I Q \
V7
*4 PRO^
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
December 2023
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
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 (email:
OAQPSeconomics@epa.gov).
ACKNOWLEDGEMENTS
In addition to U.S. EPA staff from the Office of Air and Radiation, personnel from SC&A
contributed data and analysis to this document.
iv
-------
TABLE OF CONTENTS
Table of Contents v
List of Tables vii
List of Figures ix
1 Executive Summary 1-1
1.1 Introduction 1-1
1.2 Legal and Economic Basis for this Rulemaking 1-2
1.2.1 Statutory Requirements 1-3
1.2.2 Market F ailure 1-4
1.3 Baseline and Regulatory Requirements 1-5
1.4 Summary of Projected Emissions Reductions and Benefit-Cost Analysis 1-11
1.5 Organization of RIA 1-15
1.6 References 1-16
2 Projected Compliance Costs and Emissions Reductions 2-17
2.1 Emissions Sources and Regulatory Requirements Analyzed in this RIA 2-17
2.1.1 Emissions Sources 2-18
2.1.2 Regulatory Requirements 2-24
2.2 Methodology 2-29
2.2.1 Activity Data Projections 2-31
2.2.2 Model Plant Compliance Cost and Emissions Reductions 2-49
2.2.3 State Programs 2-56
2.3 Emissions Reductions 2-56
2.4 Product Recovery 2-57
2.5 Compliance Costs 2-59
2.6 Comparison of Regulatory Alternatives 2-63
2.7 Additional Information on Use of 2016 Oil and Gas ICR Data 2-66
2.8 References 2-70
3 Benefits 3-1
3.1 Emissions Reductions 3-3
3.2 Methane Climate Effects and Valuation 3-4
3.3 Ozone-Related Health Impacts Due to VOC Emissions Changes 3-25
3.3.1 Developing Air Quality Surfaces of Ozone Impacts from VOC Emissions Changes 3-28
3.3.2 Selecting Air Pollution Health Endpoints to Quantify 3-30
3.3.3 Calculating Counts of Air Pollution Effects Using the Health Impact Function 3-33
3.3.4 Calculating the Economic Valuation of Health Impacts 3-34
3.3.5 Benefits Analysis Data Inputs 3-35
3.3.6 Quantifying Cases of Ozone-Attributable Premature Death 3-39
3.3.7 Characterizing Uncertainty in the Estimated Benefits 3-40
3.3.8 Estimated Number and Economic Value of Health Benefits 3-43
3.4 Ozone Vegetation Effects 3-52
3.5 Ozone Climate Effects 3-53
3.6 Ozone-Related Impacts Due to Methane 3-53
3.7 PM2 5-Related Impacts Due to VOC Emissions 3-54
3.7.1 PM2 5 Health Effects 3-55
3.7.2 PM25 Welfare Effects 3-56
3.8 Hazardous Air Pollutants (HAP) Impacts 3-56
3.8.1 Benzene 3-58
3.8.2 Formaldehyde 3-58
3.8.3 Toluene 3-59
v
-------
3.8.4 Carbonyl Sulfide 3-60
3.8.5 Ethylbenzene 3 -60
3.8.6 Mixed Xylenes 3-61
3.8.7 n-Hexane 3-61
3.8.8 Other Air Toxics 3-62
3.9 Secondary Air Emissions Impacts 3-62
3.10 Total Benefits 3-65
3.11 References 3-72
4 Economic Impact and Distributional Analysis 4-1
4.1 Oil and Natural Gas Market Impact Analysis 4-1
4.1.1 Crude Oil Market Model 4-2
4.1.2 Natural Gas Market Model 4-3
4.1.3 Assumptions, Data, and Parameters Used in the Oil and Natural Gas Market Models 4-4
4.1.4 Results 4-6
4.1.5 Caveats and Limitations of the Market Analysis 4-8
4.2 Financial Analysis of Marginal Wells 4-9
4.2.1 Descriptive Statistics on Marginal Wells 4-10
4.2.2 Marginal Wells Financial Analysis Model 4-12
4.2.3 Results 4-15
4.2.4 Marginal Wells Financial Analysis Caveats and Limitations 4-19
4.3 Environmental Justice Analyses 4-23
4.3.1 Analyzing EJ Impacts in this Final Action 4-25
4.3.2 Qualitative Discussion of Disparate Climate Vulnerabilities in the Baseline 4-27
4.3.3 Ozone from Oil and Natural Gas VOC Emission Impacts 4-31
4.3.4 Air Toxics Impacts 4-45
4.3.5 Demographic Characteristics of Oil and Natural Gas Workers and Communities 4-53
4.3.6 Household Energy Expenditures 4-59
4.3.7 Summary 4-62
4.4 Final Regulatory Flexibility Analysis 4-63
4.4.1 Reasons Why Action is Being Considered 4-63
4.4.2 Statement of Objectives and Legal Basis for the Final Rules 4-64
4.4.3 Significant Issues Raised 4-66
4.4.4 Small Business Administration Comments 4-67
4.4.5 Description and Estimate of Affected Small Entities 4-68
4.4.6 Compliance Cost Impact Estimates 4-70
4.4.7 Caveats and Limitations 4-74
4.4.8 Projected Reporting, Recordkeeping and Other Compliance Requirements 4-75
4.4.9 Related Federal Rules 4-76
4.4.10 Regulatory Flexibility Alternatives 4-77
4.5 Employment Impacts of Environmental Regulation 4-91
4.6 References 4-92
5 Comparison of Benefits and Costs 5-1
5.1 Comparison of Benefits and Costs 5-1
5.2 Uncertainties and Limitations 5-10
5.3 References 5-15
vi
-------
LIST OF TABLES
Table 1-1 NSPS OOOOb Emissions Sources, Baseline Requirements, and Requirements under the Final Rule . 1-6
Table 1-2 EG 0000c Emissions Sources, Baseline Requirements, and Requirements under the Final Rule 1-8
Table 1-3 Projected Emissions Reductions under the Final NSPS OOOOb and EG 0000c Option, 2024-2038ab c
1-11
Table 1-4 Projected Benefits, Compliance Costs, and Emissions Reductions for the Finalized NSPS OOOOb,
2024-2038 (million 2019$) 1-13
Table 1-5 Projected Benefits, Compliance Costs, and Emissions Reductions for the Finalized EG 0000c, 2024-
2038 (million 2019$) 1-14
Table 1-6 Projected Benefits, Compliance Costs, and Emissions Reductions for the Finalized NSPS OOOOb and
EG 0000c, 2024-2038 (million 2019$) 1-15
Table 2-1 NSPS OOOOb Emissions Sources, Baseline Requirements, and Finalized Requirements 2-25
Table 2-2 EG 0000c Emissions Sources, Baseline Requirements, and Finalized Requirements 2-27
Table 2-3 Assumed Retirement Rates and Annual New Site Counts by Site Type 2-33
Table 2-4 Decline Rate Assumptions by Production Type and Rate 2-35
Table 2-5 Distribution of Well Sites in Equipment Bins 2-42
Table 2-6 Projection of Incrementally Impacted Affected Facilities under the Final NSPS OOOOb and EG
0000c, 2024 to 2038 (Production Sources) 2-47
Table 2-7 Projection of Incrementally Impacted Affected Facilities under the Final NSPS OOOOb and EG
0000c, 2024 to 2038 (Non-Production Fugitive/Leaks and Pneumatics Sources) 2-48
Table 2-8 Projection of Incrementally Impacted Affected Facilities under the Final NSPS OOOOb and EG
0000c, 2024 to 2038 (Non-Production Compressor Sources) 2-49
Table 2-9 Projected Emissions Reductions under the Final NSPS OOOOb and EG 0000c, 2024-2038 2-57
Table 2-10 Projected Increase in Natural Gas Recovery under the Final NSPS OOOOb and EG 0000c Option,
2024-2038 2-58
Table 2-11 Projected Compliance Costs under the Final NSPS OOOOb and EG 0000c Option, 2024-2038
(millions 2019$) 2-60
Table 2-12 Undiscounted Projected Compliance Costs under the Final NSPS OOOOb and EG 0000c, 2024-2038
(millions 2019$) 2-62
Table 2-13 Discounted Projected Costs under the Final NSPS OOOOb and EG 0000c Option, 2024-2038
(millions 2019$) 2-63
Table 2-14 Summary of Regulatory Alternatives (Well Sites Only) 2-64
Table 2-15 Comparison of Regulatory Alternatives in 2024, 2028, and 2038 for the Final NSPS OOOOb and EG
0000c across Regulatory Options (millions 2019$) 2-65
Table 2-16 Well Site Equipment/Tank Category Proportions Estimated From the 2016 ICR 2-68
Table 2-17 Per-Site Average Equipment/Tank Counts Estimated From the 2016 ICR Well Site 2-69
Table 3-1 Climate and Human Health Effects of the Projected Emissions Reductions under the Final Rule 3-2
Table 3 -2 Projected Annual Reductions of Methane, VOC, and HAP Emission Reductions under the Final NSPS
OOOOb and EG 0000c, 2024-2038 3-4
Table 3-3 Estimates of the Social Cost of CH4, 2024-2038 (in 2019$ per metric ton CH4) 3-17
Table 3-4 Undiscounted Monetized Climate Benefits under the Final NSPS OOOOb and EG 0000c, 2024-2038
(millions, 2019$) 3-19
Table 3-5 Discounted Monetized Climate Benefits under the Final NSPS OOOOb and EG 0000c, 2024-2038
(millions, 2019$) 3-20
Table 3-11 Stream of Human Health Benefits from 2024 through 2038: Monetized Benefits Quantified as Sum of
Long-Term Ozone Mortality and Illness across Regulatory Options (discounted at 2 percent; millions of
2019 dollars)3 3-50
Table 3-12 Stream of Human Health Benefits from 2024 through 2038: Monetized Benefits Quantified as Sum of
Long-Term Ozone Mortality and Illness across Regulatory Options (discounted at 3 percent; millions of
2019 dollars)3 3-51
Table 3-13 Stream of Human Health Benefits from 2024 through 2038: Monetized Benefits Quantified as Sum of
Long-Term Ozone Mortality and Illness across Regulatory Options (discounted at 7 percent; millions of
2019 dollars)3 3-52
Table 3-14 Top Annual HAP Emissions as Reported in 2017 NEI for Oil and Natural Gas Sources 3-57
-------
Table
3-15
Table
3-16
Table
3-17
Table
3-18
Table
4-1
Table
4-2
Table
4-3
Table
4-4
Table
4-5
Table
4-6
Table
4-7
Table
4-8
Table
4-9
Table
4-10
Table
4-11
Table
4-12
Table
4-13
Table
4-14
Table
4-15
Table
4-16
Table
4-17
Table
4-18
Table
4-19
Table
4-20
Table
4-21
Table
4-22
Table
4-23
Table
5-1
Table
5-2
Table
5-3
Table
5-4
Table
5-5
Table
5-6
Average Oil and Natural Gas
4-11
4-12
4-14
.4-17
Low Oil and Natural Gas Pricesa4-
High Oil and Natural Gas Prices 4-
Increases in Secondary Air Pollutant Emissions, Vapor Combustion at Storage Vessels (short tons per
year) 3-63
Comparison of PV and EAV of the Projected Benefits for the Final NSPS OOOOb and EG 0000c
across Regulatory Options, 2024-2038 (millions of 2019$)a 3-67
Comparison of PV and EAV of the Projected Benefits for the Final NSPS OOOOb across Regulatory
Options, 2023-2035 (millions of 2019$)a 3-69
Comparison of PV and EAV of the Projected Benefits for the Final EG 0000c Across Regulatory
Options, 2024-2038 (millions of 2019$)a 3-71
Parameters Used in Market Analysis 4-5
Baseline Crude Oil and Natural Gas Production and Prices Used in Market Analysis 4-5
Projected Regulatory Costs for the Final NSPS OOOOb and EG 0000c Applied in the Market
Analysis (millions 2019$) 4-6
Estimated Crude Oil Production and Prices Changes under the Final NSPS OOOOb and EG OOOOc4-7
Estimated Natural Gas Production and Prices Changes under the Final NSPS OOOOb and EG 0000c
4-7
Distribution of U.S Marginal Oil Wells in 2021
Distribution of U.S Marginal Natural Gas Wells in 2021
Marginal Well Financial Analysis Parameters
Estimated Revenues, Costs, and Profits for Marginal Wells for 2021
Prices3
Estimated Revenues, Costs, and Profits for Marginal Wells for 2021
18
Estimated Revenues, Costs, and Profits for Marginal Wells for 2021
19
Demographic Populations Included in the Ozone EJ Exposure Analysis from Oil and Natural Gas VOC
Emissions 4-36
Cancer Risk and Demographic Population Estimates for 2017 NEI Nonpoint Emissions from Oil and
Natural Gas Sources 4-51
Demographic Characteristics of Oil and Natural Gas Workers and Communities 4-55
Demographic Characteristics of Oil and Natural Gas Communities by Oil and Natural Gas Intensity4-57
Hispanic Population by Oil and Natural Gas Intensity 4-58
Energy Expenditures by Quintiles of Income before Taxes, 2019 4-61
SBA Size Standards by NAICS Code 4-68
Counts and Estimated Percentages of Small Entities 4-70
Summary Statistics for Revenues of Potentially Affected Entities 4-71
Average Capital and Annual Operating Costs by Affected Facility (2019 Dollars) 4-72
Distribution of Estimated Compliance Costs across Segment and Firm Size Classes (2019$) 4-73
Compliance Cost-to-Sales Ratios for Small Entities 4-74
Projected Emissions Reductions under the Final NSPS OOOOb and EG 0000c across Regulatory
Options, 2024 2038 : • 5-2
Projected Benefits, Compliance Costs, and Emissions Reductions across Regulatory Options for the
Final NSPS OOOOb, 2024-2038 (million 2019$)
Projected Benefits, Compliance Costs, and Emissions Reductions across Regulatory Options for the
Final EG 0000c, 2024-2038 (million 2019$)
Projected Benefits, Compliance Costs, and Emissions Reductions across Regulatory Options for the
Final NSPS OOOOb and EG 0000c, 2024-2038 (million 2019$)
Projected Emissions Reductions for Incrementally Affected Sources under the Final NSPS OOOOb and
EG 0000c, 2024 to 2038 5-9
Projected Climate Benefits and Compliance Costs (millions 2019$) for Incrementally Affected Sources
under the Final NSPS OOOOb and EG 0000c Option, 2024 to 2038a-b 5-10
.5-3
.5-5
.5-7
viii
-------
LIST OF FIGURES
Figure 2-1 Projections of Cumulative Site Counts by Site Type and Vintage 2-34
Figure 3-1 Map of State-level VOC Emissions Reductions (tpy) from the Baseline to the Final Rule Scenario in
2024, 2027, 2028 and 2038 3-28
Figure 3-2 Map of Modeled Changes in April to September MDA8 Ozone Concentrations Calculated as the Final
Rule Scenario Minus the Baseline Scenario in 2024, 2027, 2028 and 2038 3-30
Figure 3-3 Data Inputs and Outputs for the BenMAP-CE Model 3-36
Figure 4-1 Heat Map of the National Average Ozone Concentrations in the Baseline and the Absolute and
Percentage Reductions under the Final Rule in 2038 (ppb) 4-38
Figure 4-2 Baseline Distributions of Ozone Concentration (ppb) Across Populations in 2038 under the Final Rule
(warm-season average of 8-hour daily maximum) 4-40
Figure 4-3 Distributions of Ozone Concentration Reductions Across Populations in 2038 under the Final Rule 4-43
Figure 4-4 Heat Map of the State Average Ozone Concentration Reductions (Purple) Due to the Final Rule Across
Demographic Groups in 2038 (ppb) 4-45
Figure 4-5 National Map of Grid Cell Median Cancer Risks for 2017 Nonpoint Oil and Natural Gas NEI Emissions
4-52
Figure 4-6 Local-Scale Map of Grid Cell Median Cancer Risks for 2017 Nonpoint Oil and Natural Gas NEI
Emissions 4-53
Figure 4-7 National-level Employment in Oil and Natural Gas Production 4-54
Figure 4-8 Continental U.S. Map of PUMAs and Oil and Natural Gas Intensive Communities 4-56
Figure 4-9 Map of Oil and Natural Gas Intensive Communities with Environmental Justice Concerns 4-59
IX
-------
1 EXECUTIVE SUMMARY
1.1 Introduction
On November 15, 2021, the Environmental Protection Agency (EPA) published a
proposed rule intended to mitigate climate-destabilizing pollution and protect human health by
reducing greenhouse gas (GHG) and VOC emissions from the Oil and Natural Gas Industry,
specifically the Crude Oil and Natural Gas source category.1'2 In the November 2021 proposal,
the EPA proposed New Source Performance Standards (NSPS) and Emissions Guidelines for
Existing Sources (EG) under CAA section 111 which would be codified in 40 CFR part 60 at
subpart OOOOb (NSPS OOOOb) and subpart OOOOc (EG 0000c). The EPA also proposed
several related actions stemming from the joint resolution of Congress, adopted on June 30,
2021, under the Congressional Review Act (CRA), disapproving the EPA's final rule titled, "Oil
and Natural Gas Sector: Emission Standards for New, Reconstructed, and Modified Sources
Review," September 14, 2020 ("2020 Policy Rule"). Lastly in November 2021, the EPA
proposed a protocol under the general provisions for optical gas imaging.
On December 6, 2022, the EPA published a supplemental proposed rule that comprised a
few distinct actions.3 First, the EPA updated, strengthened, and/or expanded on the new source
performance standards proposed in November 2021 under CAA section 111(b) for GHGs (in the
form of methane limitations) and VOC emissions from new, modified, and reconstructed
facilities that commenced construction, reconstruction, or modification after November 15, 2021.
Second, the EPA updated, strengthened, and/or expanded the presumptive standards proposed in
November 2021 as part of the CAA section 111(d) EG for GHGs emissions (in the form of
methane limitations) from existing designated facilities that commenced construction,
reconstruction, or modification on or before November 15, 2021. Third, EPA established the
implementation requirements for states to limit GHGs pollution (in the form of methane
1 86 FR63110.
2
The EPA defines the Crude Oil and Natural Gas source category to mean: (1) Crude oil production, which includes
the well and extends to the point of custody transfer to the crude oil transmission pipeline or any other forms of
transportation; and (2) natural gas production, processing, transmission, and storage, which include the well and
extend to, but do not include, the local distribution company custody transfer station, commonly referred to as the
"city-gate."
3 87 FR 74702.
1-1
-------
limitations) from existing sources (designated facilities) in the source category under CAA
section 111(d).
The purpose of this final rulemaking is to finalize these multiple actions to reduce air
emissions from the Crude Oil and Natural Gas source category. First, the EPA finalizes NSPS
regulating GHGs in the form of limitations on methane and VOCs for the Crude Oil and Natural
Gas source category pursuant the review required by the CAA. Second, the EPA finalizes the
first-ever EG under the CAA, for states to follow in developing, submitting, and implementing
state plans to establish performance standards to limit GHGs emissions (in the form of methane
limitations) from existing sources (designated facilities) in the Crude Oil and Natural Gas source
category. Third, the EPA is finalizing several related actions stemming from the joint resolution
of Congress, adopted on June 30, 2021, under the Congressional Review Act (CRA),
disapproving the EPA's final rule titled, "Oil and Natural Gas Sector: Emission Standards for
New, Reconstructed, and Modified Sources Review," Sept. 14, 2020 ("2020 Policy Rule").
Fourth, the EPA is finalizing a protocol under the general provisions for optical gas imaging.
These final actions respond to the President's January 20, 2021, Executive Order (E.O.) 13990
titled, "Protecting Public Health and the Environment and Restoring Science to Tackle the
Climate Crises," which directed the EPA to consider taking these actions finalized here.
In accordance with E.O. 12866 and E.O. 14094, the guidelines of the Office of
Management and Budget Circular A-4 and the EPA's Guidelines for Preparing Economic
Analyses (U.S. EPA, 2014), this regulatory impact analysis (RIA) analyzes the nationwide
benefits and costs associated with the projected emissions reductions under the for the final rule.4
The RIA also examines expected economic and environmental justice impacts of the final rule
requirements.
1.2 Legal and Economic Basis for this Rulemaking
In this section, we summarize the statutory requirements in the Clean Air Act that serve
as the legal basis for the final rule and the economic theory that supports environmental
4
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) may be referred to until those dates.
1-2
-------
regulation as a mechanism to enhance social welfare. The Clean Air Act requires the 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 Requirements
Clean Air Act section 111, which Congress enacted as part of the 1970 Clean Air Act
Amendments, establishes mechanisms for controlling emissions of air pollutants from stationary
sources. This provision requires the EPA to promulgate a list of categories of stationary sources
that the Administrator, in his or her judgment, finds "causes, or contributes significantly to, air
pollution which may reasonably be anticipated to endanger public health or welfare." The EPA
has listed more than 60 stationary source categories under this provision. Once the EPA lists a
source category, the EPA must, under CAA section 111(b)(1)(B), establish "standards of
performance" for emissions of air pollutants from new sources in the source categories. Under
section 111(b), EPA identifies the "best system of emission reduction" (BSER) that has been
adequately demonstrated to control emissions of a particular pollutant from a particular type of
source and sets a standard for new sources based on the application of that BSER. These
standards are known as new source performance standards (NSPS), and they are national
requirements that apply directly to the sources subject to them.
When the EPA establishes NSPS for sources in a source category under CAA section
111(b), the EPA is also required, under CAA section 111(d)(1), to prescribe regulations for states
to submit plans regulating existing sources in that source category for any air pollutant that, in
general, is not regulated under the CAA section 109 requirements for the NAAQS or regulated
under the CAA section 112 requirements for hazardous air pollutants (HAP). CAA section
111(d)'s mechanism for regulating existing sources differs from the one that CAA section 111(b)
provides for new sources because CAA section 111(d) contemplates states submitting plans that
establish "standards of performance" for the affected sources and contain other measures to
implement and enforce those standards.
"Standards of performance" are defined under CAA section 111(a)(1) as standards for
emissions that reflect the emission limitation achievable from the "best system of emission
1-3
-------
reduction," considering costs and other factors, that "the Administrator determines has been
adequately demonstrated." Under section 111(d), EPA determines the BSER, but, unlike with
new sources under 111(b), the states are the entities that establish performance standards. CAA
section 111(d)(1) grants states the authority, in applying a standard of performance, to take into
account the source's remaining useful life and other factors.
Under CAA section 111(d), a state must submit its plan to the EPA for approval, and the
EPA must approve the state plan if it is "satisfactory." If a state does not submit a plan, or if the
EPA does not approve a state's plan, then the EPA must establish a plan for that state. Once a
state receives the EPA's approval of its plan, the provisions in the plan become federally
enforceable against the entity responsible for noncompliance, in the same manner as the
provisions of an approved State Implementation Plan (SIP) under the Act.
1.2.2 Market Failure
Many regulations are promulgated to correct market failures, which otherwise lead to a
suboptimal allocation of resources within the free market. Air quality and pollution control
regulations address "negative externalities" whereby the market does not internalize the full
opportunity cost of production borne by society as public goods such as air quality are unpriced.
While recognizing that optimal social level of pollution may not be zero, methane and
VOC emissions impose costs on society, such as negative health and welfare impacts, that are
not reflected in the market price of the goods produced through the polluting process. For the
final regulatory action analyzed in this RIA, the goods produced are crude oil and natural gas. If
crude oil and natural gas producers pollute the atmosphere when extracting, processing, and
transporting products, 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 crude oil and natural gas 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. The final regulation will work towards addressing this market
failure by causing affected producers to begin internalizing the negative externality associated
with methane and VOC emissions.
1-4
-------
1.3 Baseline and Regulatory Requirements
The impacts of regulatory actions are evaluated relative to a baseline that represents the
world without the regulatory action. In this case, we present results for the final NSPS OOOOb
and EG 0000c. The baseline for the final rule incorporates changes to regulatory requirements
induced by the Congressional Review Act (CRA) resolution that disapproved the 2020 Policy
Rule. Throughout this document, we focus the analysis on the requirements that result in
quantifiable compliance cost or emissions changes compared to the baseline. We do not analyze
the regulatory impacts of all finalized requirements because we lack sufficient data, require
additional work to adapt existing data into a coherent analysis framework, or believe the
provisions would not result in compliance cost or emissions impacts; see Section for a
discussion of provisions for which impacts were not quantified.
Compared to the analyses presented in the RIAs for the November 2021 proposal and the
December 2022 supplemental proposal, this analysis reflects new methodologies to estimate and
project the universe of affected facilities and their emissions profiles, as well as the cost and
emissions impacts of applying control strategies. Most notably, the RIA estimates the cost and
emissions impacts of associated gas provisions and requirements for optical gas imaging (OGI)
monitoring of flares at controlled storage vessels at well sites, neither of which were quantified
in the proposal RIAs. The RIA also leverages region-specific information regarding well site
equipment to better characterize local air pollution impacts of the regulation, whereas the
proposal RIAs relied on national aggregates. The updated baseline represents the EPA's most
recent assessment of the current and future state of the industry absent the requirements of this
final rulemaking.
Table 1-1 and Table 1-2 summarize the baseline and finalized standards of performance
for the sources with impacts quantified in this RIA.5 In Table 1-2, requirements in the baseline
differ depending on when sources were constructed relative to previous NSPS proposal dates.
We define pre- and post-KKK as having construction dates prior to and after January 20, 1984,
respectively. The dividing dates for pre- and post-0000 and pre- and post-OOOOa are August
23, 2011, and September 18, 2015, respectively. The abbreviations used in the table are OGI
5 See the preamble for a more comprehensive description of the finalized standards.
1-5
-------
(optical gas imaging), AVO (auditory, visual, and optical), scfh (standard cubic feet per hour),
and scfm (standard cubic feet per minute).
Table 1-1 NSPS OOOOb Emissions Sources, Baseline Requirements, and
Requirements under the Final Rule
Standards of Performance
Source
In the Baseline
Under the Final Rule
Fugitive Emissions/Equipment Leaks3
Well Sites
Wellhead only, single well site
No requirement
Quarterly AVO monitoring
Wellhead only, multiple well site
No requirement
Quarterly AVO monitoring +
Semiannual OGI
Single well site with a single price of
major equipment and no tank battery
Semiannual OGI
Quarterly AVO monitoring
Multiple well site with a single piece
of major equipment, or any site with
(1) two or more pieces of major
Semiannual OGI
Bimonthly AVO monitoring +
equipment; (2) one piece of major
Quarterly OGI
equipment and a tank battery; or (3) a
controlled tank battery.
Gathering and Boosting Stations
Transmission and Storage Compressor
Stations
Quarterly OGI
Monthly AVO monitoring +
Quarterly OGI
Natural Gas Processing Plants
NSPS Subpart Wa
Bimonthly OGI
Pneumatic Pumps
Well Sites
95% control
Zero emissions'3
Gathering and Boosting Stations
No requirement
Pneumatic Controllers0
Well Sites
Gathering and Boosting Stations
Natural gas bleed rate no greater
Transmission and Storage Compressor
than 6 scfh
Zero emissions'1
Stations
Natural Gas Processing Plants
Zero emissions
Reciprocating Compressors
Gathering and Boosting Stations
Natural Gas Processing Plants
Rod-packing changeout on fixed
Volumetric flow rate of 2 scfm
Transmission and Storage Compressor
schedule
Stations
Wet-Seal Centrifugal Compressors
Gathering and Boosting Stations
No requirement
Natural Gas Processing Plants
95% control
Transmission and Storage Compressor
95% control
Stations
Liquids Unloading
Well Sites
No requirement
Zero emissions or best
management practices6
1-6
-------
Standards of Performance
Source
In the Baseline
Under the Final Rule
Storage Vessels
PTE > 6 tpy VOC
95% control, affected facility is
the tank
95% control, affected facility is
the tank battery
PTE < 6 tpy VOC
No requirement
No requirement
Associated Gas
Well Sites
No requirement
Route to sales line
a Well sites and compressor stations on the Alaska North Slope are subject to Annual OGI monitoring only.
b The zero emission standard for pumps applies to sites with electrical power and/or three or more diaphragm
pumps. Sites without access to electrical power that have fewer than three diaphragm pumps must route emissions to
a control device, provided one is onsite.
0 Specifically, the affected source is natural gas-driven controllers that vent to the atmosphere.
d The zero emissions rate standard does not apply to pneumatic controllers at sites in Alaska for which on site power
is not available. Instead, natural gas-driven continuous bleed controllers at those sites are required to achieve bleed
rates at or below 6 scfh, while natural gas-driven intermittent bleed controllers are subject to OGI monitoring and
repair of emissions from controller malfunctions.
e The final regulation requires liquids unloading events to be zero-emitting unless technical infeasibilities exist, in
which case the regulation requires that best management practices be adopted.
1-7
-------
Table 1-2 EG OOOOc Emissions Sources, Baseline Requirements, and Requirements
under the Final Rule
Presumptive Standards of Performance
Source
In the Baseline
Under the Final Rule
Fugitive Emissions/Equipment Leaksa'b
Well Sites
Wellhead only, single well site
Quarterly AVO monitoring
Wellhead only, multiple well site
No requirement
Quarterly AVO monitoring +
Semiannual OGI
Single well site with a single price of
major equipment and no tank battery
Pre-OOOOa: No requirement
Post-OOOOa: Semiannual OGI
Quarterly AVO monitoring
Multiple well site with a single piece
of major equipment, or any site with
two or more pieces of major
equipment or one piece of major
equipment and a tank battery
Pre-OOOOa: No requirement
Post-OOOOa: Semiannual OGI
Bimonthly AVO monitoring +
Quarterly OGI
Gathering and Boosting Stations
Transmission and Storage Compressor
Stations
Pre-OOOOa: No requirement
Post-OOOOa: Quarterly OGI
Monthly AVO monitoring +
Quarterly OGI
Natural Gas Processing Plants
Pre-KKK: No requirement
Post-KKK and Pre-OOOO:
NSPS Subpart W
Post-OOOO: NSPS Subpart Wa
Bimonthly OGI
Pneumatic Pumps
Well Sites
Pre-OOOOa: No requirement
Post-OOOOa: 95% control
Methane emission rate of zerob
Gathering and Boosting Stations
No requirement
Pneumatic Controllers0
Well Sites
Gathering and Boosting Stations
Pre-OOOO: No requirement
Post-OOOO: Natural gas bleed
rate no greater than 6 scfh
Methane emission rate of zerod
Transmission and Storage Compressor
Stations
Pre-OOOOa: No requirement
Post-OOOOa: Natural gas bleed
rate no greater than 6 scfh
Natural Gas Processing Plants
Pre-OOOO: No requirement
Post-OOOO: Zero emissions
Methane emission rate of zero
Reciprocating Compressors
Gathering and Boosting Stations
Natural Gas Processing Plants
Pre-OOOO: No requirement
Post-OOOO: Rod-packing
changeout on fixed schedule
Volumetric flow rate of 2 scfm
Transmission and Storage Compressor
Stations
Pre-OOOOa: No requirement
Post-OOOOa: Rod-packing
changeout on fixed schedule
Wet-Seal Centrifugal Compressors
Gathering and Boosting Stations
No requirement
Natural Gas Processing Plants
Transmission and Storage Compressor
Stations
Pre-OOOO: No requirement
Post-OOOO: 95% control
Volumetric flow rate of 3 scfm
Liquids Unloading
Well Sites
No requirement
Zero emissions or best
management practices6
1-8
-------
Presumptive Standards of Performance
Source
In the Baseline
Under the Final Rule
Storage Vessels
PTE > 20 tpy CH4
Pre-OOOO: No requirement
Post-OOOO: 95% control,
95% control, affected facility is
the tank battery
PTE < 20 tpy CH4 and >
6 tpy VOC
affected facility is the tank
PTE < 20 tpy CH4 and <
6 tpy VOC
No requirement
No requirement
Associated Gas
Well Sites
PTE > 40 tpy CH4
No requirement
Route to sales line
PTE < 40 tpy CH4
No requirement
95% control
a Well sites and compressor stations on the Alaska North Slope are subject to Annual OGI monitoring only.
b The zero emission standard for pumps applies to sites with electrical power and/or three or more diaphragm
pumps. Sites without access to electrical power that have fewer than three diaphragm pumps must route emissions to
a control device, provided one is onsite.
0 Specifically, the affected source is natural gas-driven controllers that vent to the atmosphere.
d The zero emissions rate standard does not apply to pneumatic controllers at sites in Alaska for which on site power
is not available. Instead, natural gas-driven continuous bleed controllers at those sites are required to achieve bleed
rates at or below 6 scfh, while natural gas-driven intermittent bleed controllers are subject to OGI monitoring and
repair of emissions from controller malfunctions.
e The final regulation requires liquids unloading events to be zero-emitting unless technical infeasibilities exist, in
which case the regulation requires that best management practices be adopted.
The net benefits analysis summarized in this RIA reflects a nationwide engineering
analysis of compliance cost and emissions reductions, of which there are two main components:
activity data and information on control measures. The activity data represents estimates of the
counts of affected facilities over time, and the control measure information includes data on costs
and control efficiencies for typical facilities. Both components are described briefly below, with
more detailed information provided in Section 2.2.
The first component is activity data for a set of representative or model plants for each
regulated facility.6 To project activity data for regulated facilities, we first project activity data
for oil and natural gas sites, which include well sites, natural gas processing plants, and
compressor stations (gathering and boosting, transmission, and storage). Projections include
addition of newly constructed sites and retirement of previously constructed sites, with
magnitudes based on a combination of analysis of several data sources and, where necessary,
assumptions. Using representative "per-site" factors based on the EPA's Greenhouse Gas
6 Regulated facilities include well site fugitives (including component emissions and malfunctioning storage vessel
flares), gathering and boosting station fugitives, transmission and storage compressor station fugitives, natural gas
processing plant equipment leaks, pneumatic pumps, pneumatic controllers, reciprocating compressors, centrifugal
compressors, liquids unloading, storage vessels, and associated gas.
1-9
-------
Inventory (GHGI) (U.S. EPA, 2021), regulated facilities are apportioned to sites across all
industry segments.7 We assume the per-site factors are fixed over time, so that the projected
counts of regulated facilities change in proportion to the projected counts of sites.
The regulated facility projections are combined with information on control options,
including capital costs, annual operations and maintenance costs, and control efficiencies.
Information on control options is derived from the analysis underpinning the BSER
determinations. Impacts are calculated by setting parameters on how and when affected facilities
are assumed to respond to a regulatory regime, multiplying activity data by model plant cost and
emissions estimates, differencing from the baseline scenario, and then summing to the desired
level of aggregation. In addition to emissions reductions, some control options result in natural
gas recovery, which can then be combusted in production or sold. Where applicable, we present
projected compliance costs with and without the projected revenues from product recovery.
For the analysis, we calculate the cost and emissions impacts of the final NSPS OOOOb
and EG 0000c requirements from 2024 to 2038. The initial analysis year is 2024 as the rule
will take effect early in that year. The NSPS OOOOb is assumed to take effect immediately and
impact sources that commence construction after publication of the December 2022 proposal.8
We assume that sources begin production, and thus begin generating emissions related to that
production, one year after they commence construction, so that sources assumed to be
constructed in 2023 first contribute cost and emissions impacts in 2024.9 We assume the EG
OOOOc will take longer to go into effect as states will need to develop implementation plans in
response to the rule and have them approved by the Agency. We assume that this process will
take four years, and so EG OOOOc impacts will begin in 2028. The final analysis year is 2038,
which allows us to present up to 15 post-NSPS-finalization years of regulatory impact estimates.
7
Industry segments include production, gathering and boosting, processing, transmission, and storage.
As explained in the preamble to the final rule, NSPS OOOOb would apply to all emissions sources ("affected
facilities") identified in the proposed 40 CFR 60.5365b that commenced construction, reconstruction, or
modification after December 6, 2022.
9
Due to supply chain considerations, NSPS OOOOb allows for longer compliance deadlines for process controllers
and pumps; see Sections XI.D.4 (process controllers) and XI.E.2 (pumps). To allow new wells the ability to plan to
comply with the associated gas provisions, NSPS OOOOb also allows for a longer compliance deadline for these
affected facilities; see Section XI.F.2.d (associated gas) for additional details. We do not quantify the impacts of
those extensions in the RIA, but we expect the impacts to be small relative to the overall impacts of the rule due to
the limited nature of the extensions.
1-10
-------
1.4 Summary of Projected Emissions Reductions and Benefit-Cost Analysis
A summary of the key benefit-cost analysis results is shown below. All dollar estimates
are in 2019 dollars. Also, all compliance costs, emissions changes, and benefits are estimated for
the years 2024 to 2038 relative to a baseline without the final NSPS OOOOb and EG OOOOc.
Table 1-3 summarizes the emissions reductions associated with the finalized standards
over the 2024 to 2038 period for the NSPS OOOOb, the EG OOOOc, and the NSPS OOOOb
and EG OOOOc combined. The emissions reductions are estimated by multiplying the source-
level emissions reductions associated with each applicable control and facility type by the
number of affected sources of that facility type. We present methane emissions in both short tons
and CO2 equivalents (CO2 Eq.) using a global warming potential (GWP) of 28.10
Table 1-3 Projected Emissions Reductions under the Final NSPS OOOOb and EG
OOOOc Option, 2024-2038a'bc
Emissions Changes
Methane
(million metric tons
Methane
VOC
HAP
CO2 Eq. using
Final
(million short tons)
(million short tons)
(million short tons)
GWP=28)
NSPS OOOOb
23
7.1
0.27
590
EG OOOOc
35
8.6
0.32
890
Total
58
16
0.59
1,500
a Numbers rounded to two significant digits unless otherwise noted. Totals may not appear to add correctly due to
rounding. To convert from short tons to metric tons, multiply the short tons by 0.907. Alternatively, to convert
metric tons to short tons, multiply metric tons by 1.102.
b The EG OOOOc regulates emissions of methane. Additional benefits to the regulation result from associated
reductions in VOC and HAP emissions.
0 The control techniques used to meet the storage vessel-related standards are associated with several types of
secondary emissions impacts, which may partially offset the direct benefits of this rule. Relative to the direct
emission reductions anticipated from this rule, as discussed in Section 3.9, the magnitude of these secondary air
pollutant increases is small.
Table 1-4, Table 1-5, and Table 1-6 present results for the NSPS OOOOb, EG OOOOc,
and NSPS OOOOb and EG OOOOc combined, respectively. Each table presents the present
value (PV) and equivalent annual value (EAV), estimated using discount rates of 2, 3, and 7
percent, of the changes in quantified climate and health benefits, costs, and net benefits, as well
10 Global warming potential is a measure that allows comparisons of the global warming impacts of different
greenhouse gases. Specifically, it is a measure of how much energy the emission of 1 ton of a gas will absorb over a
given period of time, relative to the emission of 1 ton of carbon dioxide (CO2).
1-11
-------
1112
as the emissions reductions relative to the baseline. ' These values reflect an analytical time
horizon of 2024 to 2038, are discounted to 2021,13 and presented in 2019 dollars. We present the
total compliance costs, the value of product recovery generated by the capture of natural gas, and
the net compliance costs, which treats the value of product recovery as an offset to the
compliance costs.14 The table includes consideration of the non-monetized benefits associated
with the emissions reductions projected under the final rule.
11 Monetized climate effects are presented under a 2 percent near-term Ramsey discount rate, consistent with EPA's
updated estimates of the SC-GHG. The 2003 version of OMB 's Circular A-4 had generally recommended 3 percent
and 7 percent as default discount rates for costs and benefits, though as part of the Interagency Working Group on
the Social Cost of Greenhouse Gases, OMB had also long recognized that climate effects should be discounted only
at appropriate consumption-based discount rates. While this RIA was being drafted and reviewed, OMB finalized an
update to Circular A-4, in which it recommended the general application of a 2.0 percent discount rate to costs and
benefits (subject to regular updates), as well as the consideration of the shadow price of capital when costs or
benefits are likely to accrue to capital (OMB 2023). Because the SC-GHG 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-GHG. See Section 3.2 for more discussion.
12
The EPA has also applied its updated estimates of the social cost of carbon dioxide (SC-CO2) in an illustrative
analysis of potential climate disbenefits from secondary CO2 emissions associated with control techniques for
storage vessels. Given that the estimated climate disbenefits from the CO2 impacts would at most offset only about 1
percent of the methane benefits, the EPA finds that the summary values shown in this table are a reasonable estimate
of the net monetized climate effects of the rule. See Section 3.9 for further discussion.
13
Costs and benefits are discounted to 2021 to maintain consistency with the November 2021 and December 2022
RIAs.
14
Almost 90 percent of revenue from the sale of captured natural gas is projected to be earned by operators in the
production segment of the industry, where we assume that the operators own the natural gas and will receive the
financial benefit from the captured natural gas. The remainder of the captured natural gas is captured within the
gathering and boosting, processing, transmission, and storage segments, where operators do not typically own the
natural gas they transport; rather, they receive payment for the service they provide. In the RIA, we treat these
revenues as an offset to projected compliance costs, while the revenues may also be considered as a benefit of the
regulatory action. However, regardless of whether the revenue from capture of natural gas is considered a
compliance cost offset or a benefit, the net benefits are equivalent. See Section 2 for further discussion regarding
firm incentives for product recovery.
1-12
-------
Table 1-4 Projected Benefits, Compliance Costs, and Emissions Reductions for the
Finalized NSPS OOOOb, 2024-2038 (million 2019$)
2 Percent Near-Term Ramsey Discount Rate
PV
EAV
PV
EAV
PV
EAV
Climate Benefits'3
$44,000
$3,400
$44,000
$3,400
$44,000
$3,400
2 Percent Discount
Rate
3 Percent Discount
Rate
7 Percent Discount
Rate
PV
EAV
PV
EAV
PV
EAV
Ozone Health Benefits0
N/A
N/A
N/A
N/A
N/A
N/A
Net Compliance Costs
$5,800
$450
$5,800
$480
$5,300
$580
Compliance Costs
$14,000
$1,100
$13,000
$1,100
$10,000
$1,100
Value of Product Recovery
$7,900
$620
$7,100
$590
$4,700
$520
Net Monetized Benefits'1
$38,000
$3,000
$38,000
$2,900
$39,000
$2,800
Ozone-related health benefits from reducing 23 million short tons of
methane from 2024 to 2038
Benefits to provision of ecosystem services from reducing 23 million short
tons of methane, 7.1 million short tons of VOC, and 270 thousand short
Non-Monetized Benefits tons of HAP from 202410 2038
PM2 5-related health benefits from reducing 7.1 million short tons of VOC
from 2024 to 2038
HAP benefits from reducing 270 thousand short tons of HAP from 2024 to
2038
a Values rounded to two significant figures. Totals may not appear to add correctly due to rounding.
b Monetized climate benefits are based on reductions in methane emissions and are calculated using three different
estimates of the social cost of methane (SC-CH4) (under 1.5 percent, 2.0 percent, and 2.5 percent near-term Ramsey
discount rates). For the presentational purposes of this table, we show the climate benefits associated with the SC-
CH4 at the 2 percent near-term Ramsey discount rate. Please see Table 3.4 and 3.5 for the full range of monetized
climate benefit estimates.
0 Due to time and resource limitations, the monetized ozone benefits under the NSPS OOOOb alone were not
quantified and are not included in the monetized benefits in this table. See Table 1-6 for an estimate of the
monetized ozone benefits under the combined NSPS OOOOb and EG OOOOc combined.
d Several categories of climate, human health, and welfare benefits from methane, VOC, and HAP emissions
reductions remain unmonetized and are thus not directly reflected in the quantified benefit estimates in the table. See
Section 3.2 for a discussion of climate effects that are not yet reflected in the SC-CH4 and thus remain unmonetized
and Sections 3.4 through 3.8 for a discussion of other non-monetized benefits.
1-13
-------
Table 1-5 Projected Benefits, Compliance Costs, and Emissions Reductions for the
Finalized EG OOOOc, 2024-2038 (million 2019$)
2 Percent Near-Term Ramsey Discount Rate
PV
EAV
PV
EAV
PV
EAV
Climate Benefits'3
$65,000
$5,100
$65,000
$5,100
$65,000
$5,100
2 Percent Discount
Rate
3 Percent Discount
Rate
7 Percent Discount
Rate
PV
EAV
PV
EAV
PV
EAV
Ozone Health Benefits0
N/A
N/A
N/A
N/A
N/A
N/A
Net Compliance Costs
$13,000
$1,000
$12,000
$1,000
$8,900
$970
Compliance Costs
$18,000
$1,400
$16,000
$1,400
$12,000
$1,300
Value of Product Recovery
$4,700
$370
$4,200
$350
$2,700
$300
Net Monetized Benefits'1
$52,000
$4,100
$53,000
$4,100
$56,000
$4,100
Ozone-related health benefits from reducing 35 million short tons of
methane from 2024 to 2038
Benefits to provision of ecosystem services from reducing 35 million short
tons of methane, 8.6 million short tons of VOC, and 320 thousand short
Non-Monetized Benefits 'ons of HAP from 2024 to 2038
PM2 5-related health benefits from reducing 8.6 million short tons of VOC
from 2024 to 2038
HAP benefits from reducing 320 thousand short tons of HAP from 2024 to
2038
a Values rounded to two significant figures. Totals may not appear to add correctly due to rounding.
b Monetized climate benefits are based on reductions in methane emissions and are calculated using three different
estimates of the social cost of methane (SC-CH4) (under 1.5 percent, 2.0 percent, and 2.5 percent near-term Ramsey
discount rates). For the presentational purposes of this table, we show the climate benefits associated with the SC-
CH4 at the 2 percent near-term Ramsey discount rate. Please see Table 3.4 and 3.5 for the full range of monetized
climate benefit estimates.
0 Due to time and resource limitations, the monetized ozone benefits under the EG OOOOc alone were not
quantified and are not included in the monetized benefits in this table. See Table 1-6 for an estimate of the
monetized ozone benefits under the combined NSPS OOOOb and EG OOOOc combined.
d Several categories of climate, human health, and welfare benefits from methane, VOC, and HAP emissions
reductions remain unmonetized and are thus not directly reflected in the quantified benefit estimates in the table. See
Section 3.2 for a discussion of climate effects that are not yet reflected in the SC-CH4 and thus remain unmonetized
and Sections 3.4 through 3.8 for a discussion of other non-monetized benefits.
1-14
-------
Table 1-6 Projected Benefits, Compliance Costs, and Emissions Reductions for the
Finalized NSPS OOOOb and EG OOOOc, 2024-2038 (million 2019$)
2 Percent Near-Term Ramsey Discount Rate
PV
EAV
PV
EAV
PV
EAV
Climate Benefits'3
$110,000
$8,500
$110,000
$8,500
$110,000
$8,500
2 Percent Discount
Rate
3 Percent Discount
Rate
7 Percent Discount
Rate
PV
EAV
PV
EAV
PV
EAV
Ozone Health Benefits0
$6,900
$540
$6,000
$510
$3,500
$380
Net Compliance Costs
$19,000
$1,500
$18,000
$1,500
$14,000
$1,600
Compliance Costs
$31,000
$2,400
$29,000
$2,400
$22,000
$2,400
Value of Product Recovery
$13,000
$980
$11,000
$950
$7,400
$820
Net Monetized Benefits'1
$97,000
$7,600
$97,000
$7,500
$98,000
$7,300
Ozone-related health benefits from reducing 58 million short tons of
methane from 2024 to 2038
Benefits to provision of ecosystem services from reducing 58 million short
tons of methane, 16 million short tons of VOC, and 590 thousand short tons
Non-Monetized Benefits HAP from 2024 to 203 8
PM2 5-related health benefits from reducing 16 million short tons of VOC
from 2024 to 2038
HAP benefits from reducing 590 thousand short tons of HAP from 2024 to
2038
a Values rounded to two significant figures. Totals may not appear to add correctly due to rounding.
b Monetized climate benefits are based on reductions in methane emissions and are calculated using three different
estimates of the social cost of methane (SC-CH4) (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-
CH4 at the 2 percent near-term Ramsey discount rate. Please see Table 3.4 and 3.5 for the full range of monetized
climate benefit estimates.
0 Monetized benefits include those related to public health associated with reductions in ozone
concentrations. The health benefits are associated with several point estimates.
d Several categories of climate, human health, and welfare benefits from methane, VOC, and HAP emissions
reductions remain unmonetized and are thus not directly reflected in the quantified benefit estimates in the table. See
Section 3.2 for a discussion of climate effects that are not yet reflected in the SC-CH4 and thus remain unmonetized
and Sections 3.4 through 3.8 for a discussion of other non-monetized benefits.
1.5 Organization of RIA
Section 2 describes the projected compliance cost and emissions impacts from the final
rule, including the PV and EAV of the projected costs over the 2024 to 2038 period and the
associated EAV. Section 3 describes the projected climate benefits, including the PV and EAV
of the projected climate benefits and ozone benefits over the 2024 to 2038 period. Section 3
additionally considers the potential beneficial climate, health, and welfare impacts that are not
quantified in this RIA. Section 4 describes the economic impact and distributional analysis
associated with the final rule. The economic impact and distributional analysis section includes
1-15
-------
analysis of oil and natural gas market impacts, marginal wells, environmental justice, small
entities, and employment. Section 5 compares the projected benefits and compliance cost
reductions of this action, as well as a summary of the net benefits with consideration of non-
monetized benefits. Section 5 also highlights uncertainties and limitations of the analysis.
1.6 References
OMB. (2003). Circular A-4: Regulatory Analysis. Washington DC. Retrieved from
https://obamavvhitehouse.archives.gov/omb/circulars_a004_a-4/
OMB. (2023). Circular A-4: Regulatory Analysis. Washington DC. Retrieved from
https://vvvvvv.vvhitehouse.gOv/vvp-content/uploads/2023/l I/CircularA-4.pdf
U.S. EPA. (2014). Guidelines for Preparing Economic Analyses. (EPA 240-R-I0-00I).
Washington DC: U.S. Environmental Protection Agency, Office of Policy, National Center
for Environmental Economics Retrieved from https://vvvvvv.epa.gov/environmental-
economics/guidelines-preparing-economic-analyses
U.S. EPA. (2021). Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019 (EPA-
430-R-2I-005). https://www.epa.gov/ghgemissions/inventory-us-greenhouse-gas-emissions-
and-sinks-1990-2019
1-16
-------
2 PROJECTED COMPLIANCE COSTS AND EMISSIONS REDUCTIONS
In this section, we present estimates of the projected engineering compliance costs and
emissions reductions associated with the final rule for the 2024 to 2038 period. These estimates
are generated by combining model plant-level cost and emissions reductions based on the BSER
analysis with activity data projections based on a combination of historical trends and third-party
projections. The methods and assumptions used to construct the activity data projections are also
documented in this section.
The cost analysis of this RIA is in part based upon on the possibility that reducing
emissions from the oil and natural gas sector also produces a financial return via preventing the
loss of marketable natural gas. Assuming financially rational producers, standard economic
theory suggests that oil and natural gas firms would incorporate all cost-effective improvements,
which they are aware of, without government intervention. In general, however, the cost of
abating emissions exceeds the potential financial returns from the captured product such that the
producer does not abate emissions absent a regulatory requirement. It is possible in certain
circumstances that the financial returns from reducing emissions exceed the abatement costs, yet
producers still do not capture the natural gas. There may be an opportunity cost associated with
the installation of environmental controls (for purposes of mitigating the emission of pollutants)
that is not reflected in the control costs. If the environmental investment displaces other
investment in productive capital, the difference between the rate of return on the marginal
investment displaced by the mandatory environmental investment is a measure of the opportunity
cost of the environmental requirement to the regulated entity. However, if firms are not capital
constrained, then there may not be any displacement of investment, and the rate of return on
other investments in the industry would not be relevant as a measure of opportunity cost. If firms
should face higher borrowing costs as they take on more debt, there may be an additional
opportunity cost to the firm. To the extent that any opportunity costs are not added to the control
costs, the compliance costs presented in this RIA may be underestimated.
2.1 Emissions Sources and Regulatory Requirements Analyzed in this RIA
A series of emissions sources and controls were evaluated as part of the NSPS OOOOb
and EG 0000c rulemaking. Section 2.1.1 provides a basic description of emissions sources and
2-17
-------
the controls evaluated for each source to facilitate the reader's understanding of the economic
analysis. Section describes the regulatory choices within the final NSPS OOOOb and EG
0000c that are examined in this RIA. Additional technical detail on the engineering and cost
basis of the analysis is available within the preamble, the Technical Support Document (TSD)
accompanying the final rulemaking, hereafter referred to as the Final Rule TSD (U.S. EPA,
2023);15 the Technical Support Document (TSD) for the December 2022 Supplemental Proposal,
hereafter referred to as the December 2022 TSD (U.S. EPA, 2022); and the TSD for the
November 2021 proposal, hereafter referred to as the November 2021 TSD (U.S. EPA, 2021).
2.1.1 Emissions Sources
The section provides brief descriptions of the emissions sources subject to NSPS OOOOb
and EG OOOOc requirements. More detailed modeling, assumptions and other crucial
information, and additional technical detail is available in the preamble, the Final Rule TSD, the
December 2022 TSD, and the November 2021 TSD.
Fugitive Emissions:16 There are several potential sources of fugitive emissions
throughout the crude oil and natural gas production source category. Fugitive emissions occur
when connection points are not fitted properly or when seals and gaskets start to deteriorate.
Changes in pressure and mechanical stresses can also cause components or equipment to emit
fugitive emissions. Poor maintenance or operating practices, such as improperly reseated
pressure relief valves (PRVs) or worn gaskets on thief hatches on controlled storage vessels are
also potential causes of fugitive emissions. Additional sources of fugitive emissions include
agitator seals, connectors, pump diaphragms, flanges, instruments, meters, open-ended lines
(OELs), pressure relief devices such as PRVs, pump seals, valves or controlled liquid storage
tanks. These fugitive emissions do not include devices that vent as part of normal operations,
such as natural gas-driven pneumatic controllers or natural gas-driven pneumatic pumps, insofar
as the natural gas discharged from the device's vent is not considered a fugitive emission (e.g.,
an intermittent pneumatic controller that is venting continuously).
15 Available at https://www.regulations.gov/ under Docket No. EPA-HQ-OAR-2021-0317.
See Chapter 5 of the December 2022 TSD and Attachments A and B of the Final Rule TSD for more information.
2-18
-------
17
Pneumatic Controllers: Pneumatic controllers are devices used to regulate a variety of
physical parameters, or process variables, using air or gas pressure to control the operation of
mechanical devices, such as valves. The valves, in turn, control process conditions such as
levels, temperatures and pressures. When a pneumatic controller identifies the need to alter a
process condition, it will open or close a control valve. In many situations across all segments of
the oil and natural gas industry, pneumatic controllers make use of the available high-pressure
natural gas to operate or control the valve. In these "gas-driven" pneumatic controllers, natural
gas may be released with every valve movement and/or continuously from the valve control.
Pneumatic controllers can be categorized based on the emissions pattern of the controller.
Some controllers are designed to have the supply-gas provide the required pressure to power the
end-device, and the excess amount of gas is emitted. The emissions of this excess gas are
referred to as "bleed," and this bleed occurs continuously. Controllers that operate in this manner
are referred to as "continuous bleed" pneumatic controllers. These controllers can be further
categorized based on the amount of bleed they are designed to have. Those that have a bleed rate
of less than or equal to 6 standard cubic feet per hour (scfh) are referred to as "low bleed," and
those with a bleed rate of greater than 6 scfh are referred to as "high bleed." Another type of
controller is designed to release gas only when the process parameter needs to be adjusted by
opening or closing the valve, and there is no vent or bleed of gas to the atmosphere when the
valve is stationary. These types of controllers are referred to as "intermittent vent" pneumatic
controllers. A third type of controller releases gas to a downstream pipeline instead of the
atmosphere. These "closed loop" types of controllers can be used in applications with very low
pressure.
Not all pneumatic controllers are natural gas-driven. At sites with electricity, electrically
powered pneumatic devices or pneumatic controllers using compressed air can be used. As these
devices are not driven by pressurized natural gas, they do not emit any natural gas to the
atmosphere. At sites without electricity provided through the grid or on-site electricity
generation, solar power can be used in some instances.
17 See Chapter 3 of the December 2022 TSD and Chapter 2 of the Final Rule TSD for more information.
2-19
-------
Pneumatic Pumps:18 Most pneumatic pumps fall into two main types: diaphragm pumps,
generally used for heat tracing and plunger/piston pumps, generally used for chemical and
methanol injection. The pneumatic pump may use natural gas or another gas to drive the pump.
These pumps can also be electrically powered. "Non-natural gas-driven" pneumatic pumps can
be mechanically operated or use sources of power other than pressurized natural gas, such as
compressed "instrument air." Because these devices are not natural gas-driven, they do not
directly release natural gas or methane emissions. However, these systems have other energy
impacts, with associated secondary impacts related to generation of the electrical power required
to drive the instrument air compressor system. Instrument air systems are feasible only at oil and
natural gas locations where the devices can be driven by compressed instrument air systems and
have electrical service sufficient and reliable enough to power an air control system.
Reciprocating Compressors:19 In a reciprocating compressor, natural gas enters the
suction manifold, and then flows into a compression cylinder where it is compressed by a piston
driven in a reciprocating motion by the crankshaft powered by an internal combustion engine.
Emissions occur when natural gas leaks around the piston rod when pressurized natural gas is in
the cylinder. The compressor rod packing system consists of a series of flexible rings that create
a seal around the piston rod to prevent gas from escaping between the rod and the inboard
cylinder head. However, over time, during operation of the compressor, the rings become worn,
and the packaging system needs to be replaced to prevent excessive leaking from the
compression cylinder.
Centrifugal Compressors:20 Centrifugal compressors use a rotating disk or impeller to
increase the velocity of the natural gas where it is directed to a divergent duct section that
converts the velocity energy to pressure energy. These compressors are primarily used for
continuous, stationary transport of natural gas in the processing and transmission systems. Some
centrifugal compressors use wet (meaning oil) seals around the rotating shaft to prevent natural
gas from escaping where the compressor shaft exits the compressor casing. The wet seals use oil
which is circulated at high pressure to form a barrier against compressed natural gas leakage. The
18 See Chapter 4 of the December 2022 TSD and Chapter 3 of the Final Rule TSD for more information.
19
See Chapter 7 of the November 2021 TSD for more information.
20
See Chapter 2 of the December 2022 TSD for more information.
2-20
-------
circulated oil entrains and adsorbs some compressed natural gas that may be released to the
atmosphere during the seal oil recirculation process. Off gassing of entrained natural gas from
wet seal centrifugal compressors is not suitable for sale and is either released to the atmosphere,
flared, or routed back to a process. Some centrifugal compressors utilize dry seal systems. Dry
seal systems minimize leakage by using the opposing force created by hydrodynamic grooves
and springs.
Storage vessels:21 Storage vessels, or storage tanks, in the oil and natural gas sector are
used to hold a variety of liquids, including crude oil, condensates, and produced water. Many
facilities operate a group of storage vessels, sometimes in series but most often in parallel, used
to store the same oil or condensate streams. This group of tanks used to store a common fluid is
typically called a tank battery.
Underground crude oil contains many light hydrocarbon gases in solution. When oil is
brought to the surface and processed, many of the dissolved lighter hydrocarbons are removed
through a series of high-pressure and low-pressure separators. The oil (or condensate or water)
from the separator is then directed to a tank battery where it is stored before being shipped off-
site. Some light hydrocarbon gases remain dissolved in the oil, condensate, or water because the
separator operates at pressures above atmospheric pressure. These dissolved hydrocarbon gases
are released from the liquid as vapors, commonly referred to as flash gas, when stored at
atmospheric pressures in the tank batteries. Typically, the larger the operating pressure of the
separator, the more flash emissions will occur in the storage stage. Temperature of the liquid
may also influence the amount of flash emissions. Lighter crude oils and condensate generally
flash more hydrocarbons than heavier crude oils.
In addition to flash gas losses, other hydrocarbons may be emitted from the storage
vessels due to working and breathing (or standing) losses. Working losses occur when vapors are
displaced due to the emptying and filling of tank batteries. When the liquid level in the tank is
lowered, ambient air is drawn into the tank's headspace. Some hydrocarbons from the liquid will
volatilize into the headspace to reach equilibrium with the new headspace gas. When the liquid
level in the tank is increased, it will expel the saturated headspace gas into the atmosphere.
21
See Chapter 6 of the November 2021 TSD for more information.
2-21
-------
Breathing losses are the release of gas associated with daily temperature fluctuations when the
liquid level remains unchanged. As temperatures drop (or atmospheric pressure increases), gas in
the headspace contracts, drawing in ambient air. Again, hydrocarbons volatilize into this new gas
due to equilibrium effects. As the temperature rises (or atmospheric pressure falls), the gas in the
tank's headspace expands, expelling a portion of the hydrocarbon-saturated gas. Working losses
increase relative to the "turnover rate" (throughput rate divided by the tank capacity) and are
typically much greater than breathing losses.
Liquids Unloading:22 In new natural gas wells, there is generally sufficient reservoir
pressure/gas velocity to facilitate the flow of water and hydrocarbon liquids through the well
head and to the separator to the surface along with produced gas. In mature gas wells, the
accumulation of liquids in the wellbore can occur when the bottom well pressure/gas velocity
approaches the average reservoir pressure (i.e., volumetric average fluid pressure within the
reservoir across the areal extent of the reservoir boundaries). This accumulation of liquids can
impede and sometimes halt gas production. When the accumulation of liquid results in the
slowing or cessation of gas production (i.e., liquids loading), removal of fluids (i.e., liquids
unloading) is required to maintain production. These gas wells therefore often need to remove or
"unload" the accumulated liquids so that gas production is not inhibited.
