79
0.22
OR
3.29
0.74
0.39
3.30
0.74
0.39
PA
0.83
0.80
0.60
0.83
0.80
0.60
A-34
-------
Baseline
Final Rule
State Tag
2028
2030
2035
2028
2030
2035
RI
0.83
0.78
0.65
0.83
0.78
0.65
SC
0.80
0.86
0.64
0.80
0.86
0.64
SD
0.73
0.73
0.25
0.73
0.73
0.25
TN
1.08
1.05
0.88
1.08
1.05
0.88
TX
0.90
0.83
0.45
0.90
0.83
0.45
UT
0.66
0.87
0.84
0.66
0.87
0.84
VA
0.81
0.73
0.47
0.81
0.73
0.48
VT
0.00
0.00
0.03
0.00
0.00
0.03
WA
0.44
0.48
0.58
0.44
0.48
0.58
WI
0.56
0.66
0.43
0.56
0.66
0.42
WV
0.51
0.38
0.10
0.51
0.38
0.10
WY
0.01
0.04
0.03
0.01
0.04
0.04
A-35
-------
Table A-ll Scaling Factors for Other EGU Tags in the Baseline and the Final Rule
Baseline
Final Rule
Pollutants
2028
2030
2035
2028
2030
2035
Seasonal NOx
1.16
1.16
1.10
1.16
1.16
1.10
Annual NOx
1.17
1.17
1.11
1.17
1.17
1.11
Annual S02
1.00
1.01
1.00
1.00
1.01
1.00
Annual PM2.5
1.37
1.37
1.32
1.37
1.37
1.32
A.5 Air Quality Surface Results
The spatial fields of model-predicted air quality changes between the baseline and the
two regulatory options in 2028, 2030, and 2035 for AS-M03 are presented in Figure A-8. It is
important to recognize that ozone is a secondary pollutant, meaning that it is formed through
chemical reactions of precursor emissions in the atmosphere. As a result of the time necessary
for precursors to mix in the atmosphere and for these reactions to occur, ozone can either be
highest at the location of the precursor emissions or peak at some distance downwind of those
emissions sources. The spatial gradients of ozone depend on a multitude of factors including the
spatial patterns of NOx and VOC emissions and the meteorological conditions on a particular
day. Thus, on any individual day, high ozone concentrations may be found in narrow plumes
downwind of specific point sources, may appear as urban outflow with large concentrations
downwind of urban source locations or may have a more regional signal. However, in general,
because the AS-M03 metric is based on the average of concentrations over more than 180 days
in the spring and summer, the resulting spatial fields are rather smooth without sharp gradients,
compared to what might be expected when looking at the spatial patterns of MDA8 ozone
concentrations on specific high ozone episode days. Air quality changes in these figures are
calculated as the final rule minus the baseline. The spatial patterns shown in the figures are a
result of (1) the spatial distribution of EGU sources that are predicted to have changes in
emissions and (2) the physical or chemical processing that the model simulates in the
atmosphere. The spatial fields used to create these maps serve as an input to the benefits analysis
and the EJ analysis. While total U.S. NOx emissions are predicted to decrease in the final rule
scenario for 2028 and 2030 when compared to the baseline, predicted NOx emissions changes
are heterogeneous across the country with increases predicted in some states. In 2035, NOx
A-36
-------
emissions across the contiguous 48 states included in this analysis are predicted to increase
compared to the baseline. In Figure A-8126 there are small predicted ozone decreases from the
final rule compared to the baseline evident in North Dakota in 2028 and Montana in 2035. There
are also small predicted ozone increases from the final rule compared to the baseline evident near
the border of Arizona and New Mexico in 2035.
Figure A-9 presents the model-predicted air quality changes between the baseline and
the final regulatory option in 2028, 2030, and 2035 for PM2.5.127 Secondary PM2.5 species sulfate
and nitrate often demonstrate regional signals without large local gradients while primary PM2.5
components often have heterogenous spatial patterns with larger gradients near emissions
sources. Air quality changes in these figures are calculated as the final rule minus the baseline.
