* O \
KWJ
*1 PRO^^
Policy Assessment for the Reconsideration of
the Ozone National Ambient Air Quality
Standards
External Review Draft

-------

-------
EPA-452/D-22-002
April 2022
Policy Assessment for the Reconsideration of the
Ozone National Ambient Air Quality Standards
External Review Draft
U.S. Environmental Protection Agency
Office of Air Quality Planning and Standards
Health and Environmental Impacts Division
Research Triangle Park, NC

-------
DISCLAIMER
This document has been prepared by staff in the U.S. Environmental Protection Agency's Office
of Air Quality Planning and Standards. Any findings and conclusions are those of the authors and
do not necessarily reflect the views of the Agency. This document does not represent and should
not be construed to represent any Agency determination or policy. Mention of trade names or
commercial products does not constitute endorsement or recommendation for use. Questions or
comments related to this document should be addressed to Dr. Mary Hutson (email:
hutson.mary@epa.gov) or Ms. Leigh Meyer (email: mever.leigh@epa.gov). U.S. Environmental
Protection Agency, Office of Air Quality Planning and Standards, C504-06, Research Triangle
Park, North Carolina 27711.
1

-------
TABLE OF CONTENTS
1	INTRODUCTION	1-1
1.1	Purpose 	1-2
1.2	Legislative Requirements	1-3
1.3	History of the O3 NAAQS, Reviews and Decisions	1-6
1.4	Review Completed in 2020 	1-12
1.5	Reconsideration of the 2020 O3 NAAQS Decision	1-13
References 	1-16
2	AIR QUALITY	2-1
2.1	O3 and Photochemical Oxidants in the Atmosphere	2-1
2.2	Sources and Emissions of O3 Precursors	2-4
2.3	Ambient Air Monitoring and Data Handling Conventions	2-10
2.3.1	Ambient Air Monitoring Requirements and Monitoring Networks	2-10
2.3.2	Data Handling Conventions and Computations for Determining Whether the
Standards are Met	2-14
2.4	O3 in Ambient Air	2-15
2.4.1	Concentrations Across the U.S	2-15
2.4.2	Trends in U.S. 03 Concentrations	2-16
2.4.3	Diurnal Patterns	2-20
2.4.4	Seasonal Patterns	2-23
2.4.5	Variation in Recent Daily Maximum 1-hour Concentrations	2-25
2.5	Background O3	2-28
2.5.1	Summary of U.S. Background 03 Sources	2-29
2.5.1.1	Stratosphere	2-31
2.5.1.2	Biogenic VOC	2-32
2.5.1.3	Wildland Fires	2-33
2.5.1.4	Lightning Nitrogen Oxides	2-33
2.5.1.5	Natural and Agricultural Soil NOx	2-34
2.5.1.6	Post-Industrial Methane	2-35
2.5.1.7	International Anthropogenic Emissions	2-36
2.5.2	Approach for Quantifying U.S. Background Ozone	2-37
2.5.2.1 Methodology: USB Attribution	2-38
ii

-------
2.5.2.2 Methodology: Strengths, Limitations and Uncertainties	2-40
2.5.3	Estimates of USB and Contributions to USB in 2016	2-42
2.5.3.1	Spatial Characterization of O3 Contributions	2-43
2.5.3.2	Seasonal and Geographic Variations in Ozone Contributions	2-45
2.5.3.3	Ozone Source Contributions as a function of Total Ozone
Concentration	2-52
2.5.3.4	Predicted USB Seasonal Mean and USB on Peak O3 Days	2-58
2.5.4	Summary of USB	2-64
References 	2-68
3 RECONSIDERATION OF THE PRIMARY STANDARD	3-1
3.1	Background on the Current Standard	3-2
3.2	General Approach and Key Issues	3-17
3.3	Health Effects Evidence	3-20
3.3.1	Nature of Effects	3-21
3.3.1.1	Respiratory Effects	3 -22
3.3.1.2	Other Effects	3-27
3.3.2	Public Health Implications and At-risk Populations	3-29
3.3.3	Exposure Concentrations Associated with Effects	3-37
3.3.4	Uncertainties in the Health Effects Evidence	3-47
3.4	Exposure and Risk Information	3-50
3.4.1	Conceptual Model and Assessment Approach	3-51
3.4.2	Population Exposure and Risk Estimates for Air Quality Just Meeting the
Current Standard	3-62
3.4.3	Population Exposure and Risk Estimates for Additional Air Quality
Scenarios	3-68
3.4.4	Key Uncertainties	3-71
3.4.5	Public Health Implications	3-77
3.5	Key Considerations Regarding the Current Primary Standard	3-81
3.5.1	Evidence-based Considerations	3-82
3.5.2	Exposure/risk-based Considerations	3-85
3.5.3	Preliminary Conclusions on the Primary Standard	3-89
3.6	Key Uncertainties and Areas for Future Research	3-102
References 	3-105
111

-------
4 RECONSIDERATION OF THE SECONDARY STANDARD	4-1
4.1	Background on the Current Standard	4-1
4.2	General Approach and Key Issues	4-15
4.3	Welfare Effects Evidence	4-18
4.3.1	Nature of Effects	4-19
4.3.2	Public Welfare Implications	4-26
4.3.3	Exposures Associated with Effects	4-33
4.3.3.1	Growth-related Effects	4-33
4.3.3.2	Visible Foliar Injury	4-41
4.3.3.3	Other Effects 	4-48
4.3.4	Key Uncertainties	4-50
4.4	Exposure and Air Quality Information	4-58
4.4.1	Influence of Form and Averaging Time of Current Standard on W126 Index and
Peak Concentration Metrics	4-62
4.4.2	Environmental Exposures in Terms of W126 Index	4-71
4.4.3	Limitations and Uncertainties	4-75
4.5	Key Considerations Regarding the Current Secondary Standard	4-77
4.5.1	Evidence and Exposure/Risk-based Considerations	4-77
4.5.1.1	Welfare Effects Evidence	4-77
4.5.1.2	General Approach for Considering Public Welfare Protection	4-85
4.5.1.3	Public Welfare Implications of Air Quality under the Current Standard
	4-100
4.5.2	Preliminary Conclusions	4-107
4.6	Key Uncertainties and Areas for Future Research	4-121
References 	4-123
iv

-------
APPENDICES
APPENDIX 2A. ADDITIONAL DETAILS ON DATA ANALYSIS PRESENTED IN PA
SECTION 2.4
APPENDIX 2B. ADDITIONAL DETAILS ON BACKGROUND OZONE MODELING AND
ANALYSIS
APPENDIX 3 A. DETAILS ON CONTROLLED HUMAN EXPOSURE STUDIES
APPENDIX 3B. AIR QUALITY INFORMATION FOR LOCATIONS OF EPIDEMIOLOGIC
STUDIES OF RESPIRATORY EFFECTS
APPENDIX 3C. AIR QUALITY DATA USED IN POPULATION EXPOSURE AND RISK
ANALYSES
APPENDIX 3D. EXPOSURE AND RISK ANALYSIS FOR THE OZONE NAAQS REVIEW
APPENDIX 4A. EXPOSURE-RESPONSE FUNCTIONS FOR 11 TREE SPECIES AND TEN
CROPS
APPENDIX 4B. U.S. DISTRIBUTION OF 11 TREE SPECIES
APPENDIX 4C. VISIBLE FOLIAR INJURY SCORES AT U.S. FOREST SERVICE
BIOSITES (2006-2010)
APPENDIX 4D. ANALYSIS OF THE W126 03 EXPOSURE INDEX AT U.S. AMBIENT
AIR MONITORING SITES
APPENDIX 4E. OZONE WELFARE EFFECTS AND RELATED ECOSYSTEM SERVICES
AND PUBLIC WELFARE ASPECTS
APPENDIX 4F ADDITIONAL ANALYSIS OF OZONE METRICS RELATED TO
CONSIDERATION OF THE SECONDARY STANDARD
v

-------
Table 2-1.
Table 2-2.
Table 2-3.
Table 2-4.
Table 2-5.
Table 2-6.
Table 3-1.
Table 3-2.
Table 3-3.
Table 3-4.
Table 3-5.
Table 3-6.
Table 3-7.
Table 3-8.
Table 4-1.
Table 4-2.
TABLE OF TABLES
Simulation names and descriptions for hemispheric-scale and regional-scale
simulations	2-39
Expressions used to calculate contributions from specific sources	2-40
Predicted USB for U.S. and U.S. regions based on averages for all U.S. grid
cells	2-62
Predicted USB for high elevation locations (>1500 m)	2-63
Predicted USB for locations within 100 km of Mexico or Canada Border	2-63
Predicted USB for low-elevation (<1500 m) that are 100 km or farther from the
border	2-64
National prevalence of asthma, 2017	3-36
Summary of 6.6-hour controlled human exposure study-findings, healthy
adults	3-43
Percent and number of simulated children and children with asthma estimated to
experience at least one or more days per year with a daily maximum 7-hour
average exposure at or above indicated concentration while breathing at an
elevated rate in areas just meeting the current standard	3-65
Percent of simulated children and children with asthma estimated to experience at
least one or more days per year with a lung function decrement at or above 10, 15
or 20% while breathing at an elevated rate in areas just meeting the current
standard	3-68
Percent and number of simulated children and children with asthma estimated
to experience one or more days per year with a daily maximum 7-hour average
exposure at or above indicated concentration while breathing at an elevated
rate - additional air quality scenarios	3-70
Percent of risk estimated for air quality just meeting the current standard in three
study areas using the E-R function approach on days where the daily maximum
7-hour average concentration is below specified values	3-76
Percent of risk estimated for air quality just meeting the current standard in three
study areas using the MSS model approach on days where the daily maximum
7-hour average concentration is below specified values	3-76
Comparison of current assessment and 2014 HREA (all study areas) for percent
of children estimated to experience at least one, or two, days with an exposure
at or above benchmarks while at moderate or greater exertion	3-89
Percent of monitoring sites during the 2018-2020 period with 4th max or W126
metrics at or below various thresholds that have N100 or D100 values above
various thresholds	4-69
Average percent of monitoring sites per year during 2016-2020 with 4th max or
vi

-------
W126 metrics at or below various thresholds that have N100 or D100 values
above various thresholds	4-70
Table 4-3. Distribution of 3-year average seasonal W126 index for sites in Class I areas and
across U.S. that meet the current standards and for those that do not	4-75
TABLE OF FIGURES
Figure 2-1. U.S. O3 precursor emissions by sector: A) NOx; B) CO; C) VOCs; D) CH4	2-6
Figure 2-2. U.S. anthropogenic O3 precursor emission trends for: A) NOx; B) CO;
C) VOCs; and D) CI I,	2-7
Figure 2-3.	U.S. county-level CO emissions density estimates (tons/year/mi2) for 2017	2-9
Figure 2-4.	U.S. county-level NOx emissions density estimates (tons/year/mi2) for 2017. ...2-9
Figure 2-5.	U.S. county-level VOC emissions density estimates (tons/year/mi2) for 2017..2-10
Figure 2-6.	Current O3 monitoring seasons in the U.S	2-12
Figure 2-7. Map of U.S. ambient air O3 monitoring sites reporting data to the EPA during
the 2018-2020 period	2-14
Figure 2-8. O3 design values in ppb for the 2018-2020 period	2-16
Figure 2-9. Trends in O3 design values based on data from 2000-2002 through
2018-2020	2-17
Figure 2-10. National trend in annual 4th highest MDA8 values, 1980 to 2020	2-18
Figure 2-11. National trend in annual 4th highest MDA8 concentrations and O3 design values
in ppb, 2000 to 2020..	2-18
Figure 2-12. Regional trends in median annual 4th highest MDA8 concentrations,
2000 to 2020	2-19
Figure 2-13. Diurnal patterns in hourly O3 concentrations at selected monitoring sites: A) an
urban site in Los Angeles; B) a downwind suburban site in Los Angeles; C) a low
elevation rural site in New Hampshire; and D) a high elevation rural site in New
Hampshire	2-22
Figure 2-14. Seasonal patterns in MDA8 O3 concentrations at selected monitoring sites
(2015-2017): A) an urban site in Baltimore, MD; B) an urban site in Baton
Rouge, LA; C) a rural site in Colorado; and D) a site in Utah experiencing
high wintertime O3	2-24
Figure 2-15. Boxplots showing the distribution of MDA1 concentrations (2018-2020), binned
according to each site's 2018-2020 design value	2-26
Figure 2-16. Number of days in 2018-2020 at each monitoring site with a MDA1
concentration greater than or equal to 120 ppb compared to its 8-hour design
value in ppb	2-26
vii

-------
Figure 2-17. National trend in the annual 2nd highest MDA1 O3 concentration,
2000 to 2020	2-27
Figure 2-18. Conceptual models for 03 sources: (a) in the U.S., and (b) at a single
location	2-30
Figure 2-19. Predicted MDA8 total O3 concentration (top left), Natural (top right),
International (bottom left), and USA (bottom right) contributions in spring
(March, April, May)	2-44
Figure 2-20. Predicted MDA8 total O3 concentration (top left), Natural (top right),
International (bottom left), and USA (bottom right) contributions in summer
(June, July, Aug)	2-45
Figure 2-21. Predicted contribution of International sources as a function of distance from
Mexico/Canada (left) and at "interior" locations (excluding border areas) by
elevation (right)	2-47
Figure 2-22. Grid cell assignments to East, West, High Elevation, Near Border, and Near and
High (i.e., both High Elevation and Near Border)	2-48
Figure 2-23. Annual time series of regional average predicted MDA8 total O3 concentration
and contributions of each source (see legend) for the West (top), and the East
(bottom)	2-49
Figure 2-24. Annual time series of regional urban area-weighted average predicted MDA8
total O3 concentration and contributions of each source (see legend) for the
High-elevation West (top), near-border West (middle), and Low/Interior West
(bottom)	2-51
Figure 2-25. Predicted contribution of Natural as a function of predicted total (Base)
MDA8 O3 concentration in the West and East	2-53
Figure 2-26. Predicted contribution of International as a function of predicted total (Base)
MDA8 O3 concentration in the West and East	2-54
Figure 2-27. Predicted contribution of USA as a function of predicted total (Base) MDA8
O3 concentration in the West and East. Sloped lines show percent contribution
as a quick reference	2-54
Figure 2-28. Annual time series of regional average predicted MDA8 O3 and contributions
of each source to predicted MDA8 total O3 (see legend) in the West (top) and
East (bottom) including only those grid-cell days with MDA8 greater than
70 ppb	2-56
Figure 2-29. Annual time series of regional average predicted MDA8 O3 and contributions of
each source to predicted MDA8 O3 (see legend) in the high-elevation West (top),
in the near-border West (middle), and in the Low/Interior West weighted toward
urban areas (bottom) including only those grid-cell days with MDA8 O3 greater
than 70 ppb	2-57
Figure 2-30. Map of predicted USB contributions by O3 season for spring average (top left),
summer average (top right), top 10 predicted total O3 days (center left),
viii