The choice of what liquids unloading technique to employ is based on a well-by-well and
reservoir-by-reservoir analysis. To address the complex science and engineering considerations
to cover well unloading requirements, many differing technologies, techniques, and practices
have been developed to address an individual well's characteristics of the well to manage liquids
and maintain production of the well. At the onset of liquids loading, techniques that rely on the
reservoir energy are typically used. Eventually a well's reservoir energy is not sufficient to
remove the liquids from the well and it is necessary to add energy to the well to continue
production. Owners and operators can choose from several techniques to remove the liquids,
including manual unloading, velocity tubing or velocity strings, beam or rod pumps, electric
submergence pumps, intermittent unloading, gas lift (e.g., use of a plunger lift), foam agents and
wellhead compression. Each of these methods/procedures removes accumulated liquids and
thereby maintains or restores gas production. Although the unloading method employed by an
22
See Chapter 11 of the November 2021 TSD for more information.
2-22
-------
owner or operator can itself be a method that mitigates/eliminates venting of emissions from a
liquids unloading event, dictating a particular method to meet a particular well's unloading needs
is a production engineering decision.
Equipment Leaks at Gas Plants:23 The primary sources of equipment leak emissions
from natural gas processing plants are pumps, valves, and connectors. The major cause of
equipment leak emissions from valves and connectors is a seal or gasket failure due to normal
wear or improper maintenance. For pumps, emissions are often a result of a seal failure. The
large number of valves, pumps, and connectors at natural gas processing plants means emissions
from these components can be significant.
Common classifications of equipment at natural gas processing facilities include
components in VOC service and in non-VOC service. "In VOC service" is defined as a
component containing or in contact with a process fluid that is at least 10 percent VOC by
weight or a component "in wet gas service," which is a component containing or in contact with
field gas before extraction. "In non-VOC service" is defined as a component in methane service
(at least 10 percent methane) that is not also in VOC service.
The most common technique to reduce emissions from equipment leaks is to implement a
leak detection and repair (LDAR) program. Implementing an LDAR program can potentially
reduce product losses, increase safety for workers and operators, decrease exposure for the
surrounding community, reduce emissions fees, and help facilities avoid enforcement actions.
The effectiveness of an LDAR program is based on the frequency of monitoring, leak definition,
frequency of leaks, percentage of leaks that are repaired, and the percentage of reoccurring leaks.
Associated Gas:24 Associated gas is natural gas that is generated by oil wells (at
wellheads). The natural gas is either naturally occurring in a discrete gaseous phase within the
liquid hydrocarbon or is released from the liquid hydrocarbons by separation. In many areas, a
natural gas gathering infrastructure may be at capacity or unavailable. In such cases, if there is
23
See Chapter 10 of the November 2021 TSD for more information.
24
See Chapter 13 of the November 2021 TSD, Chapter 6 of the December 2022 TSD, and Chapter 4 of the Final
Rule TSD for more information.
2-23
-------
not another beneficial use of the gas at the site (e.g., as fuel) the collected natural gas is either
flared or vented directly to the atmosphere.
Since unprocessed natural gas is primarily composed of methane, with additional
amounts of carbon dioxide (CO2) and volatile organic compounds (VOC), including some
hazardous air pollutant (HAP) like benzene, toluene, ethylbenzene, and xylene, these air
pollutants are vented to the atmosphere from venting associated gas. If flared, the methane and
VOC emissions are reduced, but carbon monoxide (CO), CO2, nitrogen oxides (NOx), and the
HAP formaldehyde would be generated. Several analyses conducted by the EPA have indicated
that associated gas significantly contributes to methane and VOC emissions.
Options to mitigate emissions from associated gas in order of environmental and resource
conservation benefit include:
• Capturing the associated gas from the separator and routing the gas into a gas gathering
flow line or collection system;
• Beneficially using the associated gas (e.g., onsite use, natural gas liquid processing,
electrical power generation, gas to liquid);
• Reinjecting for enhanced oil recovery; and
• Flaring.
Except for flaring, the site-specific variabilities associated with the application of these control
options are significant.
2.1.2 Regulatory Requirements
Table 2-1 and Table 2-2 summarize the baseline and finalized standards of performance
for the sources with impacts quantified in this RIA.25 In Table 2-2, requirements in the baseline
differ depending on when sources were constructed relative to previous NSPS proposal dates.
We define pre- and post-KKK as dates prior to and after January 20, 1984, respectively. The
dividing dates for pre- and post-0000 and pre- and post-OOOOa are August 23, 2011, and
September 18, 2015, respectively. The abbreviations used in the table are OGI (optical gas
25
See the preamble for a more comprehensive description of the final standards.
2-24
-------
imaging), AVO (auditory, visual, and optical), scfh (standard cubic feet per hour), and scfm
(standard cubic feet per minute).
Table 2-1 NSPS OOOOb Emissions Sources, Baseline Requirements, and Finalized
Requirements
Standards of Performance
Source
In the Baseline
Under the Final Rule
Fugitive Emissions/Equipment Leaks3
Well Sites
Wellhead only, single well site
No requirement
Quarterly AVO monitoring
Wellhead only, multiple well site
No requirement
Quarterly AVO monitoring +
Semiannual OGI
Single well site with a single price of
major equipment and no tank battery
Semiannual OGI
Quarterly AVO monitoring
Multiple well site with a single piece
of major equipment, or any site with
(1) two or more pieces of major
Semiannual OGI
Bimonthly AVO monitoring +
equipment; (2) one piece of major
Quarterly OGI
equipment and a tank battery; or (3) a
controlled tank battery.
Gathering and Boosting Stations
Transmission and Storage Compressor
Stations
Quarterly OGI
Monthly AVO monitoring +
Quarterly OGI
Natural Gas Processing Plants
NSPS Subpart Wa
Bimonthly OGI
Pneumatic Pumps
Well Sites
95% control
Zero emissions'3
Gathering and Boosting Stations
No requirement
Pneumatic Controllers0
Well Sites
Gathering and Boosting Stations
Natural gas bleed rate no greater
Transmission and Storage Compressor
than 6 scfh
Zero emissions'1
Stations
Natural Gas Processing Plants
Zero emissions
Reciprocating Compressors
Gathering and Boosting Stations
Natural Gas Processing Plants
Rod-packing changeout on fixed
Volumetric flow rate of 2 scfm
Transmission and Storage Compressor
schedule
Stations
Wet Seal Centrifugal Compressors
Gathering and Boosting Stations
No requirement
Natural Gas Processing Plants
95% control
Transmission and Storage Compressor
95% control
Stations
Liquids Unloading
Well Sites
No requirement
Zero emissions or best
management practices6
2-25
-------
Storage Vessels
PTE > 6 tpy VOC
95% control, affected facility is
the tank
95% control, affected facility is
the tank battery
PTE < 6 tpy VOC
No requirement
No requirement
Associated Gas
Well Sites
No requirement
Route to sales line
a Well sites and compressor stations on the Alaska North Slope are subject to Annual OGI monitoring only.
b The zero emission standard for pumps applies to sites with electrical power and/or three or more diaphragm
pumps. Sites without access to electrical power that have fewer than three diaphragm pumps must route emissions to
a control device, provided one is onsite.
0 Specifically, the affected source is natural gas-driven controllers that vent to the atmosphere.
d The zero emissions rate standard does not apply to pneumatic controllers at sites in Alaska for which on site power
is not available. Instead natural gas-driven continuous bleed controllers at those sites are required to achieve bleed
rates at or below 6 scfh, while natural gas-driven intermittent bleed controllers are subject to OGI monitoring and
repair of emissions from controller malfunctions.
e The final regulation requires liquids unloading events to be zero-emitting unless technical infeasibilities exist, in
which case the regulation requires that best management practices be adopted.
2-26
-------
Table 2-2 EG OOOOc Emissions Sources, Baseline Requirements, and Finalized
Requirements
Presumptive Standards of Performance
Source
In the Baseline
Under the Final Rule
Fugitive Emissions/Equipment Leaksa'b
Well Sites
Wellhead only, single well site
Quarterly AVO monitoring
Wellhead only, multiple well site
No requirement
Quarterly AVO monitoring +
Semiannual OGI
Single well site with a single price of
major equipment and no tank battery
Pre-OOOOa: No requirement
Post-OOOOa: Semiannual OGI
Quarterly AVO monitoring
Multiple well site with a single piece
of major equipment, or any site with
two or more pieces of major
equipment or one piece of major
equipment and a tank battery
Pre-OOOOa: No requirement
Post-OOOOa: Semiannual OGI
Bimonthly AVO monitoring +
Quarterly OGI
Gathering and Boosting Stations
Transmission and Storage Compressor
Stations
Pre-OOOOa: No requirement
Post-OOOOa: Quarterly OGI
Monthly AVO monitoring +
Quarterly OGI
Natural Gas Processing Plants
Pre-KKK: No requirement
Post-KKK and Pre-OOOO:
NSPS Subpart W
Post-OOOO: NSPS Subpart Wa
Bimonthly OGI
Pneumatic Pumps
Well Sites
Pre-OOOOa: No requirement
Post-OOOOa: 95% control
Methane emission rate of zerob
Gathering and Boosting Stations
No requirement
Pneumatic Controllers0
Well Sites
Gathering and Boosting Stations
Pre-OOOO: No requirement
Post-OOOO: Natural gas bleed
rate no greater than 6 scfh
Methane emission rate of zerod
Transmission and Storage Compressor
Stations
Pre-OOOOa: No requirement
Post-OOOOa: Natural gas bleed
rate no greater than 6 scfh
Natural Gas Processing Plants
Pre-OOOO: No requirement
Post-OOOO: Zero emissions
Methane emission rate of zero
Reciprocating Compressors
Gathering and Boosting Stations
Natural Gas Processing Plants
Pre-OOOO: No requirement
Post-OOOO: Rod-packing
changeout on fixed schedule
Volumetric flow rate of 2 scfm
Transmission and Storage Compressor
Stations
Pre-OOOOa: No requirement
Post-OOOOa: Rod-packing
changeout on fixed schedule
Wet Seal Centrifugal Compressors
Gathering and Boosting Stations
No requirement
Natural Gas Processing Plants
Transmission and Storage Compressor
Stations
Pre-OOOO: No requirement
Post-OOOO: 95% control
Volumetric flow rate of 3 scfm
Liquids Unloading
2-27
-------
Well Sites
No requirement
Zero emissions or best
management practices6
Storage Vessels
PTE > 20 tpy CH4
Pre-OOOO: No requirement
Post-OOOO: 95% control,
95% control, affected facility is
the tank battery
PTE < 20 tpy CH4 and >
6 tpy VOC
affected facility is the tank
PTE < 20 tpy CH4 and <
6 tpy VOC
No requirement
No requirement
Associated Gas
Well Sites
PTE > 40 tpy CH4
No requirement
Route to sales line
PTE < 40 tpy CH4
No requirement
95% control
a Well sites and compressor stations on the Alaska North Slope are subject to Annual OGI monitoring only.
b The zero emission standard for pumps applies to sites with electrical power and/or three or more diaphragm
pumps. Sites without access to electrical power that have fewer than three diaphragm pumps must route emissions to
a control device, provided one is onsite.
0 Specifically, the affected source is natural gas-driven controllers that vent to the atmosphere.
d The zero emissions rate standard does not apply to pneumatic controllers at sites in Alaska for which on site power
is not available. Instead, natural gas-driven continuous bleed controllers at those sites are required to achieve bleed
rates at or below 6 scfh, while natural gas-driven intermittent bleed controllers are subject to OGI monitoring and
repair of emissions from controller malfunctions.
e The final regulation requires liquids unloading events to be zero-emitting unless technical infeasibilities exist, in
which case the regulation requires that best management practices be adopted.
There are finalized requirements that we do not attempt to quantify regulatory impacts for
in this RIA. We do not attempt to quantify the impacts of the super-emitter response program per
se due to the unpredictable nature of super-emission events, resulting in a lack of specific data on
their frequency, intensity, and cost to mitigate. Moreover, there is a lack of specific information
on how often third parties would conduct monitoring activities, what those activities would
entail, how often they would detect super-emitters, how often those detections would provide
sufficient and timely information for companies to respond. We do, however, attempt to quantify
the impacts of a regulatory requirement to monitor flares during OGI inspections, though we
limit our analysis to controlled storage vessels because we believe the available data is sufficient
for that emissions source to conduct meaningful quantification. Process emissions from inactive
or malfunctioning flares represent a significant source of super-emission events (Cusworth et al.,
2021), and so our assessment of the flare monitoring requirements can be viewed as capturing a
portion of the impacts that the super-emitter response program might otherwise have.
In addition, we also do not attempt to quantify regulatory impacts for storage vessel
control requirements at centralized production facilities (CPFs) and in the gathering and boosting
segment, or dry seal centrifugal compressor requirements in any segment. In both instances, we
lack sufficient data to conduct an informative analysis. However, given the relatively small
2-28
-------
source-level impact estimates for storage vessel control requirements at well sites (see Table
5-6), we expect the impacts from the storage vessels provisions at CPFs and in the gathering and
boosting segment to be small relative to the overall impacts of the final rule.
Finally, we also do not account for instances in which all or some sources in Alaska are
subject to different requirements than those in the rest of the country, both in the baseline due to
previous rulemakings and in this final rule. See Section 5.2 for additional discussion.
2.2 Methodology
The compliance cost and emissions reductions analysis summarized in this RIA reflects a
nationwide engineering analysis of which there are two main components: activity data and
information on control measures. The activity data represents estimates of the counts of affected
facilities over time, and the control measure information includes data on costs and control
efficiencies for typical facilities.
The first component is activity data for a set of representative (or model) plants for each
regulated facility.26 To project activity data for regulated facilities, we first project activity data
for oil and natural gas sites, which include well sites, natural gas processing plants, and
compressor stations (gathering and boosting, transmission, and storage). Projections include
addition of newly constructed sites and retirement of previously constructed sites, with
magnitudes based on a combination of analysis of several data sources and, where necessary,
sensible assumptions. Using representative per-site factors generated from the GHGI, regulated
facilities are apportioned to sites across all industry segments.27 We assume the per-site factors
are fixed over time, so that the projected counts of regulated facilities change in proportion to the
projected counts of sites. The regulated facility projections are combined with information on
control options, including capital costs, annual operations and maintenance costs, and control
efficiencies. Information on control options is derived from the analysis underpinning the BSER
determinations. Impacts are calculated by setting parameters on how and when affected facilities
are assumed to respond to a regulatory regime, multiplying activity data by model plant cost and
26 Regulated facilities include well site fugitives (including component emissions and malfunctioning storage vessel
flares), gathering and boosting station fugitives, transmission and storage compressor station fugitives, natural gas
processing plant equipment leaks, pneumatic pumps, pneumatic controllers, reciprocating compressors, centrifugal
compressors, liquids unloading, storage vessels, and associated gas.
27
Industry segments include production, gathering and boosting, processing, transmission, and storage.
2-29
-------
emissions estimates, differencing from the baseline scenario, and then summing to the desired
level of aggregation. In addition to emissions reductions, some control options result in natural
gas recovery, which can then be combusted in production or sold. Where applicable, we present
projected compliance costs with and without the projected revenues from product recovery.
For the analysis, we calculate the cost and emissions impacts of the final NSPS OOOOb
and EG 0000c from 2024 to 2038. The initial analysis year is 2024 as the rule will take effect
early in that year. The NSPS OOOOb is assumed to take effect immediately and impact sources
that commence construction after publication of the December 2022 proposal. We assume that
sources begin production, and thus begin generating emissions related to that production, one
year after they commence construction, so that sources assumed to be constructed in 2023 first
contribute cost and emissions impacts in 2024.28 We assume the EG OOOOc will take longer to
go into effect as states will need to develop implementation plans in response to the rule and
have them approved by the Agency. We assume that this process will take four years, and so EG
OOOOc impacts will begin in 2028. The final analysis year is 2038, which allows us to present
up to 15 post-finalization years of regulatory impact estimates.
While it would be desirable to analyze impacts beyond 2038, limited information
available to model long-term changes in practices and equipment use in the oil and natural gas
industry make the choice of a longer time horizon currently infeasible with the data and
modeling currently available to the EPA. In a dynamic industry like oil and natural gas,
technological progress is likely to change control methods to a greater extent over a longer time
horizon, creating more uncertainty about impacts of the NSPS OOOOb and the EG OOOOc. For
example, the current analysis does not include potential fugitive emissions controls employing
remote sensing technologies currently under development.
28 Due to supply chain considerations, NSPS OOOOb allows for longer compliance deadlines for process controllers
and pumps; see Sections XI.D.4 (process controllers) and XI.E.2 (pumps). To allow new wells the ability to plan
ahead to comply with the associated gas provisions, NSPS OOOOb also allows for a longer compliance deadline for
these affected facilities; see Section XI.F.2.d (associated gas) for additional details. We do not quantify the impacts
of those extensions in the RIA, but we expect the impacts to be small relative to the overall impacts of the rule due
to the limited nature of the extensions.
2-30
-------
2.2.1 A ctivity Data Projections
To construct the activity data projections used in this analysis, we rely on historical data
from the GHGI,29 industry data collected by the EPA through an information collection request
(ICR) distributed in 2016 (hereafter, "2016 ICR"), information from the private firm Enverus
that provides energy sector data and analytical services,30 the Department of Homeland
Security's Homeland Infrastructure Foundation-Level Data (HIFLD),31 and projections from the
U.S. Energy Information Administration's (EIA) Annual Energy Outlook (AEO).32 Our
projections follow a two-step procedure. First, we construct projected counts of oil and natural
gas sites, such as well sites, compressor stations, and processing plants, that contain or are
themselves facilities affected by the regulations. Second, using per-well factors, we build upon
the site projections to estimate the counts of these affected facilities. The details of these
calculations are described by site/regulated facility type below.
In addition to sites and affected facilities, there is a third category of activity data that we
track. When comparing a new regulatory regime, such as this final rule, to the baseline scenario,
a subset of affected facilities is assumed to take action to comply with regulatory requirements:
we refer to these facilities as "incrementally impacted facilities." In Section 2.2.1.3 below, we
provide a table of incrementally impacted facility counts for the final rule relative to the baseline.
2.2.1.1 Projected Oil and Natural Gas Sites
There are three types of sites in our analysis of projected facilities: well sites, compressor
stations, and natural gas processing plants. Compressor stations are further subdivided into sites
located in different segments of the natural gas sector, that is, the gathering and boosting,
transmission, and storage segments. For each site type, we generate annual projections of
cumulative and new counts for four different vintage bins: the first vintage (VI) represents sites
constructed prior to NSPS OOOO, the second vintage (V2) represents sites constructed after
29
See Methodology Annexes 3.5 and 3.6 at https://www.epa.gov/ghgemissions/natural-gas-and-petroleum-systems-
ghg-inventory-additional-information-1990-2019-ghg. Activity data is presented in Tables 3.5-5 and 3.6-7,
respectively.
Enverus: https://www.enverus.com/.
31
Homeland Infrastructure Foundation-Level Data (HIFLD): https://hifld-geoplatform.opendata.arcgis.com/.
32
EIA AEO: https://www.eia.gov/outlooks/aeo/.
2-31
-------
NSPS 0000 but prior to NSPS 0000a, the third vintage (V3) represents sites constructed after
NSPS 0000a but prior to NSPS 0000b, and the fourth vintage (V4) represents sites
constructed after NSPS 0000b. Within V3, sites are further subdivided into sites constructed
after NSPS 0000a through the base year (2019) and 3 additional vintages representing sites
assumed to be constructed in each year from 2020 through 2023. V4 is subdivided into separate
vintages for each year from 2024, when the NSPS 0000b is assumed to take effect, through
2038. In the case of well sites only, VI is subdivided into sites constructed prior to 2000 and
sites constructed from 2000 on to capture differences in the characteristics of well sites
constructed in different time periods.
There are two countervailing forces that impact the overall trajectory of our estimated
sites beyond the base year: the rate at which new sites are constructed and the rate at which sites
retire (or cease operation). In our analysis, counts of newly constructed sites are based on either
analysis of historical trends from the GHGI (compressor stations), GHGI and the Department of
Homeland Security's Homeland Infrastructure Foundation-Level Data (HIFLD) (processing
plants), or projections from AEO (well sites). Estimates of retirement rates are based on analysis
of Enverus data (well sites) and assumptions underlying analysis submitted in response to the
2018 NSPS 0000a Policy Reconsideration proposal (processing plants and compressor
stations);33 along with new site counts, those rates are summarized in Table 2-3. To avoid sites
having implausibly short lifespans in the analysis, we assume site retirements only apply to well
sites that are at least five years old and processing plants and compressor stations in VI (pre-
0000).
33 See page 4 of Appendix D of Docket ID No. EPA-HQ-OAR-2017-0757-0002.
2-32
-------
Table 2-3
Assumed Retirement Rates and Annual New Site Counts by Site Type
Type of Site
Annual Retirement Rate as a
New Site Counts in Each Year Percentage of Existing Stock
Well Sites
Greater than 15 barrels of oil
equivalent (boe) per day
3-15 boe per day
Oil
Gas
Less than 3 boe per day
14,000 - 28,000
1.6%
1%
0%
Oil
6.7%
4.4%
Gas
Compressor Stations
Gathering and Boosting
Transmission
Storage
Natural Gas Processing Plants
439
102
2
7
1%
1%
1%
1%
Our projections of the cumulative counts of sites for each vintage are illustrated in Figure
2-1. While the projected total counts of wells are relatively stable over the analysis horizon, the
projected total counts of well sites decline significantly, as older, smaller sites are displaced by
newer, larger sites. The total counts of natural gas processing plants and storage compressor
stations change slightly over time, due to very few assumed annual additions and retirements.
For gathering and boosting and transmission compressor stations, the total number of sites
increase significantly over the analysis horizon. Below, we describe how those trajectories are
generated for each site type.
2-33
-------
750,000-
800,000-
600,000-
400,000
200,000
4,000
3,000-
2,000-
1,000-
Exists
Projected
2020 2025 2030 2035
2020 2025 2030 2035
Year
2020 2025 2030 2035
Figure 2-1 Projections of Cumulative Site Counts by Site Type and Vintage
(a) Well Sites
The dataset used to characterize the base year (2019) population of oil and natural gas
well sites is developed from data provided by Enverus, a private firm focused on the energy
industry that provides data and analytical seivices. The dataset includes two types of entities:
wells and leases. Whether a well is represented as its own entity or as part of a lease depends on
the state in which the well is located, as reporting requirements differ across state agencies. The
columns in the dataset include entity identifiers, well site identifiers (for wells), locations,
completion and initial production dates, well counts (for leases), and natural gas and liquids
production levels. We restricted the dataset to onshore wells with positive production values in
2019. The base year is chosen as 2019 to maintain consistency with the proposals and avoid
reflected anomalous behavior resulting from the Covid-19 pandemic.
Using the base year dataset, we perform a series of steps to convert from well- and lease-
level data to site-level data. First, we aggregate the well-level data into site-level data using the
Enverus-provided well site identifiers when available. Each data point includes information on
site location, date (based on the most recent well completion), count of oil wells, count of gas
2-34
-------
wells, and liquids and gas production levels. Wells are categorized as oil or natural gas based on
the wells' gas-to-oil ratios (GOR).34 For wells without site identifiers and leases, we track the
same information, but at the well or lease level. For each entity (site, well, or lease), we project
production forward through the end of the analysis horizon using simple decline rate
assumptions based on analysis of historical production data. Decline rates are estimated using
well-level Enverus production data from 2010-2020. For each well and production year, decline
percentages for oil/condensate and gas production are calculated as the production level in the
next year less the production in the current year, divided by current production. We then
calculate median decline rates for four production rate bins, resulting in the decline rate
assumptions in Table 2-4.
Table 2-4 Decline Rate Assumptions by Production Type and Rate
Production Rate Bin (barrels of oil equivalents/day, or BOE/d)
Production Type Greater than 100 15-100 3-15 Less than 3
Oil/Condensate 35% 18% 11% 10%
Gas 26% 13% 9% 7%
Using the projections, we aggregate entities into representative groups for each year
(2019-2038). For well sites, each group characterized by a unique combination of state, region,35
vintage (based on the bins described in the previous section), site type (oil or natural gas), well
count bin (single well or multi-well), base year production rate bin, and current year production
rate bin.36 Each group includes total counts of sites, oil and natural gas wells, and oil and natural
gas production. Likewise, the well and lease entities for which we do not have site identifiers are
aggregated analogously, but without well count bins. To fill that gap, we apportion the well
counts and production levels of the well/lease entity groups into single and multi-well bins based
on regional- or state-level proportions derived from the subset of data with well site identifiers,
34
If GOR > 100,000 mcf per bbl, then the well is designated as a gas well, otherwise, it is designated as an oil well.
35
Well sites are mapped from American Association of Petroleum Geologists (AAPG) Geologic Provinces (see
https://ngmdb.usgs.gov/Geolex/stratres/provinces) to the supply regions from EIA's National Energy Modeling
System's Oil and Gas Supply Module (OGSM) (see Figure 1-2 in the linked report from
https://www.eia.gov/analysis/pdfpages/m063index.php). Our analysis regions are nearly the same as the OGSM
supply regions, except all sites in Alaska are grouped into a single region. The full set of regions and their
abbreviations are Alaska (AK), Gulf Coast (GC), Midcontinent (MC), Northeast (NE), Northern Plains (NP), Rocky
Mountain (RM), Southwest (SW), and West Coast (WC).
Sites are grouped into the four production rate bins, based on the average BOE/d per well at the site, described in
Table 2-4.
2-35
-------
37
stratifying over site types, vintage bins, and initial production rate bins. To complete the
imputation, we calculate the number of sites within each group by dividing well counts by
national average estimates of the number of wells per site. The two sets of groups are then
combined to form one cohesive dataset with projections of production for a collection of
representative well site groups.38
In addition to the projection of the base year dataset, we also implement a series of steps
to construct projections for well sites assumed to be constructed in years beyond 2019. First, we
implement the well site grouping procedure just described, but restricted to sites with completion
dates in a recent vintage (2018). Our operating assumption is that future sites will be distributed
similarly across locations, site types, well count bins, and production rate bins as sites recently
completed. We then use the representative grouping to distribute AEO2022 nationwide-
projections of new wells drilled from 2020 to 2038 based on the relative proportions of wells in
each group, with production for each vintage projected through 2038.
The last step in the well site projection procedure is to apply the retirement percentages
presented in Table 2-3. The retirement percentages differ by production bin and site type and are
otherwise uniformly applied across groups regardless of other characteristics, such as location.39
Only low production sites (less than 15 BOE/d/well) are assumed to retire, and the bulk of
retirements come from sites with very low production (less than 3 BOE/d/well).
(a) Compressor Stations
37
Wells in Kansas, Michigan, Mississippi, Nebraska, and Oklahoma are apportioned based on region-level data,
since all data for those states is provided at the lease level. Wells in California are apportioned based on state-level
data, since most wells in that state have pad identifiers. A small number of wells in Alabama, Arizona, Florida,
Louisiana, Missouri, Nevada, Ohio, Oregon, Pennsylvania, Texas, West Virginia, and Wyoming do not have pad
identifiers: those wells are assigned as single-well pads.
The dataset, along with the analysis code used to estimate impacts, can be found in the docket.
39
Retirement percentages are estimated using well-level Enverus production data from 2010-2020. For a subset of
those years (2012-2018), we identify wells that previously had production, but have no recorded oil or gas
production records for 2 consecutive years, as retired. Retirement percentages are then calculated by dividing the
count of retired wells in each year by the total count of producing wells from the previous year. The retirement rate
percentage assumptions result from averaging the estimated retirement rates over all years.
2-36
-------
We project compressor stations for three segments (gathering and boosting, transmission,
and storage) using data from the GHGI; the approach for all three segments is analogous.40 The
first step is to estimate the number of stations in the base year, 2019. We assume that the number
of stations in 2011 are all VI stations (pre-OOOO). To get the counts of VI stations in
subsequent years, including the base year, we apply the relevant annual retirement rates to the
2011 station counts. The number of V2 stations (post-OOOO, pre-OOOOa) in 2019 is estimated
by subtracting the estimated number of VI stations in 2015 from the total station counts from
2015. The number of V3 stations (post-OOOOa) in 2019 is estimated by subtracting the
estimated number of VI and V2 stations in 2019 from the total number of stations.
To project the number of new stations constructed in the years after the base year, we
calculate a historical average number of new stations per year over a recent period (as presented
in Table 2-3) and apply it uniformly across all years. Specifically, we divide the calculated
number of V3 stations in 2019 and divide it by four, as the first V3 stations are assumed to be
operating in 2016. This yields an estimate of the average number of V3 stations added per year
through the base year, and we assume new stations are added at that same rate beyond the base
year. New stations assumed to be operating in 2020-2023 are assigned to V3, while all estimated
new stations beyond 2023 are assigned to V4.
The final step is to project station counts for those existing in the base year and combine
those projections with the new construction projections. This results in a set of projections in
which VI station counts decline over the analysis horizon due to retirements and V2 station
counts are uniform over the analysis horizon. V3 station counts are also uniform over the
analysis horizon, but they are split across five vintages (2016 to 2019, 2020, 2021, 2022, and
2023), with the last four equal to the average number of new stations described above. Finally,
V4 station counts are equal to the new station estimates in all vintage/year combinations from
2024 to 2038.
(b) Natural Gas Processing Plants
40
Station counts are extracted from the following rows: Yard Piping (gathering and boosting) and Station +
Compressor Fugitive Emissions (transmission and storage).
2-37
-------
To construct base year activity data counts for natural gas processing plants, we leverage
data from both the GHGI and HIFLD.41 The estimates of the counts of VI and V2 plants are
generated using the same process as for compressor stations: the 2011 count of plants are
assigned to VI, and the V2 count of plants in 2015 is estimated to be the 2015 count from the
GHGI minus the estimated count of VI plants in 2015 after the annual retirement rates are
applied. We use the HIFLD as a source of 2020 plant counts since plant counts have been fixed
in the GHGI in recent years due to lack of data, and the latest update date for HIFLD is from
October 2020. Estimates for the count of V3 plants in 2020 are then calculated using the 2020
total plant estimate and subtracting VI (after applying retirements) and V2 plant counts. The
estimated number of new plants in each year beyond the base year is then calculated by dividing
the number of V3 plants in 2020 by the number of years (5) assumed to have passed since the
first NSPS OOOOa-affected facilities were constructed. That estimate is used to calculate the
number of V3 plants in the base year by subtracting it from the 2020 count, as well as to populate
the counts of plants for all vintages and years beyond the base year.
2.2.1.2 Affected Facilities
In most cases, estimates of projected affected facility counts are generated by assuming
fixed proportional relationships with the site counts. This means that as site counts are projected
to expand (construction of new sources) or contract (retirement of existing sources), the counts
of affected facilities expand and contract as well such that the ratio of facilities to sites remains
constant. Details for each affected facility type are provided below.
(a) Fugitives and Leaks
The final rule features LDAR requirements across all segments. Well site requirements
are the most nuanced and depend on the equipment present at the site, which we characterize
through a series of data processing steps. Compressor station requirements are uniform across
segments, and we rely on a single representative plant in each segment to estimate the impacts of
those requirements. Requirements at natural gas processing plants distinguish the collection of
41
The dataset of processing plants is downloaded from https://hifld-
geoplatform.opendata.arcgis.com/datasets/geoplatform: :natural-gas-processing-plants/explore. We filter out plants
located in Canada and Mexico.
2-38
-------
VOC service components and the collection of non-VOC service components. Our impacts
analysis for processing plants differentiates between two model plant types representing "large"
and "small" facilities.
The final rule features different monitoring frequency requirements for well sites
depending on the equipment present at a site. The Enverus data does not provide information on
site equipment, so we assign well site groups to equipment categories in fixed proportions based
on analysis of data from the EPA's 2016 ICR and site-level survey data provided by the
American Petroleum Institute (API).42'43 The ICR data captured a survey of certain major
equipment (separators, compressors, and dehydrators) and storage tanks at more than 100,000
well sites across the U.S., which we use to separately estimate the proportions of sites in eleven
equipment categories and two tank bins for all combinations of oil and natural gas sites,
production bins, and single and multi-well sites at the regional level, where the data was
sufficient.44 For a few regions, all site observations within each region were pooled to estimate
equipment and storage vessel bin proportions.
Since the ICR data does not contain information on heater-treaters and process heaters,
we supplement with information on those equipment types from the API survey data. Grouping
the API survey sites by the presence of separators, compressors, and/or dehydrators (i.e., has
major equipment surveyed in the ICR or not); regions; site types (oil or natural gas); and well
count bins (single or multi), we calculated the proportion of sites with heater-treaters, and
process heaters for each combination. The proportions were then applied to the equipment bins
42
See https://www.epa.gov/controlling-air-pollution-oil-and-natural-gas-industry/background-information-request-
oil-and for more information on the ICR. The ICR was withdrawn in 2017, but not before significant amounts of
data were collected. The data used for this analysis were obtained from the file
"OilandGasSpreadsheetUnredacted.xlsx" found at
https://foiaonline.gov/foiaonline/action/public/submissionDetails?trackingNumber=EPA-HQ-2017-
003014&type=Request.
43
See Attachment 4 (Microsoft Excel workbook) of Docket ID No. EPA-HQ-OAR-2017-0757-0002, EPA Analysis
of Well Site Fugitive Emissions Monitoring Data Provided by API. The dataset contains survey data on 2,183 gas
well sites and 1,742 oil well sites.
44
The equipment categories are: (1) no separators, compressors, or dehydrators; (2) one separator; (3) more than one
separator; (4) one compressor; (5) more than one compressor; (6) one dehydrator; (7) more than one dehydrator; (8)
both separator(s) and compressor(s); (9) both separator(s) and dehydrator(s); (10) both compressors) and
dehydrator(s); and (11) separator(s), compressor(s), and dehydrator(s). The storage vessel categories are: (1) has
tanks, and (2) does not have tanks. Since we estimate proportions for all combinations of equipment bin, storage
vessel bin, region (AK, GC, MC, NE, NP, RM, SW, and WC), site type (oil, gas), production level (low, non-low)
and well count bin (single, multi), there are 1,408 possibilities in total.
2-39
-------
developed from the ICR data to estimate site proportions for all equipment bin combinations
including heater-treaters and process heaters.45
We also estimate average equipment and storage vessel counts for each equipment and
storage vessel bin combination. For single well sites, there are distinct estimates of separators,
compressors, and dehydrators for each combination of region, equipment bin, and storage vessel
bin at a minimum, with further disaggregation by site type and production level where possible.
For multi-well sites, we fit linear models to estimate average equipment counts as a function of
site well counts, constraining the estimated parameters to agree with the equipment bins for all
possible well counts (e.g., at least one separator in each equipment bin that denotes the presence
of a separator, at least two separators in the equipment bin that designates more than one
separator at the site).46 Average equipment counts for heater-treaters and process heaters are
calculated from the API data for each combination of region, site type, well count bin, and
equipment presence. See Section 2.7 for a detailed discussion of how the ICR data is processed
to construct equipment bin proportions and average equipment counts.
After assigning equipment proportions and average counts from the ICR and API survey
data to the well site group projections, the base year results were compared to 2019 activity data
from the 2021 GHGI. For the major equipment survey in the ICR (separators, compressors, and
dehydrators), the aggregate equipment counts were reasonably close and so no further
adjustments were made. For the additional major equipment types only found in the API survey
(heater-treaters and process heaters), the aggregate equipment counts estimated by applying the
API equipment proportions and averages to the well site group projections were significantly
larger than the estimates from the GHGI. The discrepancy is probably due to the underlying
population of the API survey, which was intended to reflect the types of sites that would be
45
The result is a total of 37 possible equipment bins: one "no equipment" bin, five "exactly one piece of major
equipment" bins (one for each major equipment type), five "exactly one major equipment type but more than one
piece of equipment" bins (one for each major equipment type), ten "exactly two major equipment types" bins (all
combinations), ten "exactly three major equipment types" bins (all combinations), five "exactly four major
equipment types" bins (all combinations), and one bin for having all five types of major equipment.
46
Due to a lack of observations of multi-well sites, the linear models were fit to higher aggregations of regions, site
types, and equipment, storage vessel, and production level bins. For separators, compressors, and dehydrators,
distinct models were run for two equipment bin categories: (1) separators-only with more than one piece of major
equipment, and (2) more than one type of major equipment. For separators, the data was sufficient to estimate
models for each region, while for compressors and dehydrators, the models were estimated at the national level. For
storage vessels, separate regional models were estimated on all sites with at least one storage vessel.
2-40
-------
impacted by NSPS 0000a requirements. Such sites would be expected to be larger and have
more equipment than the overall population of production sites. Therefore, the equipment
proportions and average counts for heater-treaters and process heaters were calibrated by
reducing the prevalence of those equipment types for older (pre-NSPS 0000) sites until the
aggregate estimates for the well site group projections in the base year matched the GHGI
activity data.
Beyond the major equipment types discussed above, we characterize two other types of
equipment that have fugitive components in the well site group projections, headers and
meters/piping. Initially, a header is assigned to all oil sites with at least one piece of major
equipment or storage vessel. As was the case for process heaters and heater-treaters, this resulted
in far more headers per well in the well site group base year projection than estimated in the
GHGI activity data. To bring the projections more in line with the GHGI, we calibrated the
proportion of pre-NSPS 0000 oil sites assumed to have headers such that the aggregate ratio of
headers per well matched the values from the GHGI. For meters/piping, we assigned one unit to
all gas well sites.
The equipment category proportions are illustrated in Table 2-5 for well sites in the base
year and newly constructed sites in subsequent vintages. The table reflects the distribution of
sites after making the adjustments for heaters and heater-treaters and applying the stratified
proportions to the well site group dataset. We assume that equipment is assigned to well sites
based on current year production levels, reflecting the reallocation of equipment away from
certain sites as production declines.
Fugitive emissions monitoring requirements differ across the equipment bins captured in
the table. In the analysis of the finalized option, single well wellhead-only sites and sites with
only one major piece of equipment and no tank battery are assumed to perform quarterly AVO
inspections. Wellhead-only sites with multiple wells are assumed to perform quarterly AVO and
semiannual OGI monitoring. Sites with two or more major pieces of equipment, one piece of
major equipment and a tank battery, or multi-wellhead sites with one piece of major equipment
or a tank battery are assumed to perform bimonthly AVO and quarterly OGI monitoring. To
calculate impacts for the fugitive monitoring requirements at well sites, we allocate the total
number of well sites to the bins defined by counts of major equipment and tank batteries.
2-41
-------
Table 2-5 Distribution of Well Sites in Equipment Bins
Site Bin
TSD
Model
Plant
Proportions
in the base
year
(2019)
Proportions in
the projected
years
(2020-2023)
Proportions in
the projected
years
(2024-2038)
Natural Gas
Single wellhead
Wellhead only
Small sites
Large sites without controlled storage vessels
Large sites with controlled storage vessels
Multi-wellhead
Wellhead only
Large sites without controlled storage vessels
Large sites with controlled storage vessels
33%
18%
44%
<1%
<1%
5%
<1%
21%
6%
23%
5%
<1%
36%
9%
21%
6%
23%
5%
<1%
15%
29%
Oil
Single wellhead
Wellhead only
Small sites
Large sites without controlled storage vessels
Large sites with controlled storage vessels
Multi-wellhead
Wellhead only
Large sites without controlled storage vessels
Large sites with controlled storage vessels
49%
11%
27%
3%
2%
3%
4%
26%
7%
7%
17%
4%
4%
34%
26%
7%
5%
19%
4%
3%
34%
Affected facility counts for compressor station fugitives are equal to the compressor
station counts detailed in the previous section. As such, compressor station fugitives affected
facility counts are binned according to segment, vintage, and year.
There are two affected facility types associated with natural gas processing plant leaks:
the collection of VOC service components and the collection of non-VOC service components.
In each case, the number of affected facilities is equal to the number of processing plants, and so
the total number of affected facilities is twice the number of processing plants. For the purposes
of calculating impacts associated with LDAR at processing plants, we assume that 80 percent of
plants are "large", and 20 percent are "small".47
47
See page 6 of Chapter 10 of the November 2021 TSD.
2-42
-------
(b) Pneumatic Controllers
Pneumatic controllers are represented in the GHGI for all segments. For well sites, we
estimate the number of controllers at sites based on equipment counts. For compressor stations,
controller counts are directly based on per-station counts from the GHGI in 2019. For processing
plants, we assume that all controllers are already powered by compressed air and therefore do not
estimate any impacts under the final rule for that segment and affected facility.
To estimate controller counts at well sites, we proceed in two steps. First, we multiply,
for each well site group, equipment counts per site by controller-per-equipment factors presented
in the Supporting Information of Zavala-Araiza et al. (2017).48 Equipment counts for separators,
compressors, dehydrators, and process heaters per site are calculated as described in the
preceding section on fugitives and leaks. Plunger lifts are assigned to low production gas sites
such that the aggregate proportion of gas sites with plunger lifts in the well site group projections
matches the 2019 proportion from the 2021 GHGI activity data. The resulting aggregate implied
controller counts for the base year are close to those from the GHGI activity data, so no further
calibration steps are taken. Second, we allocate the controllers across to three types (low-bleed,
high-bleed, and intermittent bleed) such that each type matches the corresponding GHGI
controller per-well counts in 2019, assuming that no high-bleed controllers exist at post-0000
sites in any state and at pre-0000 sites in California, Colorado, or Utah.
The estimation of controller counts at compressor stations is similar to the last step for
well sites. In that case, we assume that high-bleed controllers are only allocated to pre-0000
gathering and boosting stations and pre-OOOOa stations for transmission and storage. In
aggregate, the per-station counts for all three types of controllers in the base year match the per-
station counts from the GHGI in 2019.
48
Using data from Allen et al. (2015), the authors estimate 0.42 controllers per wellhead, 0.90 controllers per
plunger lift, 1 controller per separator at gas sites without liquids production, 2.06 controllers per separator at sites
with liquids production, 1.5 controllers per process heater, 4.3 controllers per compressor, and 2.5 controllers per
dehydrator. To prevent certain sites from having a fractional controller count, we assume 1 controller per plunger
lift. We also assume, in lieu of a specific estimate, that each heater-treater has 1.5 controllers (same as process
heaters and roughly the midpoint between separators at sites with and without liquids production). Finally, based on
a review of the underlying data from Allen et al. (2015), we assume that all controllers at wellheads qualify as
emergency shutdown devices and are thus not subject to the controller requirements.
2-43
-------
(c) Pneumatic Pumps
The GHGI provides information on the number of pneumatic pumps in the production
and gathering and boosting segments. For well sites, we assume that 30 percent of gas sites with
equipment (and no sites without equipment) have chemical injection pneumatic pumps, based on
analysis of the data underlying Allen et al. (2013).49 Likewise, we assume 25 percent of oil sites
with equipment have chemical injection pneumatic pumps, based on an analysis of the API
survey data. For each site assumed to have pumps, we initially assign one pump to the site.
Additional pumps are assigned in proportion to the number of pneumatic controllers at each site
such that number of pumps per-well matches the 2019 data from the GHGI. For the gathering
and boosting segment, we calculate the number of pumps per station implied by the GHGI in
2019 and apply the value to all stations for all vintages in all years.
(d) Reciprocating Compressors
The GHGI contains estimates of the number of reciprocating compressors in the
gathering and boosting, processing, transmission, and storage segments. In all cases, we calculate
the number of reciprocating compressors per site using the 2019 values from the GHGI and
apply those ratios to the cumulative and new station counts for all vintages and years. In the case
of gathering and boosting stations, the GHGI only includes a total count of compressors; we
assume that 89 percent of those are reciprocating.50
(e) Centrifugal Compressors
The GHGI contains estimates of the number of wet seal centrifugal compressors in the
gathering and boosting, processing, and transmission segments. For the transmission segment,
we assume that no wet seal compressors have been installed since the NSPS OOOOa due to the
routing requirements in that rule, and that no wet seal compressors will be installed at NSPS
OOOOb-affected stations either. Taking that into account, wet seal compressors are allocated on
a per-station basis such that the estimated aggregate counts of wet seal compressors per station in
49
The data can be downloaded from http://dept.ceer.utexas.edu/methane/study/datasets3.cfm. The workbook used
fortius analysis is finalSITES.xlsx.
This assumption is based on data summarized on page 28 of Zimmerle et al. (2019).
2-44
-------
the base year match the GHGI data from 2019. For gathering and boosting stations, the process is
similar except we allocate wet seal compressors to all vintages since this segment/affected
facility type has yet to be regulated. Also, the GHGI only includes a total count of compressors
for this segment; we assume that 3 percent of those are centrifugal,51 and that the proportion of
wet seal to dry seal centrifugal compressors is the same as it is in the transmission segment.
(j) Liquids Unloading
For the purposes of the RIA, liquids unloading affected facilities are defined at the event
level and apply only to natural gas well sites. To estimate impacts, we divide natural gas well site
groups into two categories: those with plunger lifts and those without plunger lifts. The process
for allocating plunger lifts across sites is described in the pneumatic controller section above; the
process for allocating sites that have liquids unloading without plunger lifts is similar, as the
GHGI contains activity data for the number of wells that perform liquids unloading events, so we
divide that number by the total number of natural gas wells in the inventory in 2019 to generate
fractions of wells performing liquids unloading for each category. Those fractions are applied to
our projections of well sites with equipment for all years and vintages. In the case of wells with
plunger lifts, we assume that 76 percent of sites perform manual unloading.52 Finally, we convert
from sites to events by multiplying by events per well values from the BSER analysis.53
(g) Associated Gas
Associated gas affected facility projections at well sites are constructed by applying base
year proportions of associated gas to the well site group projections. First, proportions of oil
wells with associated gas flaring and venting are calculated at the regional level using GHGRP
data for the 2019 reporting year. Next, the well site group data for the base year is used to
calculate proportions of oil wells both with and without gas production on site, conditional on the
51 Ibid.
52
Memorandum. Analysis of Greenhouse Gas Reporting Program Liquids Unloading Data. Prepared by SC&A
Incorporated for Amy Hambrick, SPPD/OAQPS/EPA. October 14, 2021. Docket ID No. EPA-HQ-OAR-2021-
0317-0143. As summarized in the memo, analysis of well-level data from the GHGRP for reporting years 2015-
2019 suggested that 76% of plunger lifts were manually operated.
53
See page 12 of Chapter 11 of the November 2021 TSD. We assume that wells without plunger lifts have 5.6
events per year, and wells with manually operated plunger lifts have 7.7 events per year.
2-45
-------
presence of a separator on site; gas sites and oil sites without separators are assumed not to have
associated gas venting or flaring. Oil sites with separators are then partitioned into three
categories: (1) no associated gas; (2) associated gas flaring; and (3) associated gas venting, such
that the aggregate proportion of oil wells in each region with associated gas flaring and venting
matches the GHGRP data, and no sites with gas production are assigned as having associated gas
flaring or venting before any oil site with a separator without gas production. Notably, the
proportions of oil sites with associated gas venting and flaring are assumed to be constant across
vintages and production levels after conditioning on the region and presence of separators, due to
a lack of data that would allow for more nuanced assumptions.
(h) Storage Vessels
Storage vessel-affected facility projections are generated for well sites only; projections
of tanks at centralized production facilities and in the gathering and boosting segment have been
omitted due to a lack of data. As described in Section 2.2.1.2(a)(a), proportions of sites with
tanks and tank counts per oil and natural gas well are generated from the 2016 ICR data and
merged into our well site projections. For each site assumed to have tanks, the total count of
tanks is assumed to comprise a single tank battery. All liquids production at those sites (crude at
oil sites, condensate at gas sites) is assumed to be throughput to the tank battery.
2.2.1.3 Incrementally Impacted Facilities
Estimates of incrementally impacted facility counts by year and regulated facility for the
final rule are presented in Table 2-6 through Table 2-8. The counts for well sites and compressor
stations represent fugitives requirements at those sites and the counts for natural gas processing
plants represent VOC and non-VOC service.
2-46
-------
Table 2-6 Projection of Incrementally Impacted Affected Facilities under the Final
NSPS OOOOb and EG OOOOc, 2024 to 2038 (Production Sources)
Fugitive
Emissions
Pneumatics
Associated
Gas
Liquids
Unloading
Storage Vessels
Year
Well
Sites
Flares
Well
Sites
Controllers
Pumps
Well Sites
Events
Tank
Batteries
Tanks
2024
8,800
4,700
6,500
45,000
1,900
2,300
460
570
1,900
2025
18,000
9,800
14,000
95,000
4,000
4,700
980
1,200
4,000
2026
28,000
15,000
21,000
140,000
6,200
7,200
1,500
1,900
6,300
2027
38,000
20,000
28,000
190,000
8,300
9,700
2,100
2,500
8,300
2028
520,000
55,000
350,000
1,400,000
100,000
67,000
270,000
3,000
9,700
2029
510,000
57,000
350,000
1,400,000
100,000
67,000
260,000
3,400
11,000
2030
500,000
59,000
340,000
1,400,000
100,000
67,000
260,000
3,800
12,000
2031
490,000
63,000
340,000
1,400,000
99,000
68,000
250,000
4,200
13,000
2032
480,000
66,000
340,000
1,400,000
98,000
68,000
250,000
4,600
15,000
2033
480,000
69,000
330,000
1,400,000
97,000
68,000
240,000
5,000
16,000
2034
470,000
70,000
330,000
1,400,000
96,000
69,000
240,000
5,600
19,000
2035
460,000
73,000
330,000
1,400,000
95,000
69,000
230,000
6,500
23,000
2036
460,000
76,000
320,000
1,400,000
94,000
70,000
230,000
7,300
26,000
2037
450,000
79,000
320,000
1,400,000
93,000
70,000
220,000
8,700
35,000
2038
440,000
82,000
320,000
1,400,000
92,000
71,000
220,000
9,800
38,000
2-47
-------
Table 2-7 Projection of Incrementally Impacted Affected Facilities under the Final
NSPS OOOOb and EG OOOOc, 2024 to 2038 (Non-Production Fugitive/Leaks and
Pneumatics Sources)
Fugitives/Leaks Pneumatics
Gathering
and Transmission Transmission and
Boosting Processing and Storage Gathering and Boosting Storage
Year
Stations
Plants
Stations
Stations
Controllers
Pumps
Stations
Controllers
2024
0
6
0
420
4,100
440
99
1,700
2025
0
11
0
850
8,200
880
200
3,400
2026
0
17
0
1,300
12,000
1,300
300
5,100
2027
0
23
0
1,700
16,000
1,800
400
6,800
2028
5,200
590
1,900
11,000
100,000
11,000
3,200
59,000
2029
5,100
590
1,900
11,000
110,000
11,000
3,300
61,000
2030
5,100
590
1,800
11,000
110,000
12,000
3,300
62,000
2031
5,000
590
1,800
12,000
110,000
12,000
3,400
63,000
2032
5,000
600
1,800
12,000
120,000
13,000
3,500
65,000
2033
5,000
600
1,800
13,000
120,000
13,000
3,600
66,000
2034
4,900
600
1,800
13,000
130,000
13,000
3,700
68,000
2035
4,900
600
1,800
13,000
130,000
14,000
3,800
69,000
2036
4,900
600
1,800
14,000
130,000
14,000
3,900
71,000
2037
4,800
600
1,800
14,000
140,000
15,000
3,900
72,000
2038
4,800
600
1,700
15,000
140,000
15,000
4,000
73,000
2-48
-------
Table 2-8 Projection of Incrementally Impacted Affected Facilities under the Final
NSPS OOOOb and EG OOOOc, 2024 to 2038 (Non-Production Compressor Sources)
Reciprocating Compressors Wet-Seal Centrifugal Compressors
Year
Gathering
and
Boosting
Stations
Processing
Plants
Transmission
and Storage
Stations
Gathering
and
Boosting
Stations
Processing
Plants
Transmission
and Storage
Stations
2024
970
38
290
13
0
0
2025
1,900
76
570
26
0
0
2026
2,900
110
860
39
0
0
2027
3,900
150
1,100
52
0
0
2028
24,000
4,000
9,500
320
280
740
2029
25,000
4,000
9,700
340
280
740
2030
26,000
4,000
10,000
350
280
730
2031
27,000
4,000
10,000
360
280
730
2032
28,000
4,000
10,000
370
270
720
2033
29,000
4,000
11,000
380
270
720
2034
30,000
4,000
11,000
390
270
710
2035
31,000
4,000
11,000
410
260
710
2036
31,000
4,000
11,000
420
260
700
2037
32,000
4,100
12,000
430
260
700
2038
33,000
4,100
12,000
440
260
690
2.2.2 Model Plant Compliance Cost and Emissions Reductions
The cost and emissions characteristics of the site projections used to estimate the impacts
of the final rule are derived from the technical analyses underpinning the BSER determination.
In some cases, we characterize our affected facilities to be identical to the model plants found in
the Final Rule TSD, December 2022 TSD, or the November 2021 TSD, and so the cost and
emissions estimates can be directly applied. In other cases, however, our model plants leverage
the underlying data from the TSDs and other data sources to better fit the activity data and
account for source heterogeneity.
(a) Compressor Station Fugitives, Natural Gas Processing Plant Leaks, and Compressors
We use cost and emissions information directly from the November 2021 TSD for
compressor station fugitives, natural gas processing plant leaks, and reciprocating compressors,
and from the December 2022 TSD for centrifugal compressors. Compressor station fugitives are
represented by a single model plant for each of the gathering and boosting, transmission, and
2-49
-------
storage segments.54 Processing plant leaks are divided into four different model plants: all
combinations of large and small plants, and VOC and non-VOC service.55 Reciprocating
compressors are represented by a single model plant for each of the gathering and boosting,
processing, transmission, and storage segments.56 Wet-seal and dry-seal centrifugal compressors
are each represented by a single model plant for each of the gathering and boosting, processing,
and transmission segments.57
(b) Storage Vessels
Storage vessel control costs and emissions reductions are adapted from the BSER
analysis summarized in Chapter 6 of the November 2021 TSD. For each well site group assumed
to have tanks, representative site-level tank potential to emit (PTE) is calculated by multiplying
crude or condensate throughput by an average emissions factor derived from the BSER
analysis.58 To determine which post-0000 sites are assumed to have controlled tanks in the
baseline, we use the VOC PTE estimate in the base year or initial year of construction,
whichever is later; pre-0000 sites are assumed to be uncontrolled in the baseline. In the final
rule scenario, control requirements are determined by the VOC PTE estimates in the year of
construction for NSPS OOOOb-affected facilities and CH4 PTE estimates in the year that the EG
54
See Chapter 12 of the November 2021 TSD for details on costs and emissions reductions associated with quarterly
OGI monitoring, which represents the proposed BSER for compressor station fugitives in both the NSPS OOOOb
and EG 0000c.
See Chapter 10 of the November 2021 TSD for details on costs and emissions reductions associated with NSPS
W Method 21 (the BSER established in NSPS KKK), NSPS Wa Method 21 (the BSER established in NSPS
0000), and bimonthly OGI (the BSER proposed in NSPS OOOOb and EG 0000c).
See Chapter 7 of the November 2021 TSD for details on costs and emissions reductions associated with rod-
packing replacement on a fixed schedule (the BSER established in NSPS 0000 and NSPS 0000a) and rod-
packing replacement based on emissions monitoring (the BSER proposed in NSPS OOOOb and EG 0000c).
57
See Chapter 2 of the December 2022 TSD for details on costs and emissions reductions associated with a direct
inspection and maintenance/repair program to maintain emissions below 3 scfm (the BSER proposed in NSPS
OOOOb for dry seals and EG 0000c for wet and dry seals).
Emission factors are estimated by calculating the average VOC and methane emissions per barrel across the
sample tanks on the "Condensate" and "Oil" tabs from the docketed workbook, EPA-HQ-OAR-2021-0317-
0039_attachment_21. Condensate emissions factors were applied to sites in AAPG Geological Provinces with
average API gravity greater than or equal to 40, while oil emissions factors were applied to sites in AAPG
Geological Provinces with average API gravity less than 40. Average API gravity for all AAPG Geological
Provinces was calculated by taking the average API gravity from Table J. 1 in Subpart W reporting year 2019 for all
facilities, weighted by reported oil volume; see https://enviro.epa.gov/query-
builder/ghg/PETROLEUM%20AND%20NATURAL%20GAS%20SYSTEMS%20(RY%202015-
2022)/EF_W_ATM_STG_T ANKSCALC10R2.
2-50
-------
is assumed to take effect (2028) for EG OOOOc-affected facilities.59 For sites subject to control
requirements, we assume that 95 percent control is achieved through application of flares to the
entire tank battery. The costs of control are based on the BSER analysis, with capital and annual
costs equal to a minimum value below a 50 TPY CH4 emissions threshold and following a
quadratic cost function for sites with emissions above that threshold.60
(c) Well Site Fugitives
The methodology for projecting of costs and emissions impacts from AVO and OGI
monitoring programs of different frequencies in the production segment uses counts of major
equipment in well site groups (described in Section 2.7), the results of the BSER technical
analysis performed in support of this action, and information on process emissions from the
scientific literature. The BSER analysis is used to estimate baseline fugitive emissions from
major equipment components and thief hatches at controlled storage vessels, as well as
monitoring program costs and performance. Data from the scientific literature is used to estimate
baseline emissions due to inactive flares at controlled storage vessels; OGI monitoring of flares
is a requirement in the final rule.
The BSER analysis uses simulations produced by the Fugitive Emissions Abatement
Simulation Tool (FEAST).61 FEAST calculates simulated costs and emissions reductions of
LDAR programs at model well sites under different assumptions. The BSER analysis performs
FEAST simulations using four model well sites (a single-well site with no major equipment
(MP1); a multi-well site with no major equipment (MP2); a multi-well site with a separator, an
in-line heater, and a dehydrator (MP3); and a multi-well site with a separator, an in-line heater, a
dehydrator, and a controlled storage tank battery (MP4)) and five OGI frequencies (annual,
semiannual, quarterly, bimonthly, and monthly). Each model well site has an assumed number of
components based on the number of wells and the type of major equipment present at the site. A
59
Note that V2 and V3 vintage sites are subject to the more stringent of NSPS OOOO and NSPS 0000c, which we
assume is NSPS 0000.
The cost functions above the threshold are estimated by fitting a quadratic function of methane emissions on cost
using data points for methane emissions of 50, 150, 300, and 1500 TPY in the "New" and "Existing" tabs from the
docketed workbook, EPA-HQ-OAR-2021-0317-0039_attachment_20. The fitted cost functions imposed a constraint
that the functions be equal to the cost values from the workbook at an emissions rate of 50 TPY CH4.
See Chapter 5 of the December 2022 TSD for details on the FEAST modeling and costs and emissions reductions
associated with OGI monitoring at well sites.
2-51
-------
FEAST simulation for a model well site produces an average annualized cost and emissions
reduction percentage for each OGI monitoring frequency along with a baseline emissions rate in
the absence of an LDAR program. Given that the only difference between MP3 and MP4 sites is
the presence of storage tank battery control at the latter, we assume that the difference in baseline
emissions between the two sites is due to thief hatch emissions and that those emissions are
reduced by the same proportion as component emissions for a given LDAR program.
Emissions factors for process emissions due to inactive flares at controlled storage
vessels in the production segment are estimated through a series of steps, starting with data
published as supporting information to Cusworth et al. (2021). First, we merge the plume and
source data from the study so that we can identify emissions sources in the data specific to
inactive flares at storage vessels. To ensure that we are capturing emissions that could reasonably
be reduced by an LDAR program, we limit the sample to sources that were identified by at least
three overhead flights and calculate average emissions for each source using persistence-
weighting. Then, we calculate average emissions across sources to arrive at an emissions factor
(in tons per year of methane) for an inactive flare at a production segment tank battery. Finally,
we convert the emissions factor for inactive flares to an emissions factor for all flares by
multiplying by an estimate of the ratio of inactive flares to total flares at storage vessels within
the study area. Since an estimate of the total number of flares at storage vessels in the study area
was not available, we estimated one by using our well site group data to estimate the number of
storage vessel flares per well in the southwest region (since the Cusworth et al. (2021) study was
conducted on the Permian basin) and applied it to the number of wells in the study area
(-72,000, according to the study authors).
The calculation of LDAR (including OGI monitoring at controlled storage vessels) costs
and emissions impacts from a well site group consists of five main steps. First, a well site group
is assigned major equipment as described in Section 2.2.1.2(a). Second, each well site group is
matched to a FEAST model well site based on major equipment counts. Single-well sites with no
major equipment or one piece of major equipment are matched to MP1, while multi-well sites
with no major equipment are matched to MP2. All other sites are matched, depending on
whether they are assumed to have controlled storage tank batteries, to either MP3 (without) or
MP4 (with). The presence of control of storage vessels for each well site group is estimated
based on a combination of estimated emissions, prior requirements under NSPS OOOO and
2-52
-------
NSPS 0000a, and new requirements under NSPS 0000b and EG 0000c. Third, a
component count per well site is determined for a well site group based on the counts of major
equipment from step 1. Component counts for each type of major equipment are assigned based
on Tables W-1B (for gas well sites) and W-1C (for oil well sites) from 40 CFR part 98, Subpart
W.62 Fourth, the number of components per well site are multiplied by a per-component
emissions factor and summed over well sites to determine baseline component emissions for a
well site group, while thief hatch and inactive storage vessel flare emissions factors are applied
to sites with controlled storage vessels. The per-component emissions factor is calculated by
fitting a linear model regressing baseline emissions on component counts for MP1, MP2, and
MP3 from the FEAST simulations. Finally, the cost and emissions impacts of an OGI monitoring
program of a given frequency is determined by applying the average annualized cost and
emissions reduction percentage for the matched model well site from the FEAST simulations to
the well site group.