The spatial patterns shown are a result of (1) the spatial distribution of EGU sources that are
predicted to have changes in emissions and (2) the physical or chemical processing that the
model simulates in the atmosphere. The spatial fields used to create these maps serve as an input
to the benefits analysis and the EJ analysis. Both secondary and primary PM2.5 contribute to the
spatial patterns shown in Figure A-9. In 2028, there are predicted PM2.5 decreases from the final
rule evident in Montana, North Dakota, Missouri, West Virginia, and Pennsylvania. In Montana,
West Virginia, and Pennsylvania, these PM2.5 changes coincide with predicted decreases in direct
PM2.5 emissions. In North Dakota and Missouri, emissions of NOx, SO2 and direct PM2.5 are all
predicted to decrease compared to the baseline in 2028. In 2030 and 2035, there are predicted
PM2.5 decreases from the final rule evident in Montana, West Virginia, and Pennsylvania. In
2030 those predicted PM2.5 concentration decreases coincide with direct PM2.5 emissions
decreases from all three states. In 2035 the predicted PM2.5 concentration decreases coincide with
SO2, NOx, and direct PM2.5 decreases from Montana and West Virginia and direct PM2.5
decreases from Pennsylvania in 2035.
126 Note scale change on maps compared to similar figures from the proposal RIA. Color scale presented in figure 8-
8 has a range of -0.11 ppb to 0.11 ppb. Maps in the proposal used a scale range from -0.2 ppb to 0.2 ppb.
127 Note scale change on maps compared to similar figures from the proposal RIA. Color scale presented in figure 8-
9 has a range of -0.011 |ig/m3 to 0.011 |ig/m3. Maps in the proposal used a scale range from -0.05 |ig/m3 to 0.05
|ig/m3.
A-37
-------
Final Rule 03 Impacts in 2028 Final RuleO, Impacts in 2030 Final Rule03 Impacts in 2035
Figure A-8 Maps of change in ASM-OS for the final rule compared to baseline values (ppb)
shown in 2028 (right panel), 2030 (middle panel) and 2035 (right panel)
Final Rule PM25 Impacts in 2028
1 k . v-C
Final Rule PM2 5 Impacts in 2030 Final Rule PM2 5 Impacts in 2035
j
L J H-
T- PJn.Ai?*
1>- - jr y\
/ N
If ir
Figure A-9 Maps of change in PM2.5 for the final rule compared to baseline values
(jig/m3) shown in 2028 (right panel), 2030 (middle panel) and 2035 (right panel)
A.6 Uncertainties and Limitations of the Air Quality Methodology
One limitation of the scaling methodology for creating ozone and PM2.5 surfaces
associated with the baseline or regulatory control alternatives described above is that the
methodology treats air quality changes from the tagged sources as linear and additive. It
therefore does not account for nonlinear atmospheric chemistry and does not account for
interactions between emissions of different pollutants and between emissions from different
tagged sources. The method applied in this analysis is consistent with how air quality estimations
have been made in several prior regulatory analyses (U.S. EPA, 2012, 2019, 2020a). We note
that air quality is calculated in the same manner for the baseline and for the final rule, so any
uncertainties associated with these assumptions are propagated through results for both the
A-38
-------
baseline and the final rule in the same manner. In addition, emissions changes between baseline
and the final rule are relatively small compared to modeled future year emissions that form the
basis of the source apportionment approach described in this appendix. Previous studies have
shown that air pollutant concentrations generally respond linearly to small emissions changes of
up to 30 percent (Cohan et al., 2005; Cohan and Napelenok, 2011; Dunker et al., 2002; Koo et
al., 2007; Napelenok et al., 2006; Zavala et al., 2009). A second limitation is that the source
apportionment contributions are informed by the spatial and temporal distribution of the
emissions from each source tag as they occur in the future year modeled case. Thus, the
contribution modeling results do not allow us to consider the effects of any changes to spatial
distribution of EGU emissions within a state between the future year modeled case and the
baseline and final rule scenarios analyzed in this RIA. Finally, the CAMx-modeled
concentrations themselves have some uncertainty. While all models have some level of inherent
uncertainty in their formulation and inputs, the base-year 2016 model outputs have been
evaluated against ambient measurements and have been shown to adequately reproduce spatially
and temporally varying concentrations (U.S. EPA, 2023a, 2024).