-------
4th highest total O3 simulated day (center right), and all days with total O3
greater than 70 ppb (bottom left), along with a map of the number of days with
total O3 above 70 ppb (bottom right)	2-60
Figure 3-1. Overview of general approach for the primary O3 standard	3-19
Figure 3-2. Group mean 03-induced reduction in FEV1 from controlled human exposure
studies of healthy adults exposed for 6.6 hours with quasi-continuous
exercise	3-39
Figure 3-3.	Conceptual model for exposure-based risk assessment	3-52
Figure 3-4.	Analysis approach for exposure-based risk analyses	3-53
Figure 4-1.	Overview of general approach for the secondary O3 standard	4-17
Figure 4-2.	Potential effects of O3 on the public welfare	4-32
Figure 4-3.	Established RBL functions for seedlings of 11 tree species	4-39
Figure 4-4.	Established RYL functions for 10 crops	4-40
Figure 4-5. Distribution of nonzero BI scores at USFS biosites (normal soil moisture)
grouped by assigned W126 index estimates	4-47
Figure 4-6. W126 index at monitoring sites with valid design values (2018-2020
average)	4-60
Figure 4-7. N100 values at monitoring sites with valid design values (2018-2020
average)	 4-61
Figure 4-8. D100 values at monitoring sites with valid design values (2018-2020
average) 	4-61
Figure 4-9. Relationship between the W126 index and design values for the current
standard (2018-2020). The W126 is analyzed in terms of averages across the 3-
year design value period (left) and annual values (right)	4-63
Figure 4-10. Relationship between trends in the W126 index and trends in design values
across a 21-year period (2000-2020) at U.S. monitoring sites. W126 is analyzed in
terms of averages across 3-year design value periods (left) and annual values
(right)	4-65
Figure 4-11. Distributions of MDA1 concentrations for the three design value periods in 2000-
2004 (red) and 2016-2020 (blue), binned by the design value at each monitoring
site. Boxes represent the 25th, 50th, and 75th perentiles; whiskers represent the 1st
and 99th percentiles; and circles are outlier values	4-67
Figure 4-12. Distributions of N100 (top panels) and D100 (bottom panels) values at monitoring
sites differing by design values (left panels) and W126 index values (right panels)
based on the 2018-2020 monitoring data. The boxes represent the 25th, 50th, and
75th percentiles and the whiskers extend to the 1st and 99th	4-68
Figure 4-13. Analytical approach for characterizing vegetation exposure with W126 index. 4-72
IX