(d) Pneumatic Devices
Control of pneumatic controllers and pneumatic pumps are analyzed in a unified
framework for pneumatic devices using a combination of BSER analysis, Carbon Limits'
abatement cost tool, and the GHGI. Our analysis incorporates the impacts of replacing high bleed
with low bleed pneumatic controllers,63 which reflects the BSER established in NSPS 0000 for
well sites and gathering and boosting stations and NSPS 0000a for transmission and storage
compressor stations, as well as three zero emitting control options: electronic controllers using
grid electricity, electronic controllers powered by solar photovoltaic (PV) and battery systems,
and compressed air systems using grid electricity.64 Emissions factors for low-bleed, high-bleed,
and intermittent bleed pneumatic controllers (all segments except processing) and pneumatic
pumps (production and gathering and boosting) from the BSER analysis are converted from kg
CH4 per device to tons CH4 per device and applied directly to device counts at the site level to
62 See https://www.ecfr.gov/current/title-40/chapter-I/subchapter-C/part-98.
See Chapter 8 of the November 2021 TSD for details on costs and emissions reductions associated with replacing
high bleed with low bleed pneumatic controllers.
64
See Chapter 3 of the December 2022 TSD and Attachment Q "Carbon Limits 2021 Zero Bleed Pneumatics Cost
Tool" at https://www.regulations.gov/comment/EPA-HQ-OAR-2021-0317-0845 for details on costs and emissions
reductions associated with installing zero-bleed controllers (the BSER established in NSPS 0000 for processing
plants and proposed in NSPS 0000b and EG 0000c for all other segments).
2-53
-------
calculate site-wide emissions, pre- and post-control.65 Control costs vary across control options
and are described in detail below.
For each control option, pneumatic device control costs are comprised of capital and
annual operations and maintenance costs, each of which is based on two main components: site-
level "base" costs that are independent of the number of devices at the site, and costs that scale
with the number of devices at the site. For replacement of high-bleed controllers with low-bleed
controllers, control costs scale linearly with the number of high-bleed controllers at the site and
are estimated using the BSER analysis from the 2011 TSD after updating to 2019 dollars (U.S.
EPA, 201 lc). Consistent with the Carbon Limits tool assumptions, we assume retrofit capital
costs for low-bleed controllers (not including installation labor costs) are half of the cost of new
capital costs, since only the controller and not the control valve will be required. For electronic
controllers powered by solar photovoltaic and battery systems, we calculate the base and per-
device capital costs associated with installing and replacing control panels, solar PV panels,
batteries, and devices and annual (per-device) costs associated with device maintenance. Cost
calculation for electronic controllers powered by grid electricity is similar, but with solar PV and
battery capital costs replaced by base and per-device annual electricity costs. For compressed air
systems, we calculate the base and per-device capital costs associated with installing and
replacing a compressor and base and per-device annual costs associated with compressor
maintenance and grid electricity purchases.66
For the final rule, as well as the regulatory alternatives specified in Section , control
options are applied at the site level and compared to the baseline. Costs in the baseline consist of
purchasing and installation costs (for newly constructed sites) and maintenance costs (for newly
constructed and existing sites) of natural gas-driven pneumatic devices. We assume that
electronic controllers powered by solar PV and battery systems are used to comply with the zero
emissions standard at all well sites. In contrast, we assume that gathering and boosting and
65 The emissions factor used for pumps is a blend of the diaphragm and piston pump emission factors used in the
BSER analysis, assuming 50.2% of pumps are diaphragm pumps, which is consistent with the calculation method
used in the GHGRP.
Based on the Carbon Limits tool, we assume that compressor costs (capital and maintenance) are a quadratic
function of horsepower requirements, which is a linear function of the number of each type of device at the site.
Therefore, we model a second-order polynomial relationship between site-level compressor costs and the numbers
of each type of device at the site.
2-54
-------
transmission and storage compressor stations are grid-connected and comply with the regulation
through, due to the large number of controllers assumed to be located at the model plants,
installation of a compressed air system.
(e) Liquids Unloading
We define two model plants for liquids unloading: events at wells without plunger lifts
and manual unloading events at wells with plunger lifts. In both cases, the costs per event are
taken directly out of the December 2022 TSD. However, whereas the BSER analysis evaluates a
range of emissions reductions levels associated with the final option, this analysis assumes
emissions reductions of 29 percent and 36 percent for events at wells without plunger lifts and
manual unloading events at wells with plunger lifts, respectively.67
(j) Associated Gas
Cost and emissions impacts for associated gas are estimated through a combination of
assumptions from the BSER analysis (see the 2023 Final Rule TSD) and the GHGI. Baseline
emissions are calculated by applying the 2019 basin-level, per-million-barrel emissions factors
from the 2021 GHGI to the estimated production levels of the well site groups estimated to have
associated gas flaring or venting as described in Section 2.2.1.2(g). The cost and performance
assumptions are derived from the BSER analysis. We assume that flares have 95 percent control
efficiency with capital costs of $100,579 and annual costs of $25,000 in 2019 dollars. The costs
of routing to a sales line are estimated based on assumptions of a line length of 4 miles,
compressor horsepower of 25 hp, and gathering line capital costs halfway between the costs of 4-
inch line and 6-inch line. Sites with associated gas flaring in the baseline only incur costs (and
achieve emissions reductions) if they are assumed to route to sales lines in the policy scenarios.
For our analysis of the final rule, we assume that all NSPS OOOOb-affected facilities route to
sales lines, while EG OOOOc-affected facilities only route to sales lines if estimated pre-flare
associated gas emissions (assuming 95 percent control for sites with associated gas flaring) are
greater than 10 tons of methane per year; otherwise, they are assumed to route to flare.
67 See Chapter 11 of the November 2021 TSD for details on costs associated with best management practices during
liquids unloading events, which is the compliance option we assume for this analysis. Additionally, see the memo
titled "Analysis of Greenhouse Gas Reporting Program Liquids Unloading Data," Docket ID No. EPA-HQ-OAR-
2021-0317-0143.
2-55
-------
2.2.3 State Programs
The oil and natural gas industry is subject to numerous state and local requirements.
These requirements differ greatly in scope and stringency across states. Given the difficulty in
attempting to incorporate the myriad of state regulations in the baseline, we have chosen to
incorporate state actions into the baseline for California and Colorado. Both states have
comprehensive regulatory programs for the oil and natural gas industry and contribute
significantly to national production levels. We have also incorporated fugitive monitoring
requirements in New Mexico and Pennsylvania into the baseline. By not accounting for state and
local requirements (outside of Colorado, California, New Mexico, and Pennsylvania) in the
baseline, this analysis may overestimate both the benefits and the costs of the final regulation.
Specifically, we assume that California and Colorado have requirements at least as
stringent as those in the final rule for compressor station fugitives, natural gas processing plant
leaks, pneumatic devices, reciprocating compressors, wet seal centrifugal compressors, storage
vessels; and associated gas. In addition, we assume that Colorado has requirements at least as
stringent as those in the final rule for liquids unloading. For well site fugitives, we assume
California, Colorado, Pennsylvania, and New Mexico have requirements at least as stringent as
those in the final rule.
To incorporate the California, Colorado, New Mexico, and Pennsylvania rules in the
baseline, our activity data projections for sites and affected facilities need to estimate the counts
for those states. For the production segment, the processes described in Section 2.2.1.1 already
account for state level activity counts. For the other segments, midstream data from HIFLD and
Enverus was used to calculate the proportions of natural gas processing plants and compressor
stations, respectively, in California and Colorado. We assume that those proportions hold fixed
in all analysis years, and that affected facilities are also distributed according to those
proportions.
2.3 Emissions Reductions
Table 2-9 summarizes the emissions reductions associated with the final standards. The
emissions reductions are estimated by multiplying the source-level emissions reductions
associated with each applicable control and facility type by the number of affected sources of
2-56
-------
that facility type. We present methane emissions in both short tons and CO2 Eq. using GWP of
28.
Table 2-9 Projected Emissions Reductions under the Final NSPS OOOOb and EG
OOOOc, 2024-2038
Emissions Changes
Methane
Year
Methane
(short tons)
VOC
(short tons)
HAP
(short tons)
(metric tons CO2 Eq.
using GWP=28)
2024
250,000
77,000
2,900
6,400,000
2025
490,000
150,000
5,600
12,000,000
2026
700,000
210,000
8,000
18,000,000
2027
900,000
270,000
10,000
23,000,000
2028
4,900,000
1,300,000
48,000
120,000,000
2029
4,900,000
1,300,000
49,000
130,000,000
2030
5,000,000
1,300,000
49,000
130,000,000
2031
5,000,000
1,300,000
50,000
130,000,000
2032
5,100,000
1,300,000
50,000
130,000,000
2033
5,100,000
1,400,000
51,000
130,000,000
2034
5,100,000
1,400,000
51,000
130,000,000
2035
5,100,000
1,400,000
52,000
130,000,000
2036
5,200,000
1,400,000
53,000
130,000,000
2037
5,200,000
1,400,000
54,000
130,000,000
2038
5,200,000
1,500,000
55,000
130,000,000
Total
58,000,000
16,000,000
590,000
1,500,000,000
Note: Values rounded to two significant figures. Totals may not appear to add correctly due to rounding.
2.4 Product Recovery
The projected compliance costs presented below include the revenue from natural gas
recovery projected under the final standards. Requirements for fugitive emissions monitoring,
equipment leaks at processing plants, reciprocating and centrifugal compressors, pneumatic
devices, and liquids unloading events are assumed to increase the capture of methane and VOC
emissions that would otherwise be vented to the atmosphere, and we assume that a large
proportion of the averted methane emissions can be directed into natural gas production streams
and sold; see Chapters 2-4 (pneumatic devices and associated gas) of the Final Rule TSD,
Chapters 2 (centrifugal compressors) and 5 (fugitive emissions) of the December 2022 TSD and
Chapters 7 (reciprocating compressors) and 10-11 (equipment leaks at natural gas processing
plants and liquids unloading) of the November 2021 TSD for details on the proportion of
recovered emissions associated with the compliance options.
2-57
-------
Table 2-10 summarizes the increase in natural gas recovery and the associated revenue.
The AEO2022 projects Henry Hub natural gas prices generally rising from $3.17/MMBtu in
2024 to $3.68/MMBtu in 2038 in 2021 dollars, with a low of $2.98/MMBtu in 2026.68 To be
consistent with other financial estimates in the RIA, we adjust the projected prices in AEO2022
from 2021 dollars to 2019 dollars using the GDP-Implicit Price Deflator. We also adjust prices
for the wellhead using an EIA study that indicated that the Henry Hub price is, on average, about
11 percent higher than the wellhead price (Budzik, 2002). Finally, we use a conversion factor of
1.038 MMBtu equals 1 Mcf.69 Incorporating these adjustments, wellhead natural gas prices are
assumed to rise from $2.78/Mcf in 2024 to $3.23/Mcf in 2038 in 2019 dollars, with a low of
$2.61/Mcf in 2026.
Table 2-10 Projected Increase in Natural Gas Recovery under the Final NSPS OOOOb
and EG OOOOc Option, 2024-2038
Year
Increase in Gas Recovery (Bcf)
Increased Revenue (millions 2019$)
2024
62
$170
2025
110
$290
2026
140
$380
2027
170
$460
2028
400
$1,100
2029
410
$1,200
2030
410
$1,300
2031
420
$1,300
2032
420
$1,300
2033
430
$1,400
2034
430
$1,400
2035
430
$1,400
2036
430
$1,400
2037
440
$1,400
2038
440
$1,400
Note: Values rounded to two significant figures.
Operators in the gathering and boosting, processing, and transmission and storage
segments of the industry do not typically own the natural gas they transport; rather, they receive
payment for the transportation and processing service they provide. From a social perspective,
68 Available at: https://www.eia.gov/outlooks/aeo/excel/aeotab_13.xlsx. Accessed July 25, 2022.
69
For MMbtu-Mcf conversion factor, see
https://www.eia.gov/tools/faqs/faq.php?id=45&t=8#:~:text=One%20thousand%20cubic%20feet%20(Mcf,1.036%20
MMBtu%2C%20or%2010.36%20therms. Accessed October 31, 2023.
2-58
-------
however, the increased financial returns from natural gas recovery accrues to entities somewhere
along the natural gas supply chain and should be accounted for in a national-level analysis. An
economic argument can be made that, in the long run, no single entity bears the entire burden of
compliance costs or fully appropriates the financial gain of the additional revenues associated
with natural gas recovery. The change in economic surplus resulting from natural gas recovery is
likely to be spread across different market participants. Therefore, the simplest and most
transparent option for allocating these revenues would be to keep the compliance costs and
revenues within a given source category and not make assumptions regarding the allocation of
costs and revenues across agents.
2.5 Compliance Costs
Table 2-11 summarizes the compliance costs and revenue from product recovery for the
evaluated emissions sources and points. Total costs consist of capital costs, annual operating and
maintenance costs, and revenue from product recovery. Capital costs include the capital costs
from the requirements on newly affected pneumatic devices, reciprocating compressors, storage
vessels, and associated gas facilities, as well as the planning costs associated with monitoring
requirements for fugitive emissions at well sites and compressor stations and equipment leaks at
processing plants; these costs are reincurred as operators are assumed to have to renew survey
monitoring plans or purchase new capital equipment at the end of its useful life. The annual
operating and maintenance costs are due to requirements on fugitive emissions and equipment
leaks, controllers at gas processing plants, compressors, liquids unloading events, storage
vessels, and associated gas.
Note that Table 2-11 shows a pulse of capital expenditures in 2028, the year the RIA
assumes to be the compliance year for the EG OOOOc. In practice, however, the EG 0000c
gives States and sources the flexibility to spread these installations over a period of up to three
years, or the 2027 to 2029 period.™ While we do not distribute compliance expenditures across
these years in the RIA, we believe that States and sources will avail themselves of this flexibility.
70
The compliance timeline for EG OOOOc is 36 months after the state plan submittal deadline. See Section XIII.E.
of the final rule preamble for discussion on state plan submittal and compliance timelines.
2-59
-------
Table 2-11 Projected Compliance Costs under the Final NSPS OOOOb and EG OOOOc
Option, 2024-2038 (millions 2019$)
Operating and Increased Revenue Annualized Costs with
Year
Capital Costs
Maintenance
Costs
Annualized
Costs
from Product
Recovery
Increased Revenue from
Product Recovery
2024
$1,400
$27
$180
$170
$3
2025
$1,500
$57
$370
$290
$78
2026
$1,600
$88
$560
$380
$190
2027
$1,600
$120
$760
$460
$300
2028
$13,000
$1,500
$3,600
$1,100
$2,500
2029
$1,600
$1,500
$3,800
$1,200
$2,500
2030
$1,600
$1,500
$3,900
$1,300
$2,600
2031
$1,600
$1,500
$4,000
$1,300
$2,700
2032
$1,900
$1,500
$4,000
$1,300
$2,700
2033
$1,600
$1,400
$4,100
$1,400
$2,800
2034
$1,700
$1,400
$4,100
$1,400
$2,800
2035
$1,600
$1,400
$4,300
$1,400
$2,900
2036
$2,100
$1,400
$4,400
$1,400
$3,000
2037
$1,700
$1,400
$4,500
$1,400
$3,100
2038
$1,800
$1,400
$4,600
$1,400
$3,200
Note: Values rounded to two significant figures. Totals may not appear to add correctly due to rounding.
The expected lifetimes that capital and planning costs are incurred over differs across
affected facilities. The cost of designing, or redesigning, fugitive emissions monitoring programs
at well sites and compressor stations are assumed to occur every eight years, while the planning
cost associated with equipment leak surveys at processing plants are assumed to occur every five
years. Pneumatic device lifetimes are assumed to be 15 years, the lifetimes of solar PV panels
and batteries used to power electronic controllers are assumed to be 10 and four years,
respectively, and the lifetime of compressors used to power compressed air systems is assumed
to be six years. Rod-packing replacement at reciprocating compressors is assumed to happen
about every three years in the processing segment and four years in the gathering and boosting
and transmission and storage segments. The capital costs in each year outlined in Table 2-11
includes the estimated costs for newly affected sources in that year, plus the costs for sources
affected previously that have reached the end of their assumed economic lifetime.
The calculation of total annualized costs proceeds as follows. Capital and planning costs
are annualized over their requisite expected lifetimes at an interest rate of 7 percent. These
annualized capital costs are then added to the annual operating and maintenance costs of the
requirements to get the total annualized costs without product recovery in each year. The value
2-60
-------
of product recovery is then subtracted to get the total annualized costs with product recovery in
each year. Under the final rule, over 90 percent of revenue from the sale of captured natural gas
is projected to be earned by operators in the production segment of the industry, where we
assume that the operators own the natural gas and will receive the financial benefit from the
captured natural gas. The remainder of the captured natural gas is captured within the gathering
and boosting, processing, transmission, and storage segments, where operators do not typically
own the natural gas they transport; rather, they receive payment for the transportation service
they provide. In the RIA, we treat these revenues as an offset to projected compliance costs,
though the revenues could instead be considered as a benefit of the regulatory action. However,
regardless of whether the revenue from capture of natural gas is considered a compliance cost
offset or a benefit, the net benefits are equivalent.
Table 2-12 shows the undiscounted stream of costs for each year from 2024 through 2038
due to the final standards. Capital costs are the projected capital and planning costs expected to
be incurred. Total costs are the sum of the capital costs and annual operating costs. The revenue
from the increase in product recovery is estimated using the AEO2022 natural gas price
projections, as described earlier. Total costs with revenue from product recovery equal the total
anticipated costs minus the revenue.
2-61
-------
Table 2-12 Undiscounted Projected Compliance Costs under the Final NSPS OOOOb
and EG OOOOc, 2024-2038 (millions 2019$)
Year
Capital Costs
Annual
Operating
Costs
Total Costs
(w/o Revenue)
Revenue from
Product
Recovery
Total Costs
(with Revenue)
2024
$1,400
$27
$1,400
$170
$1,300
2025
$1,500
$57
$1,600
$290
$1,300
2026
$1,600
$88
$1,600
$380
$1,300
2027
$1,600
$120
$1,700
$460
$1,200
2028
$13,000
$1,500
$14,000
$1,100
$13,000
2029
$1,600
$1,500
$3,100
$1,200
$1,900
2030
$1,600
$1,500
$3,100
$1,300
$1,800
2031
$1,600
$1,500
$3,100
$1,300
$1,800
2032
$1,900
$1,500
$3,400
$1,300
$2,000
2033
$1,600
$1,400
$3,100
$1,400
$1,700
2034
$1,700
$1,400
$3,100
$1,400
$1,800
2035
$1,600
$1,400
$3,000
$1,400
$1,700
2036
$2,100
$1,400
$3,500
$1,400
$2,100
2037
$1,700
$1,400
$3,100
$1,400
$1,700
2038
$1,800
$1,400
$3,200
$1,400
$1,800
Note: Values rounded to two significant figures. Totals may not appear to add correctly due to rounding.
We now present the compliance costs of the final NSPS OOOOb and EG OOOOc in a
PV framework. The stream of the estimated costs for each year from 2024 through 2038 is
discounted back to 2021 using 2, 3, and 7 percent discount rates and summed to get the PV of the
costs. The PV is then used to estimate the EAV of the estimated costs. The EAV is the single
annual value which, if summed in PV terms across years in the analytical time frame, equals the
PV of the original (i.e., likely time-varying) stream of costs. In other words, the EAV takes the
potentially "lumpy" stream of costs and converts them into a single value that, when discounted
and added together over each period in the analysis time frame, equals the original stream of
values in PV terms.
Table 2-13 shows the discounted stream of costs discounted to 2021 using 2, 3, and 7
percent discount rates. The PV of the stream of costs discounted to 2021 using a 2 percent
discount rate and accounting for product recovery is $19 billion, with an EAV of $2.1 billion per
year. The PV of the stream of costs discounted to 2021 using a 3 percent discount rate and
accounting for product recovery is $18 billion, with an EAV of $2 billion per year. The PV of the
stream of costs discounted to 2021 using a 7 percent discount rate and accounting for product
recovery is $14 billion, with an EAV of $1.5 billion per year.
2-62
-------
Table 2-13 Discounted Projected Costs under the Final NSPS OOOOb and EG OOOOc
Option, 2024-2038 (millions 2019$)
2 Percent
3 Percent
7 Percent
Total
Total
Total
Total
Total
Total
Annual
Annual
Annual
Annual
Annual
Annual
Cost (w/o
Revenue
Costs (w/
Cost (w/o
Revenue
Costs (w/
Cost (w/o
Revenue
Cost (w/
Product
from
Product
Product
from
Product
Product
from
Product
Recovery
Product
Recovery
Recovery
Product
Recovery
Recovery
Product
Recovery
Year
Revenue)
Recovery
Revenue)
Revenue)
Recovery
Revenue)
Revenue)
Recovery
Revenue)
2024
$130
$160
-$32
$130
$160
-$24
$140
$140
$3
2025
$270
$270
$1
$270
$260
$14
$280
$220
$60
2026
$400
$340
$61
$400
$330
$78
$400
$270
$130
2027
$530
$410
$130
$530
$380
$140
$500
$300
$200
2028
$2,800
$1,000
$1,800
$2,600
$930
$1,700
$2,300
$710
$1,600
2029
$2,800
$1,000
$1,700
$2,600
$950
$1,700
$2,200
$700
$1,500
2030
$2,800
$1,100
$1,700
$2,600
$960
$1,700
$2,100
$680
$1,400
2031
$2,800
$1,100
$1,700
$2,600
$970
$1,600
$2,000
$660
$1,400
2032
$2,700
$1,100
$1,700
$2,500
$960
$1,600
$1,900
$630
$1,300
2033
$2,700
$1,100
$1,700
$2,500
$950
$1,600
$1,800
$600
$1,200
2034
$2,700
$1,100
$1,600
$2,500
$930
$1,500
$1,700
$570
$1,200
2035
$2,700
$1,000
$1,700
$2,400
$910
$1,500
$1,700
$530
$1,100
2036
$2,700
$1,000
$1,700
$2,400
$890
$1,500
$1,600
$500
$1,100
2037
$2,700
$1,000
$1,700
$2,400
$870
$1,600
$1,500
$470
$1,100
2038
$2,700
$1,000
$1,700
$2,400
$860
$1,600
$1,500
$450
$1,000
PV
$31,000
$13,000
$19,000
$29,000
$11,000
$18,000
$22,000
$7,400
$14,000
EAV
$3,500
$1,400
$2,100
$3,200
$1,200
$2,000
$2,400
$950
$1,500
Note: Values rounded to two significant figures. Totals may not appear to add correctly due to rounding. Costs and
revenue from product recovery in each year are discounted to 2021.
2.6 Comparison of Regulatory Alternatives
In this section, we compare the compliance cost and emissions impacts projected under
the final rule with the results of two alternative regulatory scenarios, one less stringent and one
more stringent than the final rule. The alternative scenarios focus on sources that account for the
greatest quantities of estimated emissions reductions of methane for the final rule: fugitives,
pneumatic devices, and associated gas, all in the production segment.
The alternative scenarios are summarized in Table 2-14.a The NSPS OOOOa established
a standard of performance of 95% control for diaphragm pumps at well sites with existing
combustion devices, which we do not include in our baseline due to a lack of information
regarding which sites would be subject to the requirement. We believe this results in a slight
overestimate of the impacts of the final rule and more stringent scenarios. In the less stringent
scenario, we estimate the impacts of allowing sites with four or fewer controllers to convert all
2-63
-------
continuous bleed controllers to either zero or intermittent bleed controllers (our analysis assumes
the latter occurs in all cases) and perform periodic inspections to detect and repair
malfunctioning intermittent bleed controllers. In the more stringent scenario, we illustrate the
impact of the small well site OGI exemption for by requiring well sites with a single piece of
major equipment to perform semiannual OGI in addition to quarterly AVO. The more stringent
scenario also estimates the impact of the associated gas emissions threshold in the EG OOOOc
by requiring all sites to route to sales regardless of estimated associated gas methane emissions.
These alternatives reflect key regulatory design alternatives that the EPA considered while
developing the final rule.
Table 2-14 Summary of Regulatory Alternatives (Well Sites Only)
Applicable
NSPS
Less
More
Source
NSPS
Baseline
Stringent
Final
Stringent
Fugitive Emissions
Single well site with a
Quarterly
single price of major
0000a
Semiannual
Quarterly
Quarterly
AVO +
equipment and no tank
OGI
AVO
AVO
Semiannual
battery
OGI
Pneumatic Devices
Sites with four or fewer
controllers
Continuous bleed
controllers
0000
Natural gas
bleed rate no
greater than 6
scfh
Zero
emissions or
convert to
intermittent
Zero
emissions
Zero
emissions
bleed
Intermittent bleed
None
No
Inspection
Zero
Zero
controllers
requirement
program
emissions
emissions
Pumps
0000a
No
requirement3
No
requirement
Zero
emissions
Zero
emissions
Associated Gas (EG Only)
Sites with associated gas
No
requirement
emissions less than 40
TPY CH4
None
95% control
95% control
Route to sales
a The NSPS 0000a established a standard of performance of 95% control for diaphragm pumps at well sites with
existing combustion devices, which we do not include in our baseline due to a lack of information regarding which
sites would be subject to the requirement. We believe this results in a slight overestimate of the impacts of the final
rule and more stringent scenarios.
2-64
-------
A comparison of estimated costs and emissions reductions is presented in Table 2-15 for
three years: 2024 (the first year of NSPS OOOOb impacts), 2028 (the first year of EG 0000c
impacts), and 2038 (the last year of analysis) across regulatory options. Overall, the table
demonstrates that we estimate the emissions impacts to be similar across the three options. By
the time the EG 0000c is assumed to begin having an effect in 2028, we estimate that the less
stringent option would result in roughly ten percent fewer methane emissions reductions of the
finalized option, while reducing costs by a commensurate amount. On the other hand, we
estimate that the more stringent option would result in only slightly more methane emissions
reductions while costing significantly more than the finalized option.
Table 2-15 Comparison of Regulatory Alternatives in 2024, 2028, and 2038 for the Final
NSPS OOOOb and EG OOOOc across Regulatory Options (millions 2019$)
Regulatory Alternative
Less Stringent
Final Rule
More Stringent
Total ImDacts. 2024
Emissions reductions
Methane (short tons)
250,000
250,000
250,000
VOC (short tons)
76,000
77,000
77,000
Costs
Annualized Costs without
Product Recovery (3%)
Annualized Costs with
Product Recovery (3%)
$150
-$27
$150
-$27
$150
-$27
Total ImDacts. 2028
Emissions reductions
Methane (short tons)
4,500,000
4,900,000
5,000,000
VOC (short tons)
1,200,000
1,300,000
1,300,000
Costs
Annualized Costs without
Product Recovery (3%)
Annualized Costs with
$3,000
$1,900
$3,300
$2,100
$7,000
$5,800
Product Recovery (3%)
Total ImDacts. 2038
Emissions reductions
Methane (short tons)
4,900,000
5,200,000
5,300,000
VOC (short tons)
1,400,000
1,500,000
1,500,000
Costs
Annualized Costs without
$3,800
$2,500
$4,000
$2,600
$6,300
$4,900
Product Recovery (3%)
Annualized Costs with
Product Recovery (3%)
2-65
-------
2.7 Additional Information on Use of 2016 Oil and Gas ICR Data
The 2016 ICR data includes a survey of equipment at well sites and production
characteristics of the wells at those sites. The data are cleaned and processed to generate
estimates of the distribution of major equipment and storage vessels across sites which can be
directly applied to our base year well site activity data as described in Section 2.2.1. The data
processing steps, and a list of key assumptions made along the way, are summarized below.
The first step in the data cleaning procedure is to read in and remove duplicate and
incomplete entries from the raw 2016 ICR data workbook. Well-level data comes from the sheets
"PtlWellsFromWebForms" and "PtlWellsFromFileUploads", with corresponding site-level data
sheets "PtlWellSiteFromWebForms" and "PtlWellSiteFromFileUploads". Data from both types
of submissions are merged together to create two master raw data tables, one for wells and one
for sites. For both types of data, duplicate entries were removed, first on the basis of having
identical entries for all rows, and then on the basis of having identical well/wellsite IDs. For the
latter, when the only difference is the submission time entry, we assume the last entry
(chronologically) is the correct one. If submission time cannot be used to differentiate in the
well-level data, we simply pick the first entry,71 unless duplicate entries have different well types
(oil or gas), in which case we drop the wells from the sample. Finally, we remove sites that do
not produce gas or oil and choose the first entry for sites with entries that only differ by their
latitude/longitude values.
The next step is to standardize key information in the well and wellsite data tables. Wells
identified in the raw data as "active" or "producing" are grouped together. Likewise, wells
identified as producing gas (wet, dry, or unknown) or coal bed methane are designated as gas
wells, while wells identified as producing oil (light, heavy, or unknown) are designated as oil
wells. Finally, wells are designated either as low production, non-low production, or unknown.
For well sites, data on equipment counts are standardized such that counts are either left blank if
valid information has not been provided, or equal to an integer (for separators, dehydrators, and
compressors) or real (for tanks) value. Sites are further designated as having full equipment
inventories if valid entries were provided for all equipment columns and partial equipment
71
For non-unique wells with the same well type and submission time, the only difference is found in the
"pt lproductionsiteid" column, a distinction that we assume is meaningless for this exercise.
2-66
-------
inventories if at least one equipment column has a valid entry and the remaining columns were
left blank, in which case blank entries were designated as zeroes.
After data standardization, the well and wellsite data is merged to create as single dataset
with information on equipment and production. The well-level data is aggregated to create site-
level estimates of the number of wells for each combination of oil and gas and low and non-low
production level; this step removes any well entries for which the production type and level is
not known. The aggregated data is then merged with the wellsite data based on the native
"operatorname" and "well site id name" columns. To facilitate use with the well site activity
data used for the impacts analysis, sites are then characterized as single well or multi-well sites,
oil or gas sites (based on whether there are more oil or gas wells at the site, with ties designated
as oil sites), and low or non-low production sites (only sites with exclusively low production
wells were designated as low production sites). Then well sites are assigned to one of the 11
equipment categories and one of the two tank categories described in footnote .
Finally, the dataset is aggregated to calculate the proportion of sites and average
equipment counts per well in each equipment/tank category. For the aggregate calculations, we
include sites with full and partial equipment inventories. The results for well site proportions,
stratified by production type and level and well count bin, are presented in Table 2-16. A
significant portion of sites, particularly single wellhead oil sites, do not have any major
equipment. Larger sites, both in terms of production levels and well counts, tend to have more
equipment and tanks for both site types.
2-67
-------
Table 2-16 Well Site Equipment/Tank Category Proportions Estimated From the 2016
ICR
Gas Oil
Non-low Non-low
Low Production Production Low Production Production
Equipment/Tank Category Single Multi Single Multi Single Multi Single Multi
Major Equipment Count
Zero
42%
19%
35%
2%
69%
33%
49%
10%
One
51%
33%
51%
10%
25%
46%
26%
8%
Two or More
7%
49%
14%
88%
6%
21%
24%
82%
auks Present
No
61%
12%
46%
14%
61%
9%
51%
11%
Yes
39%
88%
54%
86%
39%
91%
49%
89%
The results for well site equipment averages, stratified by production level, well count
bin, and equipment category, are presented in Table 2-17. Storage vessel averages are
conditional on sites having at least one tank. Typically, non-low production sites tend to have
more equipment and tanks than low production sites, particularly when it comes to separators,
though the relationships is not unambiguous across all well type, well count bin, and equipment
category permutations.
2-68
-------
Table 2-17
Well Site
Per-Site Average Equipment/Tank Counts Estimated From the 2016 ICR
Equipment Count Per Site
Site Type
Well
Production Count
Level Bin
Major
Equipment
Count
Separators Compressors Dehydrators Tanks
Single
Low
Multi
Gas
Single
Non-Low
Multi
0
1
2+
0
1
2+
0
1
2+
0
1
2+
0.98
1.66
0.96
3.51
0.95
1.82
0.95
4.09
0.01
0.54
0.02
0.70
0.02
0.50
0.05
0.30
0.01
0.17
0.02
0.20
0.03
0.19
0.00
0.09
1.56
2.55
1.80
3.62
Single
Low
Multi
Oil
Single
Non-Low
Multi
0
1
2+
0
1
2+
0
1
2+
0
1
2+
0.98
1.93
0.99
2.81
0.98
2.22
0.95
4.39
0.02
0.36
0.01
0.27
0.01
0.51
0.03
0.48
0.00
0.05
0.00
0.03
0.00
0.06
0.02
0.07
2.39
2.80
4.36
8.27
2-69
-------
2.8 References
Allen, D. T., Pacsi, A. P., Sullivan, D. W., Zavala-Araiza, D., Harrison, M., Keen, K., . . .
Seinfeld, J. H. (2015). Methane Emissions from Process Equipment at Natural Gas
Production Sites in the United States: Pneumatic Controllers. Environmental Science &
Technology, 49(1), 633-640. https://doi.org/10.1021/es5040156
Allen, D. T., Torres, V. M., Thomas, J., Sullivan, D. W., Harrison, M., Hendler, A., . . . Seinfeld,
J. H. (2013). Measurements of methane emissions at natural gas production sites in the
United States. Proceedings of the National Academy of Sciences, 110(44), 17768-17773.
https://doi.Org/doi:10.1073/pnas.1304880110
Budzik, P. (2002). U.S. Natural Gas Markets: Relationship Between Henry Hub Spot Prices and
U.S. Wellhead Prices. Energy Information Administration Report.
Cusworth, D. H., Duren, R. M., Thorpe, A. K., Olson-Duvall, W., Heckler, J., Chapman, J. W., .
. . Miller, C. E. (2021). Intermittency of Large Methane Emitters in the Permian Basin.
Environmental Science & Technology Letters, 8(7), 567-573.
https://doi.org/10.1021/acs.estlett.lc00173
U.S. EPA. (2021). Oil and Natural Gas Sector: Emission Standards for New, Reconstructed, and
Modified Sources and Emissions Guidelines for Existing Sources: Oil and Natural Gas
Sector Climate Review Background Technical Support Document for the Proposed New
Source Performance Standards (NSPS) and Emissions Guidelines (EG). Research Triangle
Park, NC Retrieved from https://www.regulations.gov/docket/EPA-HQ-OAR-2021-0317-
0166
U.S. EPA. (2022). Oil and Natural Gas Sector: Emission Standards for New, Reconstructed, and
Modified Sources and Emissions Guidelines for Existing Sources: Oil and Natural Gas
Sector Climate Review Supplemental Background Technical Support Document for the
Proposed New Source Performance Standards (NSPS) and Emissions Guidelines (EG).
https://www.regulations.gov/document/EPA-HQ-OAR-2021-0317-1578
U.S. EPA. (2023). Oil and Natural Gas Sector: Emission Standards for New, Reconstructed, and
Modified Sources and Emissions Guidelines for Existing Sources: Oil and Natural Gas
Sector Climate Review Background Technical Support Document for the Final New Source
Performance Standards (NSPS) and Emissions Guidelines (EG).
https://www.regulations.gov/document/EPA-HQ-OAR-2021 -0317
Zavala-Araiza, D., Alvarez, R. A., Lyon, D. R., Allen, D. T., Marchese, A. J., Zimmerle, D. J., &
Hamburg, S. P. (2017). Super-emitters in natural gas infrastructure are caused by abnormal
process conditions. Nature Communications, 8(1), 14012.
https://doi.org/10.1038/ncommsl4012
Zimmerle, D., Vaughn, T., Luck, B., Lauderdale, T., Keen, K., Harrison, M., . . . Allen, D.
(2019). Methane emissions from gathering and boosting compressor stations in the U.S.
Supporting volume 3: Emission factors, station estimates, and national emissions.
https://mountainscholar.org/items/71405f94-dc24-4e47-89d5-47db6d64e756
2-70
-------
3 BENEFITS
The final NSPS 0000b and EG 0000c are projected to reduce methane, VOC, and
HAP emissions. 2 The total emissions reductions over the 2024 to 2038 period are estimated to
be about 58 million short tons of methane, 16 million tons of VOC, and 590 thousand tons of
HAP. The decrease in methane emissions in C02-equivalent (CO2 Eq.) terms is estimated to be
about 1.5 billion metric tons using a global warming potential of 28.
We monetize the impacts of methane reductions in this RIA. We estimate the climate
benefits under the final rule using updated estimates of the social cost of methane (SC-CH4), as
presented in Section 3.2. Additionally, we monetize ozone-related health impacts of VOC
reductions as presented in Section 3.3.
In addition to presenting monetized estimates of impacts from methane and VOC
reductions, we also provide a qualitative discussion of potential climate, human health, and
welfare impacts that we are unable to quantify and monetize in Sections 0 through 3.8. Table 3-1
summarizes the quantified and unquantified benefits in this analysis.
72
Some control techniques projected to be used under the final rule, such as routing emission to combustion devices,
are also anticipated to have minor disbenefits resulting from secondary emissions of carbon dioxide (CO2), nitrogen
oxides (NOx), PM, carbon monoxide (CO), and total hydrocarbons (THC).
3-1
-------
Table 3-1 Climate and Human Health Effects of the Projected Emissions Reductions
under the Final Rule
Category
Effect
Effect
Quantified
Effect
Monetized
More
Information
Environment
Climate impacts from methane (CH4)
a
Section 3.2
Climate effects
Other climate impacts (e.g., ozone, black
carbon, aerosols, other impacts)
—
—
IPCC, Ozone
ISA, PM ISA
Human Health
Mortality from exposure
Premature respiratory mortality from
short-term exposure (0-99)
~
~
Ozone ISA
to ozoneb
Premature respiratory mortality from
long-term exposure (age 30-99)
~
~
Ozone ISA
Hospital admissions—respiratory (ages
65-99)
~
~
Ozone ISA
Emergency department visits—respiratory
(ages 0-99)
~
~
Ozone ISA
Asthma onset (0-17)
~
~
Ozone ISA
Asthma symptoms/exacerbation
(asthmatics age 5-17)
V
V
Ozone ISA
Allergic rhinitis (hay fever) symptoms
(ages 3-17)
V
V
Ozone ISA
Nonfatal morbidity from
Minor restricted-activity days (age 18-65)
~
~
Ozone ISA
exposure to ozone"
School absence days (age 5-17)
~
~
Ozone ISA
Decreased outdoor worker productivity
(age 18-65)
—
—
Ozone ISA0
Metabolic effects (e.g., diabetes)
—
—
Ozone ISA0
Other respiratory effects (e.g., premature
aging of lungs)
—
—
Ozone ISA0
Cardiovascular and nervous system
effects
—
—
Ozone ISA0
Reproductive and developmental effects
—
—
Ozone ISA0
Premature mortality
Adult premature mortality from long-term
exposure (age 65-99 or age 30-99)
—
—
PM ISA
from exposure to PM2 5
Infant mortality (age <1)
—
—
PM ISA
Heart attacks (age >18)
—
—
PM ISA
Hospital admissions—cardiovascular
(ages 65-99)
—
—
PM ISA
Emergency department visits—
cardiovascular (age 0-99)
—
—
PM ISA
Hospital admissions—respiratory (ages 0-
18 and 65-99)
—
—
PM ISA
Nonfatal morbidity from
exposure to PM2 5
Emergency room visits—respiratory (all
ages)
—
—
PM ISA
Cardiac arrest (ages 0-99; excludes initial
hospital and/or emergency department
visits)
—
—
PM ISA
Stroke (ages 65-99)
—
—
PM ISA
Asthma onset (ages 0-17)
—
—
PM ISA
Asthma symptoms/exacerbation (6-17)
—
—
PM ISA
Lung cancer (ages 30-99)
—
—
PM ISA
Allergic rhinitis (hay fever) symptoms
(ages 3-17)
—
—
PM ISA
3-2
-------
Effect
Effect
More
Category
Effect
Quantified
Monetized
Information
Lost work days (age 18-65)
—
—
PM ISA
Minor restricted-activity days (age 18-65)
—
—
PM ISA
Hospital admissions—Alzheimer's
disease (ages 65-99)
—
—
PM ISA
Hospital admissions—Parkinson's disease
(ages 65-99)
—
—
PM ISA
Other cardiovascular effects (e.g., other
ages)
—
—
PM ISA0
Other respiratory effects (e.g., pulmonary
function, non-asthma ER visits, non-
bronchitis chronic diseases, other ages and
—
—
PM ISA0
populations)
Other nervous system effects (e.g., autism,
cognitive decline, dementia)
—
—
PM ISA0
Metabolic effects (e.g., diabetes)
—
—
PM ISA0
Reproductive and developmental effects
(e.g., low birth weight, pre-term births,
etc.)
—
—
PM ISA0
Cancer, mutagenicity, and genotoxicity
effects
—
—
PM ISA0
Incidence of morbidity
from exposure to HAP
Effects associated with exposure to
hazardous air pollutants such as benzene
—
—
ATSDR, IRIS'1-6
a The climate and related impacts of CH4 emissions changes, such as sea level rise, are estimated within each
integrated assessment model as part of the calculation of the SC-CH4. The resulting monetized damages, which are
relevant for conducting the benefit-cost analysis, are used in this RIA to estimate the welfare effects of quantified
changes in methane emissions.
b Ozone benefits only quantified for ozone changes resulting from VOC reductions. Benefits associated with ozone
changes resulting from CH4 reductions are discussed qualitatively in this RIA.
°Not quantified due to data availability limitations and/or because current evidence is only suggestive of causality.
d We assess these benefits qualitatively because we do not have sufficient confidence in available data or methods.
e We assess these benefits qualitatively due to data limitations for this analysis, but we have quantified them in other
analyses.
3.1 Emissions Reductions
Oil and natural gas operations in the U.S. include a variety of emission sources for
methane, VOC, and HAP, including wells, well sites, processing plants, compressor stations,
storage equipment, and natural gas transmission and distribution lines. These emission points are
located throughout much of the country, though many of these emissions sources are
concentrated in particular geographic regions. For example, wells and processing plants are
largely concentrated in the South Central, Midwest, and Southern California regions of the U.S.,
whereas natural gas compressor stations are located all over the country. Distribution lines to
customers are frequently located within areas of high population density.
Table 3-2 shows the emissions reductions projected under the final NSPS OOOOb and
EG 0000c over the 2024 to 2038 period. We present methane emissions in both short tons and
3-3
-------
C02 Eq. using a global warming potential of 28. The impacts of these pollutants accrue at
different spatial scales. HAP emissions increase exposure to carcinogens and other toxic
pollutants primarily near the emission source. VOC emissions are precursors to secondary
formation of PM2.5 and ozone on a broader regional scale. Climate effects associated with long-
lived greenhouse gases like methane generally do not depend on the location of the emission of
the gas and have global impacts. Methane is also a precursor to global background
concentrations of ozone.
Table 3-2 Projected Annual Reductions of Methane, VOC, and HAP Emission
Reductions under the Final NSPS OOOOb and EG OOOOc, 2024-2038
Methane VOC HAP Methane
Year
(short tons)
(short tons)
(short tons)
(metric tons CO2 Eq.)
2024
250,000
77,000
2,900
6,400,000
2025
490,000
150,000
5,600
12,000,000
2026
700,000
210,000
8,000
18,000,000
2027
900,000
270,000
10,000
23,000,000
2028
4,900,000
1,300,000
48,000
120,000,000
2029
4,900,000
1,300,000
49,000
130,000,000
2030
5,000,000
1,300,000
49,000
130,000,000
2031
5,000,000
1,300,000
50,000
130,000,000
2032
5,100,000
1,300,000
50,000
130,000,000
2033
5,100,000
1,400,000
51,000
130,000,000
2034
5,100,000
1,400,000
51,000
130,000,000
2035
5,100,000
1,400,000
52,000
130,000,000
2036
5,200,000
1,400,000
53,000
130,000,000
2037
5,200,000
1,400,000
54,000
130,000,000
2038
5,200,000
1,500,000
55,000
130,000,000
Total
58,000,000
16,000,000
590,000
1,500,000,000
Note: Values rounded to two significant figures. Totals may not appear to add correctly due to rounding.
3.2 Methane Climate Effects and Valuation
Methane is the principal component of natural gas. Methane is also a potent greenhouse
gas (GHG) that, once emitted into the atmosphere, absorbs terrestrial infrared radiation, which in
turn contributes to increased global warming and continuing climate change. Methane reacts in
the atmosphere to form ozone, which also impacts global temperatures. Methane, in addition to
other GHG emissions, contributes to warming of the atmosphere, which over time leads to
increased air and ocean temperatures, changes in precipitation patterns, melting and thawing of
3-4
-------
global glaciers and ice sheets, increasingly severe weather events, such as hurricanes of greater
intensity, and sea level rise, among other impacts.
According to the Intergovernmental Panel on Climate Change (IPCC) Sixth Assessment
Report (IPCC, 2021b), radiative forcing due to methane relative to the year 1750 was 0.54 W/m2
in 2019, which is about 16 percent of all global forcing due to increases in anthropogenic GHG
concentrations, and which makes methane the second leading long-lived climate forcer after
CO2. 3 After accounting for changes in other greenhouse substances such as ozone and
stratospheric water vapor due to chemical reactions of methane in the atmosphere, historical
methane emissions account for about 0.5 degrees of warming today, or about one third of the
total warming resulting from historical emissions of well-mixed GHGs.
The oil and natural gas sector emits significant quantities of methane. The U.S. Inventory
of Greenhouse Gas Emissions and Sinks: 1990-2019 (published 2021) estimates 2019 methane
emissions from Petroleum and Natural Gas Systems (not including petroleum refineries,
petroleum transportation, and natural gas distribution) to be 187 million metric tons CO2 Eq.
(U.S. EPA, 2021a). In 2019, total methane emissions from the oil and natural gas industry
represented 27 percent of the total methane emissions from all sources and account for about 3
percent of all CO2 Eq. emissions in the U.S., with the combined petroleum and natural gas
systems being the largest contributor to U.S. anthropogenic methane emissions.
We estimate the climate benefits of CH4 emissions reductions expected from the final
rule using estimates of the social cost of methane (SC-CH4) 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). The EPA presented these estimates in a sensitivity analysis in the
December 2022 RIA, solicited public comment on the methodology and use of these estimates,
and has conducted an external peer review of these estimates, as described further below.
73
Increased concentrations of methane and other well mixed greenhouse gases in the atmosphere absorb thermal
infrared emission energy, reducing the rate at which the Earth can cool through radiating heat to space. Radiative
forcing, measured as watts per square meter (W/m2), is a measure of the climate impact of greenhouse gases and
other human activities.
3-5
-------
The SC-CH4 is the monetary value of the net harm to society from emitting a metric ton
of CH4 into the atmosphere in a given year, or the benefit of avoiding that increase. In principle,
SC-CH4 is a comprehensive metric that includes the value of all future climate change impacts
(both negative and positive), including changes in net agricultural productivity, human health
effects, property damage from increased flood risk, changes in the frequency and severity of
natural disasters, disruption of energy systems, risk of conflict, environmental migration, and the
value of ecosystem services. The SC-CH4, therefore, reflects the societal value of reducing CH4
emissions by one metric ton and is the theoretically appropriate value to use in conducting
benefit-cost analyses of policies that affect CH4 emissions. In practice, data and modeling
limitations restrain the ability of SC-CH4 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, the EPA has used estimates of the social cost of various greenhouse gases
(i.e., social cost of carbon (SC-CO2), social cost of methane (SC-CH4), and social cost of nitrous
oxide (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 the EPA from 2009 to 2016,
and since 2021 — including in the November 2021 RIA and December 2022 RIA for this
rulemaking — have been consistent with those developed and recommended by the Interagency
Working Group on the SC-GHG (IWG); 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 (National Academies, 2017).
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.
3-6
-------
The EPA is a member of the IWG and is participating in the IWG's work under E.O.
13990. While that process continues, as noted in previous EPA RIAs, the 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 going forward. 4 In the December 2022 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 (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.
The EPA solicited public comment on the sensitivity analysis and the accompanying draft
technical report, EPA Report on the Social Cost of Greenhouse Gases: Estimates Incorporating
Recent Scientific Advances, which explains the methodology underlying the new set of estimates,
in the December 2022 Supplemental Proposal.75 Please see the response to comments document
for the rulemaking for summaries and responses to public comments. The response to comments
document can be found in the docket for this action.
To ensure that the methodological updates adopted in the technical report are consistent
with economic theory and reflect the latest science, the 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 towards 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
74
EPA strives to base its analyses on the best available science and economics, consistent with its responsibilities,
for example, under the Information Quality Act.
75
See https://www.epa.gov/environmental-economics/scghg for a copy of the final report and other related
materials.
3-7
-------
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 the EPA's response to each recommendation is available on EPA's website.76
The remainder of this section provides an overview of the methodological updates
incorporated into the SC-GHG estimates used in this RIA. A more detailed explanation of each
input and the modeling process is provided in the technical report, Supplementary Material for
the RIA: EPA Report on the Social Cost of Greenhouse Gases: Estimates Incorporating Recent
Scientific Advances (U.S. EPA, 2023f), included in the docket.
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
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 t, 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 t 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 the 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 & 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
6 https://www.epa.gov/environmental-economics/scghg-tsd-peer-review
3-8
-------
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. Instead, EPA uses a modular approach to
estimating the SC-GHG, consistent with the National Academies' (2017) 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
(hereafter 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' 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 (2023f), 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; Smith et al., 2018), a widely used Earth system model which
captures the relationships between GHG emissions, atmospheric GHG concentrations, and global
3-9
-------
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, offers a code base
that is fully transparent and available online, and 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 (2023f) for more
details.
The socioeconomic projections and outputs of the climate module are inputs into the
damage module to estimate monetized future damages from climate change.77 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 & Tol,
77
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. (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 EPA Report on the Social Cost of Greenhouse
Gases: Estimates Incorporating Recent Scientific Advances. Washington, DC: U.S. EPA for more details.
3-10
-------
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 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 focused on
understanding how these physical changes translate into economic impacts is still developing,
and has received less public resources, relative to the research focused on modeling and
improving our understanding of climate system dynamics and the physical impacts from climate
change (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 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, the EPA
uses three separate damage functions to form the damage module. They are:
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 the 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 allows 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
3-11
-------
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]" (National Academies 2017, p.
199), 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
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.78 See U.S. EPA (2023f) 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
8 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.
3-12
-------
partial estimates of future climate damages resulting from incremental changes in CO2, CH4, and
N20.79
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 techniques 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.80 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
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 & 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. They address 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 their analysis shows that the estimates are somewhat sensitive to defensible alternative
modeling choices. As discussed in detail in U.S. EPA (2023f), 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.
79
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).
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.
3-13
-------
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), the 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 (2003) 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).81 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 calculate the SC-GHG. Thus, EPA concludes that the use of the social rate of
return on capital (7 percent under OMB Circular A-4 (2003)), 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.82
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 (IWG, 2021), and the National Academies (2017) recommendations for updating the
discounting module, the SC-GHG estimates rely on discount rates that reflect more recent data
81 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" (OMB 2023).
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 (2023) (OMB 2023).
3-14
-------
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 EPA RIAs to date. Specifically, it
provides internal consistency within the modeling and a more complete accounting of
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 (2023f) 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 in past estimates by the EPA. The updates
incorporate a quantitative consideration of uncertainty into all modules and use a Monte Carlo
3-15
-------
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, the 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 greenhouse gases (SC-CO2, SC-CH4, and SC-N2O) for emissions years
2020 through 2080 are provided in U.S. EPA (2023f).
summarizes the resulting averaged certainty-equivalent SC-CH4 estimates under each
near-term discount rate that are used to estimate the climate benefits of the CH4 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 (2023f). The SC-CH4 increases over 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 2025 — 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.
3-16
-------
Table 3-3
Estimates of the Social Cost of CH4, 2024-2038 (in 2019$ per metric ton CH4)
Year
Near-Term Ramsey Discount Rate
1.5%
2.0%
2.5%
2024
$2,600
$1,900
$1,500
2025
$2,700
$2,000
$1,600
2026
$2,800
$2,100
$1,600
2027
$2,900
$2,200
$1,700
2028
$3,000
$2,200
$1,800
2029
$3,000
$2,300
$1,800
2030
$3,100
$2,400
$1,900
2031
$3,200
$2,500
$2,000
2032
$3,300
$2,500
$2,100
2033
$3,400
$2,600
$2,100
2034
$3,500
$2,700
$2,200
2035
$3,600
$2,800
$2,300
2036
$3,700
$2,900
$2,400
2037
$3,800
$3,000
$2,400
2038
$3,900
$3,100
$2,500
Source: U.S. EPA (2023f).
Note: These SC-CH4 values are identical to those reported in the technical report U.S. EPA (2023f) 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 CH4 and vary depending
on the year of CH4 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 (2023f) 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-
GHG estimates, including the SC-CH4 estimates presented in Table 3-3, 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 GHG emissions. The SC-CH4 estimates do not account for the
direct health and welfare impacts associated with tropospheric ozone produced by methane. As
discussed further in U.S. EPA (2023f) and in Section 3.6 of this RIA, recent studies have found
3-17
-------
the global ozone-related respiratory mortality benefits of CH4 emissions reductions, which are
not included in the SC-CH4 values presented in Table 3-3, to be, in 2019 dollars, approximately
$2,400 per metric ton of methane emissions in 2030 (McDuffie et al., 2023). In addition, the SC-
CH4 estimates do not reflect that methane emissions lead to a reduction in atmospheric oxidants,
like hydroxyl radicals, nor do they account for impacts associated with CO2 produced from
methane oxidizing in the atmosphere. 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 3-4 presents the undiscounted annual monetized climate benefits under the final
NSPS OOOOb and EG OOOOc. Projected methane emissions reductions each year are
multiplied by the SC-CH4 estimate for that year.83 Table 3-5 shows the annual climate benefits
discounted back to 2021 and the PV and the EAV for the 2024-2038 period under each discount
rate. In this analysis, to calculate the present and annualized values of climate benefits, EPA uses
the same discount rate as the near-term target Ramsey rate used to discount the climate benefits
from future CH4 reductions. That is, future climate benefits estimated with the SC-CH4 at the
near-term 2 percent Ramsey rate are discounted to the base year of the analysis using the same 2
percent rate.84
83 The EPA has also applied its updated estimates of the social cost of carbon dioxide (SC-CO2) in an illustrative
analysis of potential climate disbenefits from secondary CO2 emissions associated with particular control techniques
to meet the storage vessel-related standards. Given that the estimated climate disbenefits from the CO2 impacts
would at most offset only about 1 percent of the methane benefits, the EPA finds that the summary values shown in
Tables 3-4 and 3-5 are a reasonable estimate of the net monetized climate effects of the rule. See Section 3.9 for
further discussion.
84
As discussed in U. S. 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 EPA Report on the Social Cost of Greenhouse
Gases: Estimates Incorporating Recent Scientific Advances. Washington, DC: U.S. EPA, 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). Ibid,
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.
3-18
-------
Table 3-4 Undiscounted Monetized Climate Benefits under the Final NSPS OOOOb
and EG OOOOc, 2024-2038 (millions, 2019$)
Undiscounted3
Year 1.5% 2.0% 2.5%
2024
$600
$440
$340
2025
$1,200
$890
$700
2026
$1,800
$1,300
$1,000
2027
$2,300
$1,700
$1,400
2028
$13,000
$9,900
$7,900
2029
$14,000
$10,000
$8,200
2030
$14,000
$11,000
$8,600
2031
$15,000
$11,000
$9,000
2032
$15,000
$12,000
$9,400
2033
$16,000
$12,000
$9,900
2034
$16,000
$13,000
$10,000
2035
$17,000
$13,000
$11,000
2036
$18,000
$14,000
$11,000
2037
$18,000
$14,000
$12,000
2038
$19,000
$15,000
$12,000
a Climate benefits are based on changes (reductions) in CH4 emissions and are calculated using updated estimates of
the SC-CH4 from U.S. EPA (2023f).
3-19
-------
Table 3-5 Discounted Monetized Climate Benefits under the Final NSPS OOOOb and
EG OOOOc, 2024-2038 (millions, 2019$)
Discounted back to 2021a
Year
1.5%
2.0%
2.5%
2024
$570
$410
$320
2025
$1,100
$820
$630
2026
$1,600
$1,200
$920
2027
$2,100
$1,600
$1,200
2028
$12,000
$8,600
$6,600
2029
$12,000
$8,800
$6,700
2030
$12,000
$9,000
$6,900
2031
$13,000
$9,200
$7,000
2032
$13,000
$9,400
$7,200
2033
$13,000
$9,600
$7,300
2034
$13,000
$9,700
$7,400
2035
$14,000
$9,900
$7,500
2036
$14,000
$10,000
$7,700
2037
$14,000
$10,000
$7,800
2038
$15,000
$10,000
$7,900
PV
$150,000
$110,000
$83,000
EAV
$11,000
$8,500
$6,700
a Climate benefits are based on changes (reductions) in CH4 emissions and are calculated using updated estimates of
the SC-CH4 from U.S. EPA (2023f).
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 the EPA
centers attention on a global measure of climate benefits from CH4 reductions. Consistent with
all IWG recommended SC-GHG estimates to date, the SC-CH4 values presented in Table 3-3
provide a global measure of monetized damages from CH4 emissions, and Tables 3-4 and 3-5
present the monetized global climate benefits of the CH4 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 (2003) that states
3-20
-------
when a regulation is likely to have international effects, "these effects should be reported".85 EPA
also notes that EPA's cost estimates in RIAs, including the cost estimates contained in this RIA,
regularly do not differentiate between distinguish and segregate the share of compliance costs
expected to accrue to U.S. firms versus foreign interests, such as to foreign investors in regulated
entities.86 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), and in detail in EPA (2023 e) and in Appendix A of the Response to
Comments document for this action — why the 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 greenhouse gases 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
85 While OMB Circular A-4 (2003) 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). Circular A-4 (2023) 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:
• 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).
86 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, the 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.
3-21
-------
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 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 Supplemental Proposal RIA.87 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 (2023f) for more
discussion.
For all of these reasons, the EPA believes that a global metric is appropriate for assessing
the climate benefits of avoided methane 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 occurring 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
87 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 the EPA's updated estimates presented in the December
2022 Supplemental Proposal RIA. The Canadian interim guidance will be used across all 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.
3-22
-------
destabilization and global migration that can lead to adverse impacts on U.S. national security,
public health, and 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 greenhouse
gases.
In the case of this global pollutant, 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 the CH4 emission reductions expected from this final
rule. The EPA disagrees with commenters who suggest that the EPA can or should use a metric
focused on benefits resulting solely from changes in climate impacts occurring 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 comprehensive 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 a subset of 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 (respectively) projected to physically occur within the U.S.,
subject to caveats. As discussed at length in U.S. EPA (2023f) these damage modules are only a
partial accounting and do not capture many significant 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, this modeling only
cover a subset of potential climate change impacts. Furthermore, 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 — for example through the movement of
refugees.
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 the EPA to facilitate the characterization of net
3-23
-------
annual climate change impacts in numerous impact categories within the contiguous U.S. and
monetize the associated distribution of modeled damages (Sarofim et al., 2021; U.S. EPA,
2021c).88 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 (2023f), 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-CH4 of $590/mtCH4 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-CH4 of $280/mtCH4 and
$75/mtCH4, respectively, for 2030 emissions. 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.89 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). As one example of new research on non-climate
88 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). Technical Documentation on The Framework for Evaluating Damages and Impacts (FrEDI).
Retrieved February 16, 2023 from
https://cfpub.epa.gov/si/si_public_record_report. cfm?dirEntryId=351316&Lab=OAP&simplesearch=0&showcriteri
a=2&sortby=pubDate&searchall=fredi&timstype=&datebeginpublishedpresented=02/14/2021.
89
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. (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 EPA Report on the Social Cost of
Greenhouse Gases: Estimates Incorporating Recent Scientific Advances. Washington, DC: U.S. EPA, 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) and
focus solely on CO2. 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 ibid, would require new analysis. Finally, because total-economy empirical studies estimate market impacts,
they do not include non-market impacts of climate change (e.g., mortality impacts) and therefore are also only a
partial estimate. The EPA will continue to review developments in the literature and explore ways to better inform
the public of the full range of GHG impacts.
3-24
-------
mediated effects of methane emissions, McDuffie et al. (2023) estimate the monetized increase
in respiratory-related human mortality risk from the ozone produced from a marginal pulse of
methane emissions. Using the socioeconomics from the RFF-SPs and the 2 percent near-term
Ramsey discounting approach, this additional risk to U.S. populations is on the order of
approximately $320/mtCH4 for 2030 emissions (U.S. EPA 2023e).
Taken together, applying the U.S.-specific partial SC-CH4 estimates derived from the
evidence described above to the CH4 emissions reduction expected under the final rule would
yield substantial benefits. For example, the present value of the climate benefits of the final rule
as measured by FrEDI using additional U.S.-specific data and research on climate change
impacts in CONUS are estimated to be $27 billion (under a 2 percent near-term Ramsey discount
rate.90 However, even with these additional impact categories, the numerous explicitly omitted
damage categories and other modeling limitations discussed above and throughout U.S. EPA
(2023f) make it likely that these estimates underestimate the benefits to U.S. citizens and
residents of the CH4 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 CH4
reductions. The 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.