A.7 References
Cohan, D. S., Hakami, A., Hu, Y., & Russell, A. G. (2005). Nonlinear Response of Ozone to
Emissions: Source Apportionment and Sensitivity Analysis. Environmental Science &
Technology, 39(17), 6739-6748. doi:10.1021/es048664m
Cohan, D. S., & Napelenok, S. L. (2011). Air Quality Response Modeling for Decision Support.
Atmosphere, 2(3), 407-425. Retrieved from https://www.mdpi.eom/2073-4433/2/3/407
Ding, D., Zhu, Y., Jang, C., Lin, C.-J., Wang, S., Fu, J., . . . 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.
doi:https://doi.org/10.1016/j.jes.2015.06.003
Dunker, A. M., Yarwood, G., Ortmann, J. P., & Wilson, G. M. (2002). The Decoupled Direct
Method for Sensitivity Analysis in a Three-Dimensional Air Quality Model
Implementation, Accuracy, and Efficiency. Environmental Science & Technology,
36(13), 2965-2976. doi:10.1021/es0112691
Gold, C. M., Remmele, P. R., & Roos, T. (1997). Voronoi methods in GIS. In M. van Kreveld, J.
Nievergelt, T. Roos, & P. Widmayer (Eds.), Algorithmic Foundations of Geographic
Information Systems (pp. 21-35). Berlin, Heidelberg: Springer Berlin Heidelberg.
A-39
-------
Koo, B., Dunker, A. M., & Yarwood, G. (2007). Implementing the Decoupled Direct Method for
Sensitivity Analysis in a Particulate Matter Air Quality Model. Environmental Science &
Technology, 41(8), 2847-2854. doi: 10.102l/es0619962
Napelenok, S. L., Cohan, D. S., Hu, Y., & Russell, A. G. (2006). Decoupled direct 3D sensitivity
analysis for particulate matter (DDM-3D/PM). Atmospheric Environment, 40(32), 6112-
6121. doi:https://doi.org/10.1016/j.atmosenv.2006.05.039
Ramboll Environ. (2021). User's Guide Comprehensive Air Quality Model with Extensions
version 7.10. Retrieved from Novato, CA:
U.S. EPA. (2007). Technical Report on Ozone Exposure, Risk, and Impact Assessments for
Vegetation. (EPA 452/R-07-002). Research Triangle Park, NC: Office of Air Quality
Planning and Standards. https://nepis.epa.gov/Exe/ZyPURL.cgi?Dockey=P100PVGI.txt
U.S. EPA. (2012). Regulatory Impact Analysis for the Final Revisions to the National Ambient
Air Quality Standards for Particulate Matter. (EPA-452/R-12-005). Research Triangle
Park, NC: U.S. Environmental Protection Agency, Office of Air Quality Planning and
Standards, Health and Environmental Impact Division.
https://www3.epa.gov/ttnecasl/regdata/RIAs/finalria.pdf
U.S. EPA. (2019). 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. (EPA-452/R-19-003). Research Triangle Park, NC: U.S.
Environmental Protection Agency, Office of Air Quality Planning and Standards, Health
and Environmental Impact Division, https://www.epa.gov/sites/production/files/2019-
06/documents/utilities_ria_final_cpp_repeal_and_ace_2019-06.pdf
U.S. EPA. (2020a). Analysis of Potential Costs and Benefits for the National Emission Standards
for Hazardous Air Pollutants: Coal- and Oil-Fired Electric Utility Steam Generating
Units - Subcategory of Certain Existing Electric Utility Steam Generating Units Firing
Eastern Bituminous Coal Refuse for Emissions of Acid Gas Hazardous Air Pollutants.