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
1 INTRODUCTION
This document, Policy Assessment for the Reconsideration of the Ozone National
Ambient Air Quality Standards, External Review Draft (hereafter referred to as the draft PA),
presents the draft policy assessment for the U.S. Environmental Protection Agency's (EPA's)
reconsideration of the decision reached in the review of the ozone (O3) national ambient air
quality standards (NAAQS) completed in 2020,K 2 This draft PA considers the key policy-
relevant issues, drawing on those identified in the Integrated Review Plan for the Ozone National
Ambient Air Quality Standards (IRP; [U.S. EPA, 2019]) in light of the available evidence
assessed in the Integrated Science Assessment for Ozone and Related Photochemical Oxidants
(ISA [U.S. EPA, 2020a]) and quantitative air quality, exposure and risk analyses based on that
evidence, including any analyses updated for this reconsideration. Thus, this document will
reassess the policy implications of the scientific evidence described in the 2020 ISA and related air
quality, exposure and risk analyses. Accordingly, this document draws heavily on information
presented in the 2020 PA (U.S. EPA, 2020b), with some updates to include more recent air quality
information.
This document is organized into four chapters. Chapter 1 presents introductory
information on the purpose of the PA in the context of NAAQS reviews, legislative requirements
for NAAQS reviews, an overview of the history of the O3 NAAQS, including background
information on prior reviews, and a summary of the process for this reconsideration. Chapter 2
provides an overview of how photochemical oxidants, including O3, are formed in the
atmosphere, along with updated information on sources and emissions of important precursor
chemicals, as well as updated ambient air monitoring data. Chapter 2 also summarizes key
aspects of the ambient air monitoring requirements, and O3 air quality, including model-based
estimates of O3 resulting from natural sources and anthropogenic sources outside the U.S.
Chapters 3 focuses on policy-relevant aspects of the health effects evidence (as presented in the
2020 ISA) and exposure/risk information, identifying and summarizing key considerations
related to review of the primary (health-based) standard. Similarly, Chapter 4 focuses on policy -
1	The scope for this reconsideration, as for the 2020 decision on the O3 NAAQS, focuses on the presence in ambient
air of photochemical oxidants, a group of gaseous compounds of which ozone (the indicator for the current
standards) is the most prevalent in the atmosphere and the one for which there is a very large, well-established
evidence base of its health and welfare effects. The ozone standards that were established in 2015 (80 FR 65292,
October 26, 2015) and retained in 2020 (85 FR 87256, December 31, 2020), are referred to in this document as
the "current" or "existing" standards.
2	On October 29, 2021, the Agency announced its decision to reconsider the 2020 O3 NAAQS final action. This
announcement is available at https://www.epa.gov/ground-level-ozone-pollution/epa-reconsider-previous-
administrations-decision-retain-2015 -ozone.
April 2022
1-1 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
relevant aspects of the welfare effects evidence (as presented in the 2020 ISA) and air quality,
exposure and risk information, identifying and summarizing key considerations related to review
of the secondary (welfare -based) standard.
1.1 PURPOSE
Generally in each NAAQS review, the PA, when final, presents an evaluation, for
consideration by the EPA Administrator, of the policy implications of the available scientific
information, assessed in the ISA, any quantitative air quality, exposure or risk analyses based on
the ISA findings, and related limitations and uncertainties. Ultimately, a final decision on the
NAAQS will reflect the judgments of the Administrator. The role of the PA is to help "bridge the
gap" between the Agency's scientific assessment and quantitative technical analyses, and the
judgments required of the Administrator in determining whether it is appropriate to retain or
revise the NAAQS.
In evaluating the question of adequacy of the current standards and whether it may be
appropriate to consider alternative standards, the PA focuses on information that is most
pertinent to evaluating the standards and their basic elements: indicator, averaging time, form,
and level.3 These elements, which together serve to define each standard, must be considered
collectively in evaluating the public health and public welfare protection the standards afford.
The development of the PA is also intended to facilitate advice to the Agency and
recommendations to the Administrator from an independent scientific review committee, the
Clean Air Scientific Advisory Committee (CASAC), as provided for in the Clean Air Act
(CAA). The EPA generally makes available to the CASAC and the public one or more drafts of
the PA for CASAC review and public comment. As discussed below in section 1.2, the CASAC
is to advise on subjects including the Agency's assessment of the relevant scientific information
and on the adequacy of the current standards, and to make recommendations as to any revisions
of the standards that may be appropriate. In its review of the draft PA, the CASAC also conveys
its advice on the standards.
In this draft PA for the reconsideration of the December 2020 O3 NAAQS decision, we4
take into account the scientific evidence, as characterized in the 2020 ISA and the additional
3	The indicator defines the chemical species or mixture to be measured in the ambient air for the purpose of
determining whether an area attains the standard. The averaging time defines the period over which air quality
measurements are to be averaged or otherwise analyzed. The form of a standard defines the air quality statistic
that is to be compared to the level of the standard in determining whether an area attains the standard. For
example, the form of the annual NAAQS for fine particulate matter is the average of annual mean concentrations
for three consecutive years, while the form of the 8-hour NAAQS for carbon monoxide is the second-highest 8-
hour average in a year. The level of the standard defines the air quality concentration used for that purpose.
4	The terms "staff," "we" and "our" throughout this document refer to the staff in the EPA's Office of Air Quality
Planning and Standards (OAQPS).
April 2022
1-2 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
policy-relevant quantitative air quality, exposure and risk analyses described herein. Advice and
comments from the CASAC and the public on this draft PA will inform the final evaluation and
conclusions in the final PA.
The final PA is designed to assist the Administrator in considering the available scientific
and risk information and formulating judgments regarding the standards. Accordingly, the final
PA will inform the Administrator's decision in this reconsideration. Beyond informing the
Administrator and facilitating the advice and recommendations of the CASAC, the final PA is
also intended to be a useful reference to all interested parties. In these roles, it is intended to
serve as a source of policy-relevant information that supports the Agency's reconsideration of
the 2020 O3 NAAQS decision, and it is written to be understandable to a broad audience.
1.2 LEGISLATIVE REQUIREMENTS
Two sections of the CAA govern the establishment and revision of the NAAQS. Section
108 (42 U.S.C. 7408) directs the Administrator to identify and list certain air pollutants and then
to issue air quality criteria for those pollutants. The Administrator is to list those pollutants
"emissions of which, in his judgment, cause or contribute to air pollution which may reasonably
be anticipated to endanger public health or welfare"; "the presence of which in the ambient air
results from numerous or diverse mobile or stationary sources"; and for which he "plans to issue
air quality criteria...." (42 U.S.C. ง 7408(a)(1)). Air quality criteria are intended to "accurately
reflect the latest scientific knowledge useful in indicating the kind and extent of all identifiable
effects on public health or welfare which may be expected from the presence of [a] pollutant in
the ambient air...." (42 U.S.C. ง 7408(a)(2)).
Section 109 [42 U.S.C. 7409] directs the Administrator to propose and promulgate
"primary" and "secondary" NAAQS for pollutants for which air quality criteria are issued [42
U.S.C. ง 7409(a)], Section 109(b)(1) defines primary standards as ones "the attainment and
maintenance of which in the judgment of the Administrator, based on such criteria and allowing
an adequate margin of safety, are requisite to protect the public health."5 Under section
109(b)(2), a secondary standard must "specify a level of air quality the attainment and
maintenance of which, in the judgment of the Administrator, based on such criteria, is requisite
to protect the public welfare from any known or anticipated adverse effects associated with the
presence of [the] pollutant in the ambient air."6
5	The legislative history of section 109 indicates that a primary standard is to be set at "the maximum permissible
ambient air level.. . which will protect the health of any [sensitive] group of the population," and that for this
purpose "reference should be made to a representative sample of persons comprising the sensitive group rather
than to a single person in such a group." S. Rep. No. 91-1196, 91st Cong., 2d Sess. 10 (1970).
6	Under CAA section 302(h) (42 U.S.C. ง 7602(h)), effects on welfare include, but are not limited to, "effects on
soils, water, crops, vegetation, manmade materials, animals, wildlife, weather, visibility, and climate, damage to
April 2022
1-3 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
In setting primary and secondary standards that are "requisite" to protect public health
and welfare, respectively, as provided in section 109(b), the EPA's task is to establish standards
that are neither more nor less stringent than necessary. In so doing, the EPA may not consider the
costs of implementing the standards. See generally, Whitman v. American Trucking Ass 'ns, 531
U.S. 457, 465-472, 475-76 (2001). Likewise, "[ajttainability and technological feasibility are not
relevant considerations in the promulgation of national ambient air quality standards" (American
Petroleum Institute v. Costle, 665 F.2d 1176, 1185 [D.C. Cir. 1981], cert, denied, 455 U.S. 1034
[1982]; accord Murray Energy Corp. v. EPA, 936 F.3d 597, 623-24 [D.C. Cir. 2019]). At the
same time, courts have clarified the EPA may consider "relative proximity to peak background
... concentrations" as a factor in deciding how to revise the NAAQS in the context of
considering standard levels within the range of reasonable values supported by the air quality
criteria and judgments of the Administrator (American Trucking Ass'ns, v. EPA, 283 F.3d 355,
379 [D.C. Cir. 2002], hereafter referred to as "ATA IIF).
The requirement that primary standards provide an adequate margin of safety was
intended to address uncertainties associated with inconclusive scientific and technical
information available at the time of standard setting. It was also intended to provide a reasonable
degree of protection against hazards that research has not yet identified. See Lead Industries
Ass'n v. EPA, 647 F.2d 1130, 1154 (D.C. Cir 1980), cert, denied, 449 U.S. 1042 (1980);
American Petroleum Institute v. Costle, 665 F.2d at 1186; Coalition of Battery Recyclers Ass 'n v.
EPA, 604 F.3d 613, 617-18 (D.C. Cir. 2010); Mississippi v. EPA, 744 F.3d 1334, 1353 (D.C. Cir.
2013). Both kinds of uncertainties are components of the risk associated with pollution at levels
below those at which human health effects can be said to occur with reasonable scientific
certainty. Thus, in selecting primary standards that include an adequate margin of safety, the
Administrator is seeking not only to prevent pollution levels that have been demonstrated to be
harmful but also to prevent lower pollutant levels that may pose an unacceptable risk of harm,
even if the risk is not precisely identified as to nature or degree. The CAA does not require the
Administrator to establish a primary NAAQS at a zero-risk level or at background concentration
levels (see Lead Industries v. EPA, 647 F.2d at 1156 n.51, Mississippi v. EPA, 744 F.3d at 1351),
but rather at a level that reduces risk sufficiently so as to protect public health with an adequate
margin of safety.
In addressing the requirement for an adequate margin of safety, the EPA considers such
factors as the nature and severity of the health effects involved, the size of the sensitive
population(s), and the kind and degree of uncertainties. The selection of any particular approach
and deterioration of property, and hazards to transportation, as well as effects on economic values and on personal
comfort and well-being."
April 2022
1-4 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
to providing an adequate margin of safety is a policy choice left specifically to the
Administrator's judgment. See Lead Industries Ass 'n v. EPA, 647 F.2d at 1161-62; Mississippi v.
EPA, 744 F.3d at 1353.
Section 109(d)(1) of the Act requires periodic review and, if appropriate, revision of
existing air quality criteria to reflect advances in scientific knowledge on the effects of the
pollutant on public health and welfare. Under the same provision, the EPA is also to periodically
review and, if appropriate, revise the NAAQS, based on the revised air quality criteria.7
Section 109(d)(2) addresses the appointment and advisory functions of an independent
scientific review committee. Section 109(d)(2)(A) requires the Administrator to appoint this
committee, which is to be composed of "seven members including at least one member of the
National Academy of Sciences, one physician, and one person representing State air pollution
control agencies." Section 109(d)(2)(B) provides that the independent scientific review
committee "shall complete a review of the criteria.. .and the national primary and secondary
ambient air quality standards...and shall recommend to the Administrator any new... standards
and revisions of existing criteria and standards as may be appropriate...." Since the early 1980s,
this independent review function has been performed by the CAS AC of the EPA's Science
Advisory Board. A number of other advisory functions are also identified for the committee by
section 109(d)(2)(C), which reads:
Such committee shall also (i) advise the Administrator of areas in which
additional knowledge is required to appraise the adequacy and basis of existing,
new, or revised national ambient air quality standards, (ii) describe the research
efforts necessary to provide the required information, (iii) advise the
Administrator on the relative contribution to air pollution concentrations of
natural as well as anthropogenic activity, and (iv) advise the Administrator of any
adverse public health, welfare, social, economic, or energy effects which may
result from various strategies for attainment and maintenance of such national
ambient air quality standards.
As previously noted, the Supreme Court has held that section 109(b) "unambiguously bars cost
considerations from the NAAQS-setting process" (Whitman v. American Trucking Ass'ns, 531
U.S. 457, 471 [2001]). Accordingly, while some of the issues listed in section 109(d)(2)(C) as
those on which Congress has directed the CASAC to advise the Administrator are ones that are
relevant to the standard setting process, others are not. Issues that are not relevant to standard
setting may be relevant to implementation of the NAAQS once they are established.8
7	This section of the Act requires the Administrator to complete these reviews and make any revisions that may be
appropriate "at five-year intervals."
8	Because some of these issues are not relevant to standard setting, some aspects of CASAC advice may not be
relevant to EPA's process of setting primary and secondary standards that are requisite to protect public health
and welfare. Indeed, were the EPA to consider costs of implementation when reviewing and revising the
April 2022
1-5 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
1.3 HISTORY OF THE O3 NAAQS, REVIEWS AND DECISIONS
Primary and secondary NAAQS were first established for photochemical oxidants in
1971 (36 FR 8186, April 30, 1971) based on the air quality criteria developed in 1970 (U.S.
DHEW, 1970; 35 FR 4768, March 19, 1970). The EPA set both primary and secondary standards
at 0.08 parts per million (ppm), as a 1-hour average of total photochemical oxidants, not to be
exceeded more than one hour per year based on the scientific information in the 1970 air quality
criteria document (AQCD). Since that time, the EPA has reviewed the air quality criteria and
standards a number of times, with the most recent review being completed in 2020.
The EPA initiated the first periodic review of the NAAQS for photochemical oxidants in
1977. Based on the 1978 AQCD (U.S. EPA, 1978), the EPA published proposed revisions to the
original NAAQS in 1978 (43 FR 26962, June 22, 1978) and final revisions in 1979 (44 FR 8202,
February 8, 1979). At that time, the EPA changed the indicator from photochemical oxidants to
O3, revised the level of the primary and secondary standards from 0.08 to 0.12 ppm and revised
the form of both standards from a deterministic (i.e., not to be exceeded more than one hour per
year) to a statistical form. With these changes, attainment of the standards was defined to occur
when the average number of days per calendar year (across a 3-year period) with maximum
hourly average O3 concentration greater than 0.12 ppm equaled one or less (44 FR 8202,
February 8, 1979; 43 FR 26962, June 22, 1978).
Following the EPA's decision in the 1979 review, several petitioners sought judicial
review. Among those, the city of Houston challenged the Administrator's decision arguing that
the standard was arbitrary and capricious because natural O3 concentrations and other physical
phenomena in the Houston area made the standard unattainable in that area. The U.S. Court of
Appeals for the District of Columbia Circuit (D.C. Circuit) rejected this argument, holding (as
noted in section 1.1 above) that attainability and technological feasibility are not relevant
considerations in the promulgation of the NAAQS (American Petroleum Institute v. Costle, 665
F.2d at 1185). The court also noted that the EPA need not tailor the NAAQS to fit each region or
locale, pointing out that Congress was aware of the difficulty in meeting standards in some
locations and had addressed this difficulty through various compliance related provisions in the
CAA (id. at 1184-86).
standards "it would be grounds for vacating the NAAQS" (Whitman v. American Trucking Ass 'ns, 531 U.S. 457,
471 n.4 [2001]). At the same time, the CAA directs CASAC to provide advice on "any adverse public health,
welfare, social, economic, or energy effects which may result from various strategies for attainment and
maintenance" of the NAAQS to the Administrator under section 109(d)(2)(C)(iv). In Whitman, the Court
clarified that most of that advice would be relevant to implementation but not standard setting, as it "enable [s] the
Administrator to assist the States in carrying out their statutory role as primary implementers of the NAAQS" (id.
at 470 [emphasis in original]). However, the Court also noted that CASAC's "advice concerning certain aspects
of 'adverse public health ... effects' from various attainment strategies is unquestionably pertinent" to the
NAAQS rulemaking record and relevant to the standard setting process (id. at 470 n.2).
April 2022
1-6 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
The next periodic reviews of the criteria and standards for O3 and other photochemical
oxidants began in 1982 and 1983, respectively (47 FR 11561, March 17, 1982; 48 FR 38009,
August 22, 1983). The EPA subsequently published the 1986 AQCD (U.S. EPA, 1986) and the
1989 Staff Paper (U.S. EPA, 1989). Following publication of the 1986 AQCD, a number of
scientific abstracts and articles were published that appeared to be of sufficient importance
concerning potential health and welfare effects of O3 to warrant preparation of a supplement to
the 1986 AQCD (U.S. EPA, 1992). In August of 1992, the EPA proposed to retain the existing
primary and secondary standards based on the health and welfare effects information contained
in the 1986 AQCD and its 1992 Supplement (57 FR 35542, August 10, 1992). In March 1993,
the EPA announced its decision to conclude this review by affirming its proposed decision to
retain the standards, without revision (58 FR 13008, March 9, 1993).
In the 1992 notice of its proposed decision in that review, the EPA announced its
intention to proceed as rapidly as possible with the next review of the air quality criteria and
standards for O3 and other photochemical oxidants in light of emerging evidence of health effects
related to 6- to 8-hour O3 exposures (57 FR 35542, August 10, 1992). The EPA subsequently
published the AQCD and Staff Paper for that next review (U.S. EPA, 1996). In December 1996,
the EPA proposed revisions to both the primary and secondary standards (61 FR 65716,
December 13, 1996). With regard to the primary standard, the EPA proposed to replace the then-
existing 1-hour primary standard with an 8-hour standard set at a level of 0.08 ppm (equivalent
to 0.084 ppm based on the proposed data handling convention) as a 3-year average of the annual
third-highest daily maximum 8-hour concentration. The EPA proposed to revise the secondary
standard either by setting it identical to the proposed new primary standard or by setting it as a
new seasonal standard using a cumulative form. The EPA completed this review in 1997 by
setting the primary standard at a level of 0.08 ppm, based on the annual fourth-highest daily
maximum 8-hour average concentration, averaged over three years, and setting the secondary
standard identical to the revised primary standard (62 FR 38856, July 18, 1997).
On May 14, 1999, in response to challenges by industry and others to the EPA's 1997
decision, the D.C. Circuit remanded the O3 NAAQS to the EPA, finding that section 109 of the
CAA, as interpreted by the EPA, effected an unconstitutional delegation of legislative authority
(American Trucking Ass 'ns v. EPA, 175 F.3d 1027, 1034-1040 [D.C. Cir. 1999]). In addition, the
court directed that, in responding to the remand, the EPA should consider the potential beneficial
health effects of O3 pollution in shielding the public from the effects of solar ultraviolet (UV)
radiation, as well as adverse health effects {id. at 1051-53). In 1999, the EPA sought panel
rehearing and for rehearing en banc on several issues related to that decision. The court granted
the request for panel rehearing in part and denied it in part but declined to review its ruling with
regard to the potential beneficial effects of O3 pollution {American Trucking Ass'ns v. EPA, 195
April 2022
1-7 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
F.3d 4, 10 [D.C Cir., 1999]). On January 27, 2000, the EPA petitioned the U.S. Supreme Court
for certiorari on the constitutional issue (and two other issues) but did not request review of the
ruling regarding the potential beneficial health effects of O3. On February 27, 2001, the U.S.
Supreme Court unanimously reversed the judgment of the D.C. Circuit on the constitutional
issue (Whitman v. American Trucking Ass 'ns, 531 U.S. 457, 472-74 [2001], [holding that section
109 of the CAA does not delegate legislative power to the EPA in contravention of the
Constitution]). The Court remanded the case to the D.C. Circuit to consider challenges to the O3
NAAQS that had not been addressed by that court's earlier decisions. On March 26, 2002, the
D.C. Circuit issued its final decision on the remand, finding the 1997 O3 NAAQS to be "neither
arbitrary nor capricious," and so denying the remaining petitions for review. Sqq ATA III, 283
F.3d at 379.
Specifically, in ATA III, the D.C. Circuit upheld the EPA's decision on the 1997 O3
standard as the product of reasoned decision making. With regard to the primary standard, the
court made clear that the most important support for the EPA's decision to revise the standard
was the health evidence of insufficient protection afforded by the then-existing standard ("the
record [is] replete with references to studies demonstrating the inadequacies of the old one-hour
standard"), as well as extensive information supporting the change to an 8-hour averaging time
{id. at 378). The court further upheld the EPA's decision not to select a more stringent level for
the primary standard noting "the absence of any [emphasis in original] human clinical studies at
ozone concentrations below 0.08 [ppm]" which supported the EPA's conclusion that "the most
serious health effects of ozone are 'less certain' at low concentrations, providing an eminently
rational reason to set the primary standard at a somewhat higher level, at least until additional
studies become available" {id. at 379, internal citations omitted). The court also pointed to the
significant weight that the EPA properly placed on the advice it received from the CASAC {id. at
379). In addition, the court noted that "although relative proximity to peak background ozone
concentrations did not, in itself, necessitate a level of 0.08 [ppm], EPA could consider that factor
when choosing among the three alternative levels" {id. at 379).
Coincident with the continued litigation of the other issues, the EPA responded to the
court's 1999 remand to consider the potential beneficial health effects of O3 pollution in
shielding the public from effects of UV radiation (66 FR 57268, Nov. 14, 2001; 68 FR 614,
January 6, 2003). The EPA provisionally determined that the information linking changes in
patterns of ground-level O3 concentrations to changes in relevant patterns of exposures to UV
radiation of concern (UV-B) to public health was too uncertain, at that time, to warrant any
relaxation in 1997 O3 NAAQS. The EPA also expressed the view that any plausible changes in
UV-B radiation exposures from changes in patterns of ground-level O3 concentrations would
likely be very small from a public health perspective. In view of these findings, the EPA
April 2022
1-8 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
proposed to leave the 1997 primary standard unchanged (66 FR 57268, Nov. 14, 2001). After
considering public comment on the proposed decision, the EPA published its final response to
this remand in 2003, re-affirming the 8-hour primary standard set in 1997 (68 FR 614, January 6,
2003).
The EPA initiated the fourth periodic review of the air quality criteria and standards for
O3 and other photochemical oxidants with a call for information in September 2000 (65 FR
57810, September 26, 2000). In 2007, the EPA proposed to revise the level of the primary
standard within a range of 0.075 to 0.070 ppm (72 FR 37818, July 11, 2007). The EPA proposed
to revise the secondary standard either by setting it identical to the proposed new primary
standard or by setting it as a new seasonal standard using a cumulative form. Documents
supporting these proposed decisions included the 2006 AQCD (U.S. EPA, 2006) and 2007 Staff
Paper (U.S. EPA, 2007) and related technical support documents. The EPA completed the
review in March 2008 by revising the levels of both the primary and secondary standards from
0.08 ppm to 0.075 ppm while retaining the other elements of the prior standards (73 FR 16436,
March 27, 2008).
In May 2008, state, public health, environmental, and industry petitioners filed suit
challenging the EPA's final decision on the 2008 O3 standards. On September 16, 2009, the EPA
announced its intention to reconsider the 2008 O3 standards,9 and initiated a rulemaking to do so.
At the EPA's request, the court held the consolidated cases in abeyance pending the EPA's
reconsideration of the 2008 decision.
In January 2010, the EPA issued a notice of proposed rulemaking to reconsider the 2008
final decision (75 FR 2938, January 19, 2010). In that notice, the EPA proposed that further
revisions of the primary and secondary standards were necessary to provide a requisite level of
protection to public health and welfare. The EPA proposed to revise the level of the primary