3.3 Ozone-Related Health Impacts Due to VOC Emissions Changes
Human exposure to ambient ozone concentrations is associated with adverse health
effects, including premature respiratory mortality and cases of respiratory morbidity (U.S. EPA,
2020c). Researchers have associated ozone exposure with adverse health effects in numerous
toxicological, clinical, and epidemiological studies ozone (U.S. EPA, 2020c). When adequate
data and resources are available, the EPA has generally quantified several health effects
associated with exposure to ozone (U.S. EPA, 201 lc, 2015, 2023c). These health effects include
90
DCIM 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.
3-25
-------
respiratory morbidity, such as asthma attacks, hospital and emergency department visits, lost
school days, and premature respiratory mortality. The scientific literature is also suggestive that
exposure to ozone is associated with chronic respiratory damage and premature aging of the
lungs.
This final rulemaking is projected to reduce volatile organic compounds (VOC)
emissions, which are a precursor to ozone. Ozone is not generally emitted directly into the
atmosphere but is created when its two primary precursors, VOC and oxides of nitrogen (NOx),
react in the atmosphere in the presence of sunlight. In urban areas, compounds representing all
classes of VOC can be important for ozone formation, but biogenic VOC emitted from
vegetation tend to be more important compounds in some non-urban vegetated areas (U.S. EPA,
2013). Recent observational and modeling studies have found that VOC emissions from oil and
natural gas operations can impact ozone levels (Helmig, 2020; Kemball-Cook et al., 2010;
Lindaas et al., 2019; Lyu et al., 2021; McDuffie et al., 2016; Pozzer et al., 2020; Reddy, 2023;
Tzompa-Sosa & Fischer, 2021). As shown later in this section, VOC emissions reductions from
this rulemaking are expected decrease ozone formation, human exposure to ozone, and the
incidence of ozone-related health effects.
This section describes the methods used to estimate the benefits to human health of
reducing concentrations of ozone from this rule. This analysis uses methodology for determining
air quality changes that has been used in the RIAs from multiple previous proposed and final
rules (U.S. EPA, 2019b, 2020a, 2020b, 2021b, 2022b, 2023d, 2023e). The health benefits
analysis uses the spatial fields of air quality across the U.S. described in Section 3.3.1 in
BenMAP-CE to quantify the benefits under each regulatory alternative compared to the baseline
and for four analytical years: 2024, 2027, 2028 and 2038. Health benefit analyses were also run
for each year between 2024 and 2038, using the model surfaces for 2024, 2027, 2028 and 2038
as described in Section 3.3.1, but accounting for the change in population size in each year,
income growth and baseline mortality incidence rates at five-year increments. Specifically, the
analysis quantifies health benefits resulting from changes in ozone concentrations in 2024, 2027,
2028 and 2038 for each of the scenarios (i.e., finalized rule, less stringent scenario, and more
stringent scenario). The methods for quantifying the number and value of air pollution-
attributable premature deaths and illnesses are described in the Technical Support Document
3-26
-------
(TSD) titled EstimatingPM2.5- and Ozone-Attributable Health Benefits (U.S. EPA, 2023b) and
further referred to as the Health Benefits TSD in this RIA.
Estimating the health benefits of reductions in ozone exposure begins with estimating the
change in exposure for each individual and then estimating the change in each individual's risks
for health outcomes affected by exposure. The benefit of the reduction in 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.
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 a baseline, the finalized rule, a less stringent scenario,
and a more stringent scenario (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 rules would affect the distribution of ozone concentrations in
much of the U.S. This RIA estimates avoided ozone-related health impacts that are distinct from
those reported in the RIAs for the ozone NAAQS ozone (U.S. EPA, 2015). The ozone NAAQS
RIAs illustrates, 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 summer season average ozone concentrations
for each of the years 2024, 2027, 2028 and 2038.
3-27
-------
3.3.1 Developing Air Quality Surfaces of Ozone Impacts from VOC Emissions Changes
As described above, the final rules influence the level of VOCs, a precursor to ground-
level ozone formation, emitted in the atmosphere. The methods for determining VOC emissions
associated with the baseline, final policy, and more and less stringent regulatory alternatives are
described in Section 2.2. A map of state-level VOC emissions reductions associated with the
final rules used in the creation of the ozone surfaces is provided in Figure 3-1. 1
M/ikd iv
0 50000 100000 150000 200000
VOC emis reductions (tpy)
Figure 3-1 Map of State-level VOC Emissions Reductions (tpy) from the Baseline to the
Final Rule Scenario in 2024, 2027, 2028 and 2038
Note: the projected emissions reductions in Texas for 2028 and in Oklahoma and Texas in 2038 are above the top of
the color scale at approximately 572,000 tpy, 202,000 tpy and 684,000 tpy respectively.
The EPA used air quality modeling to estimate changes in summertime ozone
concentrations that may result from the regulatory alternatives relative to the baseline in four
91
The VOC emissions levels used to quantify ozone benefits were calculated based on the best understanding of the
policy at the time of the analysis. While some updates to the policy were made after the ozone benefits analysis was
completed, the impact on VOC emissions was less than 0.05% for the final rale scenario and therefore would not
meaningfully impact the quantified ozone benefits.
3-28
-------
92
years of analysis. Gridded spatial fields of April through September seasonal average 8-hour
daily maximum (MDA8) ozone (AS-M03) concentrations representing the baseline and
regulatory alternatives were derived from CAMx source apportionment modeling in combination
with the VOC emissions described in Section 2.2 of this RIA and shown in Figure 3-1. These
ozone 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.93 Figure 3-2 shows
the geographical distribution of ozone changes in the final rules relative to the baseline in the
four years of analysis. The basic methodology for determining air quality changes is the same as
that used in the RIAs from multiple previous rulemakings (U.S. EPA, 2019b, 2020a, 2020b,
2021b, 2022b, 2023d, 2023e). U.S. EPA (2023) provides additional details on the air quality
modeling and the methodologies EPA used to develop gridded spatial fields of summertime
ozone concentrations for this analysis.
92
The air quality modeling was conducted using emissions projected to the year 2026. The ozone surfaces
reflecting the baseline and regulatory alternatives in 2024, 2027, 2028, and 2038 adjust ozone impacts from oil and
gas sources to reflect expected VOC emissions from those sources in each of those year. Emissions from all other
sources are held constant at 2026 levels.
93
Given the regional nature of ozone pollution, it is possible, as depicted in Figure 3-2, that areas outside the
contiguous U.S. may also experience reductions in ozone concentrations, and associated health and environmental
benefits, resulting from this rule, including in portions of Mexico and Canada. This RIA does not quantify those
additional health and environmental effects.
3-29
-------
Figure 3-2 Map of Modeled Changes in April to September M DAS Ozone
Concentrations Calculated as the Final Rule Scenario Minus the Baseline Scenario in 2024,
2027, 2028 and 2038
3.3.2 Selecting Air Pollution Health Endpoints to Quantify
As a first step in quantifying ozone-related human health impacts, the Agency consults
the Integrated Science Assessment for Ozone and Related Photochemical Oxidants (Ozone ISA)
(U.S. EPA, 2020c). The Ozone ISA synthesizes the toxicological, clinical, and epidemiological
Evidence to determine whether ozone is causally related to an array of adverse human health
outcomes associated with either acute (i.e., hours or days-long) or chronic (i.e., years-long)
exposure; for each outcome, the ISA reports this relationship to be causal, likely to be causal,
suggestive of a causal relationship, inadequate to infer a causal relationship or not likely to be a
causal relationship. Historically, the Agency estimates the incidence of air pollution effects for
those health endpoints that the ISA classified as either causal or likely-to-be-causal. The analysis
also accounts for recommendations from the Science Advisory Board (U.S. EPA Science
Advisory Board, 2019, 2020a). When updating each health endpoint EPA considered: (1) the
extent to which there 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
3-30
-------
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. The Health Benefits TSD fully describes the Agency's
approach for quantifying the number and value of estimated air pollution-related impacts. 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.94
In brief, the ISA for ozone found short-term (less than one month) exposures to ozone to
be causally related to respiratory effects, a "likely to be causal" relationship with metabolic
effects and a "suggestive of, but not sufficient to infer, a causal relationship" for central nervous
system effects, cardiovascular effects, and total mortality. The ISA reported that long-term
exposures (one month or longer) to ozone are "likely to be causal" for respiratory effects
including respiratory mortality, and a "suggestive of, but not sufficient to infer, a causal
relationship" for cardiovascular effects, reproductive effects, central nervous system effects,
metabolic effects, and total mortality. Table 3-1 reports the ozone-related human health impacts
effects we quantified and those we did not quantify in this RIA. The list of benefit categories not
quantified is not exhaustive. And, among the effects quantified, it might not have been possible
to quantify completely either the full range of human health impacts or economic values.
Sections 0 through report other non-quantified health and environmental benefits expected from
the emissions and effluent changes as a result of this rule, such as health effects associated with
PM2.5, NO2 and SO2, and any welfare effects such as acidification and nutrient enrichment.
Consistent with economic theory, the willingness-to-pay (WTP) for reductions in
exposure to environmental hazards will depend on the expected impact of those reductions on
human health and other outcomes. All else equal, WTP is expected to be higher when there is
stronger evidence of a causal relationship between exposure to the contaminant and changes in a
health outcome (McGartland et al., 2017). For example, in the case where there is no evidence of
a potential relationship the WTP would be expected to be zero and the effect should be excluded
from the analysis. Alternatively, when there is some evidence of a relationship between exposure
94
The analysis was completed using BenMAP-CE version 1.5.8, which is a variant of the current publicly available
version.
3-31
-------
and the health outcome, but that evidence is insufficient to definitively conclude that there is a
causal relationship, individuals may have a positive WTP for a reduction in exposure to that
hazard (Kivi & Shogren, 2010; U.S. EPA Science Advisory Board, 2020b). Lastly, the WTP for
reductions in exposure to pollutants with strong evidence of a relationship between exposure and
effect are likely positive and larger than for endpoints where evidence is weak, all else equal.
Unfortunately, the economic literature currently lacks a settled approach for accounting for how
WTP may vary with uncertainty about causal relationships.
Given this challenge, the Agency draws its assessment of the strength of evidence on the
relationship between exposure to ozone and potential health endpoints from the ISAs that are
developed for the NAAQS process as discussed above. The focus on categories identified as
having a "causal" or "likely to be causal" relationship with the pollutant of interest is to estimate
the pollutant-attributable human health benefits in which we are most confident.95 All else equal,
this approach may underestimate the benefits of ozone exposure reductions as individuals may
be WTP to avoid specific risks where the evidence is insufficient to conclude they are "likely to
be caus[ed]" by exposure to these pollutants.96 At the same time, WTP may be lower for those
health outcomes for which causality has not been definitively established. This approach treats
relationships with ISA causality determinations of "likely to be causal" as if they were known to
be causal, and therefore benefits could be overestimated. Table 3-1 reports the effects we
quantified and those we did not quantify in this RIA. The list of benefit categories not quantified
is not exhaustive and omits welfare effects such as acidification and nutrient enrichment.
95
This decision criterion for selecting health effects to quantify and monetize PM2 5 and ozone is only applicable to
estimating the benefits of exposure of these two pollutants. This is also the approach used for identifying the
unqualified benefit categories for criteria pollutants. This decision criterion may not be applicable or suitable for
quantifying and monetizing health and ecological effects of other pollutants. The approach used to determine
whether there is sufficient evidence of a relationship between an endpoint affected by non-criteria pollutants, and
consequently a positive WTP for reductions in those pollutants, for other unqualified benefits described in this
section can be found in the source documentation for each of these pollutants (see relevant sections below). The
conceptual framework for estimating benefits when there is uncertainty in the causal relationship between a hazard
and the endpoints it potentially affects described here applies to these other pollutants.
96
EPA includes risk estimates for an example health endpoint with a causality determination of "suggestive, but not
sufficient to infer" that is associated with a potentially substantial economic value in the quantitative uncertainty
characterization (Health Benefits TSD section 6.2.3).
3-32
-------
3.3.3 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 summer season average ozone concentrations
for the years 2024, 2027, 2028, and 2038 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 PIVh.s-related total deaths (yij) during each year i among
adults aged 18 and older (a) in each county in the contiguous U.S. j (j=l,.. ,,J where J is the total
number of counties) as
yij= Ea yija
yija = moija x(ep-ACij-l) x pija3 Eq[l]
where moija is the baseline total mortality rate for adults aged a=18-99 in county j in year i
stratified in 10-year age groups, P is the risk coefficient for total mortality for adults associated
with annual average PM2.5 exposure, Cij is the annual mean PM2.5 concentration in county j in
year i, and Pija is the number of county adult residents aged a=18-99 in county j in year i
stratified into 5-year age groups.97
97
In this illustrative example, the air quality is resolved at the county level. For this RIA, we simulate air quality
concentrations at 12 km grid resolution. The BenMAP-CE tool assigns the rates of baseline death and disease stored
at the county level to the grid cell level using an area-weighted algorithm. This approach is described in greater
detail in the appendices to the BenMAP-CE user manual (U.S. EPA, 2023a).
3-33
-------
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 Integrated Science Assessments for fine particles and
ground-level ozone; and cost of illness or willingness to pay economic unit values for each
endpoint.
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, 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 Kiinzli et al. (2000) and other, more recent health impact analyses, our estimates
are based on the best available methods of benefits transfer. Benefits transfer adapts 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.
3.3.4 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.
3-34
-------
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.
3.3.5 Benefits Analysis Data Inputs
In Figure 3-3Figure , we summarize the key data inputs to the health impact and
economic valuation estimates, which were calculated using BenMAP-CE model version 1.5.1
(Sacks et al., 2020). In the sections below we summarize the data sources for each of these
inputs, including demographic projections, incidence and prevalence rates, effect coefficients,
and economic valuation.
3-35
-------
Figure 3-3 Data Inputs and Outputs for the BenMAP-CE Model
3.3.5.1 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:
First, national-level variables such as income, employment, and populations are
forecasted.
3-36
-------
Second, employment projections are made for 179 economic areas defined by the Bureau
of Economic Analysis (U.S. BEA, 2004), 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.
3.3.5.2 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 (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
3-37
-------
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
are most applicable to a national assessment of benefits. For some studies, however, the only
available incidence information comes from the studies themselves; in these cases, incidence in
the study population is assumed to represent typical incidence at the national level.
We projected mortality rates such that future mortality rates are consistent with our
projections of population growth (U.S. EPA, 2023b). 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, 2023b).
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, 2021b). 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 (AHRQ, 2016). 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.
3.3.5.3 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 ozone-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. EPA selects hazard ratios from
cohort studies to estimate counts of ozone-related premature death, following a systematic
approach detailed in the Health Benefits TSD accompanying this RIA that is generally consistent
3-38
-------
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. Further discussion of the
cohort studies and hazard ratios for quantifying ozone-attributable premature death can be found
below in Section 3.3.6.
3.3.6 Quantifying Cases of Ozone-Attributable Premature Death
Mortality risk reductions account for the majority of monetized ozone-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. As noted above, the Health Benefits TSD describes fully the
Agency's approach for quantifying the number and value of ozone-related impacts, including
additional discussion of how the Agency selected the risk studies used to quantify ozone-related
impacts 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 (National Research Council, 2008) issued a
series of recommendations to EPA regarding the procedure for quantifying and valuing ozone-
related mortality due to short-term exposures. 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)In
2008, the National Academies of Science (National Research Council, 2008) issued a series of
recommendations to EPA regarding the procedure for quantifying and valuing ozone-related
mortality due to short-term exposures. 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
3-39
-------
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
(Smith et al., 2009; Zanobetti & Schwartz, 2008) and one long-term cohort study of respiratory
mortality (Jerrett et al., 2009). The 2020 Ozone ISA included changes to the causality
relationship determinations between short-term exposures and total mortality, as well as
including more recent epidemiologic analyses of long-term exposure effects on respiratory
mortality (U.S. EPA, 2020c). Beginning with the RCU analysis 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 April-September.(Smith et al., 2009; Zanobetti & Schwartz, 2008) and
one long-term cohort study of respiratory mortality (Jerrett et al., 2009). The 2020 Ozone ISA
included changes to the causality relationship determinations between short-term exposures and
total mortality, as well as including more recent epidemiologic analyses of long-term exposure
effects on respiratory mortality (U.S. EPA, 2020c). We currently 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).
3.3.7 Characterizing Uncertainty in the Estimated Benefits
In any complex analysis using estimated parameters and inputs from numerous models,
there are likely to be many sources of uncertainty. This analysis is no exception. The Health
3-40
-------
Benefits TSD details our approach to characterizing uncertainty in both quantitative and
qualitative terms. That 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 ozone-related mortality using alternative ozone mortality effect
estimates drawn from two epidemiologic studies; and
3. Presentation of 95th percentile confidence interval around each risk estimate.
Quantitative characterization of other sources of ozone uncertainties are discussed only in
Section 7.1 of the Health Benefits TSD:
1. For adult all-cause mortality:
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).
3-41
-------
2. Plausible alternative risk estimates for asthma onset in children (Health Benefits 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);
Quantitative consideration of baseline incidence rates and economic valuation estimates
are provided in Section 7.3 and 7.4 of the TSD, respectively. Qualitative discussions of various
sources of uncertainty can be found in Section 7.5 of the TSD.
3.3.7.1 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
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.
3.3.7.2 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.
Uncertainties are associated with the projection of emissions to analytic years. For most
sectors, this process incorporates data from the base year and applying economic and other
3-42
-------
information to estimate changes in the activity for those sources in the intervening years, along
with the application of any regulatory or economic drivers that would result in a changed rate of
emissions per unit of activity in the analytic year. Uncertainties associated with applying air
quality modeling to create ozone surfaces are discussed in U.S. EPA (2023a).
3.3.8 Estimated Number and Economic Value of Health Benefits
Table 3-6 through Table 3-9 report the estimated number of reduced premature deaths
and illnesses in each year and regulatory alternative relative to the baseline along with the 95
percent confidence interval. The number of avoided estimated deaths and illnesses are calculated
from the sum of individual reduced mortality and illness risk across the population. Table 3-10
reports the estimated economic value of avoided premature deaths and illness in each year
relative to the baseline along with the 95 percent confidence interval. We also report the stream
of benefits from 2024 through 2038 for the final rule and the less- and more-stringent regulatory
alternatives, using the monetized sums of long-term ozone mortality and morbidity impacts
(Table 3-6 through Table 3-8).
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
2038 and estimates of baseline mortality incidence rates at five-year increments. Table 3-10
includes two estimates for each regulatory alternative and year. These estimates were quantified
using two different epidemiological estimates for the mortality impact of ozone. One estimate
reflects the impacts associated with short-term exposure on mortality impacts while the other
reflects long-term exposure on mortality. These estimates should not be thought of as
representing low and high bounds.
These tables include estimates based on discounting at two, three, and seven percent.
Valuations based on three and seven percent discount rates are implemented as described in the
TSD. We develop new valuations based on a two percent discount rate for effects which occur
after the exposure. This applies to long-term mortality, asthma onset, and school loss days. For
long-term mortality, we use the same lag structure for mortality as with three and seven percent,
which results in a discount factor of 0.934, compared with 0.906 when discounting at three
percent. For school loss days, we recalculate the valuation as previously done which yields a
3-43
-------
valuation of $1,186. For asthma onset, we are unable to calculate the valuation with a two
percent discount rate and thus use the values based on a three percent rate as an approximation
that is smaller in absolute value than it would be if discounted at two percent.
3-44
-------
Table 3-6 Estimated Avoided Ozone-Related Premature Respiratory Mortalities and
Illnesses in 2024 across Regulatory Options (95 percent confidence interval)3
Less Stringent
Final Rule
More Stringent
Avoided premature respiratory mortalities
Long-term exposure
Turner et al. (2016)b
2.8
(1.9 to 3.6)
2.8
(2.0 to 3.7)
2.8
(2.0 to 3.7)
Short-term exposure
Katsouyanni et al.
(2009)bc and
Zanobetti et al.
(2008)° pooled
0.13
(0.051 to 0.20)
0.13
(0.052 to 0.20)
0.13
(0.052 to 0.20)
Morbidity effects
Long-term exposure
Asthma onsetd
23
(20 to 26)
23
(20 to 27)
23
(20 to 27)
Allergic rhinitis
symptomsf
130
(69 to 190)
130
(70 to 190)
130
(70 to 190)
Hospital
admissions—
respiratory0
0.31
(-0.082 to 0.70)
0.32
(-0.083 to 0.71)
0.32
(-0.083 to 0.71)
ED visits—
respiratory6
7.1
(2.0 to 15)
7.2
(2.0 to 15)
7.2
(2.0 to 15)
Short-term exposure
Asthma symptoms
4,200
(-510 to 8,700)
4,200
(-520 to 8,800)
4,200
(-520 to 8,800)
Minor restricted-
activity days0,6
1,900
(770 to 3,000)
2,000
(780 to 3,100)
2,000
(780 to 3,100)
School absence days
1,500
(-210 to 3,100)
1,500
(-210 to 3,200)
1,500
(-210 to 3,200)
a Values rounded to two significant figures.
b Applied risk estimate derived from April-September exposures to estimates of ozone across the April-September
warm season.
0 Converted ozone risk estimate metric from maximum daily 1-hour average (MDA1) to maximum daily 8-hour
average (MDA8).
d Applied risk estimate derived from June-August exposures to estimates of ozone across the April-September warm
season.
e Applied risk estimate derived from full year exposures to estimates of ozone across the April-September warm
season.
f Converted ozone risk estimate metric from daily 24-hour average (DA24) to MDA8.
3-45
-------
Table 3-7 Estimated Avoided Ozone-Related Premature Respiratory Mortalities and
Illnesses in 2027 across Regulatory Options (95 percent confidence interval)3
Less Stringent
Final Rule
More Stringent
Avoided premature respiratory mortalities
Long-term exposure
Turner et al. (2016)b
11
(7.4 to 14)
11
(7.5 to 14)
11
(7.5 to 14)
Short-term exposure
Katsouyanni et al.
(2009)bc and
Zanobetti et al.
(2008)° pooled
0.48
(0.19 to 0.76)
0.49
(0.20 to 0.77)
0.49
(0.20 to 0.78)
Morbidity effects
Asthma onsetd
84
85
86
Long-term exposure
(72 to 95)
(73 to 97)
(73 to 97)
Allergic rhinitis
symptomsf
480
(250 to 700)
490
(260 to 710)
86
(73 to 97)
Hospital
admissions—
respiratory0
1.2
1.2
1.2
(-0.32 to 2.7)
(-0.32 to 2.7)
(-0.32 to 2.7)
ED visits—
respiratory6
26
(7.2 to 55)
27
(7.3 to 56)
27
(7.4 to 56)
Short-term exposure
Asthma symptoms
15,000
(-1,900 to 32,000)
16,000
(-1,900 to 33,000)
16,000
(-1,900 to 33,000)
Minor restricted-
activity days0,6
6,900
(2,800 to 11,000)
7,000
(2,800 to 11,000)
7,100
(2,800 to 11,000)
School absence days
5,500
(-770 to 11,000)
5,600
(-780 to 12,000)
5,600
(-780 to 12,000)
a Values rounded to two significant figures.
b Applied risk estimate derived from April-September exposures to estimates of ozone across the April-September
warm season.
0 Converted ozone risk estimate metric from maximum daily 1-hour average (MDA1) to maximum daily 8-hour
average (MDA8).
d Applied risk estimate derived from June-August exposures to estimates of ozone across the April-September warm
season.
e Applied risk estimate derived from full year exposures to estimates of ozone across the April-September warm
season.
f Converted ozone risk estimate metric from daily 24-hour average (DA24) to MDA8.
3-46
-------
Table 3-8 Estimated Avoided Ozone-Related Premature Respiratory Mortalities and
Illnesses in 2028 across Regulatory Options (95 percent confidence interval)3
Less Stringent
Final Rule
More Stringent
Avoided premature respiratory
mortalities
Long-term exposure
Turner et al. (2016)b
49
(34 to 64)
56
(39 to 73)
57
(40 to 74)
Short-term exposure
Katsouyanni et al.
(2009)bc and
Zanobetti et al.
(2008)° pooled
2.2
(0.90 to 3.5)
2.6
(1.0 to 4.0)
2.6
(1.0 to 4.1)
Morbidity effects
Long-term exposure
Asthma onsetd
400
(340 to 450)
450
(390 to 520)
460
(390 to 520)
Allergic rhinitis
symptomsf
2,300
(1,200 to 3,300)
2,600
(1,400 to 3,800)
2,600
(1,400 to 3,800)
Hospital
admissions—
respiratory0
5.9
(-1.6 to 13)
6.8
(-1.8 to 15)
6.9
(-1.8 to 15)
ED visits—
respiratory6
120
(34 to 260)
140
(39 to 300)
140
(40 to 300)
Short-term exposure
Asthma symptoms
73,000
(-9,000 to 150,000)
84,000
(-10,000 to 170,000)
85,000
(-10,000 to 180,000)
Minor restricted-
activity days0,6
33,000
(13,000 to 52,000)
38,000
(15,000 to 59,000)
38,000
(15,000 to 60,000)
School absence days
26,000
(-3,700 to 54,000)
30,000
(-4,200 to 62,000)
30,000
(-4,200 to 63,000)
a Values rounded to two significant figures.
b Applied risk estimate derived from April-September exposures to estimates of ozone across the April-September
warm season.
0 Converted ozone risk estimate metric from maximum daily 1-hour average (MDA1) to maximum daily 8-hour
average (MDA8).
d Applied risk estimate derived from June-August exposures to estimates of ozone across the April-September warm
season.
e Applied risk estimate derived from full year exposures to estimates of ozone across the April-September warm
season.
f Converted ozone risk estimate metric from daily 24-hour average (DA24) to MDA8.
3-47
-------
Table 3-9 Estimated Avoided Ozone-Related Premature Respiratory Mortalities and
Illnesses in 2038 across Regulatory Options (95 percent confidence interval)3
Less Stringent
Final Rule
More Stringent
Avoided premature respiratory
mortalities
Long-term exposure
Turner et al. (2016)b
67
(46 to 87)
74
(51 to 95)
74
(51 to 96)
Short-term exposure
Katsouyanni et al.
(2009)bc and
Zanobetti et al.
(2008)° pooled
3
(1.2 to 4.8)
3.3
(1.3 to 5.2)
3.3
(1.4 to 5.3)
Morbidity effects
Long-term exposure
Asthma onsetd
480
(410 to 540)
520
(450 to 590)
530
(450 to 600)
Allergic rhinitis
symptomsf
2,800
(1,500 to 4,100)
3,100
(1,600 to 4,400)
3,100
(1,600 to 4,500)
Hospital
admissions—
respiratory0
8.1
(-2.1 to 18)
8.9
(-2.3 to 20)
8.9
(-2.3 to 20)
ED visits—
respiratory6
160
(43 to 330)
170
(47 to 360)
170
(47 to 360)
Short-term exposure
Asthma symptoms
88,000
(-11,000 to 180,000)
97,000
(-12,000 to 200,000)
97,000
(-12,000 to 200,000)
Minor restricted-
activity days0,6
40,000
(16,000 to 64,000)
44,000
(18,000 to 70,000)
44,000
(18,000 to 70,000)
School absence days
32,000
(-4,500 to 67,000)
35,000
(-4,900 to 74,000)
35,000
(-5,000 to 74,000)
a Values rounded to two significant figures.
b Applied risk estimate derived from April-September exposures to estimates of ozone across the April-September
warm season.
0 Converted ozone risk estimate metric from maximum daily 1-hour average (MDA1) to maximum daily 8-hour
average (MDA8).
d Applied risk estimate derived from June-August exposures to estimates of ozone across the April-September warm
season.
e Applied risk estimate derived from full year exposures to estimates of ozone across the April-September warm
season.
f Converted ozone risk estimate metric from daily 24-hour average (DA24) to MDA8.
3-48
-------
Table 3-10 Estimated Discounted Economic Value of Avoided Ozone-Attributable
Premature Mortality and Illness across Regulatory Options (95 percent confidence
interval; millions of 2019 dollars)3
Disc.
Rate
Year
LessStringent
Final Rule
MoreStringent
2024
$4.1
($1.0—$8.5)
and
$32
($3.6-$84)
$4.1
($1.1—$8.6)
and
$33
($3.6-$85)
$4.10
($1.1—$8.7)
and
$33
($3.6-$86)
2%
2027
$15
($3.8-$32)
and
$120
($14-$330)
$16
($3.9-$33)
and
$130
($14-$330)
$16
($3.9-$33)
and
$130
($14-$330)
2028
$72
($18-$150)
and
$580
($64-$l,500)
$82
($20-$170)
and
$660
($73-$l,700)
$83
($21—$180)
and
$670
($74-$l,800)
2038
$93
($22-$200)
and
$820
($87-$2,200)
$100
($24-$220)
and
$900
($96-$2,400)
$100
($24-$220)
and
$900
($97-$2,400)
2024
$4.0
($l-$8.5)b
and
$31
($3.5-$82)c
$4.1
($1.1—$8.6)b
and
$32
($3.6-$83)c
$4.1
($1.1—$8.6)b
and
$32
($3.6-$83)c
3%
2027
$15
($3.8-$32)b
and
$120
($13-$320)c
$15
($3.9-$33)b
and
$120
($14-$320)c
$15
($3.9-$33)b
and
$120
($14-$320)c
2028
$71
($18—$150)b
and
$560
($62-$l,500)c
$82
($21—$170)b
and
$640
($71-$l,700)c
$83
($21—$170)b
and
$650
($72-$l,700)c
2038
$92
($22-$200)b
and
$790
($86-$2,100)c
$100
($24-$220)b
and
$870
($94-$2,300)c
$100
($24-$220)b
and
$880
($94-$2,300)c
2024
$3.6
($0.67-$8.0)b
and
$28
($2.9-$74)c
$3.7
($0.68—$8.1)b
and
$28
($2.9-$75)c
$3.7
($0.68—$8.1)b
and
$28
($2.9-$75)c
7%
2027
$14
($2.5-$30)b
and
$110
($11—$290)c
$14
($2.5—$3 l)b
and
$110
($11—$290)c
$14
($2.5—$3 l)b
and
$110
($11—$290)c
2028
$64
($12—$140)b
and
$500
($51—$l,300)c
$73
($13—$160)b
and
$580
($59-$l,500)c
$74
($13—$160)b
and
$580
($59-$l,500)c
2038
$83
($14—$190)b
and
$710
($71-$l,900)c
$91
($16—$210)b
and
$780
($78-$2,100)c
$92
($16—$210)b
and
$790
($79-$2,100)c
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 Sum of ozone mortality estimated using the pooled short-term ozone exposure risk estimate.
0 Sum of the Turner et al. (2016) long-term ozone exposure risk estimate.
3-49
-------
Table 3-6 Stream of Human Health Benefits from 2024 through 2038: Monetized
Benefits Quantified as Sum of Long-Term Ozone Mortality and Illness across Regulatory
Options (discounted at 2 percent; millions of 2019 dollars)3
Yearb
Less Stringent
Final Rule
More Stringent
2024
$30
$31
$31
2025
$30
$31
$31
2026
$110
$110
$110
2027
$110
$110
$110
2028
$500
$580
$580
2029
$500
$580
$590
2030
$510
$580
$590
2031
$510
$580
$590
2032
$510
$590
$590
2033
$510
$590
$600
2034
$590
$640
$650
2035
$590
$640
$650
2036
$590
$640
$650
2037
$590
$640
$650
2038
$580
$640
$640
PV
$6,300
$7,000
$7,100
EAV
$490
$540
$550
a For simplicity of exposition, the estimated value of the health benefits reported here are the larger of the two
benefits estimates presented in Table 3-10. Monetized benefits include those related to public health associated with
reductions in ozone concentrations. The health benefits are associated with several point estimates.
b Air quality models were run for 2024, 2027, 2028, and 2038. Benefits for all other years were extrapolated from
years with model-based air quality estimates. Benefits calculated as value of avoided: Ozone-attributable deaths
(quantified using a concentration-response relationship from the Turner et al. 2016 study); and ozone-related
morbidity effects.
3-50
-------
Table 3-7 Stream of Human Health Benefits from 2024 through 2038: Monetized
Benefits Quantified as Sum of Long-Term Ozone Mortality and Illness across Regulatory
Options (discounted at 3 percent; millions of 2019 dollars)3
Yearb
Less Stringent
Final Rule
More Stringent
2024
$29
$30
$30
2025
$29
$30
$30
2026
$100
$110
$110
2027
$100
$110
$110
2028
$470
$540
$540
2029
$470
$540
$540
2030
$460
$530
$540
2031
$460
$530
$530
2032
$460
$530
$530
2033
$460
$520
$530
2034
$520
$570
$570
2035
$510
$560
$570
2036
$510
$560
$560
2037
$500
$550
$550
2038
$490
$540
$550
PV
$5,600
$6,200
$6,300
EAV
$470
$520
$530
a For simplicity of exposition, the estimated value of the health benefits reported here are the larger of the two
benefits estimates presented in Table 3-10. Monetized benefits include those related to public health associated with
reductions in ozone concentrations. The health benefits are associated with several point estimates.
b Air quality models were run for 2024, 2027, 2028, and 2038. Benefits for all other years were extrapolated from
years with model-based air quality estimates. Benefits calculated as value of avoided: Ozone-attributable deaths
(quantified using a concentration-response relationship from the Turner et al. 2016 study); and ozone-related
morbidity effects.
3-51
-------
Table 3-8 Stream of Human Health Benefits from 2024 through 2038: Monetized
Benefits Quantified as Sum of Long-Term Ozone Mortality and Illness across Regulatory
Options (discounted at 7 percent; millions of 2019 dollars)3
Yearb
Less Stringent
Final Rule
More Stringent
2024
$26
$27
$27
2025
$25
$26
$26
2026
$86
$88
$88
2027
$83
$84
$84
2028
$360
$410
$420
2029
$340
$400
$400
2030
$330
$380
$380
2031
$310
$360
$360
2032
$300
$350
$350
2033
$290
$330
$340
2034
$310
$350
$350
2035
$300
$330
$330
2036
$290
$310
$320
2037
$270
$300
$300
2038
$260
$280
$290
PV
$3,600
$4,000
$4,100
EAV
$390
$440
$450
a For simplicity of exposition, the estimated value of the health benefits reported here are the larger of the two
benefits estimates presented in Table 3-10. Monetized benefits include those related to public health associated with
reductions in ozone concentrations. The health benefits are associated with several point estimates.
b Air quality models were run for 2024, 2027, 2028, and 2038. Benefits for all other years were extrapolated from
years with model-based air quality estimates. Benefits calculated as value of avoided: Ozone-attributable deaths
(quantified using a concentration-response relationship from the Turner et al. 2016 study); and ozone-related
morbidity effects.
3.4 Ozone Vegetation Effects
Human exposure to ambient ozone concentrations is associated with adverse health
effects, including premature respiratory mortality and cases of respiratory morbidity (U.S. EPA,
2020a). Researchers have associated ozone exposure with adverse health effects in numerous
toxicological, clinical, and epidemiological studies (U.S. EPA, 2020a). When adequate data and
resources are available, the EPA has generally quantified several health effects associated with
exposure to ozone (U.S. EPA, 2010, 201 le, U.S. EPA, 2021c). These health effects include
respiratory morbidity, such as asthma attacks, hospital and emergency department visits, lost
school days, and premature respiratory mortality. The scientific literature is also suggestive that
exposure to ozone is associated with chronic respiratory damage and premature aging of the
lungs.
3-52
-------
3.5 Ozone Climate Effects
Ozone is a well-known short-lived climate forcing GHG (U.S. EPA, 2013). Stratospheric
ozone (the upper ozone layer) is beneficial because it protects life on Earth from the sun's
harmful ultraviolet (UV) radiation. In contrast, tropospheric ozone (ozone in the lower
atmosphere) is a harmful air pollutant that adversely affects human health and the environment
and contributes significantly to regional and global climate change. Due to its short atmospheric
lifetime, tropospheric ozone concentrations exhibit large spatial and temporal variability (U.S.
EPA, 2020c, 2023b). The IPCC AR5 estimated that the contribution to current warming levels of
increased tropospheric ozone concentrations resulting from human methane, NOx, and VOC
emissions was 0.5 W/m2, or about 30 percent as large a warming influence as elevated CO2
concentrations. This quantifiable influence of ground level ozone on climate leads to increases in
global surface temperature and changes in hydrological cycles.
3.6 Ozone-Related Impacts Due to Methane
The tropospheric ozone produced by the reaction of methane in the atmosphere has
harmful effects for human health and plant growth in addition to its climate effects (Nolte et al.,
2018). In remote areas, methane is a dominant precursor to tropospheric ozone formation.
Approximately 50 percent of the global annual mean ozone increase since preindustrial times is
believed to be due to anthropogenic methane (Myhre et al., 2013). Projections of future
emissions also indicate that methane is likely to be a key contributor to ozone concentrations in
the future (Myhre et al., 2013). Unlike NOx and VOC, which affect ozone concentrations
regionally and at hourly time scales, methane emissions affect ozone concentrations globally and
on decadal time scales given methane's long atmospheric lifetime when compared to these other
ozone precursors (Myhre et al., 2013). Reducing methane emissions, therefore, will contribute to
efforts to reduce global background ozone concentrations that contribute to the incidence of
ozone-related health effects (USGCRP, 2018). The benefits of such reductions are global and
occur in both urban and rural areas. As noted above, these effects are not currently included in
estimates of the social cost of methane. However, a recent analysis by McDuffie et al. (2023)
used a combination of global model simulations from the United Nations Environment
Programme & Climate and Clean Air Coalition (UNEP/CCAC), in combination with BenMAP,
to evaluate the additional risk in respiratory-related human mortality from ozone produced per
3-53
-------
ton of methane emissions. This approach is similar to the social cost of methane and finds that
the monetized increase in respiratory-related human mortality risk from ozone produced from
methane emissions in 2020 is $1,800 per ton of methane (95 percent confidence interval: $760-
2,800 per mt CH4 in 2020 US dollars). As discussed in U.S. EPA (2023f), this monetized result is
similar to an earlier study by Sarofim et al. (2017) but smaller than a 2021 study conducted by
the UNEP/CCAC, which included additional cardiovascular mortality risk due to elevated ozone
concentrations (United Nations Environment Programme and Climate and Clean Air Coalition,
2021). Collectively, these and other prior studies suggest that there are additional risks to human
health from the methane-ozone mechanism that are not currently accounted for in the social cost
of methane. Applying the ozone-related health benefit per ton estimates from McDuffie et al.
(2023) would yield a present value of the ozone-related health benefits from the 2024-2038 CH4
emission reductions of the final rule on the order of $110 billion (2019 dollars), of which
approximately $14 billion are accruing to populations within U.S. borders.98 Because these
benefits are the result of methane, which is a global pollutant, EPA believes it is most
appropriate to focus attention on the global benefits to human health from the methane-ozone
mechanism for the same reasons discuss above with respect to climate benefits. EPA will
continue to look for opportunities to incorporate the ozone related impacts of CH4 emissions in
future updates to the SC-CH4.
3.7 PM2.5-Related Impacts Due to VOC Emissions
This final rulemaking is expected to result in emissions reductions of VOC, which are a
precursor to PM2.5, thus decreasing human exposure to PM2.5 and the incidence of PIVh.s-related
health effects, although the magnitude of this effect has not been quantified at this time. Most
VOC emitted are oxidized to CO2 rather than to PM, but a portion of VOC emissions contributes
to ambient PM2.5 levels as organic carbon aerosols (U.S. EPA, 2019a). Analysis of organic
carbon measurements suggest only a fraction of secondarily formed organic carbon aerosols are
of anthropogenic origin. The current state of the science of secondary organic carbon aerosol
formation indicates that anthropogenic VOC contribution to secondary organic carbon aerosol is
98
This estimate relies on benefit per ton numbers that use the socioeconomics from the RFF-SPs and the 2 percent
near-term Ramsey discounting approach. See McDuffie, E. E., Sarofim, M. C., Raich, W., Jackson, M., Roman, H.,
Seltzer, K.,. . . Fann, N. (2023). The Social Cost of Ozone-Related Mortality Impacts From Methane Emissions.
Earth's Future, 11(9), e2023EF003853. https://doi.Org/https://doi.org/10.1029/2023EF003853 for more details.
3-54
-------
often lower than the biogenic (natural) contribution (U.S. EPA, 2019a). The potential for an
organic compound to partition into the particle phase is highly dependent on its volatility such
that compounds with lower volatility are more prone to partition into the particle phase and form
secondary organic aerosols (SOA) (Cappa & Wilson, 2012; Donahue et al., 2012; Jimenez et al.,
2009).
Hydrocarbon emissions from oil and natural gas operations tend to be dominated by high
volatility, low-carbon number compounds that are less likely to form SOA (Helmig et al., 2014;
Koss et al., 2017; Petron et al., 2012). Given that only a fraction of secondarily formed organic
carbon aerosols is from anthropogenic VOC emissions, and the relatively volatile nature of
VOCs emitted from this sector, it is unlikely that the VOC emissions reductions projected to
occur under this final rule would have a large contribution to ambient secondary organic carbon
aerosols. Therefore, we have not quantified the PIVh.s-related benefits in this analysis. Moreover,
without modeling air quality changes, we are unable to determine how this rule might affect air
quality in downwind PM2.5 nonattainment areas. However, we note that in future regulatory
impact analyses supporting other regulations, the EPA plans to account for the emissions impacts
of the oil and natural gas NSPS OOOOb and EG OOOOc in the baseline for the analysis.
3.7.1 PM2.5 Health Effects
Decreasing exposure to PM2.5 is associated with significant human health benefits,
including reductions in respiratory mortality and respiratory morbidity. Researchers have
associated PM2.5 exposure with adverse health effects in numerous toxicological, clinical, and
epidemiological studies (U.S. EPA, 2019a). These health effects include asthma development
and aggravation, decreased lung function, and increased respiratory symptoms, such as irritation
of the airways, coughing, or difficulty breathing (U.S. EPA, 2019a). These health effects result in
hospital and ER visits, lost workdays, and restricted activity days. When adequate data and
resources are available, the EPA has quantified the health effects associated with exposure to
PM2.5 (U.S. EPA, 2023b).
When the EPA quantifies PIVh.s-related benefits, the Agency assumes that all fine
particles, regardless of their chemical composition, are equally potent in causing premature
mortality because the scientific evidence is not yet sufficient to allow differentiation of effect
3-55
-------
estimates by particle type (U.S. EPA, 2019a). Based on our review of the current body of
scientific literature, the EPA estimates PM-related premature mortality without applying an
assumed concentration threshold. This decision is supported by the data, which are quite
consistent in showing effects down to the lowest measured levels of PM2.5 in the underlying
epidemiology studies.
3.7.2 PM2.5 Welfare Effects
Suspended particles and gases degrade visibility by scattering and absorbing light.
Decreasing secondary formation of PM2.5 from VOC emissions could improve visibility
throughout the U.S. Visibility impairment has a direct impact on 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. Previous analyses
show that visibility benefits are a significant welfare benefit category (U.S. EPA, 2006, 201 la,
201 lc, 2012a). However, without air quality modeling of PM2.5 impacts, we are unable to
estimate visibility related benefits.
Separately, persistent and bioaccumulative HAP reported as emissions from oil and
natural gas operations, including polycyclic organic matter, could lead to PM welfare effects.
Several significant ecological effects are associated with the deposition of organic particles,
including persistent organic pollutants and polycyclic aromatic hydrocarbons (PAHs) (U.S. EPA,
2009). PAHs can accumulate to high enough concentrations in some coastal environments to
pose an environmental health threat that includes cancer in fish populations, toxicity to
organisms living in the sediment and risks to those (e.g., migratory birds) that consume these
organisms. Atmospheric deposition of particles is thought to be the major source of PAHs to the
sediments of coastal areas of the U.S. (U.S. EPA, 2012b).
3.8 Hazardous Air Pollutants (HAP) Impacts
Available emissions data show that several different HAP are emitted from oil and
natural gas operations. The HAP emissions from the oil and natural gas sector in the 2017
National Emissions Inventory (NEI) emissions data are summarized in Table 3-9. The table
includes either oil and natural gas nonpoint or oil and natural gas point emissions of at least 10
tons per year, in descending order of annual nonpoint emissions. Emissions of eight HAP make
3-56
-------
up a large percentage of the total HAP emissions by mass from the oil and natural gas sector:
toluene, hexane, benzene, xylenes (mixed), ethylene glycol, methanol, ethyl benzene, and 2,2,4-
trimethylpentane (U.S. EPA, 2011b).
Table 3-9 Top Annual HAP Emissions as Reported in 2017 NEI for Oil and Natural
Gas Sources
Pollutant Nonpoint Emissions (tons/year) Point Emissions (tons/year)
Benzene
26,869
502
Xylenes (Mixed Isomers)
25,410
506
Formaldehyde
23,413
222
Toluene
18,054
823
Acetaldehyde
2,722
26
Hexane
2,675
886
Ethyl Benzene
2,021
113
Acrolein
1,602
18
Methanol
1,578
342
1,3-Butadiene
337
5.80E-01
2,2,4-Trimethylpentane
252
46
Naphthalene
104
1.10E+00
Propionaldehyde
102
0.00E+00
PAH/POM - Unspecified
68
2.50E-02
1,1,2-T richloroethane
25
1.40E-03
Methylene Chloride
22
8.70E-02
1,1,2,2-Tetrachloroethane
14
1.90E-03
Ethylene Dibromide
13
1.90E-03
Methyl Tert-Butyl Ether
0
17.30
In the subsequent sections, we describe the health effects associated with the main HAP
of concern from the oil and natural gas sector: benzene (Section 3.8.13.8.1), formaldehyde
(Section 3.8.23.8.2), toluene (Section 3.8.33.8.3), carbonyl sulfide (Section 3.8.43.8.4),
ethylbenzene (Section 3.8.53.8.5), mixed xylenes (Section 3.8.63.8.6), and n-hexane (Section
3.8.73.8.7), and other air toxics (Section 3.8.83.8.8). This final rule is projected to reduce
590,000 tons of HAP emissions over the 2024 through 2038 period." With the data available, it
was not possible to estimate the change in emissions of each individual HAP.
99
The projected emissions reductions from the proposed NSPS and EG, including projections of HAP reductions,
are based upon the unit-level model plant analysis supporting this rulemaking multiplied by counts of units that are
potentially affected by this proposal. The model plants and counts are built from a different basis than the oil and
natural gas sector emissions estimated in the NEI. Comparisons between the projected emissions reductions under
this proposal and the NEI should be made with caution.
3-57
-------
Monetization of the benefits of reductions in cancer incidences requires several important
inputs, including central estimates of cancer risks, estimates of exposure to carcinogenic HAP,
and estimates of the value of an avoided case of cancer (fatal and non-fatal). Due to methodology
and data limitations, we did not attempt to monetize the health benefits of reductions in HAP in
this analysis. Instead, we are providing a qualitative discussion of the health effects associated
with HAP emitted from sources subject to control under the final NSPS OOOOb and EG
0000c. The EPA remains committed to improving methods for estimating HAP benefits by
continuing to explore additional aspects of HAP-related risk from the oil and natural gas sector,
including the distribution of that risk. This is discussed further in the context of environment
justice in Section 4.3.4.
3.8.1 Benzene
The EPA's Integrated Risk Information System (IRIS) database lists benzene as a known
human carcinogen (causing leukemia) by all routes of exposure and concludes that exposure is
associated with additional health effects, including genetic changes in both humans and animals
and increased proliferation of bone marrow cells in mice (IARC, 1982; Irons et al., 1992; U.S.
EPA, 2003a). The EPA states that data indicate a causal relationship between benzene exposure
and acute lymphocytic leukemia and suggest a relationship between benzene exposure and
chronic non-lymphocytic leukemia and chronic lymphocytic leukemia. The International Agency
for Research on Carcinogens (IARC) has determined that benzene is a human carcinogen, and
the U.S. Department of Health and Human Services has characterized benzene as a known
human carcinogen (IARC, 1987; NTP, 2004). Several adverse noncancer health effects have
been associated with chronic inhalation of benzene in humans including arrested development of
blood cells, anemia, leukopenia, thrombocytopenia, and aplastic anemia. Respiratory effects
have been reported in humans following acute exposure to benzene vapors, such as nasal
irritation, mucous membrane irritation, dyspnea, and sore throat (ATSDR, 2007a).
3.8.2 Formaldehyde
In 1989, the EPA classified formaldehyde as a probable human carcinogen based on
limited evidence of cancer in humans and sufficient evidence in animals (U.S. EPA, 1991b).
Later the IARC (2006, 2012) classified formaldehyde as a human carcinogen based upon
3-58
-------
sufficient human evidence of nasopharyngeal cancer and strong evidence for leukemia.
Similarly, in 2016, the National Toxicology Program (NTP) classified formaldehyde as known to
be a human carcinogen based on sufficient evidence of cancer from studies in humans supporting
data on mechanisms of carcinogenesis (NTP, 2016). Formaldehyde inhalation exposure causes a
range of noncancer health effects including irritation of the nose, eyes, and throat in humans and
animals. Repeated exposures cause respiratory tract irritation, chronic bronchitis and nasal
epithelial lesions such as metaplasia and loss of cilia in humans. Airway inflammation, including
eosinophil infiltration, has been observed in animals exposed to formaldehyde. In children, there
is evidence that formaldehyde may increase the risk of asthma and chronic bronchitis (ATSDR,
1999; WHO, 2002).
3.8.3 Toluene100
Under the 2005 Guidelines for Carcinogen Risk Assessment (U.S. EPA, 2005a), there is
inadequate information to assess the carcinogenic potential of toluene because studies of humans
chronically exposed to toluene are inconclusive, toluene was not carcinogenic in adequate
inhalation cancer bioassays of rats and mice exposed for life, and increased incidences of
mammary cancer and leukemia were reported in a lifetime rat oral bioassay.
The central nervous system (CNS) is the primary target for toluene toxicity in both
humans and animals for acute and chronic exposures. CNS dysfunction (which is often
reversible) and narcosis have been frequently observed in humans acutely exposed to low or
moderate levels of toluene by inhalation: symptoms include fatigue, sleepiness, headaches, and
nausea. Central nervous system depression has been reported to occur in chronic abusers exposed
to high levels of toluene. Symptoms include ataxia, tremors, cerebral atrophy, nystagmus
(involuntary eye movements), and impaired speech, hearing, and vision. Chronic inhalation
exposure of humans to toluene also causes irritation of the upper respiratory tract, eye irritation,
dizziness, headaches, and difficulty with sleep.
100 All health effects language for this section came from U.S. EPA. (2005b). Toluene; CASRN108-88-3 [Chemical
Assessment Summary] (Integrated Risk Information System (IRIS), Issue.
https://iris.epa.gov/ChemicalLanding/&substance_nmbr=118.
3-59
-------
Human studies have also reported developmental effects, such as CNS dysfunction,
attention deficits, and minor craniofacial and limb anomalies, in the children of women who
abused toluene during pregnancy. A substantial database examining the effects of toluene in
subchronic and chronic occupationally exposed humans exists. The weight of evidence from
these studies indicates neurological effects (i.e., impaired color vision, impaired hearing,
decreased performance in neurobehavioral analysis, changes in motor and sensory nerve
conduction velocity, headache, and dizziness) as the most sensitive endpoint.
3.8.4 Carbonyl Sulfide
Limited information is available on the health effects of carbonyl sulfide. Acute (short-
term) inhalation of high concentrations of carbonyl sulfide may cause narcotic effects and irritate
the eyes and skin in humans.101 No information is available on the chronic (long-term),
reproductive, developmental, or carcinogenic effects of carbonyl sulfide in humans. Carbonyl
sulfide has not undergone a complete evaluation and determination under the EPA's IRIS
program for evidence of human carcinogenic potential (U.S. EPA, 1991a).
3.8.5 Ethylbenzene
Ethylbenzene is a major industrial chemical produced by alkylation of benzene. The pure
chemical is used almost exclusively for styrene production. It is also a constituent of crude
petroleum and is found in gasoline and diesel fuels. Acute (short-term) exposure to ethylbenzene
in humans results in respiratory effects such as throat irritation and chest constriction, and
irritation of the eyes, and neurological effects such as dizziness. Chronic (long-term) exposure of
humans to ethylbenzene may cause eye and lung irritation, with possible adverse effects on the
blood. Animal studies have reported effects on the blood, liver, and kidneys and endocrine
system from chronic inhalation exposure to ethylbenzene. No information is available on the
developmental or reproductive effects of ethylbenzene in humans, but animal studies have
reported developmental effects, including birth defects in animals exposed via inhalation. Studies
in rodents reported increases in the percentage of animals with tumors of the nasal and oral
101 Hazardous Substances Data Bank (HSDB), online database. US National Library of Medicine, Toxicology Data
Network, available online at https://pubchem.ncbi.nlm.nih.gov/. Carbonyl sulfide health effects summary available
at https://pubchem.ncbi.nlm.nih.gov/compound/10039#section=Safety-and-Hazards. Accessed April 26, 2020.
3-60
-------
cavities in male and female rats exposed to ethylbenzene via the oral route (Maltoni et al., 1997;
Maltoni et al., 1985). The reports of these studies lacked detailed information on the incidence of
specific tumors, statistical analysis, survival data, and information on historical controls, thus the
results of these studies were considered inconclusive by the IARC (1982) and the NTP (1999).
The NTP (1999) carried out a chronic inhalation bioassay in mice and rats and found clear
evidence of carcinogenic activity in male rats and some evidence in female rats, based on
increased incidences of renal tubule adenoma or carcinoma in male rats and renal tubule
adenoma in females. The NTP also noted increases in the incidence of testicular adenoma in
male rats. Increased incidences of lung alveolar/bronchiolar adenoma or carcinoma were
observed in male mice and liver hepatocellular adenoma or carcinoma i female mice, which
provided some evidence of carcinogenic activity in male and female mice (NTP, 1999). The
IARC classified ethylbenzene as Group 2B, possibly carcinogenic to humans, based on the NTP
studies (IARC, 2000).
3.8.6 Mixed Xylenes
Short-term inhalation of mixed xylenes (a mixture of three closely related compounds) in
humans may cause irritation of the nose and throat, nausea, vomiting, gastric irritation, mild
transient eye irritation, and neurological effects (U.S. EPA, 2003b). Other reported effects
include labored breathing, heart palpitation, impaired function of the lungs, and possible effects
in the liver and kidneys (ATSDR, 2007b). Long-term inhalation exposure to xylenes in humans
has been associated with a number of effects in the nervous system including headaches,
dizziness, fatigue, tremors, and impaired motor coordination (ATSDR, 2007b). The EPA has
classified mixed xylenes in Category D, not classifiable with respect to human carcinogenicity.
3.8.7 tt-Hexane
The studies available in both humans and animals indicate that the nervous system is the
primary target of toxicity upon exposure of n-hexane via inhalation. There are no data in humans
and very limited information in animals about the potential effects of n-hexane via the oral route.
Acute (short-term) inhalation exposure of humans to high levels of hexane causes mild central
nervous system effects, including dizziness, giddiness, slight nausea, and headache. Chronic
(long-term) exposure to hexane in air causes numbness in the extremities, muscular weakness,
3-61
-------
blurred vision, headache, and fatigue. Inhalation studies in rodents have reported behavioral
effects, neurophysiological changes, and neuropathological effects upon inhalation exposure to
n-hexane. Under the Guidelines for Carcinogen Risk Assessment (U.S. EPA, 2005a), the
database for n-hexane is considered inadequate to assess human carcinogenic potential, therefore
the EPA has classified hexane in Group D, not classifiable as to human carcinogenicity.
3.8.8 Other Air Toxics
In addition to the compounds described above, other toxic compounds might be affected
by this rule, including hydrogen sulfide (H2S). Information regarding the health effects of those
compounds can be found in the EPA's IRIS database.102
3.9 Secondary Air Emissions Impacts
The control techniques to meet the storage vessel-related standards are associated with
several types of secondary emissions impacts, which may partially offset the direct benefits of
this rule. Table 3-10Table 3-10 shows the estimated secondary emissions associated with
combustion of emissions as a result of these requirements; this includes additional flaring
expected to occur as a result of detecting inactive flares through OGI and repairing them.
Relative to the direct emission reductions anticipated from this rule, the magnitude of these
secondary air pollutant increases is small.
102 The U.S. EPA Integrated Risk Information System (IRIS) database is available at https://www.epa.gov/iris.
3-62
-------
Table 3-10 Increases in Secondary Air Pollutant Emissions, Vapor Combustion at
Storage Vessels (short tons per year)
Year
THC
CO
NOx
PM
CO2
2024
250
650
120
5
510,000
2025
510
1,300
250
9
1,100,000
2026
770
2,000
370
14
1,600,000
2027
1,000
2,700
500
19
2,100,000
2028
3,100
8,300
1,500
57
6,500,000
2029
3,200
8,400
1,500
58
6,600,000
2030
3,300
8,700
1,600
60
6,800,000
2031
3,500
9,200
1,700
64
7,200,000
2032
3,600
9,600
1,800
67
7,600,000
2033
3,800
10,000
1,800
69
7,900,000
2034
3,800
10,000
1,800
70
7,900,000
2035
3,900
10,000
1,900
72
8,200,000
2036
4,100
11,000
2,000
74
8,500,000
2037
4,200
11,000
2,000
76
8,700,000
2038
4,300
11,000
2,100
78
8,900,000
Total
43,000
110,000
21,000
790
90,000,000
Note: Totals may not appear to add correctly due to rounding.
The CO2 impacts inTable 3-10 are the emissions that are expected to occur from vapor
combustion at affected storage vessels. However, because of the atmospheric chemistry
associated with the natural gas emissions, most of the carbon in the VOCs and CH4 emissions
expected in the absence of combustion-related emissions controls would have eventually
oxidized forming CO2 in the atmosphere and led to approximately the same long-run CO2
concentrations as with controls.103 Therefore, most of the impact of these CO2 contribution to
atmospheric concentrations from the flaring of CH4 and VOC versus future oxidization is not
additional to the impacts that otherwise would have occurred through the oxidation process.
However, there is a shift in the timing of atmospheric CO2 concentration changes under
the policy case, in which case combustion controls lead to contemporaneous increases in CO2
concentrations, compared to the baseline where the CO2 concentration increase is delayed
through the oxidation process. In the case of VOC, the oxidization time in the atmosphere is
relatively short, on the order of hours to months, so from a climate perspective the difference
between emitting the carbon immediately as CO2 during combustion or as VOC is expected to be
103
The social cost of methane (SC-CH4) used previously in this chapter to monetize the benefits of the CH4
emissions reductions does not include the impact of the carbon in CH4 emissions after it oxidizes to CO2.
3-63
-------
negligible. In the case of CH4, the oxidization time is on the order of a decade, so the timing of
the contribution to atmospheric CO2 concertation will differ between the baseline and policy
case. Because the growth rate of the SC-CO2 estimates is lower than their associated discount
rates, the present value of the estimated impact of CO2 produced in the future via oxidized
methane from these fossil-based emissions may be less than the present value of the estimated
impact of CO2 released immediately from combusting emissions, which would imply a small
disbenefit associated with the earlier release of CO2 during combustion of the CH4 emissions.
In the NSPS OOOOa rulemaking, the EPA solicited comment on the appropriateness of
monetizing: (1) the impact of CO2 emissions associated with combusting methane and VOC
emissions from oil and natural gas sites; and (2) a new potential approach for approximating this
value using the SC-CO2. The illustrative analysis in the NSPS OOOOa RIA provided a method
for evaluating the estimated emissions outcomes associated with destroying one metric ton of
methane by combusting fossil-based emissions at oil and natural gas sites (flaring) and releasing
the CO2 emissions immediately versus releasing them in the future via the methane oxidation
process.104 The analysis demonstrated that the potential disbenefits of flaring (i.e., an earlier
contribution of CO2 emissions to atmospheric concentrations) are minor compared to the benefits
of flaring (i.e., avoiding the release of and associated climate impacts from CH4 emissions).
While recognizing the challenges and uncertainties related to estimation of these
secondary emissions impacts for this final rulemaking, EPA has continued to examine this issue
in the context of this RIA and includes an illustrative analysis using the methodology from the
NSPS OOOOa final RIA. Specifically, for this illustrative analysis, EPA assumes the oxidization
process of CH4 to be consistent with the modeling that underlies the SC-CH4 estimates presented
in Section 3.2. The estimated disbenefits associated with destroying one metric ton of methane
through combustion of emissions at oil and natural gas sites and releasing the CO2 emissions in
2023 instead of being released in the future via the methane oxidation process are found to be
small relative to the benefits of flaring. Specifically, the disbenefit is estimated to be about $20
per metric ton CH4 (based on average SC-CO2 using the 2 percent near-term target discount rate)
or about one percent of the SC-CH4 estimate per metric ton for 2024 ($1,950).105 The analogous
104 See Section 4.7 of U.S. EPA (2016).
105 See Table A.5.1 in U.S. EPA (2023e).
3-64
-------
estimate for 2038 is $37 per metric ton CH4 or about one percent of the SC-CH4 estimates per
metric ton for 2038 ($3,105).106
It is important to note that there are challenges and uncertainties related to this illustrative
method and estimates, which was developed to analyze secondary fossil-based emissions from
combustion. For example, these dollar per ton CH4 estimates cannot readily be applied to the
total CH4 emissions reductions presented in Section 3.1 without additional information about the
downstream outcomes associated with the recovered gas that is not flared — e.g., whether some
of that captured gas going to be burned or leaked somewhere down the line. The EPA will
continue to study this issue and assess the complexities involved in estimating the net emissions
effects associated with secondary fossil-based emissions, including differences in the timing of
contributions to atmospheric CO2 concentrations. Given the uncertainties related to estimating
net secondary emissions effects and that the EPA has not yet received appropriate input and
review on some aspects of these calculations, the EPA is not including monetized estimates of
the impacts of small changes in the timing of atmospheric CO2 concentration increases in the
benefits estimates in this RIA. The EPA will continue to follow the scientific literature on this
topic and update its methodologies as warranted.