Research Triangle Park, NC: U.S. Environmental Protection Agency, Office of Air
Quality Planning and Standards, Health and Environmental Impact Division.
https://www.epa.gOv/sites/default/files/2020-04/documents/mats_coal_refuse_cost-
benefit_memo.pdf
U.S. EPA. (2020b). Benefit and Cost Analysis for Revisions to the Effluent Limitations
Guidelines and Standards for the Steam Electric Power Generating Point Source
Category. (EPA-821-R-20-003). Washington DC: U.S. Environmental Protection
Agency, https://www.epa.gov/sites/default/files/2020-
08/documents/steam_electric_elg_2020_final_reconsideration_rule_benefit_and_cost_an
alysis.pdf
U.S. EPA. (2021a). Flat File Generation Methodology: Version: Summer 2021 Reference Case
using EPA Platform v6. U.S. Environmental Protection Agency.
https://www.epa.gov/system/files/documents/2021-09/flat-file-methodology-epa-
platform-v6-summer-2021-reference-case.pdf
A-40
-------
U.S. EPA. (2021b). Regulatory Impact Analysis for the Final Revised Cross-State Air Pollution
Rule (CSAPR) Update for the 2008 Ozone NAAQS. (EPA-452/R-21-002). Research
Triangle Park, NC: U.S. Environmental Protection Agency, Office of Air Quality
Planning and Standards, Health and Environmental Impact Division.
https://www.epa.gov/sites/default/files/2021-
03/documents/revi sed_csapr_update_ria_final.pdf
U.S. EPA. (2022a). Regulatory Impact Analysis for Proposed Federal Implementation Plan
Addressing Regional Ozone Transport for the 2015 Ozone National Ambient Air Quality
Standard. (EPA-452/D-22-001). Research Triangle Park, NC: U.S. Environmental
Protection Agency, Office of Air Quality Planning and Standards, Health and
Environmental Impact Division, https://www.epa.gov/system/files/documents/2022-
03/transport_ria_proposal_fip_2015_ozone_naaqs_2022-02.pdf
U.S. EPA. (2022b). Software for Model Attainment Test - Community Edition (SMAT-CE)
User's Guide Software version 2.1. (EPA-454/B-22-013). Research Triangle Park, NC.
https://www.epa.gov/system/files/documents/2022-
1 l/User%27s%20Manual%20for%20SMAT-CE%202. l_EPA_Report_l l_30_2022.pdf
U.S. EPA. (2023a). Air Quality Modeling Final Rule Technical Support Document: 2015 Ozone
NAAQS Good Neighbor Plan. Research Triangle Park, NC: U.S. Environmental
Protection Agency, Office of Air Quality Planning and Standards.
https://www.epa.gov/system/files/documents/2023-
03/AQ%20Modeling%20Final%20Rule%20TSD.pdf
U.S. EPA. (2023b). Technical Support Document (TSD): Preparation of Emissions Inventories
for the 2016v3 North American Emissions Modeling Platform. (EPA-454/B-23-002).
Research Triangle Park, NC: U.S. Environmental Protection Agency, Office of Air
Quality Planning and Standards, https://www.epa.gov/system/files/documents/2023-
03/2016v3_EmisMod_TSD_January2023_l.pdf
U.S. EPA. (2024). Air Quality Modeling Technical Support Document: PM2.5 Model Evaluation
for 2016 CAMx Modeling to Support Multiple 2024 EGU Rulemakings. Research
Triangle Park, NC: U.S. Environmental Protection Agency, Office of Air Quality
Planning and Standards
Zavala, M., Lei, W., Molina, M. J., & Molina, L. T. (2009). Modeled and observed ozone
sensitivity to mobile-source emissions in Mexico City. Atmos. Chem. Phys., 9(1), 39-55.
doi: 10.5194/acp-9-3 9-2009
A-41
-------
APPENDIX B: CLIMATE BENEFITS APPENDIX
B.l Climate Benefits Estimated using the Interim SC-CO2 values used in the Proposal
This appendix presents the climate benefits of the final standards using the interim SC-
CO2 values used in the proposal of this rulemaking. The interim SC-CO2 values are presented in
Table B-l and the climate benefits using these values are presented in Table B-2.