standard from 0.075 ppm to a level within the range of 0.060 to 0.070 ppm, and to revise the
secondary standard to one with a cumulative, seasonal form. At the EPA's request, the CAS AC
reviewed the proposed rule at a public teleconference on January 25, 2010 and provided
additional advice in early 2011 (Samet, 2010, Samet, 2011). Later that year, in view of the need
for further consideration and the fact that the Agency's next periodic review of the O3NAAQS
required under CAA section 109 had already begun (as announced on September 29, 2008),10 the
EPA decided to consolidate the reconsideration with its statutorily required periodic review.11
9	The press release of this announcement is available at:
https.V/archive. epa.gov/epapages/newsroom_archive/newsreleases/85J90b 7711acb 0c88525763300617d0d. html.
10	The Call for Information initiating the new review was announced in the Federal Register (73 FR 56581,
September 29, 2008).
11	This rulemaking, completed in 2015, concluded the reconsideration process.
April 2022
1-9 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
In light of the EPA's decision to consolidate the reconsideration with the ongoing
periodic review, the D.C. Circuit proceeded with the litigation on the 2008 O3 NAAQS decision.
On July 23, 2013, the court upheld the EPA's 2008 primary standard, but remanded the 2008
secondary standard to the EPA (Mississippi v. EPA, 744 F.3d 1334 [D.C. Cir. 2013]). With
respect to the primary standard, the court first rejected arguments that the EPA should not have
lowered the level of the existing primary standard, holding that the EPA reasonably determined
that the existing primary standard was not requisite to protect public health with an adequate
margin of safety, and consequently required revision. The court went on to reject arguments that
the EPA should have adopted a more stringent primary standard. With respect to the secondary
standard, the court held that the EPA's explanation for the setting of the secondary standard
identical to the revised 8-hour primary standard was inadequate under the CAA because the EPA
had not adequately explained how that standard provided the required public welfare protection.
At the time of the court's decision, the EPA had already completed significant portions of
its next statutorily required periodic review of the O3 NAAQS. This review had been formally
initiated in 2008 with a call for information in the Federal Register (73 FR 56581, September 29,
2008). In late 2014, based on the ISA, Risk and Exposure Assessments (REAs) for health and
welfare, and PA12 developed for this review, the EPA proposed to revise the 2008 primary and
secondary standards by reducing the level of both standards to within the range of 0.070 to 0.065
ppm (79 FR 75234, December 17, 2014).
The EPA's final decision in this review was published in October 2015, establishing the
now-current standards (80 FR 65292, October 26, 2015). In this decision, based on consideration
of the health effects evidence on respiratory effects of O3 in at-risk populations, the EPA revised
the primary standard from a level of 0.075 ppm to a level of 0.070 ppm, while retaining all the
other elements of the standard (80 FR 65292, October 26, 2015). The EPA's decision on the
level for the standard was based on the weight of the scientific evidence and quantitative
exposure/risk information. The level of the secondary standard was also revised from 0.075 ppm
to 0.070 ppm based on the scientific evidence of O3 effects on welfare, particularly the evidence
of O3 impacts on vegetation, and quantitative analyses available in the review.13 The other
elements of the standard were retained. This decision on the secondary standard also
incorporated the EPA's response to the D.C. Circuit's remand of the 2008 secondary standard in
Mississippi v. EPA, 744 F.3d 1344 (D.C. Cir. 2013). The 2015 revisions to the NAAQS were
12	The final versions of these documents, released in August 2014, were developed with consideration of the
comments and recommendations from the CASAC, as well as comments from the public on the draft documents
(Frey, 2014a, Frey, 2014b, Frey, 2014c, U.S. EPA, 2014a, U.S. EPA, 2014b, U.S. EPA, 2014c).
13	These standards, set in 2015, are specified at 40 CFR 50.19.
April 2022
1-10 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
accompanied by revisions to the data handling procedures, and the ambient air monitoring
requirements14 (80 FR 65292, October 26, 2015).15
After publication of the final rule, a number of industry groups, environmental and health
organizations, and certain states filed petitions for judicial review in the D.C. Circuit. The
industry and state petitioners argued that the revised standards were too stringent, while the
environmental and health petitioners argued that the revised standards were not stringent enough
to protect public health and welfare as the Act requires. On August 23, 2019, the court issued an
opinion that denied all the petitions for review with respect to the 2015 primary standard while
also concluding that the EPA had not provided a sufficient rationale for aspects of its decision on
the 2015 secondary standard and remanding that standard to the EPA {Murray Energy Corp. v.
EPA, 936 F.3d 597 [D.C. Cir. 2019]).
In the August 2019 decision, the court additionally addressed arguments regarding
considerations of background O3 concentrations, and socioeconomic and energy impacts. With
regard to the former, the court rejected the argument that the EPA was required to take
background O3 concentrations into account when setting the NAAQS, holding that the text of
CAA section 109(b) precluded this interpretation because it would mean that if background O3
levels in any part of the country exceeded the level of O3 that is requisite to protect public health,
the EPA would be obliged to set the standard at the higher nonprotective level {id. at 622-23).
Thus, the court concluded that the EPA did not act unlawfully or arbitrarily or capriciously in
setting the 2015 NAAQS without regard for background O3 {id. at 624). Additionally, the court
denied arguments that the EPA was required to consider adverse economic, social, and energy
impacts in determining whether a revision of the NAAQS was "appropriate" under section
109(d)(1) of the CAA {id. at 621-22). The court reasoned that consideration of such impacts was
precluded by Whitman's holding that the CAA "unambiguously bars cost considerations from the
NAAQS-setting process" (531 U.S. at 471, summarized in section 1.2 above). Further, the court
explained that section 109(d)(2)(C)'s requirement that CAS AC advise the EPA "of any adverse
public health, welfare, social, economic, or energy effects which may result from various
strategies for attainment and maintenance" of revised NAAQS had no bearing on whether costs
are to be considered in setting the NAAQS (Murray Energy Corp. v. EPA, 936 F.3d at 622).
14	The current federal regulatory measurement methods for O3 are specified in 40 CFR 50, Appendix D and 40 CFR
part 53. Consideration of ambient air measurements with regard to judging attainment of the standards set in
2015 is specified in 40 CFR 50, Appendix U. The O3 monitoring network requirements are specified in 40 CFR
58.
15	This decision additionally announced revisions to the exceptional events scheduling provisions, as well as changes
to the air quality index and the regulations for the prevention of significant deterioration permitting program.
April 2022
1-11 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
Rather, as described in Whitman and discussed further in section 1.2 above, most of that advice
would be relevant to implementation but not standard setting {id).
1.4 REVIEW COMPLETED IN 2020
The EPA announced its initiation of the next periodic review of the air quality criteria for
photochemical oxidants and the O3 NAAQS in June 2018, issuing a call for information in the
Federal Register (83 FR 29785, June 26, 2018). Two types of information were called for:
information regarding significant new O3 research to be considered for the ISA for the review,
and policy-relevant issues for consideration in this NAAQS review. Based in part on the
information received in response to the call for information, the EPA developed a draft IRP
which was made available for consultation with the CASAC and for public comment (83 FR
55163, November 2, 2018; 83 FR 55528, November 6, 2018). Comments from the CASAC
(Cox, 2018) and the public were considered in preparing the final IRP (U.S. EPA, 2019).
Under the plan outlined in the IRP and consistent with revisions to the process identified
by the Administrator in his 2018 memo directing initiation of the review and completion within
the statutorily required timeframe, the O3 NAAQS review completed in 2020 progressed on an
accelerated schedule (Pruitt, 2018).16 The EPA incorporated a number of changes in various
aspects of the review process, as summarized in the IRP, to support completion within the
required period (Pruitt, 2018). For example, rather than produce separate documents for the PA
and associated quantitative analyses, the human exposure and health risk analyses (that inform
the decision on the primary standard) and the air quality and exposure analyses (that inform the
decision on the secondary standard) were included in full as appendices in the PA, along with a
number of other technical appendices.
Drafts of the ISA and PA (including the associated quantitative and exposure analyses)
were reviewed by the CASAC and made available for public comment (84 FR 50836, September
26, 2019; 84 FR 58711, November 1, 2019).17 In a divergence from recent past practice, an O3
panel was not assembled to assist the CASAC in its review. Rather, the CASAC was assisted in
its review by a pool of consultants with expertise in a number of fields (84 FR 38625, August 7,
2019).18 The approach employed by the CASAC in utilizing outside technical expertise
16	The Administrator's May 2018 direction to initiate this review of the O3 NAAQS included further direction to the
EPA staff to expedite the review, implementing an accelerated schedule aimed at completion of the review within
the statutorily required period (Pruitt, 2018).
17	The draft ISA and draft PA were released for public comment and CASAC review on September 26, 2019 and
October 31, 2019, respectively. The charges for the CASAC review summarized the overarching context for the
document review (including reference to Pruitt [2018], and the CASAC's role under section 109(d)(2)(C) of the
Act), as well as specific charge questions for review of each of the documents.
18	Rather than join with some or all of the CASAC members in a pollutant specific review panel as had been
common in previous NAAQS reviews, the consultants comprised a pool of expertise that CASAC members drew
April 2022
1-12 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
represented an additional modification of the process from past reviews. The CASAC discussed
its draft letters describing its advice and comments on the documents in a public teleconference
in early February 2020 (85 FR 4656; January 27, 2020). The letters to the Administrator
conveying the CASAC advice and comments on the draft PA and draft ISA were released later
that month (Cox, 2020a, Cox, 2020b). Comments from the CASAC and the public on the draft
ISA were considered by the EPA and led to a number of revisions in developing the final
document (ISA, Appendix 10, section 10.4.5). The ISA was completed and made available to the
public in April 2020 (85 FR 21849, April 20, 2020). The comments from CASAC and the public
were also considered in completing the PA and the advice regarding the standards was described
and considered in the final 2020 PA (85 FR 31182, May 22, 2020), and in the EPA's decision-
making. On August 14, 2020, the EPA proposed to retain both the primary and secondary O3
standards, without revision (85 FR 49830, August 14, 2020). In December 2020, the EPA issued
its final decision to retain the existing standards without revision (85 FR 87256, December 31,
2020).19
Following publication of the 2020 final action, three petitions were filed for review of the
EPA's final decision in the D.C. Circuit and the court consolidated the cases. The EPA also
received two petitions for reconsideration of the 2020 action. On October 29, 2021, the Agency
filed a motion with the court explaining that it had decided to reconsider the 2020 O3 NAAQS
final decision20 and requested that the consolidated cases be held in abeyance until December 15,
2023. On December 21, 2021, the court ordered that the consolidated cases continue to be held in
abeyance pending further order of the court and directed the parties to file motions to govern by
December 15, 2023.
1.5 RECONSIDERATION OF THE 2020 O3 NAAQS DECISION
On October 29, 2021, the EPA announced that it will reconsider the 2020 decision to
retain the 2015 O3 standards. The EPA's plans are to reconsider the decision based on the
existing scientific record and in a manner that adheres to rigorous standards of scientific integrity
and provides ample opportunities for public input and engagement.21 Consistent with the
on through the use of specific questions, posed in writing prior to the public meeting, regarding aspects of the
documents being reviewed, as a means of obtaining subject matter expertise for its document review.
19	The decision on the secondary standard also considered and addressed the 2019 remand of the secondary standard
by the D.C. Circuit such that that decision incorporated the EPA's response to that remand.
20	The Agency's October 29, 2021 announcement is available at https://www.epa.gov/ground-level-ozone-
pollution/epa-reconsider-previous-administrations-decision-retain-2015-ozone.
21	Information about the decision to reconsider the December 2020 O3 NAAQS decision is available on this
webpage: https://www.epa.gov/ground-level-ozone-pollution/epa-reconsider-previous-administrations-decision-
retain-2015 -ozone
April 2022
1-13 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
commitment to rigorous standards of scientific integrity, the EPA will receive advice and
comments from a reestablished CASAC22 assisted by an expert O3 Panel.23 This reflects EPA's
renewed commitment to a rigorous NAAQS review process, with a focus on protecting scientific
integrity.
Presentations and considerations to be included in the PA for reconsideration will be
based on the conclusions, studies and related information included in the air quality criteria for
the 2020 review. This includes the studies assessed in the 2020 ISA and PA and the integration
of the scientific evidence presented in them. The EPA has additionally provisionally considered
two sets of scientific studies on the health and welfare effects of O3 that were not included in the
ISA (" 'new' studies") and that did not go through the comprehensive review process utilized in
review of the air quality criteria. With regard to the first set of studies, the EPA provisionally
considered a set of "new" scientific studies on the health and welfare effects of O3 that were
raised and discussed in public comments on the July 2020 proposed decision (Luben et al.,
2020). In considering and responding to the comments, the EPA provisionally considered the
studies in the context of the findings of the ISA, as described in the December 2020 decision (85
FR 87262, December 31, 2020). The EPA concluded that, taken in context, the "new"
information and findings did not materially change any of the broad scientific conclusions
regarding the health and welfare effects of O3 in ambient air made in the air quality criteria, and
accordingly, reopening the air quality criteria review was not warranted (Luben et al., 2020).24
More recently, in the context of this reconsideration of the 2020 decision on the primary
standard, given the primary role of controlled human exposure studies in the most recent
decisions on the primary standard, the EPA has conducted a literature search for any "new"
controlled human exposure studies that may have been published since the literature cutoff date
for the 2020 ISA, and provisionally evaluated this small set of such newly identified studies
(Duffney et al., 2022). Based on this provisional evaluation, the EPA has concluded that, taken in
context, the "new" information and findings do not materially change any of the broad scientific
22	Consistent with his decision to reestablish the membership of the CASAC to "ensure the agency received the best
possible scientific insight to support our work to protect human health and the environment," after consideration
of a candidate list based on public request for nominations (86 FR 17146-17147, April 1, 2021) the Administrator
announced selection of the seven members to serve on the chartered CASAC on June 17, 2021
(https://www.epa.gov/newsreleases/epa-announces-selections-charter-members-clean-air-scientific-advisory-
committee). The current CASAC membership is listed here:
httos://casac.epa.gov/ords/sab/f?p=105:29:1723269351020:::RP.29:P29 COMMITTEEON:CASAC.
23	The members of the O3 CASAC panel are identified here:
https://casac.epa.gov/ords/sab/f?p=l 13:14:11923922295141::: 14:P14_COMMITTEEON:2022%20CASAC%20O
zone%20Review%20Panel.
24	As noted at that time, "new" studies may sometimes be of such significance that it is appropriate to delay a
decision in a NAAQS review and to supplement the pertinent air quality criteria so the studies can be taken into
account (58 FR at 13013- 13014, March 9, 1993).
April 2022
1-14 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
conclusions regarding the health and welfare effects of O3 in ambient air made in the air quality
criteria; thus, reopening the air quality criteria review is not warranted (Duffney et al., 2022).
This PA is being developed for consideration by the EPA Administrator in reaching his
decision on the reconsideration of the December 2020 decision to retain the existing O3 NAAQS.
In assessing the policy implications of the available scientific information, this PA for the
reconsideration, as for the 2020 PA, is intended to help "bridge the gap" between the Agency's
scientific assessment, presented in the 2020 ISA, and quantitative technical analyses, and the
judgments required of the Administrator in determining whether it is appropriate to retain or
revise the O3 NAAQS. Accordingly, the PA for reconsideration will again address policy-
relevant questions based on those identified in the 2018 IRP. With regard to considerations
related to the primary standard, the PA for the reconsideration will focus on the evidence
described in the 2020 ISA,25 and the exposure/risk analyses presented in the 2020 PA, which
will be included in full in this PA. With regard to considerations related to the secondary
standard, the PA for reconsideration will focus on the evidence documented in the 2020 ISA,
along with quantitative analyses presented in the 2020 PA and in subsequent technical memos,
which have been updated to reflect recent air quality data.
This draft PA for the reconsideration is being provided to the CASAC for review and
comment and made available for public comment. The CASAC advice and public comment on
this draft PA will inform completion of the final PA and development of the Administrator's
proposed decision. The EPA is targeting the end of 2023 to complete decision-making in this
reconsideration.
25 The ISA builds on evidence and conclusions from previous assessments, focusing on synthesizing and integrating
the newly available evidence (ISA, section IS. 1.1). Past assessments are generally cited when providing further,
still relevant, details that informed the current assessment but are not repeated in the latest assessment.
April 2022
1-15 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
REFERENCES
Cox, LA. (2018). Letter from Dr. Louis Anthony Cox, Jr., Chair, Clean Air Scientific Advisory
Committee, to Acting Administrator Andrew R. Wheeler, Re: Consultation on the EPA's
Integrated Review Plan for the Review of the Ozone. December 10, 2018. EPA-CASAC-
19-001. Available at:
https://yosemite.epa.gov/sab/sabproduct.nsf/LookupWebReportsLastMonthCASAC/A286
A 0F0151DC8238525835F00 7D348A/$File/EPA-CASAC-19-001.pdf
Cox, LA. (2020a). Letter from Louis Anthony Cox, Jr., Chair, Clean Air Scientific Advisory
Committee, to Administrator Andrew R. Wheeler. Re:CASAC Review of the EPA's
Integrated Science Assessment for Ozone and Related Photochemical Oxidants (External
Review Draft - September 2019). February 19, 2020. EPA-CASAC-20-002. Availbale at:
https: yosemite.epa.gov/sab/sabproduct.nsf/264cb!22 7d55e02c85257402007446a4/F22
8E5D4D848BBED85258515006354D0/$File/EPA-CASAC-20-002.pdf.
Cox, LA. (2020b). Letter from Louis Anthony Cox, Jr., Chair, Clean Air Scientific Advisory
Committee, to Administrator Andrew R. Wheeler. Re:CASAC Review of the EPA's
Policy Assessment for the Review of the Ozone National Ambient Air Quality Standards
(ExternalReview Draft - October 2019). February 19, 2020. EPA-CASAC-20-003.
Available at:
https://yosemite. epa.gov/sab/sabproduct. nsf/264cbl22 7d55e02c85257402007446a4/4 713
D217BC07103485258515006359BA/$File/EPA-CASAC-20-003.pdf.
Duffney, PF, Brown, JS, and Stone, SL (2022). Memorandum to the Review of the Ozone
National Ambient Air Quality Standards (NAAQS) Docket (EPA-HQ-ORD-2018-
0279). Re: Provisional Evaluation of Newly Identified Controlled Human Exposure
Studies in the context of the 2020 Integrated Science Assessment for Ozone and Related
Photochemical Oxidants. April 15, 2020.
Frey, HC. (2014a). Letter from Dr. H. Christopher Frey, Chair, Clean Air Scientific Advisory
Committee, to Administrator Gina McCarthy. Re: CAS AC Review of the EPA's Welfare
Risk and Exposure Assessment for Ozone (Second External Review Draft). June 18,
2014. EPA-CASAC-14-003. Available at:
http://nepis.epa.gov/Exe/ZyPDF.cgi?Dockey=P 100JMSY.PDF.
Frey, HC. (2014b). Letter from Dr. H. Christopher Frey, Chair, Clean Air Scientific Advisory
Committee to Honorable Gina McCarthy, Administrator, US EPA. Re: CASAC Review
of the EPA's Second Draft Policy Assessment for the Review of the Ozone National
Ambient Air Quality Standards. June 26, 2014. EPA-CASAC-14-004. Available at:
https://nepis. epa.gov/Exe/ZyPURL. cgi?Dockey=P 100JR6F. txt.
Frey, HC. (2014c). Letter from Dr. H. Christopher Frey, Chair, Clean Air Scientific Advisory
Committee, to Administrator Gina McCarthy. Re: Health Risk and Exposure Assessment
for Ozone (Second External Review Draft - February 2014). EPA-CASAC-14-005.
Available at: https://nepis.epa.gov/Exe/ZyPURL.cgi?Dockey=P 100JR8I. txt.
April 2022
1-16 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
Luben, T, Lassiter, M and Herrick, J (2020). Memorandum to Ozone NAAQS Review Docket
(EPA-HQ-ORD-2018-0279). RE: List of Studies Identified by Public Commenters That
Have Been Provisionally Considered in the Context of the Conclusions of the 2020
Integrated Science Assessment for Ozone and Related Photochemical Oxidants.
December 2020. Docket Document ID: EPA-HQ-OAR-2018-0279-0560.
Pruitt, E. (2018). Memorandum from E. Scott Pruitt, Administrator, U.S. EPA to Assistant
Administrators. Back-to-Basics Process for Reviewing National Ambient Air Quality
Standards. May 9, 2018. Office of the Administrator U.S. EPA HQ, Washington DC.
Available at: https://www.epa.gov/criteria-air-pollutants/back-basics-process-reviewing-
national-ambient-air-quality-standards.
Samet, JM. (2010). Letter from Jonathan Samet, Chair, Clean Air Scientific Advisory
Committee, to Administrator Lisa Jackson. Re: CAS AC Review of EPA's Proposed
Ozone National Ambient Air Quality Standard (Federal Register, Vol. 75, Nov. 11,
January 19, 2010). February 19, 2010. EPA-CASAC-10-007. Available at:
https://nepis.epa.gov/Exe/ZyPURL.cgi?Dockey=P10072Tl.txt.
Samet, JM. (2011). Letter from Jonathan Samet, Chair, Clean Air Scientific Advisory
Committee, to Administrator Lisa Jackson. Re: CASAC Response to Charge Questions
on the Reconsideration of the 2008 Ozone National Ambient Air Quality Standards. .
March 30, 2011. EPA-CASAC-11-004. Available at:
https://yosemite.epa.gov/sab/sabproduct.nsf/368203f97al5308a852574ba005bbd01/F08
BEB48C1139E2A8525785E006909AC/$File/EPA-CASAC-l l-004-unsigned+.pdf
U.S. DHEW (1970). Air Quality Criteria for Photochemical Oxidants. National Air Pollution
Control Administration Washington, DC. U.S. DHEW. publication no. AP-63. NTIS,
Springfield, VA; PB-190262/BA.
U.S. EPA (1978). Air Quality Criteria for Ozone and Other Photochemical Oxidants
Environmental Criteria and Assessment Office. Research Triangle Park, NC. EPA-600/8-
78-004. April 1978. Available at:
https://nepis. epa.gov/Exe/ZyPURL. cgi?Dockey=200089CW. txt.
U.S. EPA (1986). Air Quality Criteria for Ozone and Other Photochemical Oxidants (Volume I -
V). Environmental Criteria and Assessment Office. Research Triangle Park, NC. U.S.
EPA. EPA-600/8-84-020aF, EPA-600/8-84-020bF, EPA-600/8-84-020cF, EPA-600/8-
84-020dF, EPA-600/8-84-020eF. August 1986. Available at:
https://nepis. epa.gov/Exe/ZyPURL. cgi?Dockey=30001D3J. txt
https://nepis. epa.gov/Exe/ZyPURL. cgi?Dockey=30001DA V. txt
https://nepis. epa.gov/Exe/ZyPURL. cgi?Dockey=30001DNN. txt
https://nepis. epa.gov/Exe/ZyPURL. cgi?Dockey=3000lE0F.txt
https://nepis. epa.gov/Exe/ZyPURL. cgi?Dockey=30001E9R txt.
U.S. EPA (1989). Review of the National Ambient Air Quality Standards for Ozone: Policy
Assessment of Scientific and Technical Information. OAQPS Staff Paper. Office of Air
Quality Planning and Standards. Research Triangle Park, NC U.S. EPA.
April 2022
1-17 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
U.S. EPA (1992). Summary of Selected New Information on Effects of Ozone on Health and
Vegetation: Supplement to 1986 Air Quality Criteria for Ozone and Other Photochemical
Oxidants. Office of Research and Development. Washington, DC. U.S. EPA. EPA/600/8-
88/105F.
U.S. EPA (1996). Air Quality Criteria for Ozone and Related Photochemical Oxidants. Volume I
-	III. Office of Research and Development Research Triangle Park, NC. U.S. EPA. EPA-
600/P-93-004aF, EPA-600/P-93-004bF, EPA-600/P-93-004cF. July 1996. Available at:
https://nepis. epa.gov/Exe/ZyPURL. cgi?Dockey=300026GN. Ixl
https://nepis. epa.gov/Exe/ZyPURL. cgi?Dockey=300026SH. Ixl
https://nepis. epa.gov/Exe/ZyPURL. cgi?Dockey=10004RHL. Ixl.
U.S. EPA (2006). Air Quality Criteria for Ozone and Related Photochemical Oxidants (Volume I
-	III). Office of Research and Development U.S. EPA. EPA-600/R-05-004aF, EPA-
600/R-05-004bF, EPA-600/R-05-004cF February 2006. Available at:
https://cfpub. epa.gov/ncea/risk/recordisplay. cfm ?deid=149923.
U.S. EPA (2013). Integrated Science Assessment of Ozone and Related Photochemical Oxidants
(Final Report). Office of Research and Development, National Center for Environmental
Assessment. Research Triangle Park, NC. U.S. EPA. EPA-600/R-10-076F. February
2013. Available at: https://nepis.epa.gov/Exe/ZyPURL.cgi?Dockey=P 100KETF.txt.
U.S. EPA (2014a). Policy Assessment for the Review of National Ambient Air Quality
Standards for Ozone (Final Report). Office of Air Quality Planning and Standards, Health
and Environmental Impacts Divison. Research Triangle Park, NC. U.S. EPA. EPA-
452/R-14-006 August 2014. Available at:
https://nepis. epa.gov/Exe/ZyPDF. cgi?Dockey=P 100KCZ5. Ixl.
U.S. EPA (2014b). Welfare Risk and Exposure Assessment for Ozone (Final). . Office of Air
Quality Planning and Standards. Research Triangle Park, NC. U.S. EPA. EPA-452/P-14-
005a August 2014. Available at:
https://nepis. epa.gov/Exe/ZyPURL. cgi?Dockey=P 100KB9D. Ixl.
U.S. EPA (2014c). Health Risk and Exposure Assessment for Ozone. (Final Report). Office of
Air Quality Planning and Standards. Research Triangle Park, NC. U.S. EPA. EPA-452/R-
14-004a. August 2014. Available at:
https://nepis. epa.gov/Exe/ZyPURL. cgi?Dockey=P 100KBUF. Ixl.
U.S. EPA (2019). Integrated Review Plan for the Ozone National Ambient Air Quality
Standards. Office of Air Quality Planning and Standards. Research Triangle Park, NC.
U.S. EPA. EPA-452/R-19-002. Available at:
https://www.epa.gov/sites/production/files/2019-08/documents/o3-irp-aug27-
2019Jinal.pdf.
U.S. EPA (2020a). Integrated Science Assessment for Ozone and Related Photochemical
Oxidants. U.S. Environmental Protection Agency. Washington, DC. Office of Research
and Development. EPA/600/R-20/012. Available at: https://www.epa.gov/isa/integrated-
science-assessment-isa-ozone-and-related-photochemical-oxidants.
April 2022
1-18 External Review Draft - Do Not Quote or Cite