3.10 Total Benefits
Table 3-11 presents the PV and EAV of the projected climate benefits across the three
regulatory options for the final NSPS OOOOb and EG OOOOc examined in this RIA. These
106 To calculate the CO2 related impacts associated the complete destruction of a ton of CH4 emissions through
flaring for this illustrative application, EPA took the difference between the SC-CO2 at the time of the flaring and
the discounted value of the CO2 impacts assuming a geometric decay of CH4 via the oxidation process with the same
e-folding time and near-term target discount rate as used to estimate the SC-CO2. See U.S. 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 EPA Report on the Social Cost of Greenhouse Gases: Estimates Incorporating Recent
Scientific Advances. Washington, DC: U.S. EPA for discussion. This value was then scaled by 44/16 to account for
the relative mass of carbon contained in a ton of CH4 versus a ton of CO2. More specifically, the impacts of shifting
where r is the year
the CO2 impacts are calculated as: |sC-C02T - £[=T e '!t TJ (l — e SC-C02t
the CH4 is destroyed, I is the CH4 e-folding time, r is the discount rate, and T is the time horizon of the analysis.
Ideally the time horizon, T. would be sufficiently long to capture the period in which nearly all of the CH4 is
expected to have been oxidized. In this analysis we use the 2100 as the time horizon, making the assumption that the
SC-CO2 remains constant after 2080, the last year for which updated SC-GHG estimates are presented in ibid. This
methodology improves upon the one presented at proposal by updating the oxidization process of CH4 to be
dynamic and consistent with the modeling that underlies the SC-CH4 estimates.
3-65
-------
values reflect an analytical time horizon of 2024 to 2038, are discounted to 2021, and presented
in 2019 dollars. Multiple benefits estimates are presented reflecting alternative discount rates.
The table includes consideration of the non-monetized benefits associated with the emissions
reductions projected under this final rule.Table 3-12 Table 3-12 and Table 3-13,
respectively,Table 3-13 present the same information for the final NSPS OOOOb and EG
OOOOc separately.
3-66
-------
Table 3-11 Comparison of PV and EAV of the Projected Benefits for the Final NSPS
OOOOb and EG OOOOc across Regulatory Options, 2024-2038 (millions of 2019$)a
2 Percent Near-Term Ramsey Discount Rate
PV
EAV
PV
EAV
PV
EAV
Climate Benefitsb
Less Stringent
$100,000
$7,800
$100,000
$7,800
$100,000
$7,800
Final Rule
$110,000
$8,500
$110,000
$8,500
$110,000
$8,500
More Stringent
$110,000
$8,500
$110,000
$8,500
$110,000
$8,500
2 Percent Discount
3 Percent Discount
7 Percent Discount
Rate
Rate
Rate
PV
EAV
PV
EAV
PV
EAV
Ozone Health Benefits0
Less Stringent
$6,200
$480
$5,400
$450
$3,100
$340
Final Rule
$7,000
$540
$6,100
$510
$3,500
$380
More Stringent
$7,000
$550
$6,100
$510
$3,500
$390
Total Monetized Benefits
Less Stringent
$110,000
$7,800
$100,000
$7,800
$100,000
$7,800
Final Rule
$110,000
$8,500
$110,000
$8,500
$110,000
$8,500
More Stringent
$110,000
$8,500
$110,000
$8,500
$110,000
$8,500
Non-Monetized Benefits
Benefits to provision of ecosystem services and ozone health benefits from reducing methane emissions by (in
short tons):
Less Stringent
Final Rule
More Stringent
54,000,000
58,000,000
59,000,000
Benefits to provision of ecosystem services from reducing VOC emissions by (in short tons)b:
Less Stringent
Final Rule
More Stringent
14,000,000
16,000,000
16,000,000
PM2 5-related health benefits from reducing VOC emissions by (in short tons)b:
Less Stringent
Final Rule
More Stringent
14,000,000
16,000,000
16,000,000
Benefits to provision of ecosystem services and HAP-related health benefits from reducing HAP emissions by (in
short tons):
Less Stringent 540,000
Final Rule 590,000
More Stringent 590,000
a Values rounded to two significant figures. Totals may not appear to add correctly due to rounding.
b Monetized climate benefits are based on reductions in methane emissions and are calculated using three different
estimates of the social cost of methane (SC-CH4) (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-
CH4 at the 2 percent near-term Ramsey discount rate. Please see Table 3.4 and 3.5 for the full range of monetized
climate benefit estimates.
3-67
-------
0 The ozone-related health benefits estimates use the larger of the two benefits estimates presented in Table 3-10.
Monetized benefits include those related to public health associated with reductions in ozone concentrations. The
health benefits are associated with several point estimates.
d Several categories of climate, human health, and welfare benefits from methane, VOC, and HAP emissions
reductions remain unmonetized and are thus not directly reflected in the quantified benefit estimates in the table. See
Section 3.2 for a discussion of climate effects that are not yet reflected in the SC-CH4 and thus remain unmonetized
and Sections 3.4 through 3.8 for a discussion of other non-monetized benefits.
3-68
-------
Table 3-12 Comparison of PV and EAV of the Projected Benefits for the Final NSPS
OOOOb across Regulatory Options, 2023-2035 (millions of 2019$)a
2 Percent Near-Term Ramsey Discount Rate
PV
EAV
PV
EAV
PV
EAV
Climate Benefitsb
Less Stringent
$43,000
$3,300
$43,000
$3,300
$43,000
$3,300
Final Rule
$44,000
$3,400
$44,000
$3,400
$44,000
$3,400
More Stringent
$44,000
$3,400
$44,000
$3,400
$44,000
$3,400
2 Percent Discount
3 Percent Discount
7 Percent Discount
Rate
Rate
Rate
PV
EAV
PV
EAV
PV
EAV
Ozone Health Benefits0
Less Stringent
N/A
N/A
N/A
N/A
N/A
N/A
Final Rule
N/A
N/A
N/A
N/A
N/A
N/A
More Stringent
N/A
N/A
N/A
N/A
N/A
N/A
Total Monetized Benefits
Less Stringent
$43,000
$3,300
$43,000
$3,300
$43,000
$3,300
Final Rule
$44,000
$3,400
$44,000
$3,400
$44,000
$3,400
More Stringent
$44,000
$3,400
$44,000
$3,400
$44,000
$3,400
Non-Monetized Benefits
Benefits to provision of ecosystem services and ozone health benefits from reducing methane emissions by (in
short tons):
Less Stringent
Final Rule
More Stringent
23,000,000
23,000,000
23,000,000
Benefits to provision of ecosystem services from reducing VOC emissions by (in short tons)b:
Less Stringent
Final Rule
More Stringent
6,900,000
7,100,000
7,100,000
PM2 5-related health benefits from reducing VOC emissions by (in short tons)b:
Less Stringent
Final Rule
More Stringent
6,900,000
7,100,000
7,100,000
Benefits to provision of ecosystem services and HAP-related health benefits from reducing HAP emissions by (in
short tons):
Less Stringent 260,000
Final Rule 270,000
More Stringent 270,000
a Values rounded to two significant figures. Totals may not appear to add correctly due to rounding.
b Monetized climate benefits are based on reductions in methane emissions and are calculated using three different
estimates of the social cost of methane (SC-CH4) (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-
CH4 at the 2 percent near-term Ramsey discount rate. Please see Table 3.4 and 3.5 for the full range of monetized
climate benefit estimates.
3-69
-------
0 The ozone-related health benefits estimates use the larger of the two benefits estimates presented in Table 3-10.
Monetized benefits include those related to public health associated with reductions in ozone concentrations. The
health benefits are associated with several point estimates.
d Several categories of climate, human health, and welfare benefits from methane, VOC, and HAP emissions
reductions remain unmonetized and are thus not directly reflected in the quantified benefit estimates in the table. See
Section 3.2 for a discussion of climate effects that are not yet reflected in the SC-CH4 and thus remain unmonetized
and Sections 3.4 through 3.8 for a discussion of other non-monetized benefits.
3-70
-------
Table 3-13 Comparison of PV and EAV of the Projected Benefits for the Final EG
OOOOc Across Regulatory Options, 2024-2038 (millions of 2019$)a
2 Percent Near-Term Ramsey Discount Rate
PV
EAV
PV
EAV
PV
EAV
Climate Benefitsb
Less Stringent
$57,000
$4,500
$57,000
$4,500
$57,000
$4,500
Final Rule
$65,000
$5,100
$65,000
$5,100
$65,000
$5,100
More Stringent
$66,000
$5,100
$66,000
$5,100
$66,000
$5,100
2 Percent Discount
3 Percent Discount
7 Percent Discount
Rate
Rate
Rate
PV
EAV
PV
EAV
PV
EAV
Ozone Health Benefits0
Less Stringent
N/A
N/A
N/A
N/A
N/A
N/A
Final Rule
N/A
N/A
N/A
N/A
N/A
N/A
More Stringent
N/A
N/A
N/A
N/A
N/A
N/A
Total Monetized Benefits
Less Stringent
$57,000
$4,500
$57,000
$4,500
$57,000
$4,500
Final Rule
$65,000
$5,100
$65,000
$5,100
$65,000
$5,100
More Stringent
$66,000
$5,100
$66,000
$5,100
$66,000
$5,100
Non-Monetized Benefits
Benefits to provision of ecosystem services and ozone health benefits from reducing methane emissions by (in
short tons):
Less Stringent
Final Rule
More Stringent
31,000,000
35,000,000
35,000,000
Benefits to provision of ecosystem services from reducing VOC emissions by (in short tons)b:
Less Stringent
Final Rule
More Stringent
7,500,000
8,600,000
8,700,000
PM2 5-related health benefits from reducing VOC emissions by (in short tons)b:
Less Stringent
Final Rule
More Stringent
7,500,000
8,600,000
8,700,000
Benefits to provision of ecosystem services and HAP-related health benefits from reducing HAP emissions by (in
short tons):
Less Stringent 280,000
Final Rule 320,000
More Stringent 330,000
a Values rounded to two significant figures. Totals may not appear to add correctly due to rounding.
b Monetized climate benefits are based on reductions in methane emissions and are calculated using three different
estimates of the social cost of methane (SC-CH4) (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-
CH4 at the 2 percent near-term Ramsey discount rate. Please see Table 3.4 and 3.5 for the full range of monetized
climate benefit estimates.
3-71
-------
0 The ozone-related health benefits estimates use the larger of the two benefits estimates presented in Table 3-10.
Monetized benefits include those related to public health associated with reductions in ozone concentrations. The
health benefits are associated with several point estimates.
d Several categories of climate, human health, and welfare benefits from methane, VOC, and HAP emissions
reductions remain unmonetized and are thus not directly reflected in the quantified benefit estimates in the table. See
Section 3.2 for a discussion of climate effects that are not yet reflected in the SC-CH4 and thus remain unmonetized
and Sections 3.4 through 3.8 for a discussion of other non-monetized benefits.
3.11 References
AHRQ. (2016). Healthcare Cost and Utilization Project (HCUP).
http s ://www. ahrq. gov/data/hcup/index. html
Anthoff, D., & Tol, R. S. J. (2013a). Erratum to: The uncertainty about the social cost of carbon:
A decomposition analysis using FUND (vol I 17, pg 515, 2013). Climatic Change, 121(2),
413-413. https://doi org/10.1007/s10584-013-0959-1
Anthoff, D., & Tol, R. S. J. (2013b). The uncertainty about the social cost of carbon: A
decomposition analysis using FUND. Climatic Change, 117(3), 5 15-530.
https://doi.org/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.
https://doi.org/doi: 10.1 126/science. 1235665
ATSDR. (1999). Toxicological Profile for Formaldehyde.
https://vvvvvv.atsdr.cdc.gov/ToxProfiles/tp I I I pdf
ATSDR. (2007a). Toxicological Profile for Benzene.
https://vvvvvv.atsdr.cdc.gov/toxprofiles/tp3.pdf
ATSDR. (2007b). Toxicological Profile for Xylene.
https://vvvvvv.atsdr.cdc.gOv/ToxProfiles/tp71 pdf
Auffhammer, M. (2018). Quantifying Economic Damages from Climate Change. Journal of
Economic Perspectives, 32(4), 33-52. https://doi.org/10.1257/jep.32.4.33
Bauer, M. D., & Rudebusch, G. D. (2020). Interest Rates under Falling Stars. American
Economic Review, 110(5), 13 16-1354. https://doi.org/! 0.1257/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.
https://doi.org/10.1 162/rest_a_01 109
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. https://doi.org/! 0.1097/01 ede.OOOO 165817.40152.85
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.
https://doi.org/! 0.100 l/jama.292.19.2372
3-72
-------
Cappa, C. D., & Wilson, K. R. (2012). Multi-generation gas-phase oxidation, equilibrium
partitioning, and the formation and evolution of secondary organic aerosol. Atmospheric
Chemistry and Physics, 12(20), 9505-9528. https://doi.org/10.5 194/acp-12-9505-2012
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.
https://doi.org/10.1093/qje/qjac020
Climate Impact Lab (CIL). (2023). Data-driven Spatial Climate Impact Model User Manual,
Version 092023-EPA. Retrieved from https://impactlab.org/vvp-
content/uploads/2023/10/DSCIM_UserManual_V ersion092023-EPA.pdf
Cropper, M. L., Freeman, M. C., Groom, B., & Pizer, W. A. (2014). Declining Discount Rates.
American Economic Review, 104(5), 538-543. https://doi.org/IO. 1257/aer. 104.5.538
Donahue, N. M., Kroll, J. H., Pandis, S. N., & Robinson, A. L. (2012). A two-dimensional
volatility basis set - Part 2: Diagnostics of organic-aerosol evolution. Atmospheric Chemistry
and Physics, 12(2), 615-634. https://doi org/10.5 194/acp-12-615-2012
EPA Science Inventory. (2021). Technical Documentation on The Framework for Evaluating
Damages and Impacts (FrEDl). Retrieved February 16, 2023 from
https://cfpub.epa.gov/si/si_public_record_report.cfm9dirEntryId=35 13 16&Lab=OAP&simpl
esearch=0&showcriteria=2&sortby=pubDate&searchall=fredi&timstype=&datebeginpublish
edpresented=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, I-32. https://doi.Org/l0.5 l94/egusphere-2023-l 14
Helmig, D. (2020). Air quality impacts from oil and natural gas development in Colorado.
Elementa: Science of the Anthropocene, 8. https://doi .org/! 0.1525/elementa.398
Helmig, D., Thompson, C. R., Evans, J., Boylan. P., Hueber, J., & Park, J. H. (2014). Highly
Elevated Atmospheric Levels of Volatile Organic Compounds in the Uintah Basin, Utah.
Environmental Science & Technology, 48(9), 4707-4715. https://doi.org/! 0.102 l/es405046r
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), 53 I -543.
https ://doi. org/10.1007/s 10584-012-0633-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.
https ://doi. org/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.
https://doi.org/10.1002/env.721
3-73
-------
I ARC. (1982). Some Industrial Chemicals and Dyestuffs (Volume 29). (Monographs on the
Evaluation of Carcinogenic Risk of Chemicals to Humans, Issue.
https://publications.iarc.fr/47
I ARC. (1987). Overall Evaluations of Carcinogenicity: An Updating of I ARC Monographs
Volumes I to 42 (Supplement 7). (Monographs on the Evaluation of Carcinogenic Risk of
Chemicals to Humans, Issue, https://publications.iarc.fr/139
I ARC. (2000). Some Industrial Chemicals (Volume 77). (Monographs on the Evaluation of
Carcinogenic Risk of Chemicals to Humans, Issue, https://publications.iarc.fr/95
I ARC. (2006). Formaldehyde, 2-Butoxyethanol and l-tert-Butoxypropan-2-ol (Volume 88).
(Monographs on the Evaluation of Carcinogenic Risk of Chemicals to Humans, Issue.
https://publications.iarc.fr/106
I ARC. (2012). Chemical Agents and Related Occupations (Volume I OOF). (Monographs on the
Evaluation of Carcinogenic Risk of Chemicals to Humans, Issue,
https ://publ i cati ons. i arc, fr/123
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. Pi rani, W. Moufouma-Okia, C. Pean, R.
Pidcock, S. Connors, J. B. R. Matthews, Y. Chen, X. Zhou, M. I. Gomis, E. Lonnoy, T.
May cock, a. M. Tignor, & T. Waterfield, Eds), https://www.ipcc.ch/srl5/download/
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. Pi rani, S.L. Connors, C. Pean,
S. Berger, N. Caud, Y. Chen, L. Goldfarb, M.I. Gomis, M. Huang, K. Leitzell, E. Lonnoy,
J.B.R. Matthews, T.K. May cock, T. Waterfield, O. Yelekgi, R. Yu, & B. Zhou (Eds),
Climate Change 2021: The Physical Science Basis (pp. 923-1054). Cambridge University
Press. https://doi org/10.1017/9781009157896.009.
IPCC. (2021b). Short-Lived Climate Forcers. In Climate Change 2021: The Physical Science
Basis. Cambridge, UK; New York, NY: IPCC
Irons, R. D., Stillman, W. S., Colagiovanni, D. B., & Henry, V. A. (1992). Synergistic action of
the benzene metabolite hydroquinone on myelopoietic stimulating activity of
granulocyte/macrophage colony-stimulating factor in vitro. Proceedings of the National
Academy of Sciences, 89(9), 3691 -3695. https://doi.org/10.1073/pnas.89.9.3691
Ito, K., De Leon, S. F., & Lippmann, M. (2005). Associations between ozone and daily
mortality: analysis and meta-analysis. Epidemiology, 16(4), 446-457.
https://doi.org/10.1097/01 ede.OOOO 165821.901 14.7f
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 Retrieved from
https://www.epa.gov/sites/default/files/2016-12/documents/scc_tsd_2010.pdf
3-74
-------
IWG. (2013). Technical Support Document: Technical Update of the Social Cost of Carbon for
Regulatory Impact Analysis Under Executive Order 12866. U.S. Government, Interagency
Working Group (IWG) on Social Cost of Carbon.
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 Retrieved from https://vvvvvv.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
Retrieved from https://www.epa.gov/sites/default/files/20l6-
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 Retrieved from
https://www.whitehouse.gov/wp-
content/uploads/2021 /02/T echnical SupportDocumentSocialCostofCarbonMethaneNitrousO
xide.pdf?source=email
Jerrett, M., Burnett, R. T., Pope, C. A., Ito, K., Thurston, G., Krewski, D., . . . Thun, M. (2009).
Long-Term Ozone Exposure and Mortality. New England Journal of Medicine, 360(11),
1085-1095. https://doi.Org/l 0.1056/NEJMoa0803894
Jimenez, J. L., Canagaratna, M. R., Donahue, N. M., Prevot, A. S. H., Zhang, Q., Kroll, J. H., . . .
Worsnop, D. R. (2009). Evolution of Organic Aerosols in the Atmosphere. Science,
326(5959), 1525-1529. https://doi.org/IO I 126/science. I 180353
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 ( APHEN A). Res Rep Health Eff lnst( 142), 5-90.
https://www.ncbi.nlm.nih.gov/pubmed/20073322
Kemball-Cook, S., Bar-llan, A., Grant, J., Parker, L., Jung, J., Santamaria, W., . . . Yarwood, G.
(2010). Ozone Impacts of Natural Gas Development in the Haynesville Shale. Environmental
Science & Technology, 44(24), 9357-9363. https://doi.org/IO. 1021/es1021 137
Kivi, P. A., & Shogren, J. F. (2010). Second-Order Ambiguity in Very Low Probability Risks:
Food Safety Valuation. Journal of Agricultural and Resource Economics, 35(3), 443-456.
http ://www. j stor.org/ stabl e/23 243 065
Koss, A., Yuan, B., Warneke, C., Oilman, J. B., Lenier, B. M., Veres, P. R., . . . de Gouw, J.
(2017). Observations of VOC emissions and photochemical products over US oil- and gas-
producing regions using high-resolution H30+ CIMS (PTR-ToF-
MS). Atmospheric Measurement Techniques, 10(8), 2941-2968. https://doi.org/10.5 194/amt-
10-2941-2017
3-75
-------
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.
https://doi org/10.1097/01 ede.OOOO 165820.08301 b3
Lindaas, J., Farmer, D. K., Pollack, 1. B., Abeleira, A., Flocke, F., & Fischer, E. V. (2019). Acyl
Peroxy Nitrates Link Oil and Natural Gas Emissions to High Ozone Abundances in the
Colorado Front Range During Summer 2015. Journal of Geophysical Research:
Atmospheres, 124(4), 2336-2350. https://doi.Org/https://doi.org/10.1029/2018JD028825
Lvu, C., Capps, S. L., Kurashima, K., Henze, D. K., Pierce, G., Hakami, A., . . . Milford, J.
(2021). Evaluating oil and gas contributions to ambient nonmethane hydrocarbon mixing
ratios and ozone-related metrics in the Colorado Front Range. Atmospheric Environment,
246, I 181 13. https://doi org/https://doi org/10.1016/j atmosenv.2020.1 181 13
Maltoni, C., Ciliberti, A., Pinto, C., Soffritti, M., Belpoggi, F., & Menarini, L. (1997). Results of
long-term experimental carcinogenicity studies of the effects of gasoline, correlated fuels,
and major gasoline aromatics on rats. Annals of the New York Academy of Sciences, 837,
15-52. https://doi.org/IO I I I l/j. 1749-6632.1997.tb56863.x
Maltoni, C., Conti, B., Cotti, G., & Belpoggi, F. (1985). Experimental studies on benzene
carcinogenicity at the Bologna Institute of Oncology: current results and ongoing research.
American Journal of Industrial Medicine, 7(5-6), 415-446.
https://doi.org/10.1002/ajim.4700070508
McDuffie, E. E., Edwards, P. M., Oilman, J. B., Lerner, B. M., Dube, W. P., Trainer, M., . . .
Brown, S. S. (2016). Influence of oil and gas emissions on summertime ozone in the
Colorado Northern Front Range. Journal of Geophysical Research: Atmospheres, 121( 14),
8712-8729. https://doi org/https://doi org/10.1002/2016JD025265
McDuffie, E. E., Sarofim, M. C., Raich, W., Jackson, M., Roman, H., Seltzer, K., . . . Fann, N.
(2023). The Social Cost of Ozone-Related Mortality Impacts From Methane Emissions.
Earth's Future, 11(9), e2023EF003853. https://d0i.0rg/https://d0i.0rg/10.1029/2023EF003853
McGartland, A., Revesz, R., Axelrad, D. A., Dockins, C., Sutton, P., & Woodruff, T. J. (2017).
Estimating the health benefits of environmental regulations. Science, 357(6350), 457-458.
https://doi.org/doi: 10.1 126/science.aam8204
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. https://doi.org/10.5194/acp-l 7-7213-2017
Mvhre, G., Shindell, D., Breon, F.-M., Collins, W., Fuglestvedt, J., Huang, J , . . . Zhang, H.
(2013). Anthropogenic and Natural Radiative Forcing. In T. Case (Ed ), Climate Change
2013: The Physical Science Basis. Contribution of Working Group I to the Fifth Assessment
Report of the Intergovernmental Panel on Climate Change. Cambridge University Press.
https://vvvvvv.ipcc.ch/report/ar5/vvgl/anthropogenic-and-natural-radiative-forcing/
3-76
-------
National Academies. (2017). Valuing Climate Damages: Updating Estimation of the Social Cost
of Carbon Dioxide. The National Academies Press.
https://vvvvvv.nap.edu/catalog/2465 I/valuing-climate-damages-updating-estimation-of-the-
social-cost-of
National Research Council. (2008). Estimating Mortality Risk Reduction and Economic Benefits
from Controlling Ozone Air Pollution. In. National Academies Press (US),
https://doi.org/10.17226/12198
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. https://doi org/10.1086/718145
Nolte, C. G., Dolwick, P. D., Fann, N., Horowitz, L. W., Naik, V., Pinder, R. W., . . . Ziska, L.
H. (2018). Chapter 13: Air Quality. In C. Test (Ed ), Impacts, Risks, and Adaptation in the
United States: Fourth National Climate Assessment (Vol. II, pp. 512-538). U.S. Global
Change Research Program. https://nca2018.globalchange.gov/chapter/13/
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), I 1721-I 1726. https://doi org/10.1073/pnas. 1005985 107
Nordhaus, W. D. (2014). Estimates of the Social Cost of Carbon: Concepts and Results from the
DICE-20I3R Model and Alternative Approaches. Journal of the Association of
Environmental and Resource Economists, 1(1/2), 273-3 12. https://doi.org/! 0.1086/676035
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. http://www.nber.org/papers/vv23646
NTP. (1999). Toxicology and Carcinogenesis Studies of Ethylbenzene (CAS No. 100-41 -4) in
F344/N Rats and B6C3FI Mice (Inhalation Studies). National Toxicology Program
Technical Report Series, 466, I -23 I.
NTP. (2004). Report on Carcinogens, Eleventh Edition (PB2005 104914).
https://ntrl.ntis.gov/NTRL/dashboard/searchResults/titleDetail/PB2005 104914.xhtml
NTP. (2016). Report on Carcinogens, Fourteenth Edition, https://ntp.niehs.nih.gov/go/roc 14
OMB. (2003). Circular A-4: Regulatory Analysis. Washington DC. Retrieved from
https://obamawhitehouse.archives.gov/omb/circulars_a004_a-4/
OMB. (2023). Circular A-4: Regulatory Analysis. Washington DC. Retrieved from
https://www.vvhitehouse.gOv/vvp-content/uploads/2023/l I/CircularA-4.pdf
Petron, G., Frost, G., Miller, B. R., Hirsch, A. I., Montzka, S. A., Karion, A., . . . Tans, P. (2012).
Hydrocarbon emissions characterization in the Colorado Front Range: A pilot study. Journal
of Geophysical Research: Atmospheres, I 17(D4). https://doi.org/10.1029/201 IJDO16360
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. Retrieved from
https://dovvnloads.regulations.gov/EPA-HQ-OAR-2018-0283-6184/attachment_6.pdf
3-77
-------
Pozzer, A., Schultz, M. G., & Helmig, D. (2020). Impact of U.S. Oil and Natural Gas Emission
Increases on Surface Ozone Is Most Pronounced in the Central United States. Environmental
Science & Technology, 54(19), 12423-12433. https://doi org/10.1021 /acs.est.9506983
Ramsey, F. P. (1928). A Mathematical Theory of Saving. The Economic Journal, 38( 152), 543-
559. https://doi.org/l0.2307/2224098
Reddy, P. J. (2023). Synthesis of Satellite and Surface Measurements, Model Results, and
FR APPE Study Findings to Assess the Impacts of Oil and Gas Emissions Reductions on
Maximum Ozone in the Denver Metro and Northern Front Range Region in Colorado. Earth
and Space Science, 10(5), e2023EA002917.
https://doi.Org/https://doi.org/l 0.1029/2023EA002917
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. https://d0i.0rg/l 0.1038/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.
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. https://doi.org/! 0.1038/s41586-021 -03883-8
Rose, S., Turner, D., Blanford, G., Bistline, J., de la Chesnaye, F., & Wilson, T. (2014).
Understanding the Social Cost of Carbon: A Technical Assessment. EPR1 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, I I (5), 5 16.
https://www.mdpi.eom/2073-4433/l 1/5/516
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( I), 22. https://doi org/10.1007/s 10584-021 -03048-6
Sarofim, M. C., Waldhoff, S. T., & Anenberg, S. C. (2017). Valuing the ozone-related health
benefits of methane emission controls. Environmental and Resource Economics, 66(1), 45-
63.
Schwartz, J. (2005). How sensitive is the association between ozone and daily deaths to control
for temperature9 Am J Respir Crit Care Med, 171 (6), 627-63 I.
https://doi.org/10.1 164/rccm.200407-9330C
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. https://doi.org/! 0.5 194/gmd-1 I -2273-2018
3-78
-------
Smith, R. L., Xu, B., & Svvitzer, P. (2009). Reassessing the relationship between ozone and
short-term mortality in US urban communities. Inhalation toxicology, 2l(sup2), 37-61.
Tol, R. (2009). An analysis of mitigation as a response to climate change. Copenhagen
Consensus on Climate. Copenhagen Consensus Center Retrieved from
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.
https://doi.org/10.1 164/rccm.201508-16330C
Tzompa-Sosa, Z. A., & Fischer, E. V. (2021). Impacts of Emissions of C2-C5 Alkanes From the
U.S. Oil and Gas Sector on Ozone and Other Secondary Species. Journal of Geophysical
Research: Atmospheres, 126( I), e2019JD03 1935.
https://d0i.0rg/https://d0i.0rg/l 0.1029/2019JD03 1935
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
U.S. BEA. (2004). New BE A Economic Areas For 2004. Washington DC Retrieved from
https://vvvvvv.bea.gov/nevvs/2004/nevv-bea-economic-areas-2004
U.S. BEA. (2021). Table 1. 1.9. Implicit Price Deflators for Gross Domestic Product.
Washington, DC Retrieved from
https://apps.bea.gov/iTable/9reqid= 19&step=3&isuri= I & 1921 =survey& 1903= 13
U.S. EPA. (1991a). Carbonyl sulfide; CASRN 463-58-1 [Chemical Assessment
Summary](Integrated Risk Information System (IRIS), Issue.
https://iris.epa.gov/ChemicalLanding/&substance_nmbr=617
U.S. EPA. (1991b). Formaldehyde; C ASRN 50-00-0 [Chemical Assessment
Summary](Integrated Risk Information System (IRIS), Issue.
https://iris.epa.gov/ChemicalLanding/&substance_nmbr=419
U.S. EPA. (2003a). Benzene; C ASRN 71-43-2 [Chemical Assessment Summary](Integrated
Risk Information System (IRIS), Issue.
https://iris.epa.gov/ChemicalLanding/&substance_nmbr=276
U.S. EPA. (2003b). Xylenes; C ASRN 1330-20-7 [Chemical Assessment Summary](Integrated
Risk Information System (IRIS), Issue.
https://iris.epa.gov/Chemical Landing/&substance_nmbr=270
U.S. EPA. (2005a). Guidelines for Carcinogen Risk Assessment (EPA-630-P-03-001F).
https://vvvvvv.epa.gOv/sites/default/files/20l3-09/documents/cancer_guidelines_final_3-25-
05.pdf
U.S. EPA. (2005b). Toluene; C ASRN 108-88-3 [Chemical Assessment Summary](Integrated
Risk Information System (IRIS), Issue.
https://iris.epa.gov/Chemical Landing/&substance_nmbr= I 18
3-79
-------
U.S. EPA. (2006). Regulatory Impact Analysis, 2006 National Ambient Air Quality Standards
for Particulate Matter, https://www.epa.gov/sites/default/files/2020-07/documents/naaqs-
pm_ria_final_2006-10.pdf
U.S. EPA. (2009). 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 Retrieved from
https://cfpub. epa.gov/ncea/risk/recordisplay. cfm?deid=216546
U.S. EPA. (201 la). The Benefits and Costs of the Clean Air Act from 1990 to 2020.
Washington, DC: U.S. EPA, Office of Air and Radiation Retrieved from
https://www.epa.gov/sites/default/files/20l 5-07/documents/fullreport_rev_a.pdf
U.S. EPA. (201 lb). Draft Residual Risk Assessment for the Oil and Gas Production and Natural
Gas Transmission and Storage Source Categories.
https://wvvvv.regulations.gOv/document/EPA-HQ-OAR-20l0-0505-0032
U.S. EPA. (201 Ic). Regulatory Impact Analysis for the Federal Implementation Plans to Reduce
Interstate T ran sport of Fine Particulate Matter and Ozone in 27 States; Correction of SIP
Approvals for 22 States. Research Triangle Park, NC: U.S. Environmental Protection
Agency, Office of Air Quality Planning and Standards, Health and Environmental Impact
Division Retrieved from https://wvvw3.epa.gov/ttn/ecas/docs/ria/transport_ria_final-
csapr_201 I -06.pdf
U.S. EPA. (2012a). Regulatory Impact Analysis for the Final Revisions to the National Ambient
Air Quality Standards for Particulate Matter. (EPA-452/R-I2-005). Research Triangle Park,
NC: U.S. Environmental Protection Agency, Office of Air Quality Planning and Standards,
Health and Environmental Impact Division Retrieved from
https://vvvvvv3.epa.gov/ttnecasl/regdata/RlAs/finalria.pdf
U.S. EPA. (2012b). Regulatory Impact Analysis for the Proposed Revisions to the National
Ambient Air Quality Standards for Particulate Matter. (EPA-452/R-12-003). Research
Triangle Park, NC: U.S. Environmental Protection Agency, Office of Air Quality Planning
and Standards, Health and Environmental Impact Division
U.S. EPA. (2013). Integrated Science Assessment of Ozone and Related Photochemical Oxidants
(Final Report) (EPA-600-R-10-076F).
https://cfpub.epa.gOv/ncea/i sa/recordisplay.cfm?deid=247492
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 Retrieved from https://vvvvvv3 epa.gov/ttnecas I /docs/2015 1001 ria.pdf
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 Retrieved from https://www.epa.gov/naaqs/particulate-matter-pm-standards-
integrated-science-assessments-current-review
3-80
-------
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 Retrieved from https://www.epa.gov/sites/production/files/20l9-
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 Retrieved from
https://vvvvvv.epa.gOv/sites/default/files/2020-04/documents/mats_coal_refuse_cost-
b en efi t_m em opdf
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 Retrieved
from https://www.epa.gov/sites/default/files/2020-
08/documents/steam_electric_elg_2020_fmal_reconsideration_rule_benefit_and_cost_analys
is. pdf
U.S. EPA. (2020c). Integrated Science Assessment (ISA) for Ozone and Related Photochemical
Oxidants (Final Report). (EPA/600/R-20/012). Washington DC: U.S. Environmental
Protection Agency Retrieved from
https://cfpub.epa.gOv/ncea/i sa/recordisplay.cfrn?deid=348522
U.S. EPA. (2021a). Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019 (EPA-
430-R-2I-005). https://www.epa.gov/ghgemissions/inventory-us-greenhouse-gas-emissions-
and-sinks-1990-2019
U.S. EPA. (202 lb). 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 Retrieved from
https://www.epa.gov/sites/default/files/2021 -
03/documents/revisedcsaprupdateriafinal pdf
U.S. EPA. (2021c). Technical Documentation on the Framework for Evaluating Damages and
Impacts (FrEDl). (EPA 430-R-21-004). Retrieved from 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 Retrieved from
https://www.epa.gov/benmap/benmap-ce-manual-and-appendices
3-81
-------
U.S. EPA. (2022b). Regulatory Impact Analysis for Proposed Federal Implementation Plan
Addressing Regional Ozone T ran sport 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 Retrieved from https://www.epa.gov/system/files/documents/2022-
03/transport_ria_proposal_fip_2015_ozone_naaqs_2022-02.pdf
U.S. EPA. (2023a). 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).
U.S. EPA. (2023b). 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 Retrieved from
https://www.epa.gov/system/files/documents/2023-0l/Estimating%20PM2.5-
%20and%200zone-Attributable%20Health%20Benefits%20TSD_0.pdf
U.S. EPA. (2023c). Regulatory Impact Analysis for the Final Federal Good Neighbor Plan
Addressing Regional Ozone T ran sport 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 Retrieved from https://www.epa.gov/system/files/documents/2023-
03/SAN%208670%20Federal%20Good%20Neighbor%20Plan%2020230315%20RIA_Final.
pdf
U.S. EPA. (2023d). 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 Retrieved from
https://www.epa.gov/system/files/documents/2023-
04/MATS°o20RTR°o20Proposal°o20Rl A0 o20Formatted.pdf
U.S. EPA. (2023e). Regulatory Impact Analysis for the Proposed New Source Performance
Standards for Greenhouse Gas Emissions from New, Modified, and Reconstructed Fossil
Fuel-Fired Electric Generating Units; Emission Guidelines for Greenhouse Gas Emissions
from Existing Fossil Fuel-Fired Electric Generating Units; and Repeal of the Affordable
Clean Energy Rule. (EPA-452/R-23-006). Research Triangle Park, NC: U.S. Environmental
Protection Agency, Office of Air Quality Planning and Standards, Health and Environmental
Impact Division Retrieved from https://www.epa.gov/system/files/documents/2023-
05/uti 1 ities ria proposal 2023-05.pdf
U.S. 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":
EPA Report on the Social Cost of Greenhouse Gases: Estimates Incorporating Recent
Scientific Advances. Washington, DC: U.S. EPA
3-82
-------
U.S. EPA Science Advisory Board. (2019). Letter from Louis Anthony Cox, Jr., Chair, Clean Air
Scientific Advisory Committee, to Administrator Andrew R. Wheeler. Re: CAS AC Review
of the EPA's Integrated Science Assessment for Particulate Matter (External Review Draft -
October 2018). (EPA-CASAC-19-002). Washington DC
U.S. EPA Science Advisory Board. (2020a). Letter from Louis Anthony Cox, Jr., Chair, Clean
Air Scientific Advisory Committee, to Administrator Andrew R. Wheeler. Re: CAS AC
Review of the EPA's Integrated Science Assessment for Ozone and Related Photochemical
Oxidants (External Review Draft - September 2019). (EPA-CASAC-20-002). Washington
DC
U.S. EPA Science Advisory Board. (2020b). Letter from Michael Honeycutt Chair, Scientific
Advisory, to Administrator Lisa Jackson. Re: Science Advisory Board (SAB) Consideration
of the Scientific and Technical Basis of EPA's Proposed Rule titled "Increasing Consistency
and Transparency in Considering Benefits and Costs in the Clean Air Act Rulemaking
Process ". (EPA-SAB-20-012). Washington DC
United Nations Environment Programme and Climate and Clean Air Coalition. (2021). Global
Methane Assessment: Benefits and Costs of Mitigating Methane Emissions. Nairobi, Kenya
USGCRP. (2018). Impacts, Risks, and Adaptation in the United States: Fourth National Climate
Assessment, Volume 11. Washington DC: U.S. Global Change Research Program Retrieved
from http://dx.doi.org/10.7930/NC A4.2018
WHO. (2002). Concise International Chemical Assessment Document 40: Formaldehyde.
https://vvvvvv.vvho.int/ipcs/publications/cicad/en/cicad40.pdf
Woods & Poole. (2015). Complete Demographic Database, https://vvvvvv.vvoodsandpoole.com/
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. https://doi.org/IO. I 164/rccm.200706-8230C
3-83
-------
4 ECONOMIC IMPACT AND DISTRIBUTIONAL ANALYSIS
The final NSPS OOOOb and EG 0000c constitute an economically significant action
under E.O. 12866. As discussed in previous section, the emissions reductions projected under the
rule are likely to produce substantial climate benefits and ozone-related health benefits as well as
non-monetized benefits from large reductions in emissions of multiple pollutants. At the same
time, the NSPS OOOOb and EG 0000c is projected to result in substantial environmental
control expenditures by the oil and natural gas industry to comply with the rule.
While the national-level impacts demonstrate the final rule is likely to lead to significant
benefits and costs, the benefit-cost analysis does not speak directly to potential economic and
distributional impacts of the rule, which may be important consequences of the action. This
section includes five sets of economic impact and distributional analyses directed toward
complementing the benefit-cost analysis and includes an analysis of potential national-level
impacts on oil and natural gas markets, a financial analysis of marginal oil and natural gas wells,
a series of environmental justice analyses, employment impacts, and a Final Regulatory
Flexibility Analysis (FRF A) that includes an analysis of projected compliance costs of the final
NSPS OOOOb on small entities.
4.1 Oil and Natural Gas Market Impact Analysis
In addition to the engineering cost analysis that produces the compliance cost and
emissions reduction projections that inform the net benefits analysis, the EPA developed a pair
of single-market, static partial-equilibrium analyses of national crude oil and natural gas markets.
The market impact analyses are intended to provide readers some information on the economic
impacts of the final NSPS OOOOb and EG 0000c and to inform the EPA's response to EO
13211 "Actions Concerning Regulations that Significantly Affect Energy Supply, Distribution,
or Use." The partial equilibrium market impact estimates, however, do not inform the projected
engineering costs and emissions reductions used in the comparison of benefits and costs.
Our partial equilibrium analyses treat crude oil markets and natural gas markets
separately. We implement a pair of single-market analyses instead of a coupled market or
general equilibrium approach to provide broad insights into potential national-level market
impacts while providing analytical transparency.
4-1
-------
The oil market model assumes a single, aggregate U.S supplier, a single, aggregate world
consumer, and a residual world supply. We assume the U.S. supply response to a percentage
change in costs has the same effect as a percentage change in price. We do not try to model the
residual world supply precisely. Instead, we model two extreme cases — perfectly inelastic
residual world supply and perfectly elastic residual world supply. These cases bound the residual
world supply response.
The natural gas market model assumes a single, aggregate U.S. supplier, a single,
aggregate U.S. consumer, and no international trade. We assume the U.S. supply response to a
percentage change in costs has the same effect as a percentage change in price. Existing natural
gas markets are segmented in the short-term by transmission constraints, but prices are
cointegrated across the United States. Infrastructure, including new infrastructure in the long
term, joins disparate markets. The assumption of a single natural gas market is a long-term
modeling assumption.
In each market, we first use a supply elasticity to solve for the supply change that results
from the imposition of regulatory costs. Given the change in supply, we then use a demand
elasticity to solve for the change in price that balances supply and demand. We use projected
crude oil and natural gas prices and production for a select set of years of analysis to
operationalize the model. In the sections that follow, we discuss the data and parameters used to
implement the models, present results of each analysis, and conclude with a discussion of caveats
and limitations of the analyses.
4.1.1 Crude Oil Market Model
The crude oil market model is a constant elasticity model that assumes a competitive U.S.
market with a rest of world residual oil supply that is either perfectly inelastic or perfectly
elastic. To find the changes in crude oil production and prices under the NSPS OOOOb and EG
0000c, we first solve for the change in production using a supply elasticity and the regulatory
cost. The year t change in U.S. oil production Ais estimated using Eq. 4-1:
AQgst = T$jr- * « C. El- 4"!
^o,t Fo,t
4-2
-------
where Co t is the projected regulatory cost impacting oil-producing sources in year /, Qo,t is the
baseline U.S. crude oil production in year t, Po t is the baseline crude oil price, and eo s is the
supply elasticity of crude oil. The term — describes the cost change as a fraction of
revenue, akin to a percentage change in price. A key modeling assumption here is that, in
addition to a constant elasticity, a fractional change in revenue due to a cost change is equivalent
to a fractional change in output price. The term — * Eo,s then describes the fractional
change in production.
For the model assuming perfectly inelastic rest-of-world production, we use the change in
supply solved in Eq. 4-1 the find the change in crude oil prices using Eq. 4-2:
aq^s i
APo.t = 0worm * ~ *Po,t, Eq. 4-2
Qo,t eO,D
where Q™°rld is global production of crude oil and eo d is the world demand elasticity for crude
oil.
Price does not change in the alternative model; it assumes perfectly elastic rest-of-world
production, so APo t = 0.
4.1.2 Natural Gas Market Model
We model U.S. natural gas supply and demand as a closed market. For the natural gas
market, we first find the change in quantity produced AQGt using Eq 4.-3:
A(?G,t = 7 * £g,s * Qg,t, Eq. 4-3
Qc,t*Pc,t
where CGt is the projected regulatory cost impacting all segments of the natural gas industry in
year t, QG t is the baseline U.S. production forecast, PG t is the natural gas price forecast, and eg s
is the supply elasticity for natural gas.
We then use the change in quantity solved in Eq. 4.3 to solve for the natural gas price
change APG t using Eq. 4-4:
APGit=^*j-*PGit Eq. 4-4
Qc,t SC,D
4-3
-------
4.1.3 Assumptions, Data, and Parameters Used in the Oil and Natural Gas Market Models
This section presents the basics assumptions applied in this analysis. The section also
presents the data and parameters used to operationalize the model, including our choice of years
of analysis, elasticity estimates, and production and price data.
4.1.3.1 Years of Analysis
We estimate the price and quantity impacts of the final NSPS OOOOb and EG 0000c
on crude oil and natural gas markets for a subset of years within the time horizon analyzed in this
RIA. We analyze 2024 and 2027 as these years represent the first and last year the requirements
in the final NSPS OOOOb will be in effect for the purposes of the RIA before the requirements
of EG 0000c are assumed to go into effect. We then analyze market impacts in 2028, 2033,
and 2038 to examine the effects of the EG 0000c in addition to the cumulative impacts of the
NSPS OOOOb. We analyze 2033 and 2038 to project impacts in later years of the time horizon.
4.1.3.2 Elasticity Choices
The elasticity estimates used in the analysis are based on estimates from the published
economics literature (). Natural gas demand elasticity is calculated as the sector-level
consumption-weighted average of demand elasticities from Hausman and Kellogg (2015). The
consumption proportions used to weight the elasticities are derived from 2019 levels of natural
consumption by the residential, commercial, industrial, and electric power sectors, as reported in
EIA.
4-4
-------
Table 4-1 Parameters Used in Market Analysis
Parameter
Symbol
Value
Source
Oil supply
elasticity
Newell, R. G.. & B. C. Prest. 2019. The unconventional oil supply
£o,s
1.2
boom: Aggregate price response from microdata. The Energy
Journal 40(3).
Oil demand
elasticity
Coglianese , J., L. W. Davis, L. Kilian. & J. H. Stock. 2017.
£o,d
-0.37
Anticipation, tax avoidance, and the price elasticity of gasoline
demand. Journal of Applied Econometrics 32(1): 1-15.
Natural gas
supply
elasticity
Newel 1, R. G.. B. C. Prest, & A. B. Vissing. 2019. Trophv hunting
p „ „
0.9
versus manufacturing energy: The price responsiveness of shale
gas." Journal of the Association of Environmental and Resource
Economists 6(2): 391-431.
Natural gas
demand
elasticity
Sector-level consumption-weighted average of demand elasticities from
eG,D
-0.43
Hausman, C. & R. Kellogg. 2015. Welfare and Distributional
Implications of Shale Gas. Brookings Papers on Economic Activity:! 1-
125.
4.1.3.3 Production and Price Data
Baseline U.S. crude oil production, dry gas production, West Texas Intermediate (WTI)
crude oil prices, and Henry Hub natural gas prices are drawn from AEO2022. Prices are deflated
to 2019 dollars using the GDP-Implicit Price Deflator. As the NSPS OOOOb and EG 0000c
apply to onshore production but not offshore production, only onshore U.S. crude oil production
is analyzed. Dry natural gas production is the sum of onshore production from the lower 48
states and all production from Alaska. Baseline world crude oil production is from the Energy
Information Administration's 2021 International Energy Outlook, presents the baseline crude oil
and natural gas production and prices used in the market impacts analysis.
Table 4-2 Baseline Crude Oil and Natural Gas Production and Prices Used in Market
Analysis
Year
Data Resource Unit 2024 2027 2028 2033 2038
Baseline Production3
U.S. Crude Oil Production
million bbl/day
10.6
11.1
11.1
10.9
10.8
World Oil Production
million bbl/day
82.8
84.5
85.0
87.8
91.1
U.S. Onshore Production
tcf/year
35.6
35.8
36.2
37.2
37.8
Baseline Prices3
Crude Oil
2019$/bbl
60.7
64.4
65.7
71.0
75.0
Natural Gas
2019$/MMbtu
3.01
2.92
3.08
3.46
3.50
Natural Gas
2019$/Mcf
3.12
3.03
3.19
3.59
3.62
3Baseline U.S. crude oil and natural gas production and prices drawn from AEO2022. Baseline world oil production
drawn from EIA's International Energy Outlook.
4-5
-------
4.1.3.4 Regulatory Cost Impacts
As discussed earlier, we assume the projected regulatory costs associated with the final
NSPS 0000b and EG 0000c produce a fractional change in output price. We distribute the
projected regulatory costs to crude oil markets and natural gas markets according to whether the
emissions sources incurring the regulatory costs are more likely to be producing crude oil or
producing, processing, or transporting natural gas. To begin, all projected regulatory costs for
natural gas processing, storage, and transmission sources are assumed to impact the natural gas
market. Within the production segment, projected regulatory costs for natural gas-related model
plants are directed to natural gas markets and costs for oil-related model plants are assigned to
crude oil markets. For example, projected regulatory costs associated with fugitive emissions
monitoring at natural gas well sites are directed to the natural gas market, and projected
regulatory costs at oil well sites are directed to crude oil markets.
For this analysis, we use the projected regulatory costs with capital costs annualized
using a 7 percent interest rate. We also use the net regulatory costs, which include projected
revenues from natural gas recovery from emissions abatement activities. Table 4-3 presents the
results of decomposing the projected regulatory costs into crude oil and natural gas shares.
Table 4-3 Projected Regulatory Costs for the Final NSPS OOOOb and EG OOOOc
Applied in the Market Analysis (millions 2019$)
Year
Resource
2024
2027
2028
2033
2038
Crude Oil
-10.8
238.0
1,531.6
1,952.1
2,569.8
Natural Gas
14.2
61.9
967.3
806.9
637.8
4.1.4 Results
The results of incorporating the projected regulatory costs into the crude oil market
model are presented in . At its peak, the reduction is about 41.4 million barrels in 2038 or about
1.05 percent of crude oil production.
4-6
-------
Table 4-4 Estimated Crude Oil Production and Prices Changes under the Final NSPS
OOOOb and EG OOOOc
Year
Variable
Change
2024
2027
2028
2033
2038
U.S. Production
million bbls/year
0.2
-4.4
-28.0
-33.0
-41.1
%
0.01%
-0.11%
-0.69%
-0.83%
-1.05%
U.S. Prices
Assuming Perfectly Inelastic
Rest of World Supply
$/bbl
0.00
0.03
0.16
0.20
0.25
%
0.00%
0.04%
0.24%
0.28%
0.33%
Assuming Perfectly Elastic
Rest of World Supply
$/bbl
0.0
0.0
0.0
0.0
0.0
%
0.00%
0.00%
0.00%
0.00%
0.00%
We describe two models of world oil markets that bound the market price responses. The
results of incorporating the projected regulatory costs into the crude oil market model are
presented in . At its peak, the reduction is about 41.4 million barrels in 2038 or about 1.05
percent of crude oil production.
Table 4-4 describes results. Assuming perfectly inelastic world oil markets represents an
upper bound on the crude oil price change. The maximum projected oil price change in modeled
years is 0.25 dollars per barrel in 2038, an increase of 0.33 percent. The alternative model is that
world oil markets are perfectly elastic and maintain a fixed oil price. In that case the price change
would be zero. Table 4-5
presents results of entering the projected regulatory costs in the natural gas market
model. We project a maximum natural gas price increase of about $0.06 per mcf and a maximum
production reduction of about 272.5 million Mcf per year, changes of about 1.76 percent and
0.75 percent, respectively.
Table 4-5 Estimated Natural Gas Production and Prices Changes under the Final
NSPS OOOOb and EG OOOOc
Year
Variable Change 2024 2027 2028 2033 2038
U.S. Onshore Production million Mcf/year -4.1 -18.4 -272.5 -202.4 -158.3
% -0.01% -0.05% -0.75% -0.54% -0.42%
U.S. Prices
2019$/Mcf 0.00 0.00 0.06 0.05 0.04
% 0.03% 0.12% 1.76% 1.27% 0.98%
4-7
-------
We use the results in The results of incorporating the projected regulatory costs into the
crude oil market model are presented in . At its peak, the reduction is about 41.4 million barrels
in 2038 or about 1.05 percent of crude oil production.
Table 4-4The results of incorporating the projected regulatory costs into the crude oil
market model are presented in . At its peak, the reduction is about 41.4 million barrels in 2038 or
about 1.05 percent of crude oil production.
Table 4-4 and Table 4-5Table 4-5 to evaluate whether the final NSPS OOOOb and EG
OOOOc is likely to have a significant effect on the supply, distribution, or use of energy as
defined by EO 13211. To make this determination, we compare the projected change in crude oil
and natural gas production to guidance articulated in a January 13, 2021, OMB memorandum
"Furthering Compliance with Executive Order 13211, Titled "Actions Concerning Regulations
That Significantly Affect Energy Supply, Distribution, or Use.107 With respect to crude oil
production, the guidance indicates that a regulatory action produces a significant adverse effect if
it is expected to produce reductions in crude oil supply in excess of 20 million barrels per year.
With respect to natural gas production, the guidance indicates that a regulatory action produces a
significant adverse effect if it reduces natural gas production in excess of 40 million mcf per
year.108 The maximum projected annual decreases in both oil production and natural gas
production exceed benchmarks for adverse effects, so this analysis indicates the final NSPS
OOOOb and EG OOOOc constitutes a significant energy action.
4.1.5 Caveats and Limitations of the Market Analysis
The oil and natural gas market impact analysis presented in this section is subject to
several caveats and limitations, which we discuss here. As with any modeling exercise, the
market impact analysis presented here depends crucially on uncertain input parameters. These
107
See https://www.whitehouse.gov/wp-content/uploads/2021/01/M-21-12.pdf.
108 The 2021 E.O. 13211 guidance memo states that the natural gas production decrease that indicates the regulatory
action is a significant energy action is 40 mcf per year. Because this is a relatively small amount of natural gas
and previous guidance from 2001 indicated a threshold of 25 million Mcf, we assume the 2021 memo was
intended to establish 40 million Mcf as the indicator of an adverse energy effect. See
https://www.whitehouse.gOv/wp-content/uploads/2017/ll/2001-M-01-27-Guidance-for-Implementing-E.0.-
13211.pdf
4-8
-------
parameters include the cost to firms of compliance, the amount of natural gas that would be
recovered and sold because of emissions abatement requirements compliance, baseline
projections, and elasticity estimates. We note the change in price is particularly sensitive to the
demand elasticity.
This analysis considers two residual rest-of-world supply models — perfectly elastic and
perfectly inelastic. The structure of international oil markets (both supply and demand) have
shifted historically and may shift in the future. While these models bound the minimum and
maximum price changes, there is uncertainty within those bounds. One common modeling
assumption is that world oil prices are fixed relative to policy changes. This would imply
perfectly elastic residual rest-of-world supply.
This analysis uses a single-period model which is parameterized for different years,
whereas dynamic effects are important in oil and natural gas markets. Production decisions
relating to drilling and shutting-in wells affect future production, well decline curves, and
intertemporal price arbitrage (the Hotelling Rule) (Hotelling, 1931). Consideration of dynamic
effects may shift numerical results. To the extent the NSPS OOOOb and EG 0000c may
impact well drilling and shut-in decisions, the static analysis present here potentially overlooks
important distributional consequences of the final regulation.
This analysis does not distinguish between different regions of the United States. The
cost of producing oil and natural gas varies over the United States. Compliance costs may also
vary. Reductions in oil and natural gas production would be larger in regions with higher
production costs or higher compliance costs. This could result in different price changes in
different regions of the country if oil is there are bottlenecks in oil or natural gas shipping
infrastructure.
Oil and natural gas markets are linked on both the supply and demand sides. On the
supply side, individual wells generally produce a mixture of oil and natural gas, and some of the
same resources can be used to drill either oil-targeting wells or natural gas-targeting wells. On
the demand side, oil and natural gas are substitutes in some markets. Consideration of these
linkages may additionally shift numerical results.
4-9
-------
4.2 Financial Analysis of Marginal Wells
In addition to the oil and natural gas market impact analysis the EPA developed a
financial analysis of marginal oil and natural gas wells. The marginal well analysis is intended to
provide readers some information on the financial condition of marginal well owners and
operators. The financial analysis, however, does not inform the projected engineering costs and
emissions impacts used in the comparison of benefits and costs.
Marginal oil and gas wells are important to consider for several reasons. First, marginal
wells are the most numerous type of existing well by production level, comprising more than
75% of existing oil and natural gas wells in the United States. Second, while EPA does not have
data on the distribution of ownership based on firm size, there are small owners and operators
who own marginal oil and natural gas wells. Third, commenters have brought up marginal wells
in many contexts and these comments are illustrative of the importance of marginal oil and
natural gas wells to the public. Finally, states are responsible for oil and gas wells that may be
orphaned by owners within their boundaries. With the large numbers of marginal oil and gas
wells across the country, the financial and pollution burdens on states may be significant if some
owners and operators become insolvent and their marginal wells become the responsibility of the
government.
The financial analysis assumes a baseline regulatory environment along with assumptions
on operating and fixed costs and estimates profits or losses of marginal wells for a single year
based on different oil and natural gas production levels. In the subsequent section, we examine
the financial condition of marginal wells, discuss the data and parameters used to implement the
analysis, present results of the analysis, and conclude with a discussion of caveats and limitations
of the analysis. With the available data and the complexity described, we cannot estimate the
impacts of the final regulation on the owners or operators of marginal wells.
4.2.1 Descriptive Statistics on Marginal Wells
According to the EIA, the total number of producing oil and gas wells in the United
States in 2021 was 916,934, of which 403,294 were oil wells and 513,640 were natural gas
4-10
-------
wells.109 The amount of oil or gas produced from each well varies from less than one barrel of oil
equivalent per day (BOE) to over 12,800 BOE. This section focuses on marginal wells, that is,
low-producing wells. EPA uses the EIA definition of a marginal oil well, which is one that over
the course of a calendar year, produces 15 barrels per day or less. For gas wells, the equivalent
production rate is 90,000 cubic feet per day or less.
Even though many wells are considered marginal, these wells represent a relatively small
portion of total domestic production. In 2021, marginal wells comprised about 78 percent of all
producing wells in the United States. There were 318,256 marginal oil wells, or 78 percent of oil
wells, and 396,347 marginal natural gas wells, or 77 percent of natural gas wells. Marginal oil
wells produced about 7 percent of the total oil production. Of these marginal oil wells, there were
299,368 wells, about 74 percent of the total oil wells, that produced 10 BOE per day or less.
These oil wells accounted for about 5 percent of the total U.S. oil production. If marginal oil
wells are further segmented, 157,916 wells, or 39 percent of total oil wells, produced less than
one BOE per day, or about 0.46 percent of the total U.S oil production. Table 4-6 presents the
numbers of oil wells and production information.
Table 4-6 Distribution of U.S Marginal Oil Wells in 2021
Production rate
bracket (BOE/d)a
Number of Oil
Wells
Share of Total
Number of Oil
Wells (%)
Production
Amount (million
barrels)
Share of Total Oil
Production (%)
0-1
157,916
39
16.204
0.46
1-2
45,332
11
21.218
0.60
2-4
45,415
11
42.042
1.19
4-6
24,019
6
37.128
1.05
6-8
15,368
4
33.334
0.94
8-10
11,318
3
31.354
0.89
Subtotal <=10
299,368
74
181.280
5.14
10-12
8,735
2
29.204
0.83
12-15
10,153
3
41.235
1.17
Subtotal <=15
318,256
79
251.718
7.13
All Oil Wells
403,294
100
3528.769
100.00
Source: Energy Information Administration: The Distribution of U.S. Oil and Natural Gas Wells by Production
Rate, December 2022. See link to Appendix B: Selected summary tables at https://www.eia.gov/petroleum/wells/.
a Production rate is defined as barrels of oil equivalent per day (BOE/d).
109
See report available at https://www.eia.gov/petroleum/wells/. Data tables available in link to Appendix B:
Selected summary tables.
4-11
-------
Table 4-7 presents information on the number of natural gas wells and production for
2021. The patterns for natural gas wells are like those for oil wells. Marginal natural gas wells
produced about 7.5 percent of the total natural gas production. There were 365,606 natural gas
wells, about 71 percent of the total natural gas wells, that produced 10 BOE per day or less.
These wells accounted for about 5 percent of the total U.S. natural gas production. Of those
wells, 166,864 wells, or 39 percent of total natural gas wells, produced less than one BOE per
day, or about 0.38 percent of the total U.S. natural gas production.
Table 4-7 Distribution of U.S Marginal Natural Gas Wells in 2021
Production rate
bracket (BOE/d)a
Number of Natural
Gas Wells
Share of Total
Number of
Natural Gas Wells
(%)
Production Amount
(million BOE)
Share of Total
Natural Gas
Production (%)
0-1
166,864
32
122
0.38
1-2
57,265
11
168
0.52
2-4
60,332
12
351
1.09
4-6
37,163
7
363
1.13
6-8
25,358
5
344
1.07
8-10
18,624
4
323
1.01
Subtotal <=10
365,606
71
1,671
5.21
10-12
14,382
3
304
0.95
12-15
16,359
3
424
1.32
Subtotal <=15
396,347
77
2,399
7.48
All Gas Wells
513,640
100
32,092
100.00
Source: Energy Information Administration: The Distribution of U.S. Oil and Natural Gas Wells by Production
Rate, December 2022. See link to Appendix B: Selected summary tables at https://www.eia.gov/petroleum/wells/.
a Production rate is defined as barrels of oil equivalent per day (BOE/d).
While marginal wells account for a relatively small percentage of domestic production of
oil and natural gas, recent studies suggest that their contribution to methane emissions from oil
and natural gas production is much greater. For example, in a study funded by DOE's National
Energy Technology Laboratory (DOE-NETL), Bowers (2022) estimates that marginal natural
gas and oil wells account for 59 percent and 37 percent of cumulative methane emissions from
oil and natural gas production, respectively, and roughly half of cumulative methane emissions
from combined oil and natural gas production. Similarly, Omara et al. (2022) estimate that low
production well sites account for roughly half (95 percent confidence interval: 37-75 percent) of
all oil and natural gas well site methane emissions. In the analysis performed for this final
rulemaking, we estimate that marginal wells account for 47-53 percent of well site methane
4-12
-------
emissions and 49-55 percent of reductions for the quantified emissions sources once the EG is
assumed to take effect in 2028, with the ranges due to differing percentages from year to year.