Table B-l Interim SC-CO2 Values, 2028 to 2037 (2019 dollars per metric ton)
Discount Rate and Statistic
5%
3%
2.50%
3%
Emissions Year
Average
Average
Average
95th Percentile
2028
$16
$54
$79
$160
2029
$16
$55
$81
$160
2030
$17
$56
$82
$170
2031
$17
$57
$83
$170
2032
$18
$58
$85
$170
2033
$18
$59
$86
$180
2034
$19
$60
$87
$180
2035
$19
$61
$88
$180
2036
$20
$62
$90
$190
2037
$20
$63
$91
$190
Note: These SC-CO2 values are identical to those reported in the 2016 SC-GHG TSD (IWG, 2016) adjusted for
inflation to 2019 dollars using the annual GDP Implicit Price Deflator values in the U.S. Bureau of Economic
Analysis' (BEA) NIPA Table 1.1.9 (U.S. BEA, 2021). The values are stated in $/metric ton CO2 (1 metric ton equals
1.102 short tons) and vary depending on the year of CO2 emissions. This table displays the values rounded to the
nearest dollar; the annual unrounded values used in the calculations in this RIA are available on OMB's website:
https://www.whitehouse.gov/omb/information-regulatory-affairs/regulatory-mattersMscghgs.
Source: Technical Support Document: Social Cost of Carbon, Methane, and Nitrous Oxide Interim Estimates under
E.O. 13990 (IWG, 2021).
B-l
-------
Table B-2 Stream of Projected Climate Benefits under the Final Rule from 2028 to 2037
(millions of 2019 dollars, discounted to 2023)
SC-CO2 Discount Rate and Statistic
5%
3%
2.50%
3%
Emissions Year
Average
Average
Average
95th Percentile
2028*
$0.9
$3.3
$5.0
$10
2029
$0.9
$3.3
$4.9
$9.9
2030*
-$0.49
-$1.8
-$2.7
-$5.4
2031
-$0.48
-$1.8
-$2.7
-$5.4
2032
$1.3
$4.8
$7.2
$15
2033
$1.3
$4.7
$7.1
$14
2034
$1.2
$4.7
$7.1
$14
2035*
$1.2
$4.6
$7.0
$14
2036
$1.2
$4.6
$6.9
$14
2037
$1.2
$4.5
$6.9
$14
PV
$8.2
$31
$47
$94
EAV
$1.1
$3.6
$5.3
$11
Note: Climate benefits are based on reductions in CO2 emissions and are calculated using the IWG interim SC-CO2
estimates from IWG (2021).
B.2 References
IWG. (2016). Addendum to Technical Support Document on Social Cost of Carbon for
Regulatory Impact Analysis under Executive Order 12866: Application of the
Methodology to Estimate the Social Cost of Methane and the Social Cost of Nitrous
Oxide. Washington DC: U.S. Government, Interagency Working Group (IWG) on Social
Cost of Greenhouse Gases, https://www.epa.gov/sites/default/files/2016-
12/documents/addendum_to_sc-ghg_tsd_august_2016.pdf
IWG. (2021). Technical Support Document: Social Cost of Carbon, Methane, and Nitrous Oxide
Interim Estimates under Executive Order 13990. Washington DC: U.S. Government,
Interagency Working Group (IWG) on Social Cost of Greenhouse Gases.
https://www.whitehouse.gov/wp-
content/uploads/2021/02/TechnicalSupportDocument_SocialCostofCarbonMethaneNitro
usOxide.pdf?source=email
U.S. BEA. (2021). Table 1.1.9. Implicit Price Deflators for Gross Domestic Product.
Washington, DC.
https://apps.bea.gov/iTable/?reqid=19&step=3&isuri=l&1921=survey& 1903=13
B-2
-------
United States Office of Air Quality Planning and Standards Publication No. EPA-452/R-24-005
Environmental Protection Health and Environmental Impacts Division April 2024
Agency Research Triangle Park, NC
------- |