-------
1	U.S. EPA (2020b). Policy Assessment for the Review of the Ozone National Ambient Air
2	Quality Standards. U.S. Environmental Protection Agency, Office of Air Quality
3	Planning and Standards, Health and Environmental Impacts Division. Research Triangle
4	Park, NC. U.S. EPA. EPA-452/R-20-001. 2020 Available at: https://www.epa.gov/
5	naaqs/ozone-o3-standards-policyassessments-current-review.
6
April 2022
1-19 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
2 AIR QUALITY
This chapter begins with an overview of O3 and other photochemical oxidants in the
atmosphere (section 2.1). Subsequent sections summarize the sources and emissions of O3
precursors (section 2.2), ambient air monitoring and data handling conventions for determining
whether the standards are met (section 2.3), O3 concentrations measured in the U.S. ambient air
(section 2.4), and available evidence and information related to background O3 in the U.S.
(section 2.5). These focus primarily on tropospheric O3 and surface-level concentrations
occurring in ambient air1.
2.1 O3 AND PHOTOCHEMICAL OXIDANTS IN THE ATMOSPHERE
O3 is one of many photochemical oxidants formed in the troposphere2 by photochemical
reactions of precursor gases in the presence of sunlight (ISA, Appendix 1, section l.l)3 and is
generally not directly emitted from specific sources. Tropospheric O3 and other oxidants, such as
peroxyacetyl nitrate (PAN) and hydrogen peroxide, form in polluted areas through atmospheric
reactions involving two main classes of precursor pollutants: volatile organic compounds
(VOCs) and nitrogen oxides (NOx = NO and NO2). The photolysis of the primary pollutant
nitrogen dioxide (NO2) results in products of NO and a singlet oxygen radical that can
subsequently either form ozone or react with NO to reform the parent NO2 compound. The
reaction of the oxygen radical with NO to form NO2 is disrupted by the presence of VOCs4
which leads to net ozone formation in the troposphere. Thus, NOx, VOCs, CH4 and CO are
considered to be the primary precursors of tropospheric O3 (ISA, Appendix 1, section 1.3.1)
The formation of O3, other oxidants and oxidation products from these precursors is a
complex, nonlinear function of many factors including (1) the intensity and spectral distribution
of sunlight; (2) atmospheric mixing; (3) concentrations of precursors in the ambient air and the
rates of chemical reactions of these precursors; and (4) processing on cloud and aerosol particles
(ISA, Appendix 1, section 1.4; 2013 ISA, section 3.2). As a result, O3 changes in a nonlinear
fashion with the concentrations of its precursors rather than varying proportionally to emissions
1	Ambient air means that portion of the atmosphere, external to buildings, to which the general public has access
(see 40 CFR 50.1(e)).
2	Ozone also occurs in the stratosphere, where it serves the beneficial role of absorbing the sun's harmful ultraviolet
radiation and preventing the majority of this radiation from reaching the Earth's surface.
3	The only other appreciable source of O3 to the troposphere is transport from the stratosphere, as described in
section 2.5.1.1 below.
4	This reaction can also be disrupted by the radical that results from methane (CH4) oxidation or a reaction between
carbon monoxide (CO) and the hydroxyl radical (OH) in the atmosphere.
April 2022
2-1
External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
of its precursors (2013 ISA, section 3.2.4). In addition to the chemistry described above, NO can
also react with ozone directly such that emissions of NOx lead to both the formation and
destruction of O3, with the net formation or destruction depending on the local quantities of
NOx, VOCs, radicals, and sunlight.03 chemistry is often described in terms of which precursors
most directly impact formation rates. A NOx-limited regime indicates that O3 concentrations will
decrease in response to decreases in ambient NOx concentrations and vice-versa. These
conditions tend to occur when NOx concentrations are generally low compared to VOC
concentrations and during warm, sunny conditions when NOx photochemistry is relatively fast.
NOx-limited conditions are more common during daylight hours, in the summertime, in
suburban and rural areas, and in portions of the country with high biogenic VOC emissions like
the Southeast. In contrast, NOx-saturated conditions (also referred to as VOC-limited or radical-
limited) indicate that O3 will increase as a result of NOx reductions but will decrease as a result
of VOC reductions (2013 ISA, section 3.2; 2006 AQCD, chapter 2). NOx-saturated conditions
occur at times when and at locations with lower levels of available sunlight, resulting in slower
photochemical formation of O3, and when NOx concentrations are in excess compared to VOC
concentrations. NOx-saturated conditions are more common during nighttime hours, in the
wintertime, and in densely populated urban areas or industrial plumes. These varied relationships
between precursor emissions and O3 chemistry result in localized areas in which O3
concentrations are suppressed compared to surrounding areas, but which contain NO2 that
contributes to subsequent O3 formation further downwind (2013 ISA, section 3.2.4).
Consequently, O3 response to reductions in NOx emissions is complex and may include
decreases in O3 concentrations at some times and locations and increases in O3 concentrations at
other times and locations. Over the past decade, there have been substantial decreases in NOx
emissions in the U.S. (see Figure 2-2) and many locations have transitioned from NOx-saturated
to NOx-limited (Jin et al., 2017) during times of year that are conducive to O3 formation
(generally summer). As these NOx emissions reductions have occurred, lower O3 concentrations
have generally increased while the higher O3 concentrations have generally decreased, resulting
in a compressed O3 distribution, relative to historical conditions (ISA, Appendix 1, section 1.7).
Prior to 1979, the indicator for the NAAQS for photochemical oxidants was total
photochemical oxidants (36 FR 8186, April 30, 1971). Early ambient air monitoring indicated
similarities between O3 measurements and the photochemical oxidant measurements, as well as
reduced precision and accuracy of the latter (U.S. EPA, 1978). To address these issues, the EPA
established O3 as the indicator for the NAAQS for photochemical oxidants in 1979 (44 FR 8202,
February 8, 1979), and it is currently the only photochemical oxidant other than nitrogen dioxide
that is routinely monitored in a national ambient air monitoring network.
April 2022
2-2 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
O3 is present not only in polluted urban atmospheres, but throughout the troposphere,
even in remote areas of the globe. The same basic processes involving sunlight-driven reactions
of NOx, VOCs, CH4 and CO contribute to O3 formation throughout the troposphere. These
processes also lead to the formation of other photochemical products, such as PAN, HNO3, and
H2SO4, HCHO and other carbonyl compounds, as well as a number of organic particulate
compounds (ISA, Appendix 1, section 1.4; 2013 ISA, section 3.2).
As mentioned above, the formation of O3 from precursor emissions is also affected by
meteorological parameters such as the intensity of sunlight and atmospheric mixing (2013 ISA,
section 3.2). Major episodes of high O3 concentrations in the eastern U.S. are often associated
with slow-moving high-pressure systems which can persist for several days. High pressure
systems during the warmer seasons are associated with the sinking of air, resulting in warm,
generally cloudless skies, with light winds. The sinking of air results in the development of
stable conditions near the surface which inhibit or reduce the vertical mixing of O3 precursors,
concentrating them near the surface. Photochemical activity involving these precursors is
enhanced because of higher temperatures and the availability of sunlight during the warmer
seasons. In the eastern U.S., concentrations of O3 and other photochemical oxidants are
determined by meteorological and chemical processes extending typically over areas of several
hundred thousand square kilometers. Therefore, O3 episodes are often regarded as regional in
nature, although more localized episodes often occur in some areas, largely the result of local
pollution sources during summer, e.g., Houston, TX (2013 ISA, section 2.2.1; Webster et al.,
2007). In addition, in some parts of the U.S. (e.g., Los Angeles, CA), mountain barriers limit O3
dispersion and result in a higher frequency and duration of days with elevated O3 concentrations
(2013 ISA, section 3.2).
More recently, high O3 concentrations of up to 150 parts per billion (ppb)5 have been
measured during the wintertime in two western U.S. mountain basins (ISA, Appendix 1, section
1.4.1). Wintertime mountain basin O3 episodes occur on cold winter days with low wind speeds,
clear skies, substantial snow cover, extremely shallow boundary layers driven by strong
temperature inversions, and substantial precursor emissions activity from the oil and gas sector.
The results of recent modeling studies suggest that photolysis of VOCs provides the source of
reactive chemical species (radicals) needed to initiate the chemistry driving these wintertime O3
episodes. This mechanism is somewhat different from the chemistry driving summertime O3
formation, which is initiated with the photolysis of NO2 followed by the formation of the OH
radicals (ISA, Appendix 1, section 1.4.1).
5 Although the standards are specified in ppm (e.g., as described in Chapter 1), the units, ppb, are commonly used in
describing O3 concentrations throughout this document, with 0.070 ppm being equivalent to 70 ppb.
April 2022
2-3 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
O3 concentrations in a region are affected both by local formation and by transport of O3
and its precursors from upwind areas. O3 transport occurs on many spatial scales including local
transport within urban areas, regional transport over large regions of the U.S., and long-range
transport which may also include international transport. In addition, O3 can be transferred into
the troposphere from the stratosphere, which is rich in naturally occurring O3, through
stratosphere-troposphere exchange (STE). These intrusions usually occur behind cold fronts,
bringing stratospheric air with them and typically affect O3 concentrations in higher elevation
areas (e.g., > 1500 m) more than areas at lower elevations, as discussed in section 2.5.3.2 (ISA,
Appendix 1, section 1.3.2.1; 2013 ISA, section 3.4.1.1).
2.2 SOURCES AND EMISSIONS OF O3 PRECURSORS
Sources of emissions of O3 precursor compounds can be divided into anthropogenic and
natural source categories, with natural sources further divided into emissions from biological
processes of living organisms (e.g., plants, microbes, and animals) and emissions from chemical
or physical processes (e.g., biomass burning, lightning, and geogenic sources). Anthropogenic
emissions associated with combustion processes, including mobile sources and power plants,
account for the majority of U.S. NOx and CO emissions (Figure 2-1 and Figure 2-2). Emissions
of these chemicals have declined appreciably in the U.S. since 2002 (Figure 2-2). Anthropogenic
sources are also important for VOC emissions, though in some locations and times of the year
(e.g., southern states during summer) the majority of VOC emissions come from vegetation
(2013 ISA, section 3.2.1).6 In practice, the distinction between natural and anthropogenic sources
is often unclear, as human activities directly or indirectly affect emissions from what would have
been considered natural sources during the preindustrial era. Thus, precursor emissions from
plants, animals, and wildfires could be considered either natural or anthropogenic, depending on
whether emissions result from agricultural practices, forest management practices, lightning
strikes, or other types of events. There are additional challenges in distinguishing between ozone
resulting from natural versus anthropogenic sources because much O3 results from reactions of
anthropogenic precursors with natural precursors (ISA, Appendix 1, section 1.8.1.2).
The National Emissions Inventory (NEI) is a comprehensive and detailed estimate of air
emissions of criteria pollutants, precursors to criteria pollutants, and hazardous air pollutants
from air emissions sources (U.S. EPA, 2021b). The NEI is released every three years based
primarily upon data provided by State, Local, and Tribal air agencies for sources in their
6 It should be noted that the definition of VOCs used in this section does not include CH4 because it is excluded
from the EPA's regulatory definition of VOCs in 40 CFR 51.100(s). More information about this regulatory
definition of VOCs is available at https://www. epa.gov/indoor-air-quality-iaq/technical-overview-volatile-organic-
compounds.
April 2022
2-4 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
jurisdictions and supplemented by data developed by the US EPA. The NEI is built using the
EPA's Emissions Inventory System (EIS) which collects data from State, Local, and Tribal air
agencies and blends that data with other data sources.7
Anthropogenic emissions of air pollutants result from a variety of sources such as power
plants, industrial sources, motor vehicles, and agriculture. The emissions from any individual
source typically vary in both time and space. For many of the thousands of sources that make up
the NEI, there is uncertainty in both of these factors. For some sources, such as power plants,
direct emission measurements enable more certain quantification of the magnitude and timing of
emissions than from sources without such direct measurements. However, for many source
categories emission inventories necessarily contain assumptions, interpolation and extrapolation
from a limited set of sample data (U.S. EPA, 2021b).
7 More details are available from: https://www.epa.gov/enviro/nei-overview.
April 2022	2-5 External Review Draft - Do Not Quote or Cite