4.2.2 Marginal Wells Financial Analysis Model
The expected single year profit is calculated for marginal wells in each production
bracket reported by the EIA. A single year is calculated because oil and natural gas production
from a particular well naturally declines over time. Assuming the single year is the first year of
the remaining life of the well, it is very likely that this single year captures the maximum profit
for the remaining production life of the average marginal well.
The parameter values presented in Table 4-8 are obtained from comments submitted for
the November 2021 Proposal and the December 2022 Supplemental Proposal. The profit is
estimated prior to any new regulation by any government entity. Natural gas wells primarily
produce natural gas but there is also a relatively small amount of oil produced. Similarly, oil
wells also produce some natural gas. The gross revenue is calculated by adding gas and oil
production for marginal wells multiplied by their respective prices then subtracting the
landowner royalty:
Gross_Revenue = (1 — Royalty_Rate) * (Oil_Price * Oil_Production + Gas_Price *
Gas_Productiori), Eq. 4-5
here Oil_Production and Gas_Production are the respective annual production levels for each
production bracket reported by the EIA. Total costs, which include operating costs and fixed
costs, are subtracted from total revenue to obtain net profit:
Operating_Cost = (Severance + Oil_Transport + Variable_0&.M) * Oil_Production +
(Gas_Transport + Gas_Treat + GasjOperating ) * Gas_Production + WelljOverhead * 12,
Eq. 4-6
where Severance, Oil_Transport, Variable_0&M represent annual variable costs for oil
production, Gas_Transport, Gas_Treat, GasjOperating represent annual variable costs for
gas production, and Well_Overhead * 12 is the annual fixed costs. Profit is estimated using the
follow equation:
Profit = Gross_Revenue — Operating_Cost Eq. 4-7
4-13
-------
Table 4-8
Marginal Well Financial Analysis Parameters
Parameter
Description
Value
Royalty
Landowner royalty3
20%
Gas Transport
Transportation cost ($/mcf)a
$ 0.30
Gas Treat
Treating cost ($/mcf)a
$ 0.35
Gas Operating
Operating and compression cost ($/mcf)a
$ 1.10
Well Overhead
Fixed overhead ($/well/month)a
$ 250.00
Severance
Severance tax ($/bbl)b
$ 3.00
Oil Transport
Transportation cost ($/bbl)b
$ 4.00
Variable O&M
Variable operation and maintenance ($/bbl)b
$ 4.00
a Source Riverside comment on November 2021 Proposal (EPA-HQ-OAR-2021-0317-0411)
b Source Michigan Oil and Gas Association comments on December 2022 Proposal (EPA-HQ-OAR-2021-0317-
2257).
As a point of comparison for the cost information provided by commenters, Weber et al.
(2021) present a case study where they conduct a break-even analysis of marginal conventional
natural gas wells in Pennsylvania. Based on publicly available data and company disclosures
from Diversified Energy Company, the largest operator of conventional wells in Pennsylvania,
the authors estimate the fixed costs and operating costs of natural gas wells in the company's
conventional natural gas well portfolio. According to Weber et al. (2021), Diversified has more
than 60,00 wells across Appalachia, with the largest concentrations in Pennsylvania and West
Virginia. Weber et al. (2021) estimate operating costs to be about $886 annually and variable
costs to be $0.74 per Mcf for the average natural gas well. Using these values, Weber et al.
(2021) estimate that wells with a production rate below 0.5 Mcf per day were likely to be
uneconomical, assuming a high natural gas price scenario of $3.94. The variable costs used in
this analysis include the gas transport cost of $0.30 per Mcf, the gas treatment cost of $0.35 per
mcf, and operating and compression cost of $1.10 per Mcf for a total of $1.75 per Mcf. The fixed
costs are estimated at $3,000 per well annually. The costs are higher than reported in Weber et al.
(2021), however, the authors note that Diversified has an economy of scale in Pennsylvania that
likely reduces operating expenses below that incurred by other operators in the state and the
regional nature of the study make it difficult to directly compare nationwide.
Weber et al. (2021) also find evidence that operators postpone plugging marginal
conventional natural gas wells in Pennsylvania. They find that about one-third of wells
considered uneconomical in 2019 have not been plugged, which include wells with zero
production and wells producing below 0.5 Mcf per day. The authors also document that these
4-14
-------
wells have reported low- or zero-production over multiple years yet have remained unplugged
over this time period. The authors also find that more than 37 percent of the wells that reported
no production in 2019 also had no reported production from 2015 to 2018. For wells that
produced below the 0.5 Mcf per day threshold, 32 percent were below the threshold for 2015 to
2019.
4.2.3 Results
The results of the financial analysis are presented in Table 4-9, Table 4-10, and Table
4-11. The tables present the profit estimation for the average monthly price of oil and natural gas
from October 2022 to September 2023, as well as the low and high prices for the same period. In
each table, the One Year Net Profit is the estimated profit for a single year from an average
marginal well in 2021 before any additional regulatory costs are considered. In addition to
presenting the estimated profits for all oil and gas wells with production of 15 barrels of oil
equivalent per day (BOE) (listed in the last row of the table), the estimated profits are also
disaggregated into smaller production brackets to give a fuller picture of the profits of marginal
oil and gas wells. This highlights the feature that, while comprising a smaller percentage of the
overall number of marginal wells, the higher producing marginal wells generate comparatively
more profits thereby increasing the average profits of all marginal wells. This can be seen when
comparing the estimated single year profit for average marginal oil and gas wells in Table 4-9,
$42,033 and $5,648, respectively, to each disaggregated production bracket.
Focusing on the average price scenario in Table 4-9, marginal oil wells are profitable at
every production level. At the lowest production bracket, 0-1 BOE, the average oil well earns an
estimated positive profit of $2,968. At the next highest production level, 1-2 BOE, the profits for
the average oil well are estimated to be $23,138. At the highest production level for marginal
wells, 12-15 BOE, the profits are more than $221,000 per year.
Marginal gas wells can be unprofitable before the addition of any regulatory costs and
have lower profits than marginal oil wells in corresponding production bands. Again, looking at
the average price scenario in Table 4-9, the average gas wells in the 0-1 production bracket are
estimated to be unprofitable with $2,075 loss. Estimated profits are positive for 1-2 BOE at $538
4-15
-------
with the highest level of profits also occuring at the 12-15 BOE production band at over $36,444
per year.
As a sensitivity check on the average price scenario, low and high price scenarios are also
presented. In Table 4-10 the one-year profits for the low oil and natural gas prices are presented.
In the low-price scenario the oil price is assumed to be $70.25 per barrel and the natural gas price
is assumed to be $2.15 per Mcf. Profits for marginal oil wells follow a similar pattern as the
average price scenario, profitable in all production bands, with profits rising with production. For
marginal natural gas wells, profits do not become positive until production reaches the 4-6 BOE
per day range. While there are some production bands that have relatively high levels of profits,
such as the 12-15 BOE per day band, average profits for all marginal gas wells are negative.
Table 4-11 contains the one-year profit estimates for the high price scenario. In the high-
price scenario the oil price is assumed to be $89.43 per barrel and the natural gas price is
assumed to be $5.66 per Mcf. While the one-year profits for marginal oil wells behave similarly
to the average and the low-price scenarios, there is one notable result for marginal natural gas
wells. One-year profits for the 0-1 BOE per day production band are negative. While not directly
comparable to the regional study by Weber et al. (2021), this result is in general agreement with
their conclusion that the lowest producing natural gas wells are likely to be uneconomical.
Compliance costs stemming from the requirements for each production site depend on the
equipment located at the site. While data that correlates well site production levels and site
equipment is scarce, our analysis of the 2016 ICR data suggests that low production sites
generally have less equipment than non-low production sites (see, e.g., Table 2-16). When
applied to our well site projections, we estimate that, depending on analysis year, roughly 50-60
percent of sites would be classified as wellhead only or small sites for the purposes of the
fugitives monitoring requirements. These sites are subject to the least stringent standards, costing
an estimated $336-660 per site per year, depending on whether the site require additional travel.
In addition, we estimate that 45-50 percent of sites do not have pneumatic controllers. For those
that do, we estimate that roughly 60 percent have two or fewer controllers. As an example, for
existing sites with two controllers that do not have access to grid electricity, we estimate an
annualized compliance cost of $1,312 for a retrofit with solar powered electric controllers, not
including $630 in gas savings assuming both controllers are intermittent bleed with average
4-16
-------
emissions rates consistent with the assumptions in the Final Rule TSD.n° See the December 2022
TSD and the Final Rule TSD for more information on the costs of controllers.
Table 4-9 Estimated Revenues, Costs, and Profits for Marginal Wells for 2021:
Average Oil and Natural Gas Prices3
Oil Wells
Gas Wells
Production rate
bracket (BOE/d)
One Year
Gross
Revenue
Operating
Cost
One Year
Net Profit
One Year
Gross
Revenue
Operating
Cost
One Year
Net Profit
0-1
$ 7,285
$ 4,317
$ 2,968
$ 2,350
$ 4,426
$ (2,075)
1-2
$ 31,938
$ 8,799
$23,138
$ 9,085
$ 8,547
$ 538
2-4
$ 62,593
$ 14,492
$48,101
$ 18,747
$ 14,138
$ 4,609
4-6
$ 105,157
$ 22,645
$ 82,512
$ 32,632
$21,948
$ 10,684
6-8
$ 147,241
$ 30,749
$116,491
$ 46,447
$ 29,664
$ 16,783
8-10
$ 188,432
$ 38,769
$149,663
$ 60,637
$ 37,427
$23,210
Subtotal <=10b
$ 42,367
$ 10,870
$ 31,497
$ 15,294
$ 11,933
$ 3,360
10-12
$ 229,093
$ 46,776
$182,317
$ 74,245
$45,130
$29,115
12-15
$ 278,452
$ 56,594
$221,858
$91,034
$ 54,589
$ 36,444
Subtotal <=15c
$ 55,424
$ 13,392
$ 42,033
$ 20,530
$ 14,882
$ 5,648
Note: One year of financial estimates for marginal oil and natural gas wells using the average monthly price of oil
and natural gas. One Year Gross Revenue is estimated using Equation 4-5. Operating costs are fixed plus variable
costs as calculated by Equation 4-6, using parameters from Table 4-8. One Year Net Profit is calculated from
Equation 4-7.
a Source EIA. Average Monthly oil and natural gas prices from October 2022 to September 2023: Oil price is West
Texas Intermediate, $78.73/bbl; See https://www.eia.gov/dnav/pet/hist/rwtcM.htm. Natural gas price is Henry Hub,
$3.24/mcf; See https://www.eia.gov/dnav/ng/hist/rngwhhdM.htm.
b Subtotals are averages for oil and natural gas wells with the 10 BOE per day or less range.
0 Subtotals are averages for all oil and natural gas wells with 15 BOE per day or less.
110 Costs here are based on the 'Two Electric Controllers Solar' scenario illustrated on the 'electric controllers' tab
from the pneumatic controllers spreadsheet accompanying the December 2022 TSD (EPA-HQ-OAR-2021-0317-
1578_attachment_2.xlsx), applying a capital recovery factor based on a 7 percent interest rate and 15 year control
lifetime to the estimated capital cost of $15,102 and adding it to an estimated reduction in maintenance costs of
$346. Gas savings is calculated based on an assumed gas price of $3.24 per mcf.
4-17
-------
Table 4-10 Estimated Revenues, Costs, and Profits for Marginal Wells for 2021: Low Oil
and Natural Gas Prices"
Oil Wells
Gas Wells
Production rate
bracket (BOE/d)
One Year
Gross
Revenue
Operating
Cost
One Year
Net Profit
One Year
Gross
Revenue
Operating
Cost
One Year
Net Profit
0-1
$ 6,480
$ 4,317
$2,163
$ 1,633
$ 4,426
$ (2,793)
1-2
$ 28,397
$ 8,799
$ 19,597
$ 6,297
$ 8,547
$ (2,250)
2-4
$ 55,595
$ 14,492
$41,103
$ 13,134
$ 14,138
$ (1,004)
4-6
$ 93,246
$ 22,645
$ 70,601
$ 23,063
$21,948
$ 1,115
6-8
$ 130,452
$ 30,749
$ 99,703
$ 32,967
$ 29,664
$ 3,302
8-10
$ 166,829
$ 38,769
$128,060
$43,213
$ 37,427
$ 5,787
Subtotal <=10b
$ 37,588
$ 10,870
$ 26,719
$ 10,785
$ 11,933
$ (1,149)
10-12
$ 202,697
$ 46,776
$155,921
$ 52,921
$45,130
$ 7,792
12-15
$246,192
$ 56,594
$189,598
$64,919
$ 54,589
$ 10,330
Subtotal <=15c
$49,129
$ 13,392
$ 35,737
$ 14,528
$ 14,882
$ (354)
Note: One year of financial estimates for marginal oil and natural gas wells using low oil and natural gas prices. One
Year Gross Revenue is estimated using Equation 4-5. Operating costs are fixed plus variable costs as calculated by
Equation 4-6, using parameters from Table 4-8. One Year Net Profit is calculated from Equation 4-7.
a Source EIA. Lowest monthly oil and natural gas prices from October 2022 to September 2023: Oil price is West
Texas Intermediate $70.25/bbl; See https://www.eia.gov/dnav/pet/hist/rwtcM.htm. Natural gas price is Henry Hub,
$2.15/mcf; See https://www.eia.gov/dnav/ng/hist/rngwhhdM.htm.
b Subtotals are averages for oil and natural gas wells with the 10 BOE per day or less range.
0 Subtotals are averages for all oil and natural gas wells with 15 BOE per day or less.
4-18
-------
Table 4-11 Estimated Revenues, Costs, and Profits for Marginal Wells for 2021: High
Oil and Natural Gas Prices
Oil Wells
Gas Wells
Production rate
bracket (BOE/d)
One Year
Gross
Revenue
Operating
Cost
One Year
Net Profit
One Year
Gross
Revenue
Operating
Cost
One Year
Net Profit
0-1
$ 8,329
$ 4,317
$ 4,012
$ 3,909
$ 4,426
$ (517)
1-2
$ 36,549
$ 8,799
$ 27,750
$ 15,155
$ 8,547
$ 6,608
2-4
$71,784
$ 14,492
$ 57,292
$ 30,895
$ 14,138
$ 16,757
4-6
$ 121,010
$ 22,645
$ 98,366
$ 53,242
$21,948
$31,293
6-8
$ 169,735
$ 30,749
$138,986
$ 75,408
$ 29,664
$ 45,744
8-10
$217,533
$ 38,769
$178,764
$ 97,977
$ 37,427
$ 60,550
Subtotal <=10a
$ 48,699
$ 10,870
$ 37,829
$25,017
$ 11,933
$ 13,084
10-12
$ 264,825
$ 46,776
$218,049
$ 119,937
$45,130
$ 74,807
12-15
$ 322,355
$ 56,594
$265,761
$ 146,976
$ 54,589
$ 92,387
Subtotal <=15b
$ 63,826
$ 13,392
$ 50,435
$ 33,449
$ 14,882
$ 18,567
Note: One year of financial estimates for marginal oil and natural gas wells using high oil and natural gas prices.
One Year Gross Revenue is estimated using Equation 4-5. Operating costs are fixed plus variable costs as calculated
by Equation 4-6, using parameters from Table 4-8. One Year Net Profit is calculated from Equation 4-7.
a Source EIA. Highest monthly oil and natural gas prices from October 2022 to September 2023: Oil price is West
Texas Intermediate, $89.43/bbl; See https://www.eia.gov/dnav/pet/hist/rwtcM.htm. Natural gas price is Henry Hub,
$5.66/mcf; See https://www.eia.gov/dnav/ng/hist/rngwhhdM.htm.
b Subtotals are averages for oil and natural gas wells with the 10 BOE per day or less range.
0 Subtotals are averages for all oil and natural gas wells with 15 BOE per day or less.
4.2.4 Marginal Wells Financial Analysis Caveats and Limitations
The goal of this analysis is to evaluate the profitability of marginal wells under a certain
set of assumptions. We are not able to evaluate other empirical considerations with this analysis.
Marginal wells may continue to operate at low or negative profits rather than be shut-in
and plugged due to a variety of reasons. Wells that would otherwise be left in a low-producing or
idled state for many years because the costs of plugging are too high can benefit from federal
subsidies to reduce, or completely cover the costs of plugging. The Inflation Reduction Act
(IRA) provides $700 million for methane and greenhouse gas mitigation activities for
conventional marginal wells. The Department of Energy and the EPA have issued a joint Notice
of Intent to provide grants up to $350 million for conventional marginal well plugging as an
4-19
-------
initial effort to distribute the IRA funds.111 The goal of marginal well funding from the IRA is to
offset the cost of plugging low producing wells. The costs of plugging can vary widely
depending on well characteristics and by state. Estimates range from less than $1,000 to over $1
million, with an average of about $75,000 per well (Raimi et al., 2021). It has been reported that
in some cases those costs can be reduced to around $25,000, although it is not clear if the
112
methods used could be more broadly applied. Additionally, the federal money reduces the
possibility that the operator files for bankruptcy and orphans their well(s). To maintain
transparency for the responsible owner or operator and prevent well sites from becoming
orphaned, EPA is finalizing the rule with a well closure plan requirement which requires fugitive
monitoring of well sites for the life of the well and can only be discontinued if the well site is
properly plugged and an OGI survey is done to ensure there are no emissions from the site. EPA
is also requiring owners and operators to submit, through an annual ownership report, changes in
ownership at individual well sites so it is clear who the responsible owners and operators are
until the site is plugged and closed and fugitive emissions monitoring is no longer required.
Through accounting practices and discounting, the costs associated with plugging wells
can be delayed far into the future. This delay effectively reduces the present value of costs to the
owners and operators and increases the firm's profitability in the short run. Thus, by changing
the financial break-even point for when to close the low producing well, owner/operators can
extend the well's production period and push the decision to shut-in the well further into the
future.
Federal and state tax-credits are available for owners and operators of marginal oil and
gas wells in a low commodity price environment. These tax credits reduce the cost of owning
113
marginal wells. At the federal level, wells that produce no more than 25 BOE per day can
qualify for the credit. With no limit on the number of wells that can qualify for the credit, each
well can receive a tax credit on up to 1,905 BOE of total production per year. The credit value
for marginal gas wells is based on the reference natural gas price published by the Internal
Revenue Service (IRS); in 2021, the maximum value was $4,336 per well. Any unused portion
111 See: https://www.grants.gov/web/grants/view-opportunity.html?oppld=349508.
112
See: https://www.bloomberg.com/features/diversified-energy-natural-gas-wells-methane-leaks-2021/.
113
See: https://crsreports.congress.gov/product/pdf/IF/IFl 1528.
4-20
-------
of the tax credit can either be carried back five years or forward 20 years. The current tax credit
114
applies when the reference price of natural gas is below $2 per mcf. For marginal oil wells, the
reference price is $15 per barrel of oil.115 Since the amount of the credit is phased out when oil
prices are greater than the reference price, high oil prices relative to the reference price means
the marginal oil credit has not been available in recent years.
State bonding requirements are sometimes insufficient to cover the costs of plugging
wells. Blanket bonds that an operator takes out to cover all their wells in a state are especially
problematic as the fixed amount that is set aside is often too low to cover the plugging costs for
an operator with many wells (Ho et al., 2018).
Accounting practices that enable the delay of plugging costs, tax subsidies, and low
bonding requirements that may incentivize owners/operators to continue to operate a marginal
well that is barely or is not profitable are all considerations that may impact the owner/operator's
decision on whether to shut-in and plug a well. A simple break-even analysis that only evaluates
well profitability in a single time-period can help shed light on whether the well has a higher
likelihood of shutting-in, but it doesn't capture the full decision-making process. That is, a single
period of losses most likely will not cause a firm to shut-in and plug a well but larger, sustained
losses over many years could. Thus, it seems that costs must be much larger than any benefits of
delay, including expected gains from production, before a well is realistically at risk for being
shut-in and plugged. The uncertainty surrounding future oil and gas production, commodity
prices, tax treatments, and regulatory regimes, also contributes to the difficulty in predicting
when or if an operator decides to shut-in and plug any given well. This uncertainty also makes it
extremely difficult to determine the full impact of regulation on the financial status of marginal
well owners.
Further, because regulatory costs are dependent on, among other things, types of
equipment used at well sites, the full impact of regulation on the financial status of marginal well
owners cannot be determined. The age, location, and what is produced from each marginal well
114
See: https://www.irs.gov/pub/irs-irbs/irb23-23.pdf.
115 See: https://www.irs.gov/irb/2023-26_IRB.
4-21
-------
also contribute to differences in baseline regulatory costs. This adds heterogeneity to the
regulatory cost burdens on owners and adds complexity to estimating impacts.
The financial condition of the owners and operators of the marginal wells is unknown. It
is likely that some of them are financially distressed, but unavailable data makes it near
impossible to determine conclusively. Thus, the EPA is unable to assess the viability of the
owners and operators of the marginal wells, and as such, the EPA is therefore unable to predict
which firms will be most adversely affected by regulatory costs.
There is a market for buying and selling marginal wells, and it is possible that not only
could current market participants increase their holdings of marginal wells, but new participants
could also enter the market. For example, Diversified Energy, headquartered in Birmingham,
Alabama, is reported to be the largest owner oil and natural gas wells in the country, with a large
percentage being marginal wells.116 Diversified Energy describes its business strategy as
acquiring uneconomic wells that are low-cost, low decline wells that are expected to continue
producing far into the future. Diversified owns wells throughout Appalachia and has also begun
purchasing wells in Louisiana and Texas. It is conceivable that as more marginal wells become
uneconomic through market conditions, regulatory costs, and declining production other firms
will find it advantageous to buy these low-producing wells from the current owners.
Because of their low production, some marginal wells are temporarily shut-in and
production is paused. This can happen for a variety of reasons including maintenance and repair.
While these wells can restart production, the decision to do so typically relies on the geologic
properties of the reservoir. In cases where the marginal well is shut-in for an extended period,
restarting production can be difficult and even impossible. This temporary pausing of production
further complicates the analysis effort.
The decision to plug a well can be described as an optimal stopping problem. Provencher
(1997), Pindyck (2002), Kellogg (2014), and others have used a dynamic framework to model
similar types of problems in the natural resource sector. While such an analysis is not undertaken
here, the methods used to solve such problems are complex and require data that is currently
unavailable.
116 See: https://ohiorivervalleyinstitute.org/wp-content/uploads/2022/04/Diversified-Energy-Report-FINAL-l.pdf.
4-22
-------
4.3 Environmental Justice Analyses
E.O. 12898 directs the EPA to "achiev[e] environmental justice (EJ) by identifying and
addressing, as appropriate, disproportionately high and adverse human health or environmental
effects" (59 FR 7629, February 16, 1994), termed disproportionate impacts in this chapter.
Additionally, E.O. 13985 was signed to advance racial equity and support underserved
communities through Federal government actions (86 FR 7009, January 20, 2021). 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.
E.O. 14096 defines EJ as the just treatment and meaningful involvement of all people,
regardless of income, race, color, national origin, Tribal affiliation, or disability, in agency
decision-making and other Federal activities that affect human health and the environment. The
EPA further defines the term fair treatment to mean that "no group of people should bear a
disproportionate burden of environmental harms and risks, including those resulting from the
negative environmental consequences of industrial, governmental, and commercial operations or
programs and policies".117 Meaningful involvement means that: (1) potentially affected
populations have an appropriate opportunity to participate in decisions about an activity that will
affect their environment and/or health; (2) the public's contribution can influence the regulatory
Agency's decision; (3) the concerns of all participants involved will be considered in the
decision-making process; and (4) the rule-writers and decision-makers seek out and facilitate the
involvement of those potentially affected.
The term "disproportionate impacts" refers to differences in impacts or risks that are
extensive enough that they may merit Agency action.118 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
117
See, e.g., "Environmental Justice." Epa.gov, U.S. Environmental Protection Agency, 4 Mar. 2021,
https://www.epa.gov/environmentaljustice.
See https://www.epa.gov/environmentaljustice/technical-guidance-assessing-environmental-justice-regulatory-
analysis.
4-23
-------
groups of concern for both the baseline and final regulatory options, using the best available
information (both quantitative and qualitative) to inform the decision-maker and the public.
A regulatory action may involve potential EJ concerns if it could: (1) create new
disproportionate impacts on populations or communities of concern based on income, race,
color, national origin, Tribal affiliation, or disability; (2) exacerbate existing disproportionate
impacts on populations or communities of concern; or (3) present opportunities to address
existing disproportionate impacts on populations or communities of concern through the action
under development.
Under E.O. 13563, federal agencies may consider equity, human dignity, fairness, and
distributional considerations, where appropriate and permitted by law. E.O. 14094 directs
Federal agencies to recognize distributive impacts and equity in regulatory analysis, to the extent
permitted by law, as practicable and appropriate.119 For purposes of analyzing regulatory impacts,
the EPA relies upon its June 2016 "Technical Guidance for Assessing Environmental Justice in
Regulatory Analysis",120 which provides recommendations that encourage analysts to conduct the
highest quality analysis feasible, recognizing that data limitations, time, resource constraints, and
analytical challenges will vary by media and circumstance.
A reasonable starting point for assessing the need for a more detailed EJ analysis is to
review the available evidence from the published literature and from community input on what
factors may make population groups of concern more vulnerable to adverse effects (e.g.,
underlying risk factors that may contribute to higher exposures and/or impacts). It is also
important to evaluate the data and methods available for conducting an EJ analysis. EJ analyses
can be grouped into two types, both of which are informative, but not always feasible for a given
rulemaking:
1. Baseline Analysis: Describes the current (pre-control) distribution of exposures and
risk and identifies potential disparities.
119 88 FR 21879
120
See https://www.epa.gov/environmentaljustice/technical-guidance-assessing-environmental-justice-regulatory-
analysis.
4-24
-------
2. Policy Analysis: Describes the distribution of exposures and risk after the regulatory
option(s) have been applied (post-control) and identifies how potential disparities
change in response to the rulemaking.
EPA's 2016 Technical Guidance does not prescribe or recommend a specific approach or
methodology for conducting EJ analyses. However, a key consideration is consistency with the
assumptions underlying other parts of the regulatory analysis when evaluating the baseline and
regulatory options.
4.3.1 Analyzing EJ Impacts in this Final Action
In addition to the benefits assessment (Chapter 3), the EPA considers potential EJ
concerns of this rulemaking. An EJ concern is defined as the actual or potential lack of fair
treatment or meaningful involvement on the basis of income, race, color, national origin, Tribal
affiliation, or disability in the development, implementation and enforcement of environmental
laws, regulations, and policies. For analytic purposes, this concept refers more specifically to
disproportionate and adverse impacts that may exist prior to or be created by the final regulatory
action (U.S. EPA, 2015). Although EJ concerns for each rulemaking are unique and should be
considered on a case-by-case basis, the EPA's EJ Technical Guidance (U.S. EPA, 2015) states
that "[t]he analysis of potential EJ concerns for regulatory actions should address three
questions:
1. Are there potential EJ concerns associated with environmental stressors affected by the
regulatory action for population groups of concern in the baseline?
2. Are there potential EJ concerns associated with environmental stressors affected by the
regulatory action for population groups of concern for the regulatory option(s) under
consideration?
3. For the regulatory option(s) under consideration, are potential EJ concerns created [,
exacerbated,] or mitigated compared to the baseline?"
To address these questions, the EPA developed an analytical approach that considers the
purpose and specifics of this rulemaking, as well as the nature of known and potential exposures
and health impacts. Oil and natural gas operations in the U.S. include a variety of emission
sources for methane, VOC, and HAP, including wells, well sites, processing plants, compressor
4-25
-------
stations, storage equipment, and natural gas transmission and distribution lines. These emission
points are located throughout much of the country, though many of these emissions sources are
concentrated in particular geographic regions. For example, wells and processing plants are
largely concentrated in the South Central, Midwest, and Southern California regions of the U.S.,
whereas natural gas compressor stations are located all over the country. Distribution lines to
customers are frequently located within areas of high population density. The rulemaking is also
expected to reduce ambient ozone from VOC emission reductions across the continental U.S.
For purposes of this RIA, the EPA performed quantitative analyses of proximity to
emissions sources, the distribution of exposures from the regulated sector, the demographics of
employees in the regulated sector and of communities whose employment is disproportionately
in the regulated sector, and of impacts on consumers, as well as qualitative analyses of climate
vulnerabilities. The quantitative analyses use different measures of possible groups of concern
because each focuses on different channels of possible impact. The analysis of employment
impacts and the analysis of oil and gas intensive communities focus on characteristics which may
affect workers' ability to find work in other sectors. The analysis of household energy
expenditures characterizes how household energy expenditures vary across the income
distribution and for different racial and ethnic groups, with a goal to highlight which populations
may be most vulnerable to potential energy market effects caused by regulatory impacts.
While the qualitative discussion of climate EJ impacts (Section 4.3.24.3.2) and the
baseline assessments of air toxic EJ impacts (Section 4.3.44.3.44.3.4), demographic EJ impacts
of workers and communities (Section 4.3.54.3.5), and household energy expenditures by
demographic group (Section 4.3.64.3.64.3.6) remain largely similar to what was included in the
proposal and supplemental proposal RIAs, we have improved the quantitative assessment of
ozone EJ exposure impacts from oil and natural gas VOC emissions (Section 4.3.34.3.3). As
policy-specific ozone-season air quality surfaces were generated to estimate human health
benefits, we also evaluated a subset of these air quality surfaces for EJ ozone exposure from
VOC. We also evaluated EJ ozone exposure impacts in additional potentially overburdened
populations (e.g., redlined areas) and developed population subgroups that represent differences
in cumulative environmental exposures (e.g., life expectancy) and specifically identify
geographically similar communities of Indigenous populations that may have potential EJ
concerns (e.g., Tribal lands).
4-26
-------
4.3.2 Qualitative Discussion of Disparate Climate Vulnerabilities in the Baseline
In 2009, under the Endangerment and Cause or Contribute Findings for Greenhouse
Gases Under Section 202(a) of the Clean Air Act ("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 minority 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; including those that have been historically marginalized or
overburdened; individuals at vulnerable lifestages, such as the elderly, the very young, and
pregnant or nursing women; those already in poor health or with comorbidities; the disabled;
those experiencing homelessness, mental illness, or substance abuse; and/or Indigenous or
minority 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
Council, 2011; Oppenheimer et al., 2014; Porter et al., 2014; Smith et al., 2014; USGCRP, 2016,
2018). These reports conclude that less-affluent, traditionally marginalized, or predominantly
non-White communities can be especially vulnerable to climate change impacts because they
tend to have limited adaptive capacities and 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., Black and Hispanic/Latino communities; Native
Americans, particularly those living on Tribal lands and Alaska Natives), may be uniquely
vulnerable to climate change health impacts in the United States, 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).
4-27
-------
Per the Fourth National Climate Assessment, "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 greenhouse gases 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. Outdoor workers, such as construction or utility workers and agricultural
laborers, who are frequently part of already at-risk groups, are exposed to poor air quality and
extreme temperatures without relief. Furthermore, individuals within EJ populations of concern
face greater housing and clean water insecurity and bear disproportionate 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. The urban heat island effect
can add additional stress to vulnerable populations in densely populated cities who do not have
access to air conditioning. Finally, resiliency and adaptation are more difficult for economically
disadvantaged communities: They tend to have less liquidity, individually and collectively, to
move or to make the types of infrastructure or policy changes necessary to limit or reduce the
hazards they face. They frequently face systemic, institutional challenges that limit their power
to advocate for and receive resources that would otherwise aid in resiliency and hazard reduction
and mitigation.
The assessment literature cited in EPA's 2009 and 2016 Endangerment Findings, as well
as Impacts of Climate Change on Human Health, also concluded that certain populations and
people in particular 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
4-28
-------
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).
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) live with many of the factors that contribute to their
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 individuals compared to White individuals after controlling for factors such as air
conditioning use. Moreover, people of color are disproportionately exposed to air pollution based
on where they live, and disproportionately vulnerable due to higher baseline prevalence of
underlying diseases such as asthma, so climate exacerbations of air pollution are expected to
have disproportionate effects on these communities.
The recent 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, 2021c). This report
4-29
-------
found that 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 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 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. Those 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.
Indigenous communities possess unique vulnerabilities to 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 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 Fourth National Climate
Assessment (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, there can be institutional barriers 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
4-30
-------
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. Confounding
general Native American response to natural hazards are limitations imposed by policies such as
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 management decisions.
Additionally, the Fourth National Climate Assessment noted that Indigenous peoples 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, diabetes,
and obesity, which can all contribute to increased vulnerability to climate-driven extreme heat
and air pollution events. These factors also may be exacerbated by stressful situations, such as
extreme weather events, wildfires, and other circumstances.
The Fourth National Climate Assessment and IPCC AR5 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 - these impacts on archaeological sites, structures, and
objects that will lead to a loss of cultural heritage for Alaska's Indigenous people. In terms of
food security, the Fourth National Climate Assessment 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 Fourth
National Climate Assessment 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.
4.3.3 Ozone from Oil and Natural Gas VOC Emission Impacts
We evaluate the potential for EJ concerns among potentially vulnerable populations
resulting from exposure to ozone due to VOC emissions from the oil and gas sector under the
baseline and policy option in this rule. This was done by characterizing the distribution of ozone
4-31
-------
exposures both in the baseline and for following the implementation of the regulatory option in
2038. The air quality surface for this analytic year was selected from the four available air
quality surfaces as the surfaces for 2024, 2027, 2028, and 2038 were all similar in both
magnitude of predicted ozone concentrations and spatial distribution.
As this analysis is based on the same ozone spatial fields as the benefits assessment (see
showing the spatial fields), it is subject to similar types of uncertainty (see Section 3.3.7 for a
discussion of uncertainty). A particularly germane limitation for this analysis is that the expected
concentration changes are quite small, likely making uncertainties associated with the various
input data more relevant.
This EJ air pollutant exposure analysis aims to evaluate the potential for EJ concerns
related to ozone exposures among potentially vulnerable populations.121122 To assess EJ ozone
exposure impacts, we focus on the first and third of the three EJ questions from the EPA's 2016
EJ Technical Guidance.123 The first questions asks if there are potential EJ concerns associated
with stressors affected by the regulatory action for population groups of concern in the baseline.
The third question asks if those potential EJ concerns in the baseline are exacerbated, unchanged,
or mitigated under the regulatory option being considered.124
To address these questions with respect to ozone exposures, EPA developed an analytical
approach that considers the purpose and specifics of this final rulemaking, as well as the nature
121
The term exposure is used here to describe estimated average warm-season ozone concentrations and not
individual dosage.
122
Air quality surfaces used to estimate exposures are based on 12 km by 12 km grids. Additional information on air
quality modeling can be found in the air quality modeling information section.
123
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
124
EJ question 2, which 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, was not focused on for several reasons. Importantly, the total
magnitude of differential exposure burdens with respect to ozone 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. We do however use the results
from questions 1 and 3 to gain insight into the answer to EJ question 2 in the summary (Section 4.3.7).
4-32
-------
of known and potential exposures and impacts. Specifically, as 1) this final rule affects oil and
natural gas sources across the U.S., which are numerous and vary in emission levels, and 2)
ozone can undergo long-range transport, it is appropriate to conduct an EJ assessment of the
contiguous U.S. Given the availability of modeled ozone air quality surfaces under the baseline
and final regulatory option, we conduct an analysis of changes in ozone concentrations modeled
to occur under the final rule as compared to the baseline scenario, characterizing average and
distributional exposures following implementation of the regulatory option in the implementation
year 2038. However, several important caveats of this analysis are as follows:
• The baseline scenario for 2038 represents expected emissions from oil and natural gas
sources covered by these rules in 2038 but includes emissions from all other sources
that are only projected to the year 2026. The 2038 baseline therefore does not capture
any anticipated changes in ambient ozone by 2038 that would occur due to emissions
changes from sources other than oil and natural gas sources covered by this
rulemaking.
• Modeling of post-policy air quality concentration changes are based on state-level
emission data paired with baseline 2038 emissions projections for oil and natural gas
sources based on version 2 of the EPA's 2016 emissions modeling platform (U.S.
EPA, 2022). While the baseline spatial patterns represent ozone concentrations
associated with the projected 2026 oil and natural gas emissions, the post-policy air
quality surfaces will capture expected ozone changes that result from state-to-state
emissions changes but will not capture any spatially heterogenous changes in oil and
natural gas emissions within a single state except for changes captured by separating
out emissions changes occurring in the three Texas sub-regions. While the
methodology applied to create the ozone surfaces does not allow us to consider the
effects of any changes to spatial distribution of oil and natural gas emissions within a
state between the 2026 modeled case and the baseline and regulatory alternatives in
this RIA, the within-state spatial patterns do represent locations of oil and natural gas
emissions in 2026. This provides a reasonable estimate of where change in oil and gas
emissions will occur from this rulemaking since, in the timeframe of the analysis for
this rulemaking, new facilities are likely to be located in the same major basins as
facilities as represented by the 2026 projections.
• Air quality simulation input information are at a 12 km by 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 specific air pollutant metric evaluated in this assessment, warm season maximum
daily eight-hour ozone average concentrations, is focused on longer-term exposures
that have been linked to adverse health effects. This assessment does not evaluate
4-33
-------
disparities in other potentially health-relevant metrics, such as shorter-term exposures
to ozone.
• Ozone 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 ozone-attributable EJ
health impacts.
Population variables considered in this EJ exposure assessment include race, ethnicity,
educational attainment, employment status, health insurance status, linguistic isolation, poverty
status, age, and sex (Table 4-12).125 Three additional population variables were included in this
analysis that were not included in previous RIAs (historically redlined areas, Tribal lands, and
life expectancy), one of which one is a measure of cumulative impacts (life expectancy).
The variable "redlined areas" was added to this RIA to assess exposure in communities
with a legacy of discriminatory land use designations and siting decisions (i.e., historically
redlined areas). We use graded census tracts developed by Noelke et al. (2022) from digitized
Home Owners' Loan Corporation (HOLC) residential security maps overlaid onto 2010 Census
tracts. Each census tract is classified as being covered by "Mainly A," "Mainly B," "Mainly C,"
and "Mainly D" grading, corresponding to coverage of different hazard ratings from original
HOLC maps. The dataset covers 14,818 census tracts, since HOLC maps only covered certain
urban areas. This dataset was adapted to cover 72,538 census tracts for use in BenMAP, with the
remaining census tracts categorized as "redlinedna" since they were not covered by HOLC
grading. Census tracts labeled as "Mainly D" were categorized as "redlined" and census tracts
that were mainly A-C were categorized as "not redlined." The practice of redlining was in effect
across 239 cities and, although illegal now for many decades, has had lasting effects on
investments in "redlined" neighborhoods where greater proportions of low income and people of
125
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 2060, 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. Projections in each
county are determined simultaneously with every other county in the U.S. to consider patterns of economic growth
and migration. The sum of growth in county-level populations is constrained to equal a previously determined
national population growth, based on Bureau of Census estimates (Hollmann et al., 2000). According to Woods and
Poole, linking county-level growth projections together and constraining to a national-level total growth avoids
potential errors introduced by forecasting each county independently. 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).
4-34
-------
color still reside and tend to have poorer health outcomes (Lee et al., 2022; Mitchell & Franco,
2018; Noelke et al., 2022; Swope et al., 2022).
The addition of Tribal lands, as defined by the Bureau of Indian Affairs, enhanced our
analysis relative to previous RIAs by addressing previous issues with undercounting American
Indians126 and serving as a health metric for communities of Tribal populations. Additionally, the
American Indian population as a whole includes individuals who live on non-Tribal lands who
have different exposures and access to resources than those individuals who live on Tribal lands,
so the Tribal lands variable is a community health metric (based on residency in a specifically
designated Tribal lands location) rather than a population health metric (based on membership in
a specific population subgroup, regardless of location). Thus, evaluating exposures on Tribal
lands (vs. non-Tribal lands) may better characterize true disparities in exposures among
American Indians.
Lastly, as one way to assess cumulative exposures and impacts, the life expectancy
variable has been added to differentiate between populations with differing baseline health levels
and measures the average life expectancy within a census tract. For average life expectancy, low
values indicate a higher overall burden or cumulative risk, while higher values indicate a lower
overall burden or cumulative risk. The life expectancy data comes from CDC's U.S. Small-area
Life Expectancy Estimates Project (USALEEP), which produced census tract-level life
expectancy estimates at birth for the period 2010-2015 which have been used in other analysis
tools such as NEXUS and CEJST.
The data sources and processing methodology for each dataset are described below.
County-level datasets were generated for 3,109 counties in the contiguous U.S.
126 American Indians are the most under-counted group in the U.S. Census with more than 80% of reservation lands
in "hard-to-count (HTC) census tracts" U.S. Department of the Interior, Indian Affairs, see I. Collection of Tribal
Enrollment Count | Indian Affairs (bia.gov).
4-35
-------
Table 4-12 Demographic Populations Included in the Ozone EJ Exposure Analysis from
Oil and Natural Gas VOC Emissions
Spatial Scale of
Population
Groups
Ages
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
Status
Insured; Uninsured
0-64
County
Linguistic
Isolation
English "well or better"; English < "well"
0-99
Census tract
Poverty Status
Above the poverty line; Below the poverty line
0-99
Census tract
Redlined Areas
HOLCa 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
aHome Owners' Loan Corporation (HOLC)
4.3.3.1 National Aggregated Results
We begin by evaluating potential disparities in ozone exposure from VOC emission
impacts aggregated across the continental U.S. (i.e., national scale). National average baseline
ozone concentrations in parts per billion (ppb) in 2038 are shown in the Figure 4-1 heat map for
each population group. Baseline concentrations represent the total estimated ozone exposure
burden in the absence of the rulemaking averaged over the April-September warm season
average. The concentration values are colored to more easily visualize differences in average
concentrations (lighter purple coloring represents lower average concentrations and darker purple
coloring represents higher average concentrations). It should be noted that in general, national
average ozone exposures in the baseline are relatively low (about 40 ppb), and that the national
ozone disparities observed in the baseline are similar to those described by recent rules (e.g.,
Regulatory Impact Analysis for the Proposed NSPS for Greenhouse Gas Emissions from New,
Modified, and Reconstructed Fossil-Fuel Fired Electric Generating Units).127 Populations
127
See https://www.epa.gov/system/files/documents/2023-05/utilities_ria_proposal_2023-05.pdf.
4-36
-------
exposed to national average ozone concentrations higher than that faced by the reference
population of all people aged 0-99 are ordered by the largest to the smallest differences:
American Indians, Hispanics, those on Tribal land, those linguistically isolated, Asians, those in
historically redlined areas, those who are unemployed, those who are less educated, the insured,
children, adults, and Whites (Figure 4-1). Columns labeled "Absolute Reductions" and
"Percentage Reductions" provide information regarding how the final rule is projected to reduce
warm season average ozone concentrations for the various population groups in the year 2038.
Most population groups were predicted to experience national-level ozone concentration
reductions of approximately 0.025 ppb (0.06 percent) compared to their baseline ozone
concentration exposure. The magnitudes of these ozone concentration reductions across
population demographics are all similar and small in magnitude with the exception of those
living on Tribal lands who are predicted to experience nearly 3 times as large of a reduction in
ozone concentrations as other population groups.
The national-level assessment of ozone exposure with and without implementation of the
final rule rulemakings suggests that while EJ exposure disparities are present in the pre-policy
scenario, these concerns are not likely mitigated or exacerbated by the rule for the population
groups evaluated, due to the similar and small differences in magnitudes of ozone concentration
reductions across demographic groups.
4-37
-------
r
2038
Population
Groups
Populations (Ages)
Baseline
(PPb)
Absolute
Reductions
(PPb)
Percentage
Reductions
Reference
Reference (0-99}
41.1
0.024
0.058
American Indian (0-99)
43.1
0.030
0.070
Race
Asian (0-99)
42.0
0.019
0.045
Black (0-99)
39.8
0.022
0.055
White (0-99)
41.2
0.025
0.061
Ethnicity
Non-Hispanic (0-99)
40.5
0.024
0.059
Hispanic (0-99)
42.8
0.024
0.056
Educational
More educated (>24)
40.9
0.024
0.059
Attainment
Less educated (>24)
41.5
0.023
0.055
Employment
Status
Employed (0-99)
41.1
0.025
0.061
Unemployed (0-99)
41.5
0.023
0.055
Not in the labor force (0-99)
41.0
0.024
0.059
Insurance
Insured (0-64)
41.2
0.024
0.058
Status
Uninsured (0-64)
40.7
0.028
0.069
Linguistic
English "well or better (0-99)
41.0
0.024
0.059
Isolation
English < "well" (0-99)
42.3
0.021
0.050
Life
Expectancy
Top 75% (0-99)
41.3
0.023
0.056
Bottom 25% (0-99)
40.1
0.028
0.070
Life expectancy data unavailable (0-99)
41.3
0.026
0.063
Poverty Status
>Poverty line (0-99)
41.0
0.024
0.058
-------
reductions, they do not provide information on the full distribution of concentration impacts.
This is because both demographic groups and ambient concentrations can be unevenly
distributed across the spectrum of exposures, meaning that average exposures may mask
important spatially localized disparities. To evaluate how the distribution of warm-season
exposures varies within and across demographic groups at the county level, we plot the full array
of exposures projected to be experienced by the entirety of each population (Figure 4-2
Baseline Distributions of Ozone Concentration (ppb) Across Populations in 2038
under the Final Rule (warm-season average of 8-hour daily maximum)
While national average results can provide some insight when comparing across
population impacts of EJ ozone exposure impacts from oil and natural gas VOC emission
reductions, they do not provide information on the full distribution of concentration impacts.
This is because both demographic groups and ambient concentrations can be unevenly
distributed across the spectrum of exposures, meaning that average exposures may mask
important spatially localized disparities. To evaluate how the distribution of warm-season
exposures varies within and across demographic groups at the county level, we plot the full array
of exposures projected to be experienced by the entirety of each population (). Distributional
figures present the running sum as a percent of each group's total population on the y-axes (i.e.,
cumulative percent of population). By constructing the cumulative percent metric, we are able to
directly compare warm-season ozone exposures across demographic populations with different
population sizes. The x-axes show baseline warm-season ozone concentrations (ppb) from low to
high concentrations. Ozone concentrations are county-level averages from all Census tracts in
the contiguous U.S. In other words, plots compare the running sum of each population against
baseline warm-season ozone concentrations such that populations whose trendlines are further
right on the plot have a higher proportion of their population exposed to higher concentrations.
For example at the national-level, the proportion of Hispanics and the proportion of non-
Hispanics exposed to lower concentrations of ozone is roughly equal because the trendlines
overlap on the left side of the plot; however, at higher concentrations of ozone on the right side
of the plot, the trendline for Hispanics is further right than for non-Hispanics, which indicates
that Hispanics are exposed to higher concentrations of ozone in a larger proportion than non-
Hispanics. ). Distributional figures present the running sum as a percent of each group's total
population on the y-axes (i.e., cumulative percent of population). By constructing the cumulative
percent metric, we are able to directly compare warm-season ozone exposures across
4-39
-------
demographic populations with different population sizes. The x-axes show baseline warm-season
ozone concentrations (ppb) from low to high concentrations. Ozone concentrations are county-
level averages from all Census tracts in the contiguous U.S.128 In other words, plots compare the
running sum of each population against baseline warm-season ozone concentrations such that
populations whose trendlines are further right on the plot have a higher proportion of their
population exposed to higher concentrations. For example, at the national-level, the proportion of
Hispanics and the proportion of non-Hispanics exposed to lower concentrations of ozone is
roughly equal because the trendlines overlap on the left side of the plot; however, at higher
concentrations of ozone on the right side of the plot, the trendline for Hispanics is further right
than for non-Hispanics, which indicates that Hispanics are exposed to higher concentrations of
ozone in a larger proportion than non-Hispanics.
128 Distributional figures in the proposal RIA EJ exposure assessment were based on county-level averages. While
tract-level averages are preferable due to the higher resolution, they required substantial additional computing power
(about 10-fold) and generate similar results. Therefore, EPA will select the geographic resolution that is most
reasonable in future EJ assessments.
4-40
-------
a
| ^
1 i
2
g
£
'S
£
£
*
Ethnicity
Educational
Attainment
Employment
Status
Insurance Status
Life Expectancy 5-
a
T5
Linguistic
Isolation
Poverty Status
Rod 1 mod Areas
Tribal Land
o
#
o
*
Population
l wince (0-99)
American Indian (0-99)
I Asian (0 99)
I Black (0-99)
I Nw-Hiipanic (0-99)
I Hispanic (0-99)
~ More educated { >24)
L est educated ( >?4)
I employed (O-W)
I unemployed (0-99)
t Not in the labor force (0-99)
I Insured (0-64)
I Uninsured (O 6-4}
I Top7S**<0-^9)
I Bottom 2^% (0 99)
Life expectancy data unavailable (O W)
I Enftiisfi "well or better" (0-99)
I English < "well" (0-9^)
I >Poverty line (0-99)
I « Poverty line (O WJ
I HOLC Grades A-C CO-9^>
I HOLC Grade O (0^9)
Not Graded by HOLC (0-99)
\ Not Tribal land (0-99)
I Tribal land (0-99)
30 35 AO 45 SO 55
Ozone (ppb, warm «ea»on avg) *
Figure 4-2 Baseline Distributions of Ozone Concentration (ppb) Across Populations in
2038 under the Final Rule (warm-season average of 8-hour daily maximum)
4-41
-------
While national average results can provide some insight when comparing across
population impacts of EJ ozone exposure impacts from oil and natural gas VOC emission
reductions, they do not provide information on the full distribution of concentration impacts.
This is because both demographic groups and ambient concentrations can be unevenly
distributed across the spectrum of exposures, meaning that average exposures may mask
important spatially localized disparities. To evaluate how the distribution of warm-season
exposures varies within and across demographic groups at the county level, we plot the full array
of exposures projected to be experienced by the entirety of each population (). Distributional
figures present the running sum as a percent of each group's total population on the y-axes (i.e.,
cumulative percent of population). By constructing the cumulative percent metric, we are able to
directly compare warm-season ozone exposures across demographic populations with different
population sizes. The x-axes show baseline warm-season ozone concentrations (ppb) from low to
high concentrations. Ozone concentrations are county-level averages from all Census tracts in
the contiguous U.S.129 In other words, plots compare the running sum of each population against
baseline warm-season ozone concentrations such that populations whose trendlines are further
right on the plot have a higher proportion of their population exposed to higher concentrations.
For example at the national-level, the proportion of Hispanics and the proportion of non-
Hispanics exposed to lower concentrations of ozone is roughly equal because the trendlines
overlap on the left side of the plot; however, at higher concentrations of ozone on the right side
of the plot, the trendline for Hispanics is further right than for non-Hispanics, which indicates
that Hispanics are exposed to higher concentrations of ozone in a larger proportion than non-
Hispanics.
As the baseline scenario shown in Figure 4-2 Baseline Distributions of Ozone
Concentration (ppb) Across Populations in 2038 under the Final Rule (warm-season
average of 8-hour daily maximum)
While national average results can provide some insight when comparing across
population impacts of EJ ozone exposure impacts from oil and natural gas VOC emission
reductions, they do not provide information on the full distribution of concentration impacts.
129
Distributional figures in the proposal RIA EJ exposure assessment were based on county-level averages. While
tract-level averages are preferable due to the higher resolution, they required substantial additional computing power
(aboutlO-fold) and generate similar results. Therefore, EPA will select the geographic resolution that is most
reasonable in future EJ assessments.
4-42
-------
This is because both demographic groups and ambient concentrations can be unevenly
distributed across the spectrum of exposures, meaning that average exposures may mask
important spatially localized disparities. To evaluate how the distribution of warm-season
exposures varies within and across demographic groups at the county level, we plot the full array
of exposures projected to be experienced by the entirety of each population (). Distributional
figures present the running sum as a percent of each group's total population on the y-axes (i.e.,
cumulative percent of population). By constructing the cumulative percent metric, we are able to
directly compare warm-season ozone exposures across demographic populations with different
population sizes. The x-axes show baseline warm-season ozone concentrations (ppb) from low to
high concentrations. Ozone concentrations are county-level averages from all Census tracts in
the contiguous U.S. In other words, plots compare the running sum of each population against
baseline warm-season ozone concentrations such that populations whose trendlines are further
right on the plot have a higher proportion of their population exposed to higher concentrations.
For example at the national-level, the proportion of Hispanics and the proportion of non-
Hispanics exposed to lower concentrations of ozone is roughly equal because the trendlines
overlap on the left side of the plot; however, at higher concentrations of ozone on the right side
of the plot, the trendline for Hispanics is further right than for non-Hispanics, which indicates
that Hispanics are exposed to higher concentrations of ozone in a larger proportion than non-
Hispanics. is similar to that described by other RIAs (e.g., the Regulatory Impact Analysis for
the Proposed NSPS for Greenhouse Gas Emissions from New, Modified, and Reconstructed
Fossil-Fuel Fired Electric Generating Units),130 we will now discuss the ozone concentration
reductions due to this proposed rulemaking. Results allow evaluation of what percentage of each
subpopulation (e.g., Hispanics) in the contiguous U.S. experience what concentration reduction
of ozone (in ppb) post-policy in 2038. The small difference in impacts shown in the 2038
distributional analyses of ozone concentrations under the final rule suggests that the rule is not
likely to meaningfully exacerbate or mitigate ozone exposure concerns for population groups
evaluated (Figure 4-3). However, while the impacts may be small in magnitude, the differences
in impacts between groups still exist and should be noted. At the national level, those who live
on Tribal lands are expected to experience higher reductions in ozone concentrations than those
130
See https://www.epa.gOv/system/files/documents/2023-05/utilities_ria_proposal_2023-05.pdf.
4-43
-------
who live on non-Tribal lands. Because those who live on Tribal lands also have higher baseline
ozone exposures (Figure 4-1), this larger reduction in ozone exposure can be expected to
somewhat mitigate existing disparities in ozone exposure between those who live on Tribal lands
and those who do not, even if only to a small degree.
4-44
-------
Ethnicity
Educational
Attainment
I unpkiymuni
Status
Hfw
Expccrancy
4Tl»QUIi*kt «C
I iQfH
Povorty
Status £
T5
J
To
RpfHlrinrl
Areas £
Tr»r>ril I And
PopulAt'on
I WlMltt (0 99)
An»»rir«n lnd*'i
Asian (0-99)
Black (0 99)
I Nofi* HisfMnic {O-
Hi;fk3nic(0-99)
i Mors oducated { >24)
I«$« «Klu^4)
i ffinpj|6y»d (O W>)
I Unamployvd <0 *
) Nk»l in ih« l*tK>r i
i Top ?S% (C-99)
Bottom 21i% (0 99)
Life p«n«cty HOIC (0 99)
I Mot Tribal Und (O OO)
I Tral land (0'99)
201S
Policy
D,ao c>,uo r.YA o ot» o.on a. jo ci, i
O/ono {ppb. warm ioa«on stvg) *¦
Figure 4-3 Distributions of Ozone Concentration Reductions Across Populations in 2038
under the Final Rule
4-45
-------
4.3.3.3 State Aggregated Results
Due to the spatial heterogeneity of EJ ozone exposure from VOC emission impacts, we
also provide state-level results that show ozone concentration reductions by state and
demographic population in 2038 for the 48 states and the District of Columbia (D.C.) in the
contiguous U.S. for the final rule (Figure 4-4). In this heat map, a darker purple shade indicates
larger ozone reductions, with demographic groups shown as rows and each state as a column.
White cells indicate that certain populations do not exist within a particular state (e.g., there are
no Tribal lands or historically redlined areas). The state-specific demographic populations are
projected to experience reductions in ozone concentrations of up to 0.185 ppb (observed for
those living on non-Tribal lands in Oklahoma in 2038). It is also important to note that there are
no observed ozone increases for any population groups across states, only reductions. Figure 4-4
shows that within most states, demographic groups are predicted to experience very similar
exposure impacts as the state reference populations.
When comparing exposure impacts across demographic groups within states, most states
display similar impacts across demographic groups in 2038. However, some states with higher
baseline exposures have larger differences in reductions between groups. For example, in
Colorado, which has the fourth highest baseline exposure at 52 ppb in 2038, the largest
difference in reductions between the ozone exposure for a population and for the reference
population is 0.112 ppb for those living on Tribal land. Uncertainties about the precision of the
air quality surfaces in capturing changes of spatial patterns in ozone concentrations within states
is discussed above.
Therefore, the state-level assessment of ozone exposure changes due to the final rule
suggests that while the final rule will not meaningfully mitigate or exacerbate ozone EJ exposure
disparities for most population groups evaluated in 2038, for some population groups, there may
be mitigated or exacerbated ozone EJ exposure disparities although at small magnitudes of
change. In particular, because baseline ozone concentrations are higher for those who live on
Tribal lands, and the expected reduction in ozone exposure is larger for those who live on Tribal
lands, the existing disparities in ozone exposure between those who do and don't live on Tribal
lands will be mitigated somewhat, even if only slightly.
4-46
-------
Ozone(ppb)
a«uuuQDu.u2Si±jiS3522S552522ZzZzz2oooaaiyi1fli:i-D»$§gg
Year Scenario Population Groups
2038 Policy Reference
Race
Ethnicity
Educational Attainm.
Employment Status
Insurance Status
Life Expectancy
Linguistic isolation
Poverty Status
Red lined Areas
Ages
Sex
Tribal Land
Population
Reference (0-99)
American Indian (0-99)
Asian (0-99)
Black (0-99)
White (0-99)
Non-Hispanic (0-99)
Hispanic (0-99)
More educated (>24)
Less educated (>24)
Employed (0-99)
Unemployed (0-99)
Not in the labor force (0-99)
insured (0-64)
Uninsured (0-64)
Top 75% (0-99)
Bottom 25% (0-99)
Life expectancy data unavailable (0-99)
English "well or better" (0-99)
English < "well" (0-99)
>Povertyline (0-99)
-------
result of this is that we cannot estimate risks from the source category alone, but rather only from
the larger industry sector. This means the source category risks are likely overestimated.
Another implication of the data limitations is that the assessment is considered a screen — it is
an estimate of potential risks over a broad area. More refined emissions data would be needed to
conduct an assessment where we could draw more accurate conclusions about risk to specific
areas and populations.
Most of these emissions (97 percent) are treated as "nonpoint" emissions which are
allocated from county-level data down to grid cells (4 km in the continental U.S. (CONUS), 9
km in Alaska) based on emissions surrogates. This means that we are making assumptions about
the spatial distribution of these emissions that may not be accurate. The approximately 3 percent
of emissions that are categorized as "point" in the NEI are emitted from about 400 facilities
across the country. For these sources, we can estimate potential exposures and impacts more
precisely. Also, we note that some sources categorized as oil and natural gas sources in the NEI
are not in the source category for this final rule.
The oil and natural gas sector was one of the sectors assessed in the 2014 National Air
Toxics Assessment (NATA). In that assessment, the nonpoint emissions were also modeled as 4
km grid cells in CONUS (9 km grid cells in Alaska) and the point emissions were modeled as
point sources in the American Meteorological Society/Environmental Protection Agency
Regulatory Model (AERMOD) using census blocks as model receptors. However, NATA risk
estimates were not presented at census block level because of uncertainties associated with the
analysis, such as not knowing exactly where in each grid cell the emissions are occurring.
Instead, NATA risk results were presented at census tract level by population-weighting the
block risks up to the tract level. Because census tracts can have large areas, the tract-level risks
may not reflect potential elevated risks present at a finer scale. The highest tract-level cancer risk
from nonpoint oil and natural gas emissions in the 2014 NATA was 30-in-l million, and only
about 30 tracts (out of approximately 74,000 tracts nationwide) had risks greater than 10-in-l
million. For comparison, the nationwide median total cancer risk estimate from the 2014 NATA
(considering contributions from all source types) was about 30-in-l million across all census
tracts.
4-48
-------
Here, using updated emissions and population data, we have conducted a new analysis of
HAP-related exposures and risks across the United States. In this analysis, to assess the potential
for elevated risks at a scale finer than the census tract level, we aggregated the block-level
AERMOD results from the modeling of the 2017 NEI nonpoint HAP emissions to the same 4 km
and 9 km grid cells that nonpoint emissions are allocated to. There are about 500,000 4 km grid
cells in CONUS, compared to about 74,000 census tracts so, on average, grid cells are at a finer
scale than census tract. For each grid cell, we used the median cancer risk of all the blocks that
have their internal point (or centroid) located within the grid cell. Census block demographic
data were also aggregated to each 4 km grid cell and risks were calculated at the census blocks
from the approximately 400 sources included in the 2017 NEI as point sources and added the
highest block-level risk for each point source "facility" to the median cell nonpoint risk for the
cell containing the block.
The data used in this analysis include spatial data of the grid cells, 2010 census block
location and population data,131 AERMOD-modeled oil and natural gas 2017 HAP concentrations
at census block level for the nonpoint and point sources, and 2015-2019 block-group
demographic data. There are separate files for the 4 km grid cells that cover CONUS and the 9
km grid cells for Alaska, each using a Lambert Conformal Conic projected coordinate system.
These are the same grid definition used for the 2014 NATA nonpoint oil and natural gas
emissions. The census data are for the year 2010, with a small number of changes made to the
locations (and sometimes deletions) of specific census blocks based on the RTR pre-modeling
review of specific source categories since the 2010 census data were first available (the current
oil and natural gas AERMOD modeling is based on the census block receptor file as of May
2019). The AERMOD modeling performed (version 19191) using 2017 NEI and meteorology
data followed the same methodology used in the 2014 NATA (U.S. EPA, 2018). Demographic
data on total population, race, ethnicity, age, education level, low household income, poverty
status and linguistic isolation were obtained from the Census' American Community Survey
(ACS) 5-year averages for 2015-2019.132
131
Data Summary File 1 available at http://www2.census.gov/census_2010/04-Summary_File_l/. See also
Technical Documentation for the 2010 Census Summary File 1.
132
Data available at https://www2.census.gov/programs-surveys/acs/summary_file/2019/data/5_year_entire_sf/.
4-49
-------
The AERMOD-modeled census block concentrations are based on the 2017 NEI
emissions data (see Table 3-9Table 3-9). The process by which emissions were calculated and
allocated to grid cells in the case of nonpoint emissions is discussed in the technical support
document for the 2017 NEI and the emissions modeling summary for 2017, respectively (U.S.
EPA, 2020b). Emissions data are publicly available online.133 These emissions were modeled in
AERMOD (version 19191), and the resulting block-level annual concentrations of each pollutant
were used to calculate cancer risks. The pollutant cancer unit risk estimates used to calculate
risks are from the toxicity value files available on the Human Exposure Model website.134 For
each census block, the cancer risks were summed over all pollutants to obtain a total cancer risk.
The demographic data from the ACS were joined to each census block based on the block group
ID (the first 12 characters of the census block ID).
For nonpoint sources, the census blocks were spatially joined to the grid cells (4 km
CONUS, 9 km Alaska), and the block data were aggregated at the cell level, using the median
cancer risk of the blocks in each cell, and the sum of block populations and the individual
demographic group populations (using QGIS version 3.16.3). For point sources, the highest
modeled block risk for each facility was added to the median nonpoint risk for the cell containing
the block, to provide a measure of total point and nonpoint combined risk.
As noted previously, and discussed further in the next paragraph, there is significant
uncertainty in the resulting risk estimates. They are overly conservative due to the inclusion of
sources categorized as oil and natural gas sources in the NEI that are not in the source category
for this final rule. We estimate that there are approximately 146 million people with nonzero
total risk (3 million census blocks) in the gridded AERMOD modeling domain of the CONUS
and Alaska nonpoint oil and natural gas sources.135 The maximum cell risk estimate from oil and
natural gas sources is 200-in-l million, which occurs in two grid cells with an estimated 10
people (3 census blocks); Carbon County, Wyoming (with an estimated 3 people) and Weld
County, Colorado (with an estimated 7 people). The CONUS results are summarized in Table
4-13Table 4-13. The population exposed to cancer risk greater than 100-in-l million for oil and
133
Data available at https://gaftp.epa.gov/Air/emismod/2017/AERMOD_inputs/.
134
See https://www.epa.gov/fera/download-human-exposure-model-hem.
135
Blocks are within approximately 159,000 4 km CONUS grid cells and approximately 240 9 km Alaska grid cells.
4-50
-------
natural gas emissions is approximately 10 people (in two grid cells),136 and the population
exposed to cancer risk greater than or equal to 100-in-l million is approximately 40,000 people
(in 36 grid cells). There are about 140,000 people living (in 122 cells) where the cell risk
estimate is greater than or equal to 50-in-l million and about 6.8 million people living (in 9500
cells) where the cell risk estimate is greater than 1-in-l million. None of the cells in Alaska has
estimated cell cancer risk greater than 1-in-l million.
It is important to reiterate that these risk estimates are based on emissions from the entire
oil and natural gas sector, which includes sources outside the scope of this regulation. To provide
some context for how these sources relate to sources impacted by this final regulation, we
categorized the fraction of oil and natural gas HAP emissions in the 2017 NEI that were
attributed to different source types. For this exercise, we specifically focused on formaldehyde
and benzene emissions (the two pollutants that accounted for most of the calculated oil and
natural gas HAP risk) in the 36 grid cells with 2017 oil and natural gas HAP risk greater than or
equal to 100-in-l million. It is likely that most of the formaldehyde emissions and about a
quarter of the benzene emissions that were categorized as coming from oil and natural gas
sources in the 2017 NEI are from sources outside of this source category. Therefore, it also
follows that most of the estimated risk is likely being driven by sources not impacted by this
final regulation. It bears repeating that this is a screening assessment and full modeling would be
required to quantitatively split out risk of sources impacted by this rule from other sources
categorized in the NEI as oil and natural gas. Risk in grid cells of interest may not scale directly
to emissions within the grid cells.
For the point sources, there were 33 sources with estimated census block maximum
cancer risk greater than 1-in-l million, and only 6 sources with estimated risk greater than 10-in-
1 million (highest was 40-in-l million). There was only a single case where the maximum census
block risk from a point source, and the median cell risk from nonpoint sources (containing the
census block), were both greater than 10-in-l million. In that case, the point risk of 20-in-l
136 Demographic information for grid cells with risks greater than 100-in-l million are not included in Table 4-13.
There are too few census blocks (three) and people (10) to appropriately characterize this population based on the
ACS data available.
4-51
-------
million and the nonpoint cell risk of 40-in-l million combined for an estimated 60-in-l million
risk.
Figure 4-5 shows the cell cancer risk estimates in CONUS and Alaska. As indicated in
the map, most of the cells in the country (about 150,000 of them) have estimated risk less than 1-
in-1 million. Figure 4-6 is a larger-scale map that shows where the estimated cell risks are the
highest. The cells with estimated risk greater than or equal to 30-in-l million are in Colorado,
Utah, Wyoming, and North Dakota, and the cells with the highest estimated risk are all in
Colorado.
Table 4-13 also contains estimated numbers of people within various demographic
groups who live in areas above the specified risk levels. For nearly all the demographic groups
the percentage of people in the cells with estimated risk above the specified levels is at or below
the national average. Above a risk level of 50-in-l million, the percent people of color is about
the same as the national average, but the Hispanic/Latino demographic group is about 10
percentage points higher than the national average. The overall minority percentage is not
elevated compared to the national average because the Black percentage is much lower than the
national average. The demographic group of people aged 0-17 is slightly higher than the national
average. For people with estimated risk greater than 1-in-l million, Hispanic/Latino populations
and the age 0-17 group are below the national average, but the percentage of Native American
populations is higher than the national average. Given that disparities in the baseline are limited,
and overall, the regulation is reducing HAP substantially from oil and natural gas sector, we do
not expect that any population would experience disbenefits or that disparities would be
exacerbated by the final rule.
4-52
-------
Table 4-13 Cancer Risk and Demographic Population Estimates for 2017 NEI Nonpoint
Emissions from Oil and Natural Gas Sources
Number of Cells
Total Population
Risks > 100-in-l
million
Risks > 50-in-l
million
Risks > 1-in-l
million
36
38,885
(936 census blocks)
122
142,885
(3,204 census
blocks)
Population
%
%
9,499
6,804,691
(172,878 census
blocks)
'opulation %
Nationwide
%
People of Color
13,268
34.1
52,154
36.5
2,010,161
29.5
39.9
Black
140
0.4
1,434
1.0
535,055
7.9
12.2
Native American
77
0.2
465
0.3
59087
0.9
0.7
Other and Multiracial
1,443
3.7
5,148
3.6
323,397
4.8
8.2
Hispanic / Latino
11,608
29.9
45,107
31.6
1,092,621
16.1
18.8
Age 0-17
10,679
27.5
37,487
26.2
1,463,907
21.5
22.6
Age >65
4,272
11.0
17,188
12.0
1,085,067
15.9
15.7
Below the Poverty Level
2,000
5.1
13,455
9.4
902,472
13.2
13.4
Over 25 Without a High
School Diploma
2,788
7.2
11,320
7.9
488,372
7.2
12.1
Linguistically Isolated
808
2.1
4,418
3.1
179,739
2.6
5.4
4-53
-------
Median Cell Cancer Risks
(in a million)
¦I < 1
¦ 20-40
¦I 40-60
60-80
80- 100
> 100
Figure 4-5 National Map of Grid Cell Median Cancer Risks for 2017 Nonpoint Oil and
Natural Gas NEI Emissions
4-54
-------
Figure 4-6 Local-Scale Map of Grid Cell Median Cancer Risks for 2017 Nonpoint Oil
and Natural Gas NEI Emissions
4.3.5 Demographic Characteristics of Oil and Natural Gas Workers and Communities
The oil and natural gas industry directly employs approximately 140,000 people in oil
and natural gas extraction, a figure which varies with market prices and technological change, in
addition to many workers in related sectors that provide materials and services. Figure 4-7Figure
4-7 shows employment since 2001. " We see a dramatic increase in employment with the rapid
expansion in hydraulic fracturing from 2005 to 2014, a decrease after oil prices fell in 2014-
2015, and volatility in employment.
137
Data was obtained from the Bureau of Labor Statistics Current Employment Statistics program for N AICS code
211.
4-55
-------
220
200
g* 180
o
o
¦M
| 160
>-
_o
Q.
J] 140
120
100
Figure 4-7 National-level Employment in Oil and Natural Gas Production
The EPA also conducted a baseline analysis to characterize potential distributional
impacts on employment. A reduction in oil and natural gas activity could have a negative effect
on employment among oil and natural gas workers. This could also reduce employment,
earnings, and tax revenues in oil and natural gas intensive communities.138 Any effect on oil and
natural gas workers or oil and natural gas intensive locations would be a local and partial
equilibrium effect. In general equilibrium, there could be other and potentially offsetting effects
in other regions and sectors.
For the distribution of employment effects, we assessed the demographic characteristics
of (1) workers in the oil and natural gas sector and (2) people living in oil and natural gas
intensive communities. Comparing workers in the oil and natural gas sector to workers in other
sectors, oil and natural gas workers may have higher than average incomes, be more likely to
have completed high school, and be disproportionately Hispanic. People living in some oil and
138 For this analysis, oil and natural gas intensive communities are defined as the top 20 percent of communities with
respect to the proportion of oil and natural gas workers. Some analyses break the top 20 percent into subgroups
which are the 80lh-95lh percentiles, the 95th-97.5th percentiles, and above the 97.5th percentile by proportion of oil
and natural gas workers.
*—1
(N
m
LO
tiD
r-.
00
CT)
O
1
(N
m
LO
05
r-.
00
<71
O
1
o
O
o
o
o
O
o
o
O
*—1
*—1
*—1
*—i
*—1
*—1
*—1
*—i
*—1
*—1
-------
natural gas-intensive communities concentrated in Texas, Oklahoma, and Louisiana, may have
disproportionate income levels, rates of high school completion, and demographic composition.
Table 4-14Table 4-14 provides summaries of average income, the percentage of
population that is non-Hispanic White, the percentage of population that speaks only English in
the home, and the percentage of the population with four years of high school education, all
among people with reported income. The table lists these data for the United States, for oil and
natural gas workers, for other people, for people in oil and natural gas intensive communities,
and for people in other locations. We see that oil and natural gas workers are more highly paid,
more likely to be non-Hispanic White individuals, and have higher rates of only speaking
English, more likely to have four years of high school, and less likely to be over 50 years old
than workers in other sectors. People in oil and natural gas communities are demographically
similar to people in other communities. This suggests that, on average, reductions in oil and
natural gas drilling or production are unlikely to disproportionately impact marginalized
communities either via direct labor channels or spillover channels.
Table 4-14 Demographic Characteristics of Oil and Natural Gas Workers and
Communities
Sectors
Places
Overall
Oil and
Natural Gas
Other
Oil and Natural
Other
Workers
People
Gas Communities
Communities
All U.S.
Average Income
$115,000
$44,000
$42,000
$44,000
$44,000
% Non-Hispanic White
81%
71%
68%
69%
71%
% Non-Hispanic Native
American
0.97%
0.86%
1.5%
0.56%
0.86%
% English Only
87%
82%
80%
81%
82%
4 years of High School
97%
88%
86%
88%
88%
At Least 50 years old
44%
50%
49%
50%
50%
Note: Calculations based on United States Census Bureau American Community Survey public use microdata from
2015-2019.
This analysis uses 5-year ACS data from 2015-2019 retrieved from IPUMS. This is
approximately 16 million individual ACS responses. Oil and natural gas workers are identified
by working in industries with a NAICS code that begins with "211." Those are "Oil and natural
gas Extraction," as well as the sub-industries "Crude Petroleum Extraction" and "Natural Gas
Extraction."
4-57
-------
The level of communities is the Public Use Microdata Area (PUMA). PUMAs are
districts defined by the United States Census Bureau. PUMA data is procured from IPUMS.
They generally have 100,000-200,000 people with an average of about 140,000 people. The
average spatial area of a PUMA is 1,692 square miles. We analyze PUMAs because economic
spillovers in this sector occur at a multicounty scale. The oil and natural gas sector includes both
substantial intercounty commuting and regional supply chains. Additionally, PUMAs are the
smallest geographic unit for which detailed individual data are available. In Table 4-14Table
4-14, oil and natural gas communities are defined as the 20 percent of PUMAs with the highest
percentage of oil and natural gas workers. Figure 4-8 shows all PUMAs in the continental United
States. Oil and natural gas communities as defined in Table 4-14 are highlighted.
Figure 4-8 Continental U.S. Map of Pl'MAs and Oil and Natural Gas Intensive
Communities
Table 4-15 describes demographics by a region's oil and natural gas (O&G) intensity.
Non-oil and natural gas intensive regions (column (1)) are the bottom 80 percent by portion of
workers in the oil and natural gas industry. Most of these have no reported oil and natural gas
workers. Low oil and natural gas intensive regions (column (2)) are between the 80th and 95th
percentiles of oil and natural gas industry employment, high (column (3) are the 95th-97.5th, and
very high (column (4)) are above the 97.5th percentile. People in oil and natural gas communities
4-58
-------
of Table 4-15 are divided between columns (2)-(4). The trimmed comparison group (column (5)
is people in non-oil and natural gas intensive regions in states that contain any PUMAs with high
or very high intensity. The group of states with high oil and natural gas intensity may be a more
appropriate comparison by removing regions of the country which do not resemble oil and
natural gas intensive areas, such as the Atlantic coast states.
We see in Block A that people in oil and natural gas intensive communities (columns
(2)-(4)) are more likely to be White and Indigenous than people in non-oil and natural gas
intensive areas (column (1)). In Block B, we see that people in oil and natural gas-intensive
areas' more likely to be Hispanic than people in non-O&G intensive areas. In Block C, we see
income, percentage of population with four years of high school education, fraction working in
the oil and natural gas industry, and fraction age 50 or older. Comparing people in high and very
high oil and natural gas intensity regions (columns (3) and (4)) to people in the trimmed
comparison group (column (5)), we see that people in in high oil and natural gas intensity
regions are more likely to be White, non-Hispanic, Native American, and less likely to be Asian
American or Pacific Islanders.
Table 4-15 Demographic Characteristics of Oil and Natural Gas Communities by Oil
and Natural Gas Intensity
Category
Non-O&G
Intensive
(1)
Low O&G
Intensity
(2)
High O&G
Intensity
(3)
Very High
O&G
Intensity
(4)
Trimmed
Comparison
Group
(5)
Block A
White
77%
81%
84%
78%
73%
Black
10%
8%
8%
7%
8%
Native American
1%
2%
2%
3%
1%
Asian American or Pacific
Islander
6%
3%
2%
5%
9%
Other Race
4%
3%
2%
4%
7%
Multiple races
2%
2%
2%
3%
3%
Block B
Non-Hispanic
88%
84%
86%
81%
80%
Hispanic
12%
16%
14%
19%
20%
Block C
Income
$44,000
$40,000
$41,000
$47,000
$44,000
Four years of High School
88%
87%
87%
86%
87%
Working in O&G
0.006%
0.1%
0.4%
1%
0.008%
At Least 50 Years Old
50%
50%
50%
47%
48%
4-59
-------
Note: Calculations based on United States Census Bureau American Community Survey public use microdata from
2014-2019. Totals may not appear to add correctly due to rounding.
Table 4-16Table 4-16 shows the percentage of people by racial group identification for
Hispanics and non-Hispanics, across oil and natural gas intensity. We see that people in high and
very high intensity communities are more likely to be Hispanic Whites and non-Hispanic Native
Americans, and less likely to be non-Hispanic Asian American and Pacific Islanders than people
in non-oil and natural gas intensive communities.
Table 4-16 Hispanic Population by Oil and Natural Gas Intensity
Category
Non-
O&G
Intensive
(1)
Low
O&G
Intensity
(2)
High
O&G
Intensity
(3)
Very High
O&G
Intensity
(4)
Trimmed
Comparison
Group
(5)
Non-Hispanic White
69%
69%
73%
65%
60%
Non-Hispanic Black and African-American
10%
8%
7%
7%
8%
Non-Hispanic Native American
1%
2%
1%
3%
0%
Non-Hispanic Asian American or Pacific
Islander
6%
3%
2%
5%
9%
Non-Hispanic Other Race
0%
0%
0%
0%
0%
Non-Hispanic Multiple Races
2%
2%
2%
2%
2%
Hispanic White
8%
12%
11%
14%
12%
Hispanic Black and African-American
0%
0%
0%
0%
0%
Hispanic Native American
0%
0%
0%
0%
0%
Hispanic Asian American or Pacific Islander
0%
0%
0%
0%
0%
Hispanic Other Race
3%
3%
2%
4%
6%
Hispanic Multiple Races
1%
1%
0%
1%
1%
Note: Calculations based on United States Census Bureau American Community Survey public use microdata from
2014-2019. Totals may not appear to add correctly due to rounding.
Marginalized communities are overrepresented in some oil and natural gas intensive
communities. Figure 4-9Figure 4-9 highlights oil and natural gas intensive communities with
substantial EJ concerns in darker blue. These communities are in the bottom twenty-five percent
by income or high-school graduate or non-Hispanic White population percentage. They are
concentrated in Texas, Louisiana, and Oklahoma.
4-60
-------
Figure 4-9 Map of Oil and Natural Gas Intensive Communities with Environmental
Justice Concerns
4.3.6 Household Energy Expenditures
Energy provides many services to households that are necessary for a basic standard of
living. The final regulatory requirements will obligate affected sources to incur costs to reduce
emissions, which impact the supply and prices of oil and natural gas and generate energy market
impacts, though these impacts are expected to be minimal (see Section 4.1). This section
characterizes how household energy expenditures vary across the income distribution and for
different racial and ethnic groups. The goal of this section is to highlight which populations and
communities may be most vulnerable to potential energy market effects caused by regulatory
impacts on the oil and natural gas industry.
Energy insecurity, poverty, and access are important concepts in the discussion of energy
burden. Energy insecurity occurs when households lack certainty that they will be able to
consume adequate and sufficient energy to meet basic needs. Energy poverty exists when
households need to pay disproportionate costs for energy use due to low income, higher energy
bills, or inefficient energy use. Energy access barriers exist when households lack access to
affordable, reliable energy. Energy insecurity and poverty are persistent problems facing many
4-61
-------
households across the U.S. (Bednar & Reames, 2020; Kaiser & Pulsipher, 2006; U.S. EIA, 2018)
and they have many consequences for human health and wellbeing (Hall, 2013; Jessel et al.,
2019; Karpinska & Smiech, 2020). The EIA found that nearly a third of U.S. households faced
challenges paying their energy bills or could not maintain adequate heating or cooling in 2015.
For purposes of this section, "energy burden" focuses primarily on energy poverty.
Low-income and minority households tend to face disproportionately high energy
burdens (Hernandez et al., 2014; Wang et al., 2021) and thus are particularly vulnerable when
energy prices increase. Although these households consume less energy, energy tends to
represent a larger share of their budgets. (Drehobl et al., 2020) find that low-income, Black,
Hispanic, Native American, and older adult households have disproportionally higher energy
burdens than the average household. Lyubich (2020) finds that Black households spend more on
residential energy than White households even after controlling for income, household size, city,
and homeowner status. Wang et al. (2021) find that Black households spent more on energy than
other households at every point on the income distribution, suggesting that energy efficiency
issues may be more problematic in Black households. They identify geographic location,
climate, the characteristics of dwellings, and socioeconomic characteristics as primary drivers of
residential energy use and energy burden.
To investigate baseline energy expenditures and potential distributional impacts of
possible increases in energy costs, we assessed expenditure and income data stratified by pre-tax
income quintiles and race/ethnicity from the 2019 Consumer Expenditure Survey (CES) from the
U.S. Bureau of Labor Statistics. We combined expenditures in the following four categories to
approximate "energy expenditures": (1) Natural gas, (2) Electricity, (3) Fuel oil and other fuels,
and (4) Gasoline, other fuels, and motor oil (transportation). The first three categories are
residential energy expenditures, and the fourth category represents transportation energy
expenditures. These categories are assumed to potentially experience price impacts due to
regulatory costs affecting the oil and natural gas industry, though we expect impacts to be
minimal (see Section 4.1).
We examined energy expenditures, the ratio of household energy expenditures to total
household expenditures, and the ratio of household energy expenditures to after-tax income
across income quintiles and racial groups. It is important to note that energy burden is sensitive
4-62
-------
to the particular energy services and expenditures are included and how income is defined (e.g.,
whether transfer payments or taxes are included in income calculation; the inclusion of
transportation-related energy expenditures).
Table 4-17Table 4-17 shows energy expenditures by quintiles of pre-tax income. The
data indicate that the highest income group consumes the most energy and spends the most per
household on it, but energy expenditures represent a smaller percentage of their total
expenditures and a much smaller percentage of their income than the lowest income quintile.
Energy expenditures as a share of total household expenditures were 8.3 percent for the lowest
income quintile and 4.9 percent for the highest income quintile. For energy expenditures as a
share of average after-tax income, the distribution is more unequal, ranging from 19.4 percent for
the lowest income quintile to 3.4 percent for the highest income quintile. This means the lowest
income households are spending over five times more of their income on energy than the highest
income households.
Table 4-17 Energy Expenditures by Quintiles of Income before Taxes, 2019
Lowest
Second
Third
Fourth
Highest
Metric
All
20%
20%
20%
20%
20%
Income after taxes
71,487
12,236
32,945
53,123
83,864
174,777
Annual expenditures
63,036
28,672
40,472
53,045
71,173
121,571
Natural gas
416
259
355
367
455
644
Electricity
1,472
1,049
1,351
1,446
1,587
1,924
Fuel oil and other fuels
113
69
101
86
121
189
Gasoline, other fuels, and
2,094
998
1,601
2,079
2,593
3,193
motor oil (transportation)
Energy expenditures
4,095
2,375
3,408
3,978
4,756
5,950
Energy expenditures as share of
total expenditures
6.5%
8.3%
8.4%
7.5%
6.7%
4.9%
Energy expenditures as share of
income
5.7%
19.4%
10.3%
7.5%
5.7%
3.4%
Quintile share of all energy
expenditures
11.6%
16.7%
19.4%
23.2%
29.1%
Source: Consumer Expenditure Survey, U.S. Bureau of Labor Statistics, September 2020.
https://www.bls.gOv/cex/tables/calendar-year/mean-item-share-average-standard-error.htm#cu-income. Accessed
5/27/2021.
Note: Income includes wages, self-employment income, Social Security and retirement payments, interest,
dividends, rental income and other property income, public assistance, unemployment and workers' compensation,
veterans' benefits, and regular contributions for support.
The EPA also examined the household energy expenditure data by race and ethnicity.
The data indicate that Black households' energy expenditures represent a higher share of their
total expenditures and income than for households of other races, yet their energy expenditures
4-63
-------
were lower. Hispanic households' energy expenditures comprise a larger share of their total
expenditures and income than non-Hispanic households, though they spent slightly more per
household on energy than non-Hispanic households.
The CES data summarized in this section highlight the disproportionately high energy
burdens experienced particularly by low-income households, as well as Black and Hispanic
households to some extent. These households must allocate a greater share of their incomes and
expenditures to energy, reducing disposable income that could be used for other essentials (e.g.,
housing, healthcare, and food) and other non-essential preferences. Thus, low income, Black, and
Hispanic households are expected to be most likely to be adversely affected by any potential
increases in energy costs due to this final rule because they face higher energy burdens under the
baseline. Nonetheless, since energy cost impacts are expected to be minimal, this rule is not
expected to significantly alter existing levels of inequality in energy burden.
4.3.7 Summary
EJ concerns for each rulemaking are unique and should be considered on a case-by-case
basis. For this final rulemaking, we quantitatively and qualitatively evaluated the potential for
several EJ concerns, although data availability limitations and the large number of oil and natural
gas locations make it quite possible that disparities may exist that our analysis did not identify.
This is especially relevant for potential EJ characteristics that were not evaluated, such as those
living in close proximity to an affected source.
Some commonalities emerged across the array of EJ analyses. Notably, more Hispanic
people may reside in communities with potentially elevated cancer risk from oil and natural gas-
related toxic emissions (Section 4.3.44.3.4). Similarly, Hispanic populations may be more likely
to reside in communities of higher oil and natural gas intensity (Section 4.3.54.3.5).
Additionally, Hispanic households' energy expenditures may comprise a disproportionate share
of their total expenditures and income as compared to non-Hispanic households (Section
4.3.64.3.6). However, the reductions in ozone concentrations due to the policy option are similar
in magnitude across most demographic groups and small such that it is unlikely that the policy
option will exacerbate or mitigate any disproportional exposures to ozone that were present at
baseline (Section 4.3.3). In some states, those who reside on Tribal land may experience larger
4-64
-------
reductions in ozone concentrations, but the total magnitude of the change would still be
relatively small. Uncertainties associated with the input data, as well as the meaningfulness of
any differences, should be taken into consideration when interpreting these results. Additionally,
we lack key information that would be needed to characterize post-control risks under the final
NSPS OOOOb and EG 0000c or the regulatory alternatives analyzed in the RIA, preventing
the EPA from analyzing spatially differentiated outcomes. Based on the assessment of the
impacts of this final rule on minority populations, low-income populations, and/or Indigenous
peoples, the EPA believes that this action will achieve substantial methane, VOC, and HAP
emissions reductions and will further improve environmental justice community health and
welfare.
4.4 Final Regulatory Flexibility Analysis
The Regulatory Flexibility Act (RFA; 5 U.S.C.§ 601 et seq.), as amended by the Small
Business Regulatory Enforcement Fairness Act (Public Law No. 104-121), provides that
whenever an agency is required to publish a general notice of final rulemaking, it must prepare
and make available an final regulatory flexibility analysis (FRFA), unless it certifies that the
final rule, if promulgated, will not have a significant economic impact on a substantial number of
small entities (5 U.S.C. § 605[b]). Small entities include small businesses, small organizations,
and small governmental jurisdictions. A FRFA describes the economic impact of the final rule
on small entities and the steps taken to minimize the significant impact on small entities
consistent with the stated objectives of applicable statutes (5 U.S.C. § 604[a]). The scope of the
FRFA is limited to the NSPS OOOOb. The impacts of the EG 0000c are not evaluated here
because the EG 0000c does not place explicit requirements on the regulated industry. Those
impacts will be evaluated pursuant to the development of a Federal plan.
4.4.1 Reasons Why Action is Being Considered
The final rulemaking takes a significant step forward in mitigating climate change and
The final rulemaking takes a significant step forward in mitigating climate change and improving
human health by reducing greenhouse gases (GHG) and volatile organic compounds (VOCs)
emissions from the oil and natural gas industry, specifically the Crude Oil and Natural Gas
source category. The oil and natural gas industry is the United States' largest industrial emitter of
4-65
-------
methane. Human emissions of methane, a potent GHG, are responsible for about one third of the
warming due to well-mixed GHGs, the second most important human warming agent after
carbon dioxide. According to the Intergovernmental Panel on Climate Change (IPCC), strong,
rapid, and sustained methane reductions are critical to reducing near-term disruption of the
climate system and a vital complement to carbon dioxide (CO2) reductions critical in limiting the
long-term extent of climate change and its destructive impacts. The oil and natural gas industry
also emits other health-harming pollutants in varying concentrations and amounts, including
CO2, VOC, sulfur dioxide (SO2), nitrogen oxide (NOx), hydrogen sulfide (H2S), carbon disulfide
(CS2), and carbonyl sulfide (COS), as well as benzene, toluene, ethylbenzene and xylenes (this
group is commonly referred to as "BTEX"), and n-hexane.
The EPA finalizing the actions described in the preamble in accordance with its legal
obligations and authorities following a review directed by EO 13990, "Protecting Public Health
and the Environment and Restoring Science to Tackle the Climate Crisis," issued on January 20,
2021. The EPA intends for the final actions to address the far-reaching harmful consequences
and real economic costs of climate change. According to the IPCC, "It is unequivocal that human
influence has warmed the atmosphere, ocean and land. Widespread and rapid changes in the
atmosphere, ocean, cryosphere and biosphere have occurred." These changes have led to
increases in heat waves and wildfire weather, reductions in air quality, more intense hurricanes
and rainfall events, and rising sea level. These changes, along with future projected changes,
endanger the physical survival, health, economic well-being, and quality of life of people living
in America, especially those in the most vulnerable communities.
In the final action, the EPA has taken a comprehensive analysis of the most attainable
data from emission sources in the Crude Oil and Natural Gas source category and the latest
available information on control measures and techniques to identify achievable, cost-effective
measures to significantly reduce emissions, consistent with the requirements of section 111 of
the CAA. The final actions will lead to significant and cost-effective reductions in climate and
health-harming pollution and encourage development and deployment of innovative technologies
to further reduce pollution in the Crude Oil and Natural Gas source category.
4-66
-------
4.4.2 Statement of Objectives and Legal Basis for the Final Rules
The EPA is revising certain NSPS and to promulgate additional NSPS for both methane
and VOC emissions from new oil and natural gas sources in the production, processing,
transmission and storage segments of the industry; and to promulgate EG to require states to
regulate methane emissions from existing sources in those segments. The large amount of
methane emissions from the oil and natural gas industry — by far, the largest methane-emitting
industry in the nation — coupled with the adverse effects of methane on the global climate
compel immediate regulatory action.
The final rule is in line with our 2016 NSPS OOOOa Rule, which likewise regulated
methane and VOCs from all three segments of the industry. The 2016 NSPS OOOOa Rule
explained that these three segments should be regulated as part of the same source category
because they are an interrelated sequence of functions in which pollution is produced from the
same types of sources that can be controlled by the same techniques and technologies. That Rule
further explained that the large amount of methane emissions, coupled with the adverse effects of
GHG air pollution, met the applicable statutory standard for regulating methane emissions from
new sources through NSPS. Furthermore, the Rule explained, this regulation of methane
emissions from new sources triggered the EPA's authority and obligation to regulate the
overwhelming majority of oil and natural gas sources, which the CAA categorizes as "existing"
sources. In the 2020 Policy Rule, the Agency reversed course, concluding based upon new legal
interpretations that it was not authorized to regulate the transmission and storage segment or to
regulate methane. In 2021, Congress adopted a joint resolution to disapprove the EPA's 2020
Policy rule under the CRA. According to the terms of CRA, the 2020 rule is "treated as though
[it] had never taken effect," 5 U.S.C. 801(f), and as a result, the 2016 rule is reinstated.
In disapproving the 2020 Policy Rule under the CRA, Congress explicitly rejected the
2020 Policy Rule interpretations and embraced the EPA's rationales for the 2016 NSPS OOOOa
Rule. The House Committee on Energy & Commerce emphasized in its report (House Report)
that the source category "is the largest industrial emitter of methane in the U.S.," and directed
that "regulation of emissions from new and existing oil and gas sources, including those located
in the production, processing, and transmission and storage segments, is necessary to protect
human health and welfare, including through combatting climate change, and to promote
4-67
-------
environmental justice." House Report at 3-5. A statement from the Senate cosponsors likewise
underscored that "methane is a leading contributing cause of climate change," whose "emissions
come from all segments of the Oil and Gas Industry," and stated that "we encourage EPA to
strengthen the standards we reinstate and aggressively regulate methane and other pollution
emissions from new, modified, and existing sources throughout the production, processing,
transmission and storage segments of the Oil and Gas Industry under section 111 of the CAA."
Senate Statement at S2283. The Senators concluded with a stark statement: "The welfare of our
planet and of our communities depends on it." Id.
The final rule comports with the EPA's CAA section 111 obligation to reduce dangerous
pollution and responds to the urgency expressed by the current Congress. With the final rule, the
EPA is taking additional steps in the regulation of the Crude Oil and Natural Gas source category
to protect human health and the environment. Specifically, the agency is revising NSPS for some
sources, adding NSPS for additional sources, and finalizing Emissions Guidelines that would
impose a requirement on states to regulate methane emissions from existing sources. As the EPA
explained in the 2016 rule, this source category collectively emits massive quantities of the
methane emissions that are among those driving the grave and growing threat of climate change,
particularly in the near term.139 Since that time, the science has repeatedly confirmed that climate
change is already causing dire health, environmental, and economic impacts in communities
across the United States.
Because the 2021 CRA resolution automatically reinstated the 2016 rule, which itself
determined that the Crude Oil and Natural Gas Source Category included the transmission and
storage segment and that regulation of methane emissions was justified, the EPA is authorized to
take the regulatory actions finalized in the rule. In addition, in this action, we are reaffirming
those determinations as clearly authorized under any reasonable interpretation of section 111.
Further information can be found in Section VIII of the preamble.
139 81 FR 3584
4-68
-------
4.4.3 Significant Issues Raised
The only significant issues raised in public comments specifically in response to the
initial regulatory flexibility analysis came from the Office of Advocacy within the Small
Business Administration (SBA). A summary of those comments and our response is provided in
the next section.
4.4.4 Small Business A dministration Comments
The SBA's Office of Advocacy (hereafter "Advocacy") provided substantive comments
on the IRFA published in the RIA accompanying the November 2021 Proposal. Those comments
made the following claims: (1) EPA failed to adequately account for additional burdens of the
proposed Appendix K; (2) EPA did not provide burden estimates for the proposed NSPS
OOOOb; (3) the IRFA lacked a sufficient discussion of regulatory alternatives that would
minimize the impacts on small businesses, and instead merely repeated the SBREFA panel report
recommendations as the description of alternatives in the IRFA; and (4) the IRFA did not reflect
significant changes to the proposed rule that occurred during and/or after the conclusion of the
SBREFA panel. Based on those claims, the Office of Advocacy insisted that EPA issue a revised
IRFA that included alternatives reflective of the November 2021 Proposal and December 2022
Supplemental Proposal.
In response to the Advocacy's comments, EPA agreed that issuing a revised IRFA with
the December 2022 Supplemental Proposal was warranted, and the revision was published as
Section 4.3 in the December 2022 Supplemental Proposal RIA. The revised IRFA addressed
Advocacy's critiques of the IRFA contained in the November 2021 Proposal RIA by providing a
robust discussion of regulatory alternatives related to provisions for the following elements:
fugitive emissions requirements, alternative technologies, associated gas requirements,
pneumatic device requirements, and reciprocating compressor requirements. For the final
regulatory flexibility analysis, EPA is also including discussion of regulatory alternatives for
centrifugal compressor and liquids unloading requirements. Taken together, this discussion
addresses Advocacy's concerns about the insufficiency of the discussion of regulatory
alternatives in the November 2021 Proposal IRFA. In addition, the revised IRFA noted that the
December 2022 Supplemental Proposal did not require OGI in accordance with the proposed
4-69
-------
Appendix K for production sites. While equipment leaks at gas plants were still proposed to be
monitored using OGI in accordance with Appendix K in the December 2022 Supplemental
Proposal, the burden estimates summarized in the revised IRFA reflected burden associated with
Appendix K. Finally, the burden estimates were updated to reflect the proposed NSPS OOOOb.
Following the issuance of the December 2022 Supplemental Proposal, Advocacy
provided additional comments. While noting that it continued to have significant concerns about
the impact the rule would have on small businesses in the oil and gas production sector,
Advocacy acknowledged the work that EPA did to improve its RFA compliance through the
IRFA between proposals. More detailed responses to Advocacy's comments can be found in
Chapter 21 of Volume I of the Response to Comments document.
4.4.5 Description and Estimate of Affected Small Entities
The Regulatory Flexibility Act (RFA) defines small entities as including "small
businesses," "small governments," and "small organizations" (5 USC 601). The regulatory
revisions being considered by EPA for this rulemaking are expected to affect a variety of small
businesses but would not affect any small governments or small organizations. The RFA
references the definition of "small business" found in the Small Business Act, which authorizes
the Small Business Administration (SBA) to further define "small business" by regulation. The
detailed listing of SBA definitions of small business for oil and natural gas industries or sectors,
by NAICS code, that are potentially affected by this rule is included in Table 4-18. The EPA
conducted this initial regulatory flexibility analysis at the ultimate (i.e., highest) level of
ownership, evaluating ultimate parent entities.
4-70
-------
Table 4-18 SBA Size Standards by NAICS Code
NAICS
Codes
NAICS Industry Description
Size Standards Size Standards
(in millions of dollars) (in no. of employees)
211120
211130
213111
213112
486210
Crude Petroleum Extraction
Natural Gas Extraction
Drilling Oil and Gas Wells
Support Activities for Oil and Gas Operations
Pipeline Transportation of Natural Gas
$41.5
$36.5
1,250
1,250
1,000
Sources: U.S. Small Business Administration, Table of Standards, Effective July 14, 2022.
https://www.sba.gov/document/support--table-size-standards. Accessed July 27, 2022.
To estimate the number of small businesses potentially impacted by the rule, EPA
developed a list of operators of oil and natural gas wells, natural gas processing plants, and
natural gas compressor stations. The list of well operators is based on data from Enverus and
consists of all operators that completed wells producing oil or natural gas in 2019, which serves
as an approximation of the universe of operators that might be affected by the NSPS. The list of
processing plant operators is from the Department of Homeland Security (DHS) Homeland
Infrastructure Foundation-Level Data.140 The compressor stations operator data is from the
Enverus Midstream database. The DHS data and Enverus Midstream data did not contain
information on when facilities were constructed, and therefore could not be restricted to only
those facilities completed in 2019. The initial list of operators included 1,451 well site operators
that completed a well in 2019, 297 processing plant operators, and 574 compressor station
operators.
The list of operators was combined with data from the D&B Hoovers and Zoomlnfo
business databases in a two-step process. D&B Hoovers and Zoomlnfo are proprietary,
subscription-based databases of business information (such as revenue, employment, and
ownership structure) gleaned from sources such as financial statements, news reports, and
industry trade group publications. Using an approximate string-matching algorithm, the list of
operators was first merged with business information from D&B Hoovers. The remaining
unmatched operators were matched to the Zoomlnfo business database when possible. This
matching process added information on the ultimate parent companies, NAICS codes, number of
employees, and annual revenues of the operators. The matches from D&B Hoovers and
140
Department of Homeland Security. (2020). Homeland Infrastructure Foundation-Level Data. Found at:
https://hifld-geoplatform.opendata.arcgis.com/datasets/geoplatform: :natural-gas-processing-plants/about.
4-71
-------
Zoomlnfo were examined and, when necessary, manual adjustments were made to the matched
list of ultimate parent companies to standardize company names, revenue, and employment
information across the two matched lists. Each matched ultimate parent company, or firm, was
classified "small business" or "not small business" based on the SBA size classification
threshold associated with the relevant NAICS code. The results of this small business coding
exercise are displayed by NAICS code in Table 4-19. In total, 998 of the 1,451 well site
operators (69 percent) matched to 914 ultimate parent companies; 270 of 297 processing plant
operators (91 percent) matched to 146 ultimate parent companies; and 519 of 574 compressor
station operators (90 percent) matched to 315 ultimate parent companies.
4-72
-------
Table 4-19 Counts and Estimated Percentages of Small Entities
Estimated
Estimated
Percentage of
Small Entities
NAICS
Number of
Number of
for Identified
Codes
NAICS Industry Description
Firms Identified
Small Entities
Firms
211120
Crude Petroleum Extraction
333
283
85%
211130
Natural Gas Extraction
22
20
91%
213111
Drilling Oil and Gas Wells
39
37
95%
213112
Support Activities for Oil and Gas
Operations
305
265
87%
486210
Pipeline Transportation of Natural Gas
47
14
30%
Many3
Other
427
267
63%
a Not all owner/operators in the Enverus well database produced a match in the D&B Hoovers database under an oil
and natural gas industry-related NAICS as presented in Table 4-18.
4.4.6 Compliance Cost Impact Estimates
To estimate the compliance cost impacts of the final rule on small entities, we use the
dataset of operators matched to ultimate parent companies discussed in the previous section and
apply the sum of incremental costs for all relevant affected facility categories. Because the
incremental costs depend on unknown characteristics of operator-specific well sites, processing
plants, and compressor stations, we use average equipment counts at each facility type to derive
estimates of average impacts at each facility type. Ultimately, we estimate cost-to-sales ratios
(CSR) for each small entity to summarize the impacts of the final NSPS.
4.4.6.1 Methodology for Estimating Impacts on Small Entities
The two main pieces of information we use to assess impacts on small entities are
ultimate parent company revenues and expected compliance costs. For most ultimate parent
companies in the dataset described in the previous section, revenue is generated from the match
with either the D&B Hoovers or Zoomlnfo database. For owners of well site operators, we also
estimated revenues from calculating total operator-level production in 2019 from Enverus,
multiplying by assumed oil and natural gas prices at the wellhead and summing over all
operators owned by a parent company. For natural gas prices, we assumed the projected price
from AEO2022 in 2022 (adjusted to approximate a wellhead price, as described in Section 2.4)
of $3.40/Mcf. For oil prices, we used the projected AEO2022 price for Brent Crude in 2022,
$66.40/barrel. Both prices are measured in 2019 dollars. For owners of well site operators,
4-73
-------
revenue was calculated as the minimum of the matched revenue from D&B Hoovers/Zoomlnfo
and the estimated revenue based on production. Operators of compressor stations were divided
into two groups: those that own gathering and boosting stations, and those that own transmission
and storage stations. While there is overlap between the two segments, they are treated as distinct
groups in this analysis and results are presented by segment. Summary statistics for firm revenue
by segment are presented in Table 4-20.
Table 4-20 Summary Statistics for Revenues of Potentially Affected Entities
Mean Revenue
Median Revenue
Segment
Size
No. of Firms
(million 2019$)
(million 2019$)
Production
Small
634
$180
$11
Not Small
67
$21,000
$1,800
Processing
Small
78
$200
$13
Not Small
60
$28,000
$6,200
Small
123
$510
$24
Gathering and Boosting
Not Small
77
$19,000
$3,200
Small
50
$260
$22
Transmission and Storage
Not Small
82
$22,000
$3,200
To calculate expected compliance costs for ultimate parent companies, we first
constructed an estimate of the number of sites for each firm in each segment. For well site
operators, the number of well sites is calculated by summing the number of well pads containing
a well that the operator completed in 2019.141 The number of well sites owned by an ultimate
parent company is calculated by summing over the well site counts of the operators it owns. For
processing plants, gathering and boosting compressor stations, and transmission and storage
compressor stations, the number of sites for each operator is obtained by summing the number of
entries of each type in the DHS data for processing plant operators and in the Enverus Midstream
data for compressor station operators. Again, the number of facilities of each type owned by an
ultimate parent company is calculated by summing over the facility counts of the processing
plant or compressor station operators it owns. Those sums are then multiplied by the ratio of the
annual number of new sites we project in each year (see Table 2-3) to the total number of sites
we project in 2024 to approximate the number of NSPS OOOOb sites each company might
141
For a small number of operators, well pad-level identifiers were not available. Those operators were dropped
from the analysis.
4-74
-------
construct each year. Companies estimated to have fewer than one new site after applying this
procedure are assumed to have exactly one new site for the purposes of the analysis.
Once site counts were assigned, we estimated compliance costs for each ultimate parent
company by assigning annualized costs (both with and without expected revenue from natural
gas recovery) from all relevant affected facilities: fugitive emissions, pneumatic devices, storage
vessels, liquids unloading, and associated gas for well sites; equipment leaks, reciprocating
compressors, and wet seal centrifugal compressors for processing plants; and reciprocating
compressors, wet seal centrifugal compressors, and pneumatic devices for gathering and
boosting and transmission and storage compressor stations. Since the precise equipment counts
at the facility level were necessary to estimate compliance costs relative to baseline, and this
information was not present in the operator data, average costs per facility were used to estimate
site-level compliance costs for this analysis. Median per-site capital and annual operating costs
for each affected facility are presented in Table 4-21. Median compliance costs by segment and
firm size are presented in Table 4-22, both with and without expected revenue from natural gas
recovery included.
Table 4-21 Average Capital and Annual Operating Costs by Affected Facility (2019
Dollars)
Annual
Segment
Source
Unit
Capital Cost
Operating Cost
Fugitives
Site
-$3
$1,185
Pneumatics
Site
$16,385
-$251
Production
Storage Vessels
Battery
$79,352
$22,840
Associated Gas
Site
$545,611
$1,659
Liquids Unloading
Event
$0
$65
Gathering and
Boosting
Pneumatics
Reciprocating Compressors
Station
Compressor
$50,024
$0
$1,977
$597
Wet-Seal Centrifugal Compressors
Compressor
$0
$25,000
Natural Gas
Leaks
Plant
-$15,255
-$28,913
Processing
Reciprocating Compressors
Compressor
$0
$597
Transmission
Pneumatics
Reciprocating Compressors
Station
Compressor
$61,329
$0
$360
$597
Storage
Pneumatics
Station
$96,269
-$2,210
Reciprocating Compressors
Compressor
$0
$597
4-75
-------
Table 4-22 Distribution of Estimated Compliance Costs across Segment and Firm Size
Classes (2019$)
Segment
Size
No. of Firms
Average Cost without
Product Recovery
Average Cost with
Product Recovery
Production
Small
Not Small
634
67
$26,000
$48,000
$17,000
$4,500
Processing
Small
Not Small
78
60
-$28,000
-$29,000
-$34,000
-$35,000
Gathering and Boosting
Small
Not Small
123
77
$12,000
$24,000
$6,800
$14,000
Transmission and Storage
Small
Not Small
50
82
$11,000
$11,000
$8,500
$8,500
Note: Totals may not appear to add correctly due to rounding.
4.4.6.2 Results
This section presents results of the cost-to-sales ratio analysis for the production,
processing, gathering and boosting, and transmission and storage segments. The cost-to-sales
ratios presented approximate the impact of the NSPS requirements on ultimate parent companies
of well site, processing plant, and compressor station operators. In the processing segment,
average annualized costs relative to baseline are expected to be negative, and no entity has a CSR
greater than either 1 percent or 3 percent.142 In the production segment, when expected revenues
from natural gas product recovery are included, 200 small entities (31 percent) have cost-to-sales
ratios greater than 1 percent, and of those, 59 have cost-to-sales ratios greater than 3 percent (9
percent). When expected revenues from natural gas product recovery are excluded, the number
of small entities with cost-to-sales ratios greater than 1 percent increases to 210 (33 percent); 62
of those small entities (10 percent) also have cost-to-sales ratios greater than 3 percent. In the
gathering and boosting segment, no parent companies have cost-to-sales ratios greater than 3
percent regardless of whether expected revenues from natural gas recovery are included. Seven
parent companies (6 percent) have a cost-to-sales ratio greater than 1 percent when expected
revenues from natural gas recovery are excluded (only one does when they are included). In the
transmission and storage segment, only one entity has a CSR greater than 1 percent, and only if
142
The net compliance costs for leak detection at natural gas processing plants decrease primarily because OGI
surveys under this proposal can be conducted much more quickly and at approximately half the cost of EPA Method
21 surveys under the current requirements in NSPS Wa, so the increased flexibility under the proposal is likely cost
saving for affected facilities.
4-76
-------
expected revenues from product recovery are included. The results for all segments are
summarized in Table 4-23.
Table 4-23 Compliance Cost-to-Sales Ratios for Small Entities
Without Product Recovery
Included
With Product Recovery
Included
Segment
CSR Ratio
Category
No. of Small °/
Entities
o of Small
Entities
No. of Small °/
Entities
o of Small
Entities
All
634
634
Production
Greater than 1%
148
23%
141
22%
Greater than 3%
62
10%
59
9%
All
78
78
Processing
Greater than 1%
0
0%
0
0%
Greater than 3%
0
0%
0
0%
Gathering and
Boosting
All
Greater than 1%
123
7
6%
123
1
1%
Greater than 3%
0
0%
0
0%
Transmission and
Storage
All
Greater than 1%
50
1
2%
50
0
0%
Greater than 3%
0
0%
0
0%
4.4.7 Caveats and Limitations
The analysis above is subject to several caveats and limitations, many of which we
discussed in the presentation of methods and results. It is useful, however, to present a compiled
list of the caveats and limitation here.
• Not all owner/operators could be identified in either the D&B Hoovers or Zoomlnfo
database. In addition, the matching procedure used to link the operator database to the
D&B Hoovers and Zoomlnfo database is imperfect, so there may be misspecified
matches or duplicate entries for the same entity.
• The analysis assumes the same population of entities completing wells in 2019 are also
completing wells at the same rate in 2024 and beyond and assumes facility counts are
stable over time. These firms may operate more or fewer facilities in the future depending
on economic and technological factors that are largely unpredictable. The analysis also
assumes the population of entities operating processing plants and compressor stations in
2019 is the same population that will construct new processing plant and compressor
stations in 2024 and beyond.
4-77
-------
• The approach used to estimate sales for the cost-to-sales ratios might over-estimate or
under-estimate sales depending upon the accuracy of the information in the underlying
databases and the market prices ultimately faced when the final requirements are in
effect.
• It is unknown what equipment is present at each site. The use of cost averages to estimate
costs may under- or over-estimate costs at the site level for any given entity, which adds
uncertainty to the calculated cost-to-sales ratios.
4.4.8 Projected Reporting, Recordkeeping and Other Compliance Requirements
The information to be collected for the final NSPS is based on notification, performance
tests, recordkeeping and reporting requirements which will be mandatory for all operators
subject to the final standards. Recordkeeping and reporting requirements are specifically
authorized by section 114 of the CAA (42 U.S.C. 7414). The information will be used by the
delegated authority (state agency, or Regional Administrator if there is no delegated state
agency) to ensure that the standards and other requirements are being achieved. Based on review
of the recorded information at the site and the reported information, the delegated permitting
authority can identify facilities that may not be in compliance and decide which facilities,
records, or processes may need inspection. All information submitted to the EPA pursuant to the
recordkeeping and reporting requirements for which a claim of confidentiality is made is
safeguarded according to Agency policies set forth in 40 CFR part 2, subpart B.
Potential respondents under subpart OOOOb are owners or operators of new, modified,
or reconstructed oil and natural gas affected facilities as defined under the rule. Few, if any, of
the facilities in the United States are owned or operated by state, local, tribal or the Federal
government. The regulated facilities are privately owned for-profit businesses. The requirements
in this action result in industry recording keeping and reporting burden associated with review of
the requirements for all affected entities, gathering relevant information, performing initial
performance tests and repeat performance tests if necessary, writing and submitting the
notifications and reports, developing systems for the purpose of processing and maintaining
information, and train personnel to be able to respond to the collection of information.
4-78
-------
The estimated average annual burden (averaged over the first 3 years after the effective
date of the standards) for the recordkeeping and reporting requirements in subpart OOOOb for
the estimated 1,849 owners and operators that are subject to the rule is approximately 883,625
labor hours, with an annual average cost of about $58 million. The annual public reporting and
recordkeeping burden for this collection of information is estimated to average about 60 hours
per respondent. Respondents must monitor all specified criteria at each affected facility and
maintain these records for 5 years. Burden is defined at 5 CFR 1320.3(b).
4.4.9 Related Federal Rules
There are two National Emission Standards for Hazardous Air Pollutants (NESHAP)
rules that apply to certain equipment and processes in the oil and natural gas sector. These rules,
listed below, address air toxics, primarily benzene, toluene, ethylbenzene, and xylenes
(collectively referred to as BTEX) and n-hexane. These two rules were promulgated under
section 112 of the Clean Air Act and are codified in 40 CFR Part 63 Subpart HH and Subpart
HHH.
Aside from the EPA, several other Federal agencies have jurisdiction over the oil and
natural gas sector.
• The Bureau of Land Management (BLM) within the Department of the Interior
regulates the extraction of oil and natural gas from federal lands. BLM manages the
Federal government's onshore subsurface mineral estate, about 700 million acres.
BLM also oversees oil and natural gas operations on many Tribal leases and
maintains an oil and natural gas leasing program. BLM does not directly regulate
emissions for the purposes of air quality but does regulate venting and flaring of
natural gas for the purposes of preventing waste. An operator may also be required to
control/mitigate emissions as a condition of approval on a drilling permit.
• The Bureau of Ocean Energy Management (BOEM) within the Department of the
Interior manages the development of America's offshore energy and mineral
resources. BOEM has certain air quality regulatory authority over activities that
BOEM authorizes on the Outer Continental Shelf of the United States in the Gulf of
4-79
-------
Mexico, west of 87.5 degrees longitude, and adjacent to the North Slope Bureau of
the State of Alaska.
• The Pipeline and Hazardous Materials Safety Administration (PHMSA) within the
Department of Transportation ensures safety in the design, construction, operation,
maintenance, and spill response planning of America's 2.8 million miles of natural
gas and hazardous liquid transportation pipelines. This includes data and risk
analysis, outreach, research and development, regulations and standards, training,
inspections and enforcement and accident investigations. Section 113 of the
Protecting our Infrastructure of Pipelines and Enhancing Safety Act of 2020 (PIPES
Act of 2020) mandates that PHMSA promulgate a final rule concerning gas pipeline
leak detection and repair programs no later than one year after the enactment of the
law.
• The Federal Energy Regulatory Commission (FERC) within the Department of
Energy (DOE) regulates natural gas pipeline, storage, and liquefied natural gas
facility construction. FERC also issues environmental assessments or draft and final
environmental impact statement for comment on most projects.
• The Internal Revenue Service (IRS), in the Internal Revenue Code (IRC), defines a
stripper well property as "a property where the average daily production of domestic
crude oil and natural gas produced from the wells on the property during a calendar
year divided by the number of such wells is 15 barrel equivalents or less." See IRC
613A(c)(6)(E).
4.4.10 Regulatory Flexibility Alternatives
Prior to the November 2021 Proposal, the EPA convened a Small Business Advocacy
Review (SBAR) Panel to obtain recommendations from small entity representatives (SERs) on
elements of the regulation. The Panel identified significant alternatives for consideration by the
Administrator of the EPA, which were summarized in a final report.143 Based on the Panel
recommendations, as well as comments received in response to the November 2021 Proposal and
143 See document ID EPA-HQ-OAR-2021-0317-0074.
4-80
-------
December 2022 Supplemental Proposal, the EPA is finalizing several regulatory alternatives that
could accomplish the stated objectives of the Clean Air Act while minimizing any significant
economic impact of the final rule on small entities. Discussion of those alternatives is provided
below.
4.4.10.1 Fugitive Emissions Requirements
As described in the preamble to the to the final rule,144 the EPA is finalizing certain
changes, proposed in the December 2022 Supplemental Proposal,145 to the fugitive emissions
standards proposed in November 2021 for NSPS OOOOb. The EPA believes that two of these
finalized changes will reduce impacts on small businesses: (1) requiring OGI monitoring for well
sites and centralized production facilities following the monitoring plan required in 40 CFR
60.5397b instead of requiring the procedures proposed in Appendix K for these sites and (2)
defining monitoring technique and frequency based on the equipment present at a well site. The
EPA describes these two changes below.
In the final rule, the EPA is not requiring OGI monitoring in accordance with the
proposed Appendix K for well sites or centralized production facilities, as was proposed in the
November 2021 proposal. Instead, the EPA is requiring OGI surveys following the procedures
specified in the regulatory text for NSPS OOOOb (at 40 CFR 60.5397b) or according to EPA
Method 21. This change is consistent with the requirements for OGI surveys found in NSPS
OOOOa at 40 CFR 60.5397a. This final change is a result of the extensive comments the EPA
received from oil and natural gas operators and other groups on the numerous complexities
associated with following the proposed Appendix K, especially considering the remoteness and
size of many of these well sites.146 In addition, commenters pointed out that OGI has always been
the BSER for fugitive monitoring at well sites and was never designed as a replacement for EPA
Method 21, while Appendix K was designed for use at more complex processing facilities that
have historically been subject to monitoring following EPA Method 21. The EPA agrees with the
commenters and is final requirements within NSPS OOOOb at 40 CFR 60.5397b in lieu of the
144
See final rule preamble Section XI. A.
145
See December 2022 Supplemental Proposal preamble Section IV.A.
146
See final rule preamble Section XI.A. and see Document ID Nos. EPA-HQ-OAR-2021-0317-0579, -0743, -0764,
-0777, -0782, -0786, -0793, -0802, -0807, -0808, -0810, -0814, -0817, -0820, -0831, -0834, and -0938.
4-81
-------
procedures in Appendix K for fugitive emissions monitoring at well sites or centralized
production facilities. See section X.I. V of the preamble for additional information on what the
EPA is finalizing for Appendix K related to other sources (e.g., natural gas processing plants).
The EPA believes this will particularly benefit small entities because it will streamline the
requirements for conducting and documenting OGI surveys at these smaller, less complex sites.
Additionally, this change provides a uniform set of requirements for regulated entities that may
have assets subject to different subparts within the same region, which leads to increased
regulatory certainty and eases the compliance burden. At the same time, the EPA believes this
does not compromise the stated objectives of the Clean Air Act because these same requirements
are already allowed in NSPS OOOOa and outline many of the same data elements required by
Appendix K.
Next, the final rule includes fugitive monitoring frequencies and detection techniques that
are based on the type of equipment located at a well site, instead of the baseline methane
emissions threshold that was included in the November 2021 Proposal. Specifically, the EPA is
finalizing four distinct subcategories of well sites:
• Well sites with only a single wellhead,
• Small well sites with a single wellhead and only one piece of major production and
processing equipment,147
• Well sites with only two or more wellheads and no other major production and
processing equipment, and
• Well sites or centralized production facilities with one or more controlled storage
vessels, control devices, natural gas-driven pneumatic controllers or pumps, or two or
more other major production and processing equipment.
147
Small well sites are defined as single wellhead well sites that do not contain any controlled storage vessels,
control devices, pneumatic controller affected facilities, or pneumatic pump affected facilities, and include only one
other piece of major production and processing equipment. Major production and processing equipment that would
be allowed at a small well site would include a single separator, glycol dehydrator, centrifugal and reciprocating
compressor, heater/treater, and storage vessel that is not controlled. By this definition, a small well site could only
potentially contain a well affected facility (for well completion operations or gas well liquids unloading operations
that do not utilize a CVS to route emissions to a control device) and a fugitive emissions components affected
facility. No other affected facilities, including those utilizing CVS (such as pneumatic pumps routing to control) can
be present for a well site to meet the definition of a small well site.
4-82
-------
The EPA is finalizing these distinct subcategories of well sites after consideration of
comments on the November 2021 Proposal that stated the original baseline methane emissions
threshold approach would be difficult to implement, especially for small businesses that may be
less familiar with the use of emissions factors from the EPA's Greenhouse Gas Reporting
Program. The EPA believes that owners and operators, including small entities, can readily
identify the number and types of major equipment located at a well site without the need for
complicated calculations of emissions.
Further, the EPA is finalizing specific monitoring frequencies and techniques as the
BSER for each well site subcategory individually. For example, the EPA is finalizing the use of
audible, visual, and olfactory (AVO) inspections at well sites containing only a single wellhead
and at small well sites. This monitoring technique does not require specialized equipment or
operator training, but does allow the identification of large leaks, which are of the most concern
from an environmental standpoint. Further, AVO monitoring can easily be built into regular
maintenance activities that are designed to keep the equipment at the site in good working order.
The final requirements are responsive to a SER recommendation that the EPA allow AVO and
soap bubble tests as an option for finding fugitive emissions, particularly because they are low
cost and easy to implement alternatives for detecting leaks, and an Advocacy recommendation
that the EPA allow AVO as an alternative in limited circumstances, such as part of an off-ramp
for facilities unlikely to emit more than insignificant methane or with a demonstrated history of
insignificant emissions. The EPA believes this will particularly benefit small entities because
AVO surveys at these types of well sites are effective at identifying the types of large emissions
from sources located at these well sites at a much lower cost than OGI surveys. For example, the
costs associated with the quarterly AVO inspections are estimated at $660/year, whereas the
costs associated with an annual OGI survey for this type of well site are estimated at
approximately $2,000/yr. Inspections via AVO allow for more frequent inspections for large
emissions events at these well sites, which results in faster emissions mitigation, than a single
OGI survey each year.
4-83
-------
4.4.10.2 Alternative Technology
As described in the preamble to the December 2022 Supplemental Proposal,148 the EPA is
finalizing changes to the November 2021 alternative technology requirements for NSPS
OOOOb. The changes are the result of overwhelming support that the EPA received for the
inclusion of an option to use advanced technologies for periodic screenings as an alternative to
the fugitive emissions monitoring and repair program in NSPS OOOOb. The EPA believes these
changes will reduce impacts on small businesses.