-------
A) NOx (11,786 kTon/yr)
B) VOCs (43,073 kTon/yr)
Stationary Fuel Combustion 22%

Biogenics 60%	

Industrial / \
Processes 10% 7 \
;>v All Fires 3%
/y y Biogenics 12%



	1- Other 1 %

\nT"~—Other 4%
Highway Vehicles 30% I
/Non-Road
Mobile 22%
Wildfires 11 % --^/ _
\ / Mobile
\ y Sources 7%
\ ^Industrial
Processes 7%


Agricultural & Prescribed Fires 5%
Solvent Utilization 7%
C) CO {70,794 kTon/yr)
D) CH4 (29,696 kTon/yr)

^Wildfires 28%

___Energy/Fossil Fuels 40%
Agricultural & / \
Prescribed Fires 13%/ \



Stationary Fuel J —-—
C^nmhi lotion fi% 1 		—
v Biogenics 6%


VJUI 1 IUUoLIUI 1 U / O 1 ^\
Industrial \ \
Processes 2% \ ^
		_J- Other 2%

~~	—	i- Other 2%
Highway Vehicles 28% 	
\ /.Non-Road
\ z7 Mobile 16%
Agriculture 38% \
\ yOWaste Disposal/
Landfills 20%
2	Sources: The 2017 National Emissions Inventory (U.S. EPA, 2021b) for panels A-C, and the Inventory of U.S. Greenhouse Gas
3	Emissions and Sinks: 1990-2019 {U.S. EPA, 2021a) for panel D. Categories contributing less than 2% each have been summed
4	and are represented by the "other" category.
5	Figure 2-1. U.S. O3 precursor emissions by sector: A) NOx; B) CO; C) VOCs; D) CH4.
6
April 2022
2-6 External Review Draft - Do Not Quote or Cite

-------
A) NOx
10000-
S 8000-
ฃ=
o
h-
ฃ 6000-
4000-
x
O
2000-
B) VOCs
10000-
15000-
-12000"
9000-
6000-
3000-
50000
^40000
in 30000
ฃ20000
uj
o
o
1
Inventory Year
C) CO
Inventory Year
~T~
xr
Inventory Year
Legend: CO, NOx, VOC
Inventory Year
Legend: CH4
	 Highway Vehicles
	 Non-Road Mobile
Stationary Fuel Combustion
	 Industrial and Other Processes
Other Anthropogenic Sources
	 Energy/Fossil Fuels
	 Agriculture
Waste Disposal/Landfills
	 Other Anthropogenic Sources
1
2	Sources: EPA's Air Pollutant Emissions Trends Data webpage (https://www,epa,gov/air-emissions-inventories/air-pollutant-
3	emissions-trends-data) for panels A-C, and the Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019 (U.S. EPA,
4	2021a) for panel D,
5	Figure 2-2. U.S. anthropogenic Os precursor emission trends for: A) NOx; B) CO; C)
6	VOCs; and D) CH4.
April 2022
2-7 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
Figure 2-3, Figure 2-4, and Figure 2-5 show county-level estimates of U.S. emissions
densities (in tons/year/mi2) for CO, NOx, and VOCs, respectively. In general, CO and NOx
emissions tend to be highest in urban areas which typically have the most anthropogenic sources,
however, CO emissions may be higher in some rural areas due to fires, and similarly, NOx
emissions may be higher in some rural areas due to sources such as electricity generation, oil and
gas extraction, and traffic along major highways. While there are some significant anthropogenic
sources of VOC emissions in urban areas, in rural areas the vast majority of VOC emissions
come from plants and trees (biogenics), particularly in the southeastern U.S. In other areas of the
U.S., such as the Great Plains region and parts of the inter-mountain west, areas with higher
levels of VOC emissions are largely due to oil and gas extraction (U.S. EPA, 2021b).
It should be noted that O3 levels in a given area are impacted by both local emissions that
form O3 in the area as well as remote emissions that form O3 that is then transported into the
area. Biogenic VOC emissions that lead to O3 formation may vary greatly depending on the type
and amount of vegetation, which is generally much lower in urban areas than in rural areas.
However, biogenic VOC emissions that are upwind of an urban area can have a significant
impact on urban O3 levels. Thus, while the county-level maps shown in Figure 2-3, Figure 2-4,
and Figure 2-5 illustrate the variability in precursor emissions in the U.S., it is not sufficient to
look only at the patterns in local emissions when considering the impact on O3 concentrations.
April 2022
2-8 External Review Draft - Do Not Quote or Cite

-------
Total Carbon Monoxide Emissions Density (tons/year/miA2)
2	~ 0-9(955)	~ 10-19 (945) ~ 20-49 (840) ฆ 50-99(279) ฆ 100-3635(201)
2	Source: 2017 National Emissions Inventory, January 2021 Updated Release (U.S. EPA, 2021b; data downloaded from
3	https://www. epa. aov/air-emissions-inventories/2017-national-emissions-inventorv-nei-data)
4	Figure 2-3. U.S. county-level CO emissions density estimates (tons/year/mi2) for 2017.
Total Nitrogen Oxides Emissions Density (tons/year/miA2)
~ 0-1 (1024) ~ 2-4(1264) P 5-9 (499) ฆ 10-19(251) ฆ 20-826(182)
6	Source: 2017 National Emissions Inventory, January 2021 Updated Release (U.S. EPA, 2021b; data downloaded from
7	https://www. epa. qov/air-emissions-inventories/2017-national-emissions-inventorv-nei-data)
8	Figure 2-4. U.S. county-level NOx emissions density estimates (tons/year/mi2) for 2017.
April 2022	2-9 External Review Draft - Do Not Quote or Cite

-------
Total Volatile Organic Compounds Emissions Density (tons/year/mi^)
~ 0-4 (557) ~ 5-9(718) ~ 10-19(777) ~ 20-49 (1053) ฆ 50-704(115)
2	Source: 2017 National Emissions Inventory, January 2021 Updated Release (U.S. EPA, 2021b; data downloaded from
3	https://www.epa.ciov/air-emissions-inventories/2017-national-emissions-inventorv-nei-clata)
4	Figure 2-5. U.S. county-level VOC emissions density estimates (tons/year/mi2) for 2017.
5	2.3 AMBIENT AIR MONITORING AND DATA HANDLING
6	CONVENTIONS
7	2.3.1 Ambient Air Monitoring Requirements and Monitoring Networks
8	State and local environmental agencies operate a network of O3 monitors at state or local
9	air monitoring stations (SLAMS). The requirements for the SLAMS network depend on the
10	population and most recent O3 design values8 in an area. The minimum number of O3 monitors
11	required in a metropolitan statistical area (MSA) ranges from zero for areas with a population
12	less than 350,000 and no recent hi story of an O3 design value greater than 85 percent of the level
13	of the standard, to four for areas with a population greater than 10 million and an O3 design value
14	greater than 85 percent of the standard level.9 At least one monitoring site for each MSA must he
15	situated to record the maximum concentration for that particular metropolitan area. Siting criteria
8	A design value is a statistic that summarizes the air quality data for a given area in terms of the indicator, averaging
time, and form of the standard. Design values can be compared to the level of the standard and are typically used
to designate areas as meeting or not meeting the standard and assess progress towards meeting the NAAQS.
9	The SLAMS minimum monitoring requirements to meet the O3 design criteria are specified in 40 CFR Part 58,
Appendix D. The minimum O3 monitoring network requirements for urban areas are listed in Table D-2 of
Appendix D to 40 CFR Part 58 (accessible at https://www.ecfr.gov).
April 2022
2-10 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
for SLAMS includes horizontal and vertical inlet probe placement; spacing from minor sources,
obstructions, trees, and roadways; inlet probe material; and sample residence times.10 Adherence
to these criteria ensures uniform collection and comparability of O3 data. Since the highest O3
concentrations tend to be associated with a particular season for various locations, the EPA
requires O3 monitoring during specific O3 monitoring seasons (shown in Figure 2-6) which vary
by state from five months (May to September in Oregon and Washington) to all twelve months
(in 11 states), with the most common season being March to October (in 27 states).11
Most of the state, local, and tribal air monitoring stations that report data to the EPA use
ultraviolet Federal Equivalent Methods (FEMs). The Federal Reference Method (FRM) was
revised in 2015 to include a new chemiluminescence by nitric oxide (NO-CL) method. The
previous ethylene (ET-CL) method, while still included in the CFR as an acceptable method, is
no longer used due to lack of availability and safety concerns with ethylene.12 The NO-CL
method is beginning to be implemented in the SLAMS network.13
Ambient air quality data and associated quality assurance (QA) data are reported to the
EPA via the Air Quality System (AQS). Data are reported quarterly and must be submitted to
AQS within 90 days after the end of the quarterly reporting period. Each monitoring agency is
required to certify data that is submitted to AQS from the previous year. The data are certified,
taking into consideration any QA findings, and a data certification letter is sent to the EPA
Regional Administrator. Data must be certified by May 1st of the following year. Data collected
by FRM or FEM monitors that meet the QA requirements must be certified.14 To provide
decision makers with an assessment of data quality, the EPA's QA group derives estimates of
both precision and bias for O3 and the other gaseous criteria pollutants from quality control (QC)
checks using calibration gas, performed at each site by the monitoring agency. The data quality
goal for precision and bias is 7 percent.15
10	The probe and monitoring path siting criteria for ambient air quality monitoring are specified in 40 CFR, Part 58,
Appendix E.
11	The required O3 monitoring seasons for each state are listed in 40 CFR Part 58, Appendix D, Table D-3.
12	The current FRM for O3 (established in 2015) is a chemiluminescence method, which is fully described in 40 CFR
Part 50, Appendix D.
13	The EPA is currently participating in an international effort to implement a globally coordinated change in the
parameter (the absorption cross-section value) used in the determination of atmospheric ozone for ozone
monitoring, which will require an update of this parameter in the ozone monitoring regulations (40 CFR Part 50,
Appendix D, section 4). The global implementation target date for this change is the beginning of the 2024 ozone
season.
14	Quality assurance requirements for monitors used in evaluations of the NAAQS are provided in 40 CFR Part 58,
Appendix A.
15	Annual summary reports of precision and bias can be obtained for each monitoring site at
https://www.epa.gov/outdoor-air-quality-data/single-point-precision-ancl-bias-report.
April 2022
2-11 External Review Draft - Do Not Quote or Cite

-------
1
2
3	Figure 2-6. Current Os monitoring seasons in the U.S. Numbers in each state indicate the months of the year the state is required
4	to monitor for O3 (e.g., 3-10 means O3 monitoring is required from March through October).
April 2022
2-12
External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
In 2020, there were over 1,300 federal, state, local, and tribal ambient air monitors
reporting O3 concentrations to the EPA. Figure 2-7 shows the locations of such monitoring sites
that reported data to the EPA at any time during the 2018-2020 period. Nearly 80% of this
network are SLAMS monitors operated by state and local governments to meet regulatory
requirements and provide air quality information to public health agencies; these sites are largely
focused on urban and suburban areas.
Two important subsets of SLAMS sites separately make up the National Core (NCore)
multi-pollutant monitoring network and the Photochemical Assessment Monitoring Stations
(PAMS) network. Each state is required to have at least one NCore station, and O3 monitors at
NCore sites are required to operate year-round. At each NCore site located in a MSA with a
population of 1 million or more (based on the most recent census), a PAMS network site is
required.16 In addition to reporting O3 concentrations, the NCore and PAMS networks provide
data on O3 precursor chemicals. The NCore sites feature co-located measurements of chemical
species such as nitrogen oxide and total reactive nitrogen, along with various meteorological
measurements. At a minimum, monitoring sites in the PAMS network are required to measure
certain O3 precursors, such as NOx and a target set of VOCs, during the months of June, July and
August, although some precursor monitoring may be required for longer periods of time to
improve the usefulness of data collected during an area's O3 season (U.S. EPA, 2018a).The
enhanced monitoring at sites in these two networks informs our understanding of local O3
formation.
While the SLAMS network has a largely urban and population-based focus, there are
monitoring sites in other networks that can be used to track compliance with the NAAQS in rural
areas. For example, the Clean Air Status and Trends Network (CASTNET) monitors are located
in rural areas. There were 84 CASTNET monitors operating in 2020, with most of the sites in the
eastern U.S. being operated by the EPA, and most of the sites in the western U.S. being operated
by the National Park Service (NPS). Finally, there are also a number of Special Purpose
Monitoring Stations (SPMs), which are not required but are often operated by air agencies for
short periods of time (less than 3 years) to collect data for human health and welfare studies, as
well as other types of monitoring sites, including monitors operated by tribes and industrial
sources. The SPMs are typically not used to assess compliance with the NAAQS.17
16	The requirements for PAMS, which were most recently updated in 2015, is fully described in section 5 of
Appendix D to 40 CFR Part 58.
17	However, SPMs that use federal reference or equivalent methods, meet all applicable requirements in 40 CFR Part
58, and operate continuously for more than 24 months may be used to assess compliance with the NAAQS.
April 2022
2-13 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
SLAMS (961)	• NCORE/PAMS (126) • CASTNET (84)	SPM/OTHER (191)
Figure 2-7. Map of U.S. ambient air O3 monitoring sites reporting data to the EPA during
the 2018-2020 period.
2.3.2 Data Handling Conventions and Computations for Determining Whether the
Standards are Met
To assess whether a monitoring site or geographic area (usually a county or urban area)
meets or exceeds a NAAQS, the monitoring data are analyzed consistent with the established
regulatory requirements for the handling of monitoring data for the purposes of deri ving a desi gn
value. A design value summarizes ambient air concentrations for an area in terms of the
indicator, averaging time and form for a given standard such that its comparison to the level of
the standard indicates whether the area meets or exceeds the standard. The procedures for
calculating design values for the current O3 NAAQS (established in 2015) are detailed in
Appendix U to 40 CFR Part 50 and are summarized below.
Hourly average O3 concentrations at the monitoring sites used for assessing whether an
area meets or exceeds the NAAQS are required to be reported in ppm to the third decimal place,
with additional digits truncated, consistent with the typical measurement precision associated
with most O3 monitoring instruments. Monitored hourly O3 concentrations flagged by the States
as having been affected by an exceptional event, having been the subject of a demonstration
submitted by the State, and having received concurrence from the appropriate EPA Regional
April 2022
2-14 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
Office, are excluded from design value calculations consistent with 40 CFR 50.14.18 The hourly
concentrations are used to compute moving 8-hour averages, which are stored in the first hour of
each 8-hour period (e.g., the 8-hour average for the 7:00 AM to 3:00 PM period is stored in the
7:00 AM hour), and digits to the right of the third decimal place are truncated. Each 8-hour
average is considered valid if 6 or more hourly concentrations are available for the 8-hour period.
Next, the daily maximum 8-hour average (MDA8) concentration for each day is
identified as the highest of the 17 consecutive, valid 8-hour average concentrations beginning at
7:00 AM and ending at 11:00 PM (which includes hourly O3 concentrations from the subsequent
day). MDA8 values are considered valid if at least 13 valid 8-hour averages are available for the
day, or if the MDA8 value is greater than the level of the NAAQS. Finally, the O3 design value is
calculated as the 3-year average of the annual 4th highest MDA8 value19. An O3 design value less
than or equal to the level of the NAAQS is considered to be valid if valid MDA8 values are
available for at least 90% of the days in the O3 monitoring season (as defined for each state and
shown in Figure 2-6) on average over the 3 years, with a minimum of 75% data completeness in
any individual year. Design values greater than the level of the NAAQS are always considered to
be valid.
An O3 monitoring site meets the NAAQS if it has a valid design value less than or equal
to the level of the standard, and it exceeds the NAAQS if it has a design value greater than the
level of the standard. A geographic area meets the NAAQS if all ambient air monitoring sites in
the area have valid design values meeting the standard. Conversely, if one or more monitoring
sites has a design value exceeding the standard, then the area exceeds the NAAQS.
2.4 O3 IN AMBIENT AIR
2.4.1 Concentrations Across the U.S.
Figure 2-8 below shows a map of the O3 design values at U.S. ambient air monitoring
sites based on data from the 2018-2020 period. From the figure it is apparent that many
monitoring sites have recent design values exceeding the current NAAQS, and that most of these
sites are located in or near urban areas. The highest design values are located in California,
Texas, along the shoreline of Lake Michigan, and near large urban areas in the northeastern and
western U.S. There are also high design values associated with wintertime O3 in the Uinta Basin
in Utah. The lowest design values are located in the north central region of the U.S., rural parts
18	A variety of resources and guidance documents related to identification and consideration of exceptional events in
design value calculations are available at https://www.epa.gov/air-quality-analysis/final-2016-exceptional-events-
rule-supporting-guidance-documents-updated-faqs.
19	Design values are reported in ppm to the third decimal place, with additional digits truncated. This truncation step
also applies to the initially calculated 8-hour average concentrations (Appendix 2A, section 2A.1).
April 2022
2-15 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
of New England and the southeastern U.S., and along the Pacific Ocean, including Alaska and
Hawaii.
• 31 - 60 ppb (287 sites) G 66 -70 ppb (248 sites) • 76 -114 ppb (93 sites)
ฉ 61 - 65 ppb (342 sites) ฉ71-75 ppb (120 sites)
Figure 2-8. O3 design values in ppb for the 2018-2020 period.
2.4.2 Trends in U.S. O3 Concentrations
Figure 2-9 shows a map of the site-level trends in the O3 design values at U.S. monitoring
sites having complete data20 from 2000-2002 through 2018-2020. The trends were computed
using the Thiel-Sen estimator (Sen, 1968; Thiel, 1950), and tests for significance were computed
using the Mann-Kendall test (Kendall, 1948; Mann, 1945). From this figure it is apparent that
design values have decreased significantly over most of the eastern U.S. during this period.
These decreases are in part due to EPA regulations aimed at reducing NOx emissions from
EGUs, such as the Clean Air Interstate Rule and the Cross-State Air Pollution Rule, with the goal
of achieving broad, regional reductions in summertime NOx emissions; as well as mobile
emission reductions from federal motor vehicle emissions and fuel standards, and; local controls
resulting from implementation of the existing O3 standards. Other areas of the country have also
20 The data completeness criteria for Figure 2-8 through Figure 2-14 are listed in Table 2A-1 of Appendix 2A.
April 2022
2-16 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
experienced decreases in design values, most notably in California and near urban areas in the
intermountain west.
~ Decreasing > 1 ppb/yr (363 sites) ฐ No Significant Trend (57 sites)
v Decreasing < 1 ppb/yr (232 sites) A increasing < 1 ppb/yr (6 sites)
Figure 2-9. Trends in O3 design values based on data from 2000-2002 through 2018-2020.
Figure 2-10 shows the national trend in the annual 4th highest MDA8 values based on 188
ambient air monitoring sites with complete data from 1980 to 2020. This figure shows that, on
average, there has been a 33% decrease in U.S. annual 4th highest MDA8 levels since 1980.
Since relatively few sites have been monitoring continuously since 1980, Figure 2-11 shows the
national trend in the annual 4lh highest MDA8 values and the design values based on the 822
monitoring sites with complete data from 2000 to 2020. The U.S. median annual 4th highest
MDA8 values decreased by 25% nationally from 2002 (88 ppb) to 2013 (66 ppb), with some
variability among individual years in this period which can partially be attributed to changes in
meteorological conditions. Similarly, the U.S. median design value decreased by 20% from
2000-2002 (84 ppb) to 2013-2015 (67 ppb). The trend in the annual 4th highest MDA8
concentrations was relatively flat from 2013 to 2018, with decreases occurring in 2019 and 2020.
The design values have been relatively constant since 2015, though there are slight decreases in
2019 and 2020. In general, the design value metric is more stable and therefore better reflects
long-term changes in Oa than the annual 4th highest MDA8 metric.
April 2022
2-17 External Review Draft - Do Not Quote or Cite

-------
Ozone Air Quality, 1980 - 2020
(Annual 4th Maximum of Daily Max 8-Hour Average)
National Trend based on 188 Sites
015
111111111111111111112 2 2222222222222222222
99999999999 9 999999990000000 0 0000000000000
8 8 8 8 8 8 8 8 8 8 9 9 9 9 9 9 9 9 9 9 0 0 0 0 0 0 0 0 0 0 11111111112
012345 6 78 9 0123456789 0 12345676 901234567890
1
2
3
4
5
1980 to 2020 : 33% decrease in National Average
Source: EPA's Air Trends website (https://www.epa.gov/air-trends/ozone-trends/).
Figure 2-10. National trend in annual 4th highest MDA8 values, 1980 to 2020. The white
center line is the average while the filled area represents the range between the
10th and 90th percentiles. The dotted line is the level of the standard.
100-
90
X2
Q_
CL
1 80-

X
tt 70-
60-
50


/










10th/90th Percentile Annual Valu
Median Annual Value
3


/
- \
\








10th/90th Percentile Design Valu
Median Design Value
e




\
\
k
>






—
Current Standard








\



**
>
v
\



ฆ











\
V
\




ฆB
> \
\


s
*
•v*












\
>
/

0*






V
%
V /



\
~r"











>
v


\v












\



ฆs.
ฆs.
V



-V
V












V
/



ฆV




	—
















ฆs.
<

— -1/


>w




"—-










































6
7
o
o
o
o
o
C\J CO
o o
o o
lo to
o o o
o o o
o
o
CO o
o o
o o
CMCMCMCvJCNJCvJCMCMCMCMfMCMCNJCMCMCNJCMCMCNJ
O
CM
O
Figure 2-11. National trend in annual 4th highest MDA8 concentrations and O3 design
values in ppb, 2000 to 2020.
April 2022
2-18 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
Figure 2-12 shows regional trends in the median annual 4th highest MDA8 values for the
9 National Oceanic and Atmospheric Administration (NOAA) climate regions21 based on
ambient air monitoring sites with complete O3 monitoring data for 2000-2020. The five eastern
U.S. regions (Central, East North Central, Northeast, Southeast, South) have all shown decreases
of at least 10 ppb in median annual 4th highest MDA8 values since the early 2000's, with the
Southeast region in particular showing the largest decrease of over 20 ppb. In contrast, the
median annual 4th highest MDA8 values have changed by less than 10 ppb in each of the four
western U.S. regions (Northwest, Southwest, West, West North Central). The large increase in
the Northwest region in 2017 and 2018 correspond to years with historically high wildfire
activity.
000000000000000000000
CMCNJCMCMCMCMCMCMCnJCMCMCMCMCMCMOJCMCMCMCMCM
Figure 2-12. Regional trends in median annual 4th highest MDA8 concentrations, 2000 to
2020.
21 These regions are defined per Karl and Koss (1984) as illustrated in Appendix 2B, Figure 2B-1.
April 2022
2-19 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
Trends presented in this section have focused on annual 4th high MDA8 concentrations
and design values. Additional information from the published literature has examined trends in
MDA8 concentrations across the distribution of high and low O3 days. Simon et al., 2015) found
that, similar to results presented in this section for DVs and annual 4th high MDA8
concentrations, the 95th percentile of summertime MDA8 concentrations decreased significantly
at most sites across the U.S. between 1998 and 2013. In contrast, trends over that time period for
the 5th percentile, median and mean of MDA8 varied with location and time of year. Similarly,
Lefohn et al. (2017) reported that between 1980 and 2014 there was a compression of the
distribution of measured hourly O3 values with extremely high and extremely low concentrations
becoming less common. As a result, O3 metrics impacted by high hourly O3 concentrations, such
as the annual 4th highest MDA8 value, decreased at most U.S. sites across this period.
Concurrently, metrics that are impacted by averaging longer time periods of hourly O3
measurements, such as the 6-month (April-September) average of daytime (8am-7pm) O3
concentrations, were more varied with only about half of the sites exhibiting decreases in this
metric and most other sites exhibiting no trend (Lefohn et al., 2017).
2.4.3 Diurnal Patterns
Tropospheric O3 concentrations in most locations exhibit a diurnal pattern due to the
photochemical reactions that drive formation and destruction of O3 molecules. Figure 2-13
shows boxplots of O3 concentrations in ambient air, by hour of the day for four monitoring sites
that represent diurnal patterns commonly observed in the U.S. The boxes represent the 25th
percentile, median, and 75th percentiles and each box has "whiskers" which extend up to 1.5 times
the interquartile range (i.e., the 75th percentile minus the 25th percentile) from the box, and dots which
represent outlier values. The top panels show diurnal patterns, based on available data from 2015-
2017, at urban (panel A) and downwind suburban (panel B) monitoring sites in the Los Angeles
metropolitan area. Both sites generally measure their highest O3 concentrations during the early
afternoon hours, and their lowest concentrations during the early morning hours, as is typical of
most urban and suburban areas in the U.S. However, higher levels of NOx emissions near the
urban site may suppress O3 formation throughout the day and increase the O3 titration rate at
night, resulting in lower O3 concentrations than those typically observed at the downwind site.
Ozone concentrations are generally lower in rural areas than in urban and suburban areas,
with less pronounced diurnal patterns. However, elevation and transport also play a larger role in
influencing concentrations in rural areas than in urban areas. The bottom panels in Figure 2-13
show diurnal patterns at low elevation (panel C) and high elevation (panel D) rural monitoring
sites in New Hampshire. The low elevation site experiences O3 concentrations that are 10-20 ppb
lower, on average, than at the high elevation site. Ozone concentrations at the low elevation site
April 2022
2-20 External Review Draft - Do Not Quote or Cite

-------
1	exhibit a slight diurnal pattern similar to that seen at the urban and suburban sites (generally
2	related to photochemical O3 formation that increases concentrations in the late morning and
3	afternoon), while O3 concentrations at the high elevation site do not exhibit any diurnal pattern.
4	The lack of a diurnal pattern observed at the high elevation site is typical of high elevation rural
5	sites throughout the U.S., suggesting that observed O3 concentrations at such sites are primarily
6	driven by transport from upwind areas rather than being formed from local precursor emissions.
7	The presence of peak O3 concentrations that are higher at the high elevation site than at the low
8	elevation site at all hours of the day indicates that the high elevation site may be influenced by
9	transport from the free troposphere to a greater extent than the low elevation site.
10
April 2022
2-21 External Review Draft - Do Not Quote or Cite