Specifically, the EPA is finalizing the use of alternative screening technologies as a
compliance option rather than an additional regulatory requirement. Through the SBAR Panel
outreach, SERs supported the use of aerial, satellite, and other forms of monitoring for fugitive
emissions requirements beyond traditional fugitive emissions monitoring and repair, but only as
an alternative and not as an additional requirement. In addition, the Panel recommended that the
EPA consider the cost and scope of alternative technologies and alternative screening
technology, and that the EPA try to minimize significant additional reporting and recordkeeping
requirements. In accordance with these recommendations, the EPA is finalizing changes that are
intended to support and incentivize the deployment and utilization of a broader spectrum of
advanced measurement technologies and, ultimately, enable more cost-effective reductions in
emissions. These changes include a "matrix" which would specify several different screening
frequencies corresponding to a range of minimum detection levels, in contrast to the single
screening frequency and detection level permitted under the November 2021 Proposal. In
addition, the EPA will allow owners and operators the option of using continuous monitoring
technologies as an alternative to periodic screening in conjunction with long- and short-term
emissions rate thresholds that would trigger investigation as well as monitoring plan
requirements for owners and operators that choose this approach. Entities interested in using an
alternative methane detection technology to comply with the rule must submit a request for
alternative test method approval to EPA. The EPA believes this approach will particularly
benefit small entities because they will be allowed flexibility to determine which screening
technology works best for their needs without the need to undertake the application of an
148
See preamble section XI.B.
4-84
-------
alternative means of emissions limitation (AMEL), which would be especially burdensome for
small entities with less ability to perform extensive field testing of technologies or conduct
sophisticated modeling simulations. Furthermore, this approach incorporates the use of
alternative test methods, which allows for broad application of technologies after approval,
without the need for individual applications from owners and operators for their specific sites.
4.4.10.3 Associated Gas
As described in the final rule,149 the EPA is finalizing certain changes to the requirements
for oil wells with associated gas that were proposed in November 2021 for NSPS OOOOb and
revised in the supplemental proposal of December 2022. These changes include adjustments to
the hierarchy of the standard and compliance options. The EPA believes these final changes will
increase the burden on some small businesses and reduce impacts on other small businesses.
Specifically, the EPA is requiring flaring of all associated gas from existing, modified,
and reconstructed wells where a determination has been made that it is not feasible to route the
associated gas to a sales line or use it for another beneficial purpose due to technical or safety
reasons. This demonstration would need to not only address the lack of availability or access to a
sales line but would also need to demonstrate why all potential beneficial uses are not feasible
due to technical or safety reasons. This demonstration, which would require certification by a
professional engineer or other qualified individual, would be submitted in the first annual report
for the well affected facility. The EPA is has described what this demonstration should entail and
what qualifications constitute an "other qualified individual" in the preamble for this final action.
Second, the EPA has subcategorized wells by the level of production above and below 10
tons per year of methane, and reduced the administrative burden for those sources below 10 tons
per year, by allowing that associated gas be directed to a flare without a demonstration that
routing the gas to a sales line, using the gas for another useful purpose, reinjecting the gas into
the subsurface, or using the gas for an onsite fuel source are each and all infeasible. The EPA
believes this approach will benefit small entities because it allows for the flaring of associated
gas when the amount of gas to be handled is unlikely to be used or disposed cost effectively in
149
See preamble section XI.F.
4-85
-------
another way. Installation of a sales pipeline or other infrastructure necessary to use associated
gas in a beneficial way is very costly, especially where well sites are located at great distances
from other necessary infrastructure, such as natural gas processing plants. These costs can
disproportionally affect small businesses who may not produce a large enough quantity of
associated gas to offset the capital necessary to install such infrastructure. The allowance of
flaring in these situations without a technical demonstration of infeasibility for another use or
disposal of the gas provides for a way to reduce emissions of methane to the atmosphere (in
contrast to direct venting of associated gas), but at a lower cost than the cost for new
infrastructure.
4.4.10.4 Pneumatic Controller and Pneumatic Pump Requirements
As described in the preamble to the final rule,150 the EPA is finalizing certain
clarifications and changes to the pneumatic controller and pneumatic pumps emissions
requirements included in the November 2021 Proposal and revised in the December 2022
Supplemental Proposal
Through the SBAR Panel outreach, SERs stated that zero emission controllers are not
feasible at wells sites or other locations without reliable electricity, installing gas-fired
compressors to provide sufficient air for instrument air systems may defeat the purpose by
ultimately increasing emissions, and the installation of electric service would be extremely
expensive. The EPA and Advocacy recommended that the EPA only propose zero emission
standards for process controllers at sites with reliable and consistent onsite power available and
clearly state that the intent is not to require the installation of electric services for this purpose.151
For process controllers, the EPA maintains that there is a technically feasible option
available for zero-emitting controllers for all production, processing, and transmission and
storage sites, except for sites in Alaska without access to electricity. The EPA further identifies
compliance options for process controllers other than using electricity. Therefore, the final NSPS
OOOOb does not include any alternative non-zero emission standards for process controllers,
except at sites in Alaska without access to electrical power. At those Alaskan sites,
150 See preamble Sections XI.D and XI.E.
4-86
-------
owners/operators may use low-emitting process controllers or route emissions to a control device
achieving a 95 percent reduction in emissions.
For pumps, the final rule recognizes that at sites without access to electricity, there could
be situations where it is technically infeasible to use zero-emitting pumps. As a result, the EPA
is finalizing a tiered structure in the rule that provides flexibility based on site-specific
conditions. At sites without access to electricity and that have fewer than three diaphragm
pumps, owners/operators may route emissions to a process through an onsite vapor recovery unit
(VRU), or if a VRU is not onsite, they may route emissions to a control device already onsite. If
there is no control device onsite, control of emissions from the pumps affected facility is not
required.
The final requirements are responsive to SER's statements and concerns about technical
feasibility and have considered the potential impacts and feasibility challenges for small
businesses.
4.4.10.5 Reciprocating Compressors
In the November 2021 Proposal, the EPA proposed that an owner or operator of a
reciprocating compressor affected facility would be required to monitor the rod packing
emissions annually by conducting flow rate measurements. When the measured flow rate
exceeded 2 scfm (in pressurized mode), replacement of the rod packing would have been
required. Alternatively, the November 2021 Proposal would have also provided owners and
operators the option of routing rod packing emissions to a process via a closed vent system under
negative pressure in order to comply with the rule. The proposed option to route to a process
would have been allowed as an alternative under NSPS OOOOb because implementing this
option, where feasible, would achieve greater emission reductions than the proposed
performance-based emissions threshold standard. The December 2022 Supplemental Proposal
proposed changes and specific clarifications to the November 2021 Proposal standards for NSPS
OOOOb. For the proposed replacement of the rod packing based on an emission limit and annual
measurement requirement, we proposed: (1) To clarify that the standard of performance is a
numeric standard (not a work practice standard) of 2 scfm, (2) to allow for repair (in addition to
replacement) of the rod packing in order to maintain an emission rate at or below 2 scfm, and (3)
4-87
-------
to allow for monitoring based on 8,760 hours of operation instead of based on a calendar year.
The EPA also proposed regulatory text that defined the required flow rate measurement methods
and/or procedure requirements, and recordkeeping and reporting requirements. For the
alternative option of routing rod packing emissions to a process via a closed vent system under
negative pressure, the EPA proposed to remove the negative pressure requirement.
As described in the preamble to the final rule,152 the EPA is finalizing changes to the
proposed requirements for reciprocating compressors in for NSPS OOOOb as a result of
comments received on the November 2021 Proposal and December 2022 Supplemental Proposal.
The EPA believes the following rule changes will reduce impacts on small businesses.
Concerns were expressed regarding the EPA's November 2021 Proposal and December
2022 Supplemental Proposal that shifted rod packing changeout requirements from a designated
schedule of once every 3 years to a performance standard based on an annual flow rate
measurement. While the November 2021 Proposal format of the performance standard based on
volumetric flow rate measurements was as a work practice standard, the December 2022
Supplemental Proposal format of the performance standard was as a numeric standard.
Commenters on the December 2022 Supplemental Proposal expressed that, as a numeric
standard, the performance standard based on flow measurements was unworkable. It was also
noted that a performance standard is often more expensive than a fixed equipment change-out
standard because of the additional monitoring and recordkeeping necessary to demonstrate
compliance with the performance standard, which they believed could negatively impact small
businesses. These commenters also supported the fixed schedule rod packing change-out
standard because this is the standard owners and operators have implemented for reciprocating
compressors under NSPS OOOOa and stated that the annual flow rate performance work
practice standard would lead to more rod packing changeouts than would be required based on
the November OOOOa fixed-schedule packing change out requirements.
The EPA is finalizing the following requirement changes associated with the
reciprocating compressor rod packing volumetric flow rate measurement performance standard
based on November 2021 Proposal and December 2022 Supplemental Proposal comments: (1) a
152
See final rule preamble Section XI.I.
4-88
-------
2 scfm volumetric flow rate per cylinder performance work practice standard, (2) repair (in
addition to replacement) of the rod packing is allowed to maintain an emission rate at or below 2
scfm per cylinder; and (3) monitoring based on 8,760 hours of operation instead of based on a
calendar year. These final requirements for reciprocating compressors are responsive to
comments and concerns expressed by industry (including small businesses).
The EPA believes the final rule 2 scfm volumetric flow rate per cylinder performance
work practice standard approach benefits small entities because facilities can use monitoring data
to determine emission levels at which it is necessary to repair or replace rod packing. This
approach can result in operational benefits, including a longer life for existing equipment,
improvements in operating efficiencies, and long-term cost savings. Allowing an owner or
operator to repair the rod packing (in addition to replacement of the rod packing) to maintain an
emission rate at or below 2 scfm per cylinder alleviates the need to replace the rod packing when
only a simple repair may be needed to maintain volumetric flow rate at or below 2 scfm per
cylinder. Requiring owners and operators to conduct volumetric flow rate monitoring based on
8,760 hours of operation instead of based on a calendar year reduces the burden on owners and
operators where compressors are not operational for multiple months or are used intermittently.
Additionally, by requiring that monitoring frequency based on hours of operation, owners and
operators have the flexibility to stagger maintenance activity throughout the year. The final rule
defines the required flow rate measurement methods and/or procedures, repair and replacement
requirements, and recordkeeping and reporting requirements.
In addition, the following regulatory options have been added to the final rule: (1) owners
and operators are allowed to change out reciprocating compressor rod packing every 8,760 hours
of operation in lieu of conducting volumetric flow rate monitoring every 8,760 hours; and (2)
owners and operators are allowed to route emissions to a control device via a closed vent system
in addition to routing emissions via a closed vent system to a process. For the alternative option
of routing rod packing emissions to a process via a closed vent system under negative pressure,
the EPA is finalizing the removal of the negative pressure requirement. By allowing owners and
operators to change out rod packing every 8,760 hours of operation in lieu of conducting
volumetric flow rate monitoring every 8,760 hours, owners and operators have the option to
choose a more-stringent rod packing change out schedule (on or before every 8,760 hours of
operation) and avoid the need to conduct volumetric flow rate monitoring. Lastly, by the final
4-89
-------
rule allowing owners and operators to route emissions to a control device in addition to routing
emissions to a process, the EPA has added flexibility to the compliance options available for
owners and operators.
4.4.10.6 Centrifugal Compressors
For the NSPS OOOOb, the December 2022 Supplemental Proposal required that
centrifugal compressor affected facilities with wet seals comply with the GHG and VOC
standards by reducing methane and VOC emissions from each centrifugal compressor wet seal
fluid degassing system by 95 percent. As an alternative to routing the closed vent system to a
control device, an owner or operator was provided the option to route the closed vent system to a
process or utilize a self- contained wet seal centrifugal compressor. If an owner or operator
chooses to comply with this requirement either by using a control device to reduce emissions or
by routing to a process to reduce emissions, an owner was required to equip the wet seal fluid
degassing system with a cover and the cover must be connected through a closed vent system
meeting specified requirements, such as design and operation with no identifiable emissions.
For owners or operators of self-contained wet seal centrifugal compressors or centrifugal
compressors equipped with dry seals, the EPA proposed that owners or operators comply with
the GHG and VOC standards by reducing methane and VOC emissions by ensuring a volumetric
flow rate at or below 3 scfm. In addition to the flow rate monitoring being required every 8,760
hours of operation.
As described in the preamble to the final rule,153 the EPA is finalizing changes to the
proposed requirements for centrifugal compressors in for NSPS OOOOb as a result of comments
received on the November 2021 Proposal and December 2022 Supplemental Proposal. The EPA
believes the following rule changes associated with the centrifugal compressor volumetric flow
rate measurement performance standards will reduce impacts on small businesses: (1) volumetric
flow rate per seal standards of performance finalized as work practice standards (not as numeric
standards), (2) a 3 scfm volumetric flow rate per seal performance work practice standard for
self-contained wet seal centrifugal compressors (including centrifugal compressors equipped
153
See section XI.G of the final rule preamble.
4-90
-------
with mechanical seals), (3) a 10 scfm volumetric flow rate per seal performance work practice
standard for dry seal compressors, (4) a 9 scfm volumetric flow rate per seal performance work
practice standard for Alaska North Slope centrifugal compressors equipped with sour seal oil
separator and capture system, and (5) monitoring based on 8,760 hours of operation instead of
based on a calendar year.
The EPA believes the final rule volumetric flow rate per seal performance work practice
standard approach for self-contained wet seal centrifugal compressors, Alaska North Slope
centrifugal compressors equipped with sour seal oil separator and capture system, and centrifugal
compressors equipped with dry seals benefits small entities because facilities can use monitoring
data to determine if repair of a seal is necessary. For self-contained wet seal centrifugal
compressors (including centrifugal compressors equipped with mechanical seals) and Alaska
North Slope centrifugal compressors equipped with sour seal oil separator and capture system,
small entities benefit by allowing owners and operators to meet the work practice standards in
lieu of requiring that they reduce methane and VOC emissions from each centrifugal compressor
wet seal fluid degassing system by 95 percent. By tailoring the final rule volumetric flow rate per
seal performance work practice standards for self-contained wet seal centrifugal compressors and
Alaska North Slope centrifugal compressors equipped with sour seal oil separator and capture
system, EPA is addressing industry concerns (including small business concerns) that it would be
cost-ineffective for these low-emitting wet seal centrifugal compressors to reduce emissions
from wet seal fluid degassing system by 95 percent. By raising the final rule volumetric flow rate
per seal performance work practice standard for centrifugal compressors equipped with dry seals
from 3 scfm to 10 scfm per seal, EPA is addressing industry concerns (including small business
concerns) that a 3 scfm per seal standard is not supported and a 10 scfm volumetric flow rate
standard represents a maximum volumetric flow rate standard that is applicable to all dry seals
based on manufacturer information.
Requiring owners and operators to conduct volumetric flow rate monitoring based on
8,760 hours of operation instead of based on a calendar year reduces the burden on owners and
operators where compressors are not operational for multiple months or are used intermittently.
Additionally, by requiring that monitoring frequency based on hours of operation, owners and
operators have the flexibility to stagger maintenance activity throughout the year. The final rule
4-91
-------
defines the required flow rate measurement methods and/or procedures, repair requirements, and
recordkeeping and reporting requirements.
4.4.10.7 Liquid Unloading Operations
In the December 2022 Supplemental Proposal, the EPA proposed regulatory text where
all gas well liquids unloading operations would be subject to the regulatory requirements. The
BSER proposed was to employ techniques or technologies that eliminate methane and VOC
emissions. Where it was technically infeasible or not safe to meet the zero emissions standard,
the EPA proposed to require the employment of best management practices to minimize methane
and VOC emissions during well liquids unloading operations to the maximum extent possible.
The December 2022 Supplemental Proposal specifically requested further comment and any
additional information regarding whether would only apply to well affected facilities that
undergo well liquids unloading that result in vented emissions.
In addition, the December 2022 Supplemental Proposal specified recordkeeping and
reporting requirements related to well liquids unloading operations. Wells that utilized a non-
venting method would have been required to maintain records of the number of well liquids
unloading operations that occur within the reporting period and the method(s) used for each well
liquids unloading operation. A summary of this information would also have been required to be
reported in the annual report. In recognition that under some circumstances, venting could occur
when a selected liquids unloading method that is designed to not vent to the atmosphere is not
properly applied (e.g., a technology malfunction or operator error). Under the proposed rule,
owners and operators in this situation would have been required to record and report these
instances, as well as document and report the length of venting and what actions were taken to
minimize venting to the maximum extent possible. Additionally, for wells that utilize methods
that vent to the atmosphere, the proposed rule would have required: (1) Documentation
explaining why it is infeasible to utilize a non-venting method due to technical, safety, or
economic reasons; (2) development of best management practices that ensure that emissions
during liquids unloading are minimized; (3) employment of the best management practices
during each well liquids unloading operation and maintenance of records demonstrating that the
best management practices were followed; (4) reporting in the annual report both the number of
4-92
-------
well liquids unloading operations and any instances where the well liquids unloading operations
did not follow the best management practices.
As described in the preamble to the final rule,154 the EPA is finalizing certain changes to
the proposed requirements for gas well liquids unloading operations in the final rule. The EPA
believes these changes will reduce impacts on small businesses.
Several commenters opposed the EPA's proposed zero-emission standard or asserted that
the EPA should only regulate unloading operations that vent emissions. Another commenter
expressed that BSER must be technically feasible for the source category. The commenter stated
that the proposed standard is based on a determination that non-emitting techniques constitute
BSER for liquids unloading operations. At the same time, the commenter pointed out that the
EPA acknowledged that non-emitting techniques are not always feasible or safe and EPA
provides alternative standards to cover those situations.
Several of the commenters also requested that the EPA not require recordkeeping and
reporting of non-venting liquids unloading events. Commenters suggested that operators
conducting liquids unloading operations with zero methane and VOC emissions should not be
subject to burdensome recordkeeping, reporting and other requirements. Another commenter
noted that the non-vented liquids unloading reporting requirements proposed in the December
2022 Supplemental Proposal are not feasible due to the nature of those events and because of the
administrative burden associated with the reporting requirements with no net gain in emission
reductions.
In the final rule, the EPA requires that each well affected facility gas well that unloads
liquids to employ techniques or technologies that minimize or eliminate venting of emissions
during liquids unloading events to the maximum extent. For unloading technologies or
techniques that result in venting to the atmosphere, the final rule requires that owners or
operators employ best management practices that meet minimum specified criteria to minimize
venting of methane and VOC emissions for each gas well liquids unloading operation. This
reduces the burden on industry (including small businesses) from what was proposed in the
December 2022 Supplemental Proposal by not requiring a zero methane and VOC emissions
154
See section XI.F.3 of the final rule preamble.
4-93
-------
standard be met unless an owner or operator can document that it was technically infeasible or
not safe to meet the zero emissions standard.
The EPA also evaluated the commenters' concerns and examples provided regarding the
burden associated with the proposed liquids unloading operations recordkeeping and reporting
requirements. The EPA has determined that requiring an owner or operator to comply with some
of the proposed recordkeeping and reporting requirements in instances where an unloading event
does not result in venting to the atmosphere would impose a burden without any added benefit
environmentally (e.g., requiring that the number of liquids unloading events that occurred when
implementing a non-venting liquids unloading technology or technique be tracked and reported).
As a result, the final rule has been changed so that an owner or operator of a well affected facility
that employs non-venting liquids unloading technologies and techniques is only required to
comply with minimal recordkeeping and reporting requirements (i.e., identification of the well
affected facility and non-venting technology or technology employed; number of unplanned
venting events). In instances where there may be an unplanned venting event, that event would
be subject to the best management practices to minimize venting of emissions and the associated
recordkeeping and reporting requirements.
The EPA believes that the final work practice standard and associated recordkeeping and
reporting requirements will result in the minimization or elimination of venting of emissions to
the maximum extent possible during liquids unloading events, while streamlining the
recordkeeping and reporting requirements to minimize burden to industry (including small
businesses).
4.5 Employment Impacts of Environmental Regulation
This section presents an overview of the various ways that environmental regulation can
affect employment.155 Employment impacts of environmental regulations are generally composed
of a mix of potential declines and gains in different areas of the economy over time. Regulatory
employment impacts can vary across occupations, regions, and industries; by labor and product
demand and supply elasticities; and in response to other labor market conditions. Isolating such
155 Additionally, see Section 4.3.5 for a discussion of the demographic characteristics of oil and natural gas workers
and communities.
4-94
-------
impacts is a challenge, as they are difficult to disentangle from employment impacts caused by a
wide variety of ongoing, concurrent economic changes. The EPA continues to explore the
relevant theoretical and empirical literature and to seek public comments to ensure that the way
the EPA characterizes the employment effects of its regulations is reasonable and informative.
Environmental regulation "typically affects the distribution of employment among
industries rather than the general employment level" (Arrow et al., 1996). Even if impacts are
small after long-run market adjustments to full employment, many regulatory actions have
transitional effects in the short run (OMB, 2015). These movements of workers in and out of jobs
in response to environmental regulation are potentially important and of interest to policymakers.
Transitional job losses have consequences for workers that operate in declining industries or
occupations, have limited capacity to migrate, or live in communities or regions with high
unemployment rates.
As indicated by the potential impacts on oil and natural gas markets discussed in Section
0, the final NSPS OOOOb and EG 0000c are projected to cause small changes in oil and
natural gas production and prices. As a result, demand for labor employed in oil and natural gas-
related activities and associated industries might experience adjustments as there may be
increases in compliance-related labor requirements as well as changes in employment due to
quantity effects in directly regulated sectors and sectors that consume oil and natural gas
products.
4.6 References
Arrow, K. J., Cropper, M. L., Eads, G. C., Hahn, R. J., Lave, L. B., Noll, R. G., . . . Stavins, R.
N. (1996). Benefit-Cost Analysis in Environmental, Health, and Safety Regulation: A
Statement of Principles. American Enterprise Institute, the Annapolis Center, and Resources
for the Future; AE1 Press, https://www.aei.org/research-products/book/benefit-cost-analysis-
in-environmental-health-and-safety-regulation/
Bednar, D. J., & Reames, T. G. (2020). Recognition of and response to energy poverty in the
United States. Nature Energy, 5(6), 432-439. https://doi.org/10.1038/s41560-020-0582-0
Bovvers, R. L. (2022). Quantification of Methane Emissions from Marginal (Low Production
Rate) Oil and Natural Gas Wells, https://vvvvvv.osti.gov/biblio/1865859
https://www osti.gov/servlets/purl/1865859
Coglianese, J., Davis, L. W., Kilian, L., & Stock, J. H. (2017). Anticipation, Tax Avoidance, and
the Price Elasticity of Gasoline Demand. Journal of Applied Econometrics, 32(1), 1-15.
https://doi.org/10.1002/jae.2500
4-95
-------
Drehobl, A., Ross, L., & Ayala, R. (2020). How High Are Household Energy Burdens9 An
Assessment of National and Metropolitan Energy Burdens across the U.S.
https://vvvvvv.aceee.org/research-report/u2006
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. https://doi.org/10.1088/1748-9326/aac4bd
Hall, S. M. (2013). Energy justice and ethical consumption: comparison, synthesis and lesson
drawing. Local Environment, 18(4), 422-437. https://doi.org/! 0.1080/13549839.2012.748730
Hausman, C., & Kellogg, R. (2015). Welfare and Distributional Implications of Shale Gas.
Brookings Papers on Economic Activity, 71-125. https://www.brookings.edu/bpea-
articles/welfare-and-distributional-implications-of-shale-gas/
Hernandez, D., Aratani, Y., & Jiang, Y. (2014). Energy Insecurity Among Families with
Children [Brief], https://www.nccp.org/wp-content/uploads/2020/05/text_l086.pdf
Ho, J. S., Shih, J.-S., Muehlenbachs, L. A., Munnings, C., & Krupnick, A. J. (2018). Managing
Environmental Liability: An Evaluation of Bonding Requirements for Oil and Gas Wells in
the United States. Environmental Science & Technology, 52(7), 3908-3916.
https://doi.org/10.1021 /acs.est.7b06609
Hotel ling, H. (1931). The economics of exhaustible resources. Journal of Political Economy,
39(2), 137-175.
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. Pi rani, W. Moufouma-Okia, C. Pean, R.
Pidcock, S. Connors, J. B. R. Matthews, Y. Chen, X. Zhou, M. I. Gomis, E. Lonnoy, T.
May cock, a. M. Tignor, & T. Waterfield, Eds), https://www.ipcc.ch/sr 15/download/
Jessel, S., Sawyer, S., & Hernandez, D. (2019). Energy, Poverty, and Health in Climate Change:
A Comprehensive Review of an Emerging Literature. Frontiers in Public Health, 7, 357.
https://doi.org/10.3389/fpubh.2019.00357
Kaiser, M. J., & Pulsipher, A. G. (2006). Concerns Over the Allocation Methods Employed in
the US Low-In come Home Energy Assistance Program. INFORMS Journal on Applied
Analytics, 36(4), 344-358. https://doi.org/10.1287/inte. 1060.0223
Karpinska, L., & Smiech, S. (2020). Invisible energy poverty9 Analysing housing costs in
Central and Eastern Europe. Energy Research & Social Science, 70, 101670.
https://doi.org/IO. 1016/j erss.2020.101670
Kellogg, R. (2014). The effect of uncertainty on investment: evidence from Texas oil drilling.
American Economic Review, 104(6), 1698-1734.
Lee, E. K., Donley, G., Ciesielski, T. H., Gill, I., Yamoah, O., Roche, A., . . . Freedman, D. A.
(2022). Health outcomes in redlined versus non-redlined neighborhoods: A systematic
review and meta-analysis. Social science & medicine (1982), 294, I 14696.
https://doi org/10.1016/j .socscimed.202 III 4696
4-96
-------
Lyubich, E. (2020). The Race Gap in Residential Energy Expenditures (306). (Energy Institute
WP, Issue. http://vvvvvv.haas.berkeley.edu/vvp-content/uploads/WP306.pdf
Mitchell, B., & Franco, J. (2018). HOLC "redlining" maps: The persistent structure of
segregation and economic inequality, https://ncrc.org/vvp-
content/uploads/dlm_uploads/2018/02/NCRC-Research-HOLC-10.pdf
National Academies. (2017). Valuing Climate Damages: Updating Estimation of the Social Cost
of Carbon Dioxide. The National Academies Press.
https://www.nap.edu/catalog/2465 I/valuing-climate-damages-updating-estimation-of-the-
social-cost-of
National Academy of Sciences. (2019). Climate Change and Ecosystems. The National
Academies Press, https://doi.org/doi: 10.17226/25504
National Research Council. (2011). America's Climate Choices. The National Academies Press.
https://www.nap.edu/catalog/12781 /americas-climate-choices
Newell, R. G., & Prest, B. C. (2019). The Unconventional Oil Supply Boom: Aggregate Price
Response from Microdata. The Energy Journal, 40(3).
https://doi.org/! 0.5547/01956574.40.3.mew
Newell, R. G., Prest, B. C., & Vissing, A. B. (2019). Trophy Hunting versus Manufacturing
Energy: The Price Responsiveness of Shale Gas. Journal of the Association of Environmental
and Resource Economists, 6(2), 391-431. https://doi.org/10.1086/70153 I
Noelke, C., Outrich, M., Baek, M., Reece, J., Osypuk, T. L., McArdle, N., . . . Acevedo-Garcia,
D. (2022). Connecting past to present: Examining different approaches to linking historical
redlining to present day health inequities. PLoS ONE, 17(5), e0267606.
https://doi.org/ 10.1371/journal pone.0267606
Omara, M., Zavala-Araiza, D., Lyon, D. R., Hmiel, B., Roberts, K. A., & Hamburg, S. P. (2022).
Methane emissions from US low production oil and natural gas well sites. Nature
Communications, 13(1), 2085. https://doi.org/10.1038/s41467-022-29709-3
OMB. (2015). 2015 Report to Congress on the Benefits and Costs of Federal Regulations and
Agency Compliance with the Unfunded Mandates Reform Act.
https://www.vvhitehouse.gov/sites/vvhitehouse.gov/files/omb/inforeg/inforeg/20l 5_cb/2015-
cost-benefit-report.pdf
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 11 to the Fifth Assessment Report of the
Intergovernmental Panel on Climate Change (pp. 1039-1099). Cambridge University Press.
https://www.ipcc.ch/report/ar5/vvg2/
Pindyck, R. S. (2002). Optimal timing problems in environmental economics. Journal of
Economic Dynamics and Control, 26(9-10), 1677-1697.
4-97
-------
Porter, J. R., Xie, L., Challinor, A. J., Cochrane, K., Hovvden, 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 11 to the Fifth Assessment Report of the
Intergovernmental Panel on Climate Change (pp. 485-533). Cambridge University Press.
https://vvvvvv.ipcc.ch/report/ar5/vvg2/
Provencher, B. (1997). Structural versus reduced-form estimation of optimal stopping problems.
American Journal of Agricultural Economics, 79(2), 357-368.
Raimi, D., Krupnick, A. J., Shah, J.-S., & Thompson, A. (2021). Decommissioning orphaned and
abandoned oil and gas wells: New estimates and cost drivers. Environmental Science &
Technology, 55(15), 10224-10230.
Smith, K. R., Woodward, A., Campbell-Lendrum, D., Chadee, D. D., Honda, Y., Liu, Q., . . .
Sauerboni, 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 University Press, https://vvvvvv.ipcc.ch/report/ar5/vvg2/
Svvope, C. B., Hernandez, D., & Cushing, L. J. (2022). The Relationship of Historical Redlining
with Present-Day Neighborhood Environmental and Health Outcomes: A Scoping Review
and Conceptual Model. Journal of Urban Health, 99(6), 959-983.
https://doi org/10.1007/s I 1524-022-00665-z
U.S. EI A. (2018). One in three U.S. households faces a challenge in meeting energy needs
(Today in Energy, Issue, https://www.eia.gov/todayinenergy/detail.php?id=37072
U.S. EPA. (2015). Guidance on Considering Environmental Justice During the Development of
Regulatory Actions. Retrieved from https://www.epa.gov/sites/default/files/2015-
06/docum ent s/con si deri ng-ej -i n-rul em aki ng-gui de-fi nal. pdf
U.S. EPA. (2018). Technical Support Document: EPA's 2014 National Air Toxics Assessment.
https://www.epa.gov/national-air-toxics-assessment/2014-nata-technical-support-document
U.S. EPA. (2022). Technical Support Document (TSD): Preparation of Emissions Inventories for
the 2016v2 North American Emissions Modeling Platform. (EPA-454/B-22-001). Research
Triangle Park, NC: U.S. Environmental Protection Agency, Office of Air Quality Planning
and Standards Retrieved from https://www.epa.gov/system/files/documents/2022-
02/2016v2_emi smod_tsd_february2022 pdf
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 Retrieved
from http://dx.doi.org/10.7930/J0R49NQX
4-98
-------
USGCRP. (2018). Impacts, Risks, and Adaptation in the United States: Fourth National Climate
Assessment, Volume 11. Washington DC: U.S. Global Change Research Program Retrieved
from http://dx.doi.org/10.7930/NC A4.2018
Wang, Q., Kvvan, M.-P., Fan, J., & Lin, J. (2021). Racial disparities in energy poverty in the
United States. Renewable and Sustainable Energy Reviews, 137, I 10620.
https://doi.org/10.1016/j .rser.2020.1 10620
Weber, J. G., Ercoli, T., Fitzgerald, W., Nied, P., Penderville, M., & Raabe, E. (2021).
Identifying the end: Minimum production thresholds for natural gas wells. Resources Policy,
74, 102404. https://doi org/https://doi org/10.1016/j resourpol.2021.102404
4-99
-------
5 COMPARISON OF BENEFITS AND COSTS
5.1 Comparison of Benefits and Costs
A comparison of quantified benefits and costs is presented below. All estimates are in
2019 dollars. Also, all compliance costs, emissions changes, and benefits are estimated for the
years 2024 to 2038 relative to a baseline without the final NSPS OOOOb and EG OOOOc.
Table 5-lTable 5-1 summarizes the emissions reductions associated with the final
standards over the 2024 to 2038 period for the NSPS OOOOb, the EG OOOOc, and the NSPS
OOOOb and EG OOOOc combined. Tables 5-2Table 5-2, 5-3Table 5-3, and 5-4Table 5-4
present the present value (PV) and equivalent annual value (EAV), estimated using discount
rates of 2, 3, and 7 percent, of the changes in quantified climate and health benefits, costs, and
net benefits, as well as the emissions reductions relative to the baseline for the NSPS OOOOb,
for the EG OOOOc, and the NSPS OOOOb and EG OOOOc, respectively.156 157 These values
reflect an analytical time horizon of 2024 to 2038, are discounted to 2021, and presented in 2019
dollars. These tables include consideration of the non-monetized benefits associated with the
emissions reductions projected under this final rule.
156 Monetized climate effects are presented under a 2 percent near-term Ramsey discount rate, consistent with EPA's
updated estimates of the SC-GHG. The 2003 version of OMB 's Circular A-4 had generally recommended 3 percent
and 7 percent as default discount rates for costs and benefits, though as part of the Interagency Working Group on
the Social Cost of Greenhouse Gases, OMB had also long recognized that climate effects should be discounted only
at appropriate consumption-based discount rates. While this RIA was being drafted and reviewed, OMB finalized an
update to Circular A-4, in which it recommended the general application of a 2.0 percent discount rate to costs and
benefits (subject to regular updates), as well as the consideration of the shadow price of capital when costs or
benefits are likely to accrue to capital (OMB 2023). Because the SC-GHG 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-GHG. See Section 3.2 for more discussion.
157
The EPA has also applied its updated estimates of the social cost of carbon dioxide (SC-CO2) in an illustrative
analysis of potential climate disbenefits from secondary CO2 emissions associated with control techniques for
storage vessels. Given that the estimated climate disbenefits from the CO2 impacts would at most offset only about 1
percent of the methane benefits, the EPA finds that the summary values shown in this table are a reasonable estimate
of the net monetized climate effects of the rule. See Section 3.9 for further discussion.
5-1
-------
Table 5-1 Projected Emissions Reductions under the Final NSPS OOOOb and EG
OOOOc across Regulatory Options, 2024-2038a'b'c
Emissions Changes
Methane
Methane (million metric
Regulatory (millions short VOC (millions HAP (millions tons CO2 Eq.
Option Final Rule tons) short tons) short tons) using GWP=28)
Less Stringent
NSPS OOOOb 23 6.9 0.26 580
EG OOOOc 31 7.5 0.28 780
Total 54 14 0.54 1,400
Final Rule
NSPS OOOOb
EG OOOOc
Total
More Stringent
NSPS OOOOb 23 7.1 0.27 590
EG OOOOc 35 8.7 0.33 890
Total 59 16 (L59 1,500
a Numbers rounded to two significant digits unless otherwise noted. Totals may not appear to add correctly due to
rounding. To convert from short tons to metric tons, multiply the short tons by 0.907. Alternatively, to convert
metric tons to short tons, multiply metric tons by 1.102.
b The EG OOOOc regulates emissions of methane. Additional benefits to the regulation result from associated
reductions in VOC and HAP emissions.
0 The control techniques to meet the storage vessel-related standards are associated with several types of secondary
emissions impacts, which may partially offset the direct benefits of this rule. As discussed in Section 3.9, the
magnitude of these secondary air pollutant increases is small relative to the direct emission reductions anticipated
from this rule.
23 7.1
35 8.6
58 16
0.27 590
0.32 890
0.59 1,500
5-2
-------
Table 5-2 Projected Benefits, Compliance Costs, and Emissions Reductions across
Regulatory Options for the Final NSPS OOOOb, 2024-2038 (million 2019$)
2 Percent Near-Term Ramsey Discount Rate
PV
EAV
PV
EAV
PV
EAV
Climate Benefitsb
Less Stringent
$43,000
$3,300
$43,000
$3,300
$43,000
$3,300
Final Rule
$44,000
$3,400
$44,000
$3,400
$44,000
$3,400
More Stringent
$44,000
$3,400
$44,000
$3,400
$44,000
$3,400
2 Percent Discount
3 Percent Discount
7 Percent Discount
Rate
Rate
Rate
PV
EAV
PV
EAV
PV
EAV
Ozone Health Benefits0
Less Stringent
N/A
N/A
N/A
N/A
N/A
N/A
Final Rule
N/A
N/A
N/A
N/A
N/A
N/A
More Stringent
N/A
N/A
N/A
N/A
N/A
N/A
Net Compliance Costs
Less Stringent
$5,800
$450
$5,700
$480
$5,200
$570
Final Rule
$5,800
$450
$5,800
$480
$5,300
$580
More Stringent
$6,000
$460
$5,900
$490
$5,300
$590
Compliance Costs
Less Stringent
$14,000
$1,100
$13,000
$1,100
$9,900
$1,100
Final Rule
$14,000
$1,100
$13,000
$1,100
$10,000
$1,100
More Stringent
$14,000
$1,100
$13,000
$1,100
$10,000
$1,100
Value of Product Recovery
Less Stringent
$7,800
$610
$7,000
$590
$4,700
$510
Final Rule
$7,900
$620
$7,100
$590
$4,700
$520
More Stringent
$7,900
$620
$7,100
$600
$4,700
$520
Net Monetized Benefits'1
Less Stringent
$37,000
$2,900
$37,000
$2,900
$38,000
$2,800
Final Rule
$38,000
$3,000
$38,000
$2,900
$39,000
$2,800
More Stringent
$38,000
$3,000
$38,000
$2,900
$39,000
$2,800
Non-Monetized Benefits
Benefits to provision of ecosystem services and ozone-related health benefits from reducing methane emissions
by (in short tons):
Less Stringent
Final Rule
More Stringent
23,000,000
23,000,000
23,000,000
Benefits to provision of ecosystem services from reducing VOC emissions by (in short tons):
Less Stringent
Final Rule
More Stringent
6,900,000
7,100,000
7,100,000
PM2 5-related health benefits from reducing VOC emissions by (in short tons)b:
Less Stringent 6,900,000
5-3
-------
Final Rule
More Stringent
7,100,000
7,100,000
Benefits to provision of ecosystem services and HAP-related health benefits from reducing HAP emissions by (in
short tons):
a Values rounded to two significant figures. Totals may not appear to add correctly due to rounding.
b Monetized climate benefits are based on reductions in methane emissions and are calculated using three different
estimates of the social cost of methane (SC-CH4) (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-
CH4 at the 2 percent near-term Ramsey discount rate. Please see Tables 3.4 and 3.5 for the full range of monetized
climate benefit estimates.
0 The ozone-related health benefits estimates use the larger of the two benefits estimates presented in Table 3-10.
Monetized benefits include those related to public health associated with reductions in ozone concentrations. The
health benefits are associated with several point estimates.
d Several categories of climate, human health, and welfare benefits from methane, VOC, and HAP emissions
reductions remain unmonetized and are thus not directly reflected in the quantified benefit estimates in the table. See
Section 3.2 for a discussion of climate effects that are not yet reflected in the SC-CH4 and thus remain unmonetized
and Sections 3.4 through 3.8 for a discussion of other nonmonetized benefits.
Less Stringent
Final Rule
More Stringent
260,000
270,000
270,000
5-4
-------
Table 5-3 Projected Benefits, Compliance Costs, and Emissions Reductions across
Regulatory Options for the Final EG OOOOc, 2024-2038 (million 2019$)
2 Percent Near-Term Ramsey Discount Rate
PV
EAV
PV
EAV
PV
EAV
Climate Benefitsb
Less Stringent
$57,000
$4,500
$57,000
$4,500
$57,000
$4,500
Final Rule
$65,000
$5,100
$65,000
$5,100
$65,000
$5,100
More Stringent
$66,000
$5,100
$66,000
$5,100
$66,000
$5,100
2 Percent Discount
3 Percent Discount
7 Percent Discount
Rate
Rate
Rate
PV
EAV
PV
EAV
PV
EAV
Ozone Health Benefits0
Less Stringent
N/A
N/A
N/A
N/A
N/A
N/A
Final Rule
N/A
N/A
N/A
N/A
N/A
N/A
More Stringent
N/A
N/A
N/A
N/A
N/A
N/A
Net Compliance Costs
Less Stringent
$12,000
$940
$11,000
$930
$7,900
$870
Final Rule
$13,000
$1,000
$12,000
$1,000
$8,900
$970
More Stringent
$38,000
$2,900
$35,000
$3,000
$27,000
$2,900
Compliance Costs
Less Stringent
$16,000
$1,300
$15,000
$1,200
$10,000
$1,100
Final Rule
$18,000
$1,400
$16,000
$1,400
$12,000
$1,300
More Stringent
$43,000
$3,300
$40,000
$3,300
$30,000
$3,300
Value of Product Recovery
Less Stringent
$4,100
$320
$3,700
$310
$2,400
$260
Final Rule
$4,700
$370
$4,200
$350
$2,700
$300
More Stringent
$5,000
$390
$4,400
$370
$2,900
$320
Net Monetized Benefits'1
Less Stringent
$45,000
$3,500
$46,000
$3,500
$50,000
$3,600
Final Rule
$52,000
$4,100
$53,000
$4,100
$56,000
$4,100
More Stringent
$28,000
$2,200
$31,000
$2,200
$39,000
$2,200
Non-Monetized Benefits
Benefits to provision of ecosystem services and ozone-related health benefits from reducing methane emissions
by (in short tons):
Less Stringent
Final Rule
More Stringent
31,000,000
35,000,000
35,000,000
Benefits to provision of ecosystem services from reducing VOC emissions by (in short tons):
Less Stringent
Final Rule
More Stringent
7,500,000
8,600,000
8,700,000
PM2 5-related health benefits from reducing VOC emissions by (in short tons)b:
Less Stringent 7,500,000
5-5
-------
Final Rule
More Stringent
8,600,000
8,700,000
Benefits to provision of ecosystem services and HAP-related health benefits from reducing HAP emissions by (in
short tons):
a Values rounded to two significant figures. Totals may not appear to add correctly due to rounding.
b Monetized climate benefits are based on reductions in methane emissions and are calculated using three different
estimates of the social cost of methane (SC-CH4) (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-
CH4 at the 2 percent near-term Ramsey discount rate. Please see Table 3.4 and 3.5 for the full range of monetized
climate benefit estimates.
0 The ozone-related health benefits estimates use the larger of the two benefits estimates presented in Table 3-10.
Monetized benefits include those related to public health associated with reductions in ozone concentrations. The
health benefits are associated with several point estimates.
d Several categories of climate, human health, and welfare benefits from methane, VOC, and HAP emissions
reductions remain unmonetized and are thus not directly reflected in the quantified benefit estimates in the table. See
Section 3.2 for a discussion of climate effects that are not yet reflected in the SC-CH4 and thus remain unmonetized
and Sections 3.4 through 3.8 for a discussion of other nonmonetized benefits.
Less Stringent
Final Rule
More Stringent
280,000
320,000
330,000
5-6
-------
Table 5-4 Projected Benefits, Compliance Costs, and Emissions Reductions across
Regulatory Options for the Final NSPS OOOOb and EG OOOOc, 2024-2038 (million
2019$)
2 Percent Near-Term Ramsey Discount Rate
PV
EAV
PV
EAV
PV
EAV
Climate Benefitsb
Less Stringent
$100,000
$7,800
$100,000
$7,800
$100,000
$7,800
Final Rule
$110,000
$8,500
$110,000
$8,500
$110,000
$8,500
More Stringent
$110,000
$8,500
$110,000
$8,500
$110,000
$8,500
2 Percent Discount
3 Percent Discount
7 Percent Discount
Rate
Rate
Rate
PV
EAV
PV
EAV
PV
EAV
Ozone Health Benefits0
Less Stringent
$6,300
$490
$5,400
$450
$3,100
$340
Final Rule
$7,000
$540
$6,100
$510
$3,500
$380
More Stringent
$7,100
$550
$6,100
$510
$3,500
$390
Net Compliance Costs
Less Stringent
$18,000
$1,400
$17,000
$1,400
$13,000
$1,400
Final Rule
$19,000
$1,500
$18,000
$1,500
$14,000
$1,600
More Stringent
$44,000
$3,400
$41,000
$3,400
$32,000
$3,500
Compliance Costs
Less Stringent
$30,000
$2,300
$27,000
$2,300
$20,000
$2,200
Final Rule
$31,000
$2,400
$29,000
$2,400
$22,000
$2,400
More Stringent
$57,000
$4,400
$53,000
$4,400
$40,000
$4,400
Value of Product Recovery
Less Stringent
$12,000
$930
$11,000
$900
$7,100
$780
Final Rule
$13,000
$980
$11,000
$950
$7,400
$820
More Stringent
$13,000
$1,000
$12,000
$970
$7,600
$840
Net Monetized Benefits'1
Less Stringent
$89,000
$6,900
$89,000
$6,900
$90,000
$6,700
Final Rule
$97,000
$7,600
$97,000
$7,500
$98,000
$7,300
More Stringent
$73,000
$5,700
$75,000
$5,600
$81,000
$5,400
Non-Monetized Benefits
Benefits to provision of ecosystem services and ozone-related health benefits from reducing methane emissions
by (in short tons):
Less Stringent
Final Rule
More Stringent
31,000,000
35,000,000
35,000,000
Benefits to provision of ecosystem services from reducing VOC emissions by (in short tons):
Less Stringent
Final Rule
More Stringent
7,500,000
8,600,000
8,700,000
PM2 5-related health benefits from reducing VOC emissions by (in short tons)b:
Less Stringent 7,500,000
5-7
-------
Final Rule
More Stringent
8,600,000
8,700,000
Benefits to provision of ecosystem services and HAP-related health benefits from reducing HAP emissions by (in
short tons):
a Values rounded to two significant figures. Totals may not appear to add correctly due to rounding.
b Monetized climate benefits are based on reductions in methane emissions and are calculated using three different
estimates of the social cost of methane (SC-CH4) (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-
CH4 at the 2 percent near-term Ramsey discount rate. Please see Table 3.4 and 3.5 for the full range of monetized
climate benefit estimates.
0 The ozone-related health benefits estimates use the larger of the two benefits estimates presented in Table 3-10.
Monetized benefits include those related to public health associated with reductions in ozone concentrations. The
health benefits are associated with several point estimates.
d Several categories of climate, human health, and welfare benefits from methane, VOC, and HAP emissions
reductions remain unmonetized and are thus not directly reflected in the quantified benefit estimates in the table. See
Section 3.2 for a discussion of climate effects that are not yet reflected in the SC-CH4 and thus remain unmonetized
and Sections 3.4 through 3.8 for a discussion of other nonmonetized benefits.
The following two tables show the total emissions reductions and the PV and EAV of net
compliance costs over the 2024 to 2038 period. The projected net compliance costs for
reciprocating compressors (in all segments) and wet seal centrifugal compressors (processing
segment only) are negative, as the projected revenue from product recovery exceeds the
projected cost increases. This observation may typically support an assumption that operators
would continue to perform the emissions abatement activity, regardless of whether a requirement
is in place, because it is in their private self-interest. However, the reciprocating compressors are
in the gathering and boosting, processing, transmission, and storage segments. As discussed in
previous oil and natural gas NSPS RIAs, operators in those segments of the industry do not
typically own the natural gas they transport; rather, the operators receive payment for the
transportation service they provide. As a result, financial incentives to reduce emissions may be
minimal because operators are not able to recoup the financial value of captured natural gas that
may otherwise be emitted. Alternatively, there may also be an opportunity cost associated with
the installation of environmental controls (for purposes of mitigating the emission of pollutants)
that is not reflected in the control costs. In the event that the environmental investment displaces
investment in productive capital, the difference between the rate of return on the marginal
investment displaced by the mandatory environmental investment is a measure of the opportunity
cost of the environmental requirement to the regulated entity. However, if firms are not capital
constrained — as is likely to be the case for firms that can access liquid capital markets — then
Less Stringent
Final Rule
More Stringent
280,000
320,000
330,000
5-8
-------
there may not be any displacement of investment, and the rate of return on other investments in
the industry would not be relevant as a measure of opportunity cost.158 If firms face higher
borrowing costs as they take on more debt, there may be an additional opportunity cost to the
firm. To the extent that any opportunity costs are not added to the control costs, the compliance
cost reductions presented above may be underestimated.
Table 5-5 Projected Emissions Reductions for Incrementally Affected Sources under
the Final NSPS OOOOb and EG OOOOc, 2024 to 2038
Nationwide Emissions Reductions
Methane
(metric
Methane
voc
HAP
tons CO2
Segment
Source
(short tons)
(short tons)
(short tons)
Eq.)
Fugitives: Components
5,500,000
1,500,000
58,000
140,000,000
Fugitives: Storage Vessel Flares
10,000,000
2,900,000
110,000
260,000,000
Production
Pneumatics
26,000,000
7,300,000
270,000
660,000,000
Storage Vessels
150,000
700,000
26,000
3,900,000
Associated Gas
4,900,000
1,400,000
51,000
120,000,000
Liquids Unloading
260,000
73,000
2,800
6,700,000
Fugitives
730,000
200,000
7,600
18,000,000
Gathering and
Pneumatics
2,400,000
670,000
25,000
61,000,000
Boosting
Reciprocating Compressors
2,100,000
580,000
22,000
53,000,000
Wet-Seal Centrifugal Compressors
970,000
270,000
10,000
25,000,000
Natural Gas
Processing
Leaks
Reciprocating Compressors
180,000
710,000
21,000
20,000
610
580
4,500,000
18,000,000
Wet-Seal Centrifugal Compressors
400,000
11,000
330
10,000,000
Fugitives
890,000
25,000
730
23,000,000
Transmission
Pneumatics
560,000
16,000
460
14,000,000
and Storage
Reciprocating Compressors
1,700,000
48,000
1,400
44,000,000
Wet-Seal Centrifugal Compressors
280,000
7,700
230
7,100,000
Note: Values rounded to two significant figures.
158 See Circular A-4: Explanation and Response to Public Input 81-87 (OMB 2023).
5-9
-------
Table 5-6 Projected Climate Benefits and Compliance Costs (millions 2019$) for
Incrementally Affected Sources under the Final NSPS OOOOb and EG OOOOc Option,
2024 to 2038a b
Costs and Benefits (PV, million 2019$)
Annualized
Cost,
Annualized
without
Cost, with
Climate
Capital
Product
Product
Segment
Source
Benefits
Cost
Recovery
Recovery
Fugitives
$30,000
$300
$8,200
$7,400
Pneumatics
$49,000
$7,200
$4,400
$730
Production
Storage Vessels
$290
$620
$1,600
$1,600
Associated Gas
$9,100
$20,000
$15,000
$8,600
Liquids Unloading
$490
$0
$140
$100
Fugitives
$1,400
$25
$580
$470
Gathering
Pneumatics
$4,500
$880
$1,100
$760
and Boosting
Reciprocating Compressors
$4,000
$130
$280
-$20
Wet-Seal Centrifugal Compressors
$1,800
$0
$86
-$52
Natural Gas
Processing
Leaks
$330
$0
$36
$12
Reciprocating Compressors
$1,300
$44
$62
-$34
Wet-Seal Centrifugal Compressors
$760
$0
$59
$4
Fugitives
$1,700
$70
$310
$200
Transmission
Pneumatics
$1,100
$320
$280
$210
and Storage
Reciprocating Compressors
$3,200
$92
$140
-$77
Wet-Seal Centrifugal Compressors
$520
$0
$160
$120
a Values rounded to two significant figures. Totals may not appear to add correctly due to rounding. Costs and
climate benefits in each year are discounted to 2021.
b Due to time and resource limitations, we are unable to estimate the monetized ozone benefits on a source-by-
source basis. Several categories of climate, human health, and welfare benefits from methane, VOC, and HAP
emissions reductions remain unmonetized and are thus not directly reflected in the quantified benefit estimates in the
table.
0 Climate benefits are based on reductions in methane emissions and are calculated using three different estimates of
the social cost of methane (SC-CH4) (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-CH4 at the
2 percent near-term Ramsey discount rate.
5.2 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.
Source-level compliance costs and emissions impacts: As discussed in Section 2.2, the
first step in the compliance cost analysis is the development of per-facility national-average
representative costs and emissions impacts using a model plant approach. The model plants are
5-10
-------
designed based upon the best information available to the Agency at the time of the rulemaking.
By emphasizing facility averages, geographic variability, and heterogeneity across producers in
the industry is masked, and regulatory impacts at the facility-level may vary from the model
plant averages. This assumption is particularly important when assessing the impacts of
requirements that depend on thresholds, such as for storage vessels. For a well site group, which
represents a collection of well sites and their average characteristics, all sites within the group are
categorized as either being below or above the emissions limit, though it may be the case that
some sites within the group exceed the limit while others do not. Misspecifications of this sort
may average out across the full set of well site groups.
There may also be an opportunity cost associated with the installation of environmental
controls (for purposes of mitigating the emissions of pollutants) that is not reflected in the
control costs. In the event that investment in environmental compliance displaces other
investment in productive capital, the difference between the rate of return on the investment
displaced by the mandatory environmental investment is a measure of the opportunity cost of the
environmental requirement. To the extent that such opportunity costs of capital are not accounted
for in the estimated compliance cost reductions, the cost reductions may be underestimated.
Projection methods and assumptions: As discussed in Section 2.2.1, the second
component in estimating national impacts is the projection of affected facilities. Uncertainties in
the projections informing this RIA results include: 1) choice of projection method; 2) data
sources and drivers; 3) limited information about rate of modification and turnover of sources; 4)
behavioral responses to regulation; and 5) unforeseen changes in industry and economic shocks.
The list of assumptions required to inform the analysis is too numerous to provide a
comprehensive list, but a few key drivers of the results warrant specific mention.
2016 ICR Data: As discussed previously, the 2016 Oil and Natural Gas ICR was
withdrawn in 2017. Therefore, the data represent an incomplete, and possibly unrepresentative,
survey of operators and well sites. Even so, we believe that it represents the best available data to
use for this analysis, as it includes additional variables beyond, and many more well site
observations than, other equipment surveys that we are aware of (e.g., the API well site survey
discussed in Section 2.2.1.2, which was used to estimate the distribution of fugitive emissions
from components at well sites for the November 2021 RIA). To date, we have not formally
5-11
-------
analyzed the representativeness of the data collected. Informal benchmarks, such as the
proportions of single-well versus multi-well sites and low production versus non-low production
sites and average equipment counts, when compared to outside data sources that attempt to
capture the universe of well sites (such as Enverus and GHGI), did not suggest significant issues
with the representativeness of the ICR data.
Site/Equipment Retirement and Modification: Our assumptions on non-well site
retirement rates are based on impressions stemming from conversations with, and comments
from, industry stakeholders and are not derived from data sources due to a lack of information.
Our assumptions for well site retirement rates are based on an analysis of Enverus data, but we
are still assessing improvements to our methods for estimating those rates. In all cases, we
assume that, prior to implementation of the NSPS and EG, equipment at sites shares the same
vintage as the sites themselves. For example, if a well site was constructed prior to the
promulgation of NSPS OOOO, we assume that all controllers at the site pre-date the NSPS
0000 as well and are not replaced until the EG goes into effect in 2028. By not accounting for
the possibility of equipment replacements and site modifications, we may be overstating the
impacts for some sources that were constructed prior to the NSPS 0000 and/or NSPS 0000a
but are now subject to those rules.
Years of analysis: The years of analysis are 2024, to represent the full first-year facilities
are affected by this action, through 2038, to represent impacts of the rule over a longer period, as
discussed in Section 2.2. While it would be desirable to analyze impacts beyond 2038 in this
RIA, the EPA has chosen not to do this largely because of the limited information available on
the turnover rate of emissions sources and controls. Extending the analysis beyond 2038 would
introduce substantial and increasing uncertainties in the projected impacts of the rule. That said,
some amount of both benefits and costs would likely continue after 2038, and we note that
toward the end of our analytical time horizon, undiscounted net costs are relatively steady from
year to year (Table 2-12) while undiscounted monetized climate benefits (Table 3-4) are rising
each year. It is therefore plausible that significant net benefits would continue in the years after
2038, though for the reasons given, we do not currently attempt to monetize these effects.
Treatment of sources in Alaska: The RIA does not account for instances in which all or
some sources in Alaska are subject to different requirements than those in the rest of the country,
5-12
-------
both in the baseline due to previous rulemakings and in this final rule. For example, the 2018
amendments to the 2016 NSPS OOOOa ("Alaska Amendments") reduced fugitives monitoring
frequency requirements for well sites and compressor stations on the Alaska North Slope.159 We
do not reflect those reduced requirements in the baseline in this RIA, nor do we reflect that the
same reduced requirements are being finalized for the NSPS OOOOb and EG 0000c. In
addition, for sites in Alaska, the NSPS OOOOb and EG 0000c only requires non-emitting
pneumatic controllers to be installed at sites where onsite power is available; otherwise, the
requirement is to replace high-bleed controllers with low-bleed controllers and to monitor
intermittent bleed controllers for malfunctions. In both cases, these omissions suggest that our
analysis may overestimate the impacts of the final regulation.
State rules and voluntary action in the baseline: As discussed in Section 2.2.3, while
we accounted for state regulations in California, Colorado, and (to a more limited degree) New
Mexico and Pennsylvania, there are many other state and local requirements that may be in the
baseline that we are unable to account for. In addition, the baseline does not reflect voluntary
actions firms may take to reduce emissions in the oil and natural gas sector. By not accounting for
state and local requirements (outside of Colorado, California, New Mexico, and Pennsylvania) and
voluntary actions in the baseline, this analysis may overestimate both the benefits and costs of the
final regulation.
Wellhead natural gas prices used to estimate revenues from natural gas recovery:
The compliance cost estimates presented in this RIA include the estimates of the revenue
associated with the increase in natural gas recovery resulting from compliance actions. As a
result, the national compliance cost impacts depend on the price of natural gas. As explained in
Section 2.4 natural gas prices used in this analysis are from the projection of the Henry Hub price
in the AEO2022. To the extent actual natural gas prices diverge from the AEO projections, the
actual impacts will diverge from our estimates.
Oil and natural gas market impact analysis: The oil and natural gas market impact
analysis presented in this RIA is subject to several caveats and limitations. As with any modeling
exercise, the market impact analysis presented here depends crucially on uncertain input
159 83 FR 10628.
5-13
-------
parameters and assumptions regarding market structure. A more detailed discussion of the
caveats and limitations of the oil and natural gas market impacts analysis can be found in Section
0.
Monetized methane-related climate benefits: The EPA considered the uncertainty
associated with the social cost of methane (SC-CH4) estimates, which were used to calculate the
monetized climate benefits of the decrease in methane emissions projected because of this action.
Section 3.2 provides a detailed discussion of the limitations and uncertainties associated with the
SC-CH4 estimates used in this analysis and describes ways in which the modeling addresses
quantified sources of uncertainty.
Monetized VOC-related ozone benefits: The analysis of monetized VOC-related ozone
benefits described in Section 3.3 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 ozone-related premature deaths.
The estimated number and value of avoided ozone-attributable deaths are subject to
uncertainty. When estimating the economic value of avoided premature mortality from long-term
exposure to ozone, we use a 20-year segmented lag as there is no alternative empirical estimate
of the cessation lag for long-term exposure to ozone. The 20-year segmented lag accounts for the
onset of cardiovascular related mortality, an outcome which is not relevant to the long-term
respiratory mortality estimated here. We use a log-linear health impact function without a
threshold in modeling both long- and short-term ozone-related mortality. However, we
acknowledge reduced confidence in specifying the shape of the concentration-response
relationship in the range of 40 ppb and below (U.S. EPA, 2020). Thus, estimates include health
benefits from reducing ozone in areas with concentrations of ozone down to the lowest modeled
concentrations.
5-14
-------
Our estimate of the total monetized ozone-attributable benefits is based on the EPA's
interpretation of the best available scientific literature and methods and supported by the SAB-
HES and the National Academies of Science (2002, 2008). Since the publication of these reports,
the EPA has continued improving its techniques for characterizing uncertainty in the estimated
air pollution-attributable benefits. Where possible, we quantitatively assess uncertainty in each
input parameter (for example, statistical uncertainty is characterized by performing Monte Carlo
simulations). However, in some cases, this type of quantitative analysis is not possible due to
lack of data, so we instead characterize the sensitivity of the results to alternative plausible input
parameters. And, for some inputs into the benefits analysis, such as the air quality data, we lack
the data to perform either a quantitative uncertainty analysis or sensitivity analysis. Additional
detail regarding specific uncertainties associated with ozone health benefit estimates can be
found in the Health Benefits TSD (U. S. EPA, 2023).
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 3.
Non-quantified regulatory impacts: We do not attempt to quantify regulatory impacts
for all finalized requirements in this RIA. For a discussion of these requirements, see Section .
Environmental justice analyses: the EPA performed quantitative EJ assessments of
baseline HAP cancer risks, ozone exposure and health risks, employment, and household energy
expenditures. Each of these analyses are subject to various types of uncertainty related to input
parameters and assumptions. Qualitatively, assessments that further subdivide the populations
are subject to increased uncertainty as compared to overall exposure and risk estimates.
5.3 References
National Research Council. (2002). Estimating the Public Health Benefits of Proposed Air
Pollution Regulations. The National Academies Press, https://doi.org/doi: 10.17226/105 I I
National Research Council. (2008). Estimating Mortality Risk Reduction and Economic Benefits
from Controlling Ozone Air Pollution. In. National Academies Press (US),
https://doi.org/10.17226/12198
OMB. (2003). Circular A-4: Regulatory Analysis. Washington DC. Retrieved from
https://obamavvhitehouse.archives.gov/omb/circulars_a004_a-4/
OMB. (2023). Circular A-4: Regulatory Analysis. Washington DC. Retrieved from
https://vvvvvv.vvhitehouse.gOv/vvp-content/uploads/2023/l I/CircularA-4.pdf
5-15
-------
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
U.S. EPA. (2020). Integrated Science Assessment (ISA) for Ozone and Related Photochemical
Oxidants (Final Report). (EPA/600/R-20/012). Washington DC: U.S. Environmental
Protection Agency Retrieved from
https://cfpub.epa.gov/ncea7i sa/recordisplay.cfrn?deid=348522
5-16
-------
United States Office of Air Quality Planning and Standards Publication No. EPA-452/R-23-013
Environmental Protection Health and Environmental Impacts Division December 2023
Agency Research Triangle Park, NC
------- |