-------
AQS Site ID: 06-037-1302 Site Name: Compton
AQS Site ID: 06-037-0016 Site Name: Glendora
150
140
9 10 11 12 13 14 15 16 17 18 19 20 21 22 23
Hour (Local Standard Time)
> Site ID: 33-007-4002 Site Name: CAMP DODGE, GREENS GRANT
n 50
O
itttYt
1
8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23
Hour (Local Standard Time)
: i
i
g
150
140
130
Bo-
no-
s'
Q-100
Q.
o 90-
"5
ฃ 80-

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
2.4.4 Seasonal Patterns
Tropospheric O3 concentrations also tend to experience seasonal patterns due to seasonal
changes in meteorological conditions and the length and intensity of daylight. High O3
concentrations are most commonly observed on hot, sunny, and stagnant days during the spring
and summer. Figure 2-14 shows boxplots of MDA8 O3 concentrations by month of the year for
four monitoring sites that represent different kinds of seasonal patterns commonly observed in
the U.S. This figure is based on data from 2015-2017. The boxes represent the 25th percentile,
median, and 75th percentiles and each box has "whiskers" which extend up to 1.5 times the
interquartile range (i.e., the 75th percentile minus the 25thpercentile) from the box, and dots which
represent outlier values. Panel A shows the seasonal pattern for an urban site in Baltimore, MD,
which reflects the typical seasonal pattern observed at many urban and suburban monitoring sites
across the U.S. The highest O3 concentrations are observed during May to September, when the
days are the longest and solar radiation is strongest.
Panel B shows the seasonal pattern for an urban site in Baton Rouge, LA. In parts of the
southeastern U.S., the highest O3 concentrations are often observed in April and May due to the
onset of warm temperatures combined with abundant emissions of biogenic VOCs at the start of
the growing season. This is often followed by lower concentrations during the summer months,
which is associated with high humidity levels that tend to suppress O3 formation in the region
(Camalier et al., 2007). Some areas, particularly in the states bordering the Gulf of Mexico, may
experience a second peak in O3 concentrations in September and October.
Panel C shows the seasonal pattern for a high elevation rural site in Colorado. The
highest O3 concentrations in rural areas are typically observed in the spring. This can be due to
several factors, including those mentioned previously, and additionally, long-range transport
from Asia is most prevalent at this time of year. Stratospheric Tropospheric Exchange events,
which most often affect high elevation areas in the western U.S., are also most common during
the spring.
Finally, Panel D shows the seasonal pattern for a monitoring site in Utah where high
wintertime O3 concentrations were observed. Over the past decade, high O3 concentrations have
been observed in two mountain basins in the western U.S. during the winter months (December
to March). These wintertime O3 episodes require a unique set of conditions, including a shallow
inversion layer, snow cover, calm or light winds, and pervasive local NOx and VOC emissions
(in these cases, from oil and gas extraction). These conditions are relatively uncommon, and
elevated wintertime O3 levels may not occur in some years.
April 2022
2-23 External Review Draft - Do Not Quote or Cite

-------
AQS Site ID: 24-005-3001 Site Name: Essex
AOS Site ID: 22-033-0009 Site Name: Capitol
120 -
110 -
2T100 -
Q-
Q-
o 90-
to
ง 80-
c
o
O



/()-
ฃ_

N

O
60-


.ฃ=

00
50-
s-

>

h


40-


s

J>N
CO
30-
ฃJ


20-

10-

0
























B














































•
•

'










• —


t
•















E
5








PE


	i





t—L






—




L_
L


' -











•

Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec
AQS Site ID: 49-047-2003 Site Name: Ouray
120
110
r 100 -
; 90-
?> 80-
70-
60-
50-
40-
f 30-
Q
20-
10-
Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec
2
3
4
5
Figure 2-14. Seasonal patterns in MDA8 O3 concentrations at selected monitoring sites
(2015-2017): A) an urban site in Baltimore, MD; B) an urban site in Baton
Rouge, LA; C) a rural site in Colorado; and D) a site in Utah experiencing
high wintertime O3.
April 2022
2-24 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
2.4.5 Variation in Recent Daily Maximum 1-hour Concentrations
To provide a characterization of recent O3 concentrations in the U.S. for periods shorter
than 8 hours, this section presents recent O3 monitoring data in terms of daily maximum 1-hour
average (MDA1) concentrations, and their variation across monitoring sites that vary with regard
to design values for the current O3 standards.
Figure 2-15 shows boxplots of MDA1 values at U.S. monitoring sites based on 2018-
2020 data stratified by each site's 8-hour O3 design value. The boxes representing the 25th
percentile, median, and 75th percentile MDA1 values increase slightly with higher design values.
Although the overall range (minimum and maximum) of observed MDA1 values does not appear
to change much, there is an increasing presence of higher MDA1 values extending up to around
160 ppb for the rightmost bin which includes only sites that exceed the current standards. The
upper percentiles, including the 75th and the 99th percentiles (represented by top of box and upper
whisker, respectively), in particular, are increased for the sites that do not meet the current
standards (up to nearly 80 ppb and 120 ppb in the rightmost bin). In contrast, the boxplots show
that there are only a small fraction of MDA1 values above 120 ppb for sites that meet the current
standards.
Figure 2-16 shows a scatter plot of the number of days at each monitoring site that have a
MDA1 value of 120 ppb or greater based on 2018-2020 data compared to the site's 2018-2020
design value. According to the figure, a small proportion of O3 monitoring sites in the U.S.
observe MDA1 values at or above 120 ppb more than once per year, but these sites all exceed the
current 8-hour standards. There are no sites that were meeting the current standards based on
2018-2020 data that had MDA1 values above 120 ppb more than three times over the same 3-
year period (Appendix 2A, Table 2A-2).
Figure 2-17 shows the national trend in the annual 2nd highest MDA1 O3 concentration,
which was the metric used to track progress towards meeting the 1-hour O3 NAAQS, originally
set in 1979 and later replaced by the current 8-hour metric in 1997 (62 FR 38856, July 18,
1997) 22 monitoring sites represented in Figure 2-17 are the 834 sites with complete data
from 2000 to 2020 (as summarized in Appendix 2A, Section 2A.2). The shapes of the trend lines
in Figure 2-17 are similar to those shown for the annual 4th highest MDA8 values in Figure 2-11.
The national median annual 2nd highest MDA1 value decreased by 27% from 2002 (105 ppb) to
2013 (77 ppb), which is comparable to the decrease observed in the national median annual 4th
highest MDA8 value (25%) during the same period.
22 The 1-hour O3 standards were formally revoked in 2005 (70 FR 44470, August 3, 2005).
April 2022
2-25 External Review Draft - Do Not Quote or Cite

-------
200
1
2
3
<60
61-70
71-84
>84
8-hour 03 Design Value (ppb)
Figure 2-15. Boxplots showing the distribution of MDA1 concentrations (2018-2020),
binned according to each site's 2018-2020 design value.
50-
_Q
Q.
Q.
o 40 -
ro
>
^ 30-
o
JZ
'a 20"
O

-------
OT-c\Jco*3-LOtor-eoG")OT-c\Jco*3-mcDr-coo")0
OOOOOOOOOOt-t— t— t— t— r— t— t— t— t— CnJ
ooo oooooooooooooooooo
1	C\ICNCNJCNJCMC\|CMCMC\ICNJCNJCMC\IC\1CM(MCMC\ICMCMCM
2	Figure 2-17. National trend in the annual 2nd highest MDA1 O3 concentration, 2000 to
3	2020. The solid blue line represents the median value, dotted blue lines
4	represent the 25th and 75th percentile values, and the light blue shaded area
5	represents the range from the 10th to the 90th percentile values.
6
7
April 2022
2-27 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
2.5 BACKGROUND 03
There are a number of definitions of background O3 used in various contexts that differ
by the specific emissions sources and/or natural processes the definition includes (e.g., see ISA,
Appendix 1, section 1.2.2). In this reconsideration, as in past reviews, the EPA generally
characterizes O3 concentrations that would exist in the absence of U.S. anthropogenic emissions
as U.S. background (USB). An alternative phrasing for USB is the O3 concentrations created
collectively from global natural sources and from anthropogenic sources existing outside of the
U.S. Such a definition helps distinguish the O3 that can be controlled by precursor emissions
reductions within the U.S. from O3 originating from global natural and foreign precursor sources
that cannot be controlled by U.S. regulations (ISA, section 1.2.2).
Because monitors cannot distinguish the origins of the O3 they measure,23 photochemical
grid models have been widely used to estimate the contribution of background sources to
observed surface O3 concentrations. This section summarizes results of a state-of-the-science
modeling analysis to estimate the magnitude of present-day USB and its various components.
Conceptually, these USB estimates represent O3 concentrations that occur as a result of global
natural sources (or processes, see section 2.5.1 for more details) and those anthropogenic sources
existing outside the U.S., i.e., the O3 concentrations that would occur in the absence of any U.S.
anthropogenic O3 precursor emissions. Modeling results summarized in this section include
average estimates of MDA8 USB concentrations for several temporal periods including seasons.
Average USB estimates are also presented for days on which the total model-predicted MDA8
O3 concentration was greater than either 60 ppb or 70 ppb, and for the days on which the 4th-
highest MDA8 O3 concentration was predicted to occur. Additionally, this modeling analysis
investigated the contributions to USB of some specific groups of sources, such as international
anthropogenic sources, and how those contributions vary by season and by location.
The section, which presents the information and analysis that were also presented in the
parallel section of the 2020 PA, is organized as follows. Section 2.5.1 provides an overview of
the various sources that contribute to USB, including currently available information on the
magnitude, seasonal variability, and spatial variability of their contributions to USB. Section
2.5.2 summarizes the methodology for the modeling analyses used to quantify USB and
component contributions. More detailed information about the modeling methodology is
presented in Appendix 2B. Section 2.5.3 summarizes USB estimates using methodology
23 Ozone concentrations that do not include contributions from U.S. anthropogenic emissions cannot be determined
exclusively from O3 measurements because even relatively remote monitoring sites in U.S. receive transport of
U.S. anthropogenic O3 from other locations.
April 2022
2-28 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
described in section 2.5.2, including estimates specific to certain subgroups of sources. Section
2.5.4 summarizes key findings of the analyses.
2.5.1 Summary of U.S. Background O3 Sources
Jaffe et al. (2018) reviewed the literature on sources that contribute to USB. While the
term "background" may imply a low concentration well-mixed24 environment, background
sources can create well-defined plumes and/or contribute to the well-mixed environment. The
USB definition, which is based on sources, includes both the well-mixed environment and more
well-defined plumes. Figure 2-18a (adapted from Jaffe et al. (2018)) illustrates sources of USB
O3 (blue) and U.S. anthropogenic sources of O3 (yellow). Figure 2-18b shows two theoretical
examples where background sources contribute to the total ground-level O3. The first example
(Ex 1) highlights a typical monitoring site with lower USB, and the second example (Ex 2)
presents a scenario in which USB is a large contributor. Both examples oversimplify methane,
which has both natural and anthropogenic and both domestic and foreign contributions. Source
contributions to USB vary in space and time, and the stacked bar plot in this figure
oversimplifies the complex relationship between USB and total O3. Even so, USB sources can
broadly be discussed as global natural sources (see sections 2.5.1.1 to 2.5.1.6) and international
anthropogenic sources (see section 2.5.1.7). In the simplest interpretation, the natural sources are
background regardless of where they occur, or which definition of background is being used
(e.g., USB or natural background25). By contrast, ozone formed from anthropogenic emissions is
only considered as background when the emissions sources are not from sources within the focus
area. However, this paradigm is complicated by the fact that many sources of O3 precursors are
the result of interactions between human and natural systems (for instance forest management
practices can impact both biogenic VOC emissions from trees and wildfires). In the context of
USB, anthropogenic background is synonymous with O3 originating from international
anthropogenic emission sources. The relative contribution of international and natural
background sources can vary dramatically from place to place and are most notably larger at
locations near borders (international) or high elevation (natural). At non-border locations and
many border locations, the natural background is usually the dominant background source.
24	We use the term "well-mixed" here to refer to conditions when the contributions from various types of sources are
mixed due to chemistry or physical processes to the point where it is not possible to discern the contribution to O3
from each individual source.
25	Natural background is the O3 that would exist in the absence of anthropogenic emission sources.
April 2022
2-29 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
Stratospheric Intrusion
Lightning |
Foreign Pollution
Wildfire Impacts j
| Agricultural Emissions [
Recirculated
Domestic Pollution
Interstate Transport,
Export of ~
Domestic Pollution
Local Photochemical
Pollution
| Prescribed Burning^
Biogenic Emissions
(a) U.S. Oa sources shown with yellow boxes or arrows represent domestic sources.
Sources shown with blue boxes or arrows represent USB sources. Note that locations for
each process are not specific to any one region. The base map shows satellite-observed
tropospheric NO2 columns for 2014 from the Ozone Monitoring Instrument (OMI) onboard
the NASA Aura satellite (Credit: NASA Goddard's Scientific Visualization Studio/T.
Schindler). NO2 column amounts are relative with red colors showing highest values,
followed by yellow then blue. We use the OMI NO2 columns as a proxy to show local O3
precursor emission sources, (b) The bar chart shows two theoretical examples of USB Os
contributions combine with domestic sources to produce elevated O3 at a specific location
on any given day. Each source varies daily and there are also nonlinear interactions
between USB O3 sources and anthropogenic sources that can further add to O3
formation, e.g., wildfires and urban anthropogenic emissions (e.g., Singh etal., 2012).
Minor adaptation from DOI: https://doi.org/10.1525/elernenta.309.f1
models for 03 sources: (a) in the U.S., and (b) at a single location.
The natural and anthropogenic sources of background O3 vary by location and by season.
Emissions from anthropogenic sources largely occur in the same areas year after year. Natural
sources of O3 and precursors, on the other hand, vaiy both in magnitude and in location from day
to day and year to year. As a result, certain types of natural sources may have large O3
contributions measured at a monitor at one point in time but not at other times. The combination
of varying proximity and magnitude means that natural sources can contribute to background in
the form of localized plumes of elevated O3 that contribute to O3 at monitoring sites on an
episodic basis. In the absence of locally well-defined plumes, global natural and international
anthropogenic sources are constantly contributing to the well-mixed background.
80
70
60
50
jQ
Ql40
Q.
30
20
10
0
US Anthro
Intl Anthro
Wildfires
Lightning
Methane
Biogenics
Stratospheric
rH fN
X X
LU LU
(b)
Figure 2-18. Conceptual
April 2022
2-30 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
USB varies by location and by season due to both the nature of sources and the loss
processes. The nature of emission sources leads to seasonal and spatial patterns that will be
described further below. The contribution of these sources is modulated by transport patterns that
interact with deposition and chemical losses. For illustration, two emission sources of identical
magnitudes may have different contributions if one emits near the surface in summer and the
other emits in the free troposphere in spring. Warmer moister air in the summer at the surface
enhances O3 chemistry losses and deposition of O3 to the surface increases losses further. In
contrast, cooler, drier temperatures in the spring and free troposphere lengthen O3 lifetimes and
faster winds in the free troposphere enable longer transport. The seasonality of temperature and
transport patterns gives O3 USB a distinct seasonal cycle that results from both sinks and
sources.
The sections below summarize the state of the science estimates of USB contributions.
Each source type is described with respect to its seasonality as well as its local vs well-mixed
contribution potential. Jaffe et al. (2018) reviewed contributions of various sources to USB O3
from modeling studies and the references therein are used to illustrate the range of O3
contributions from each source. The literature-based estimate ranges provide context to the
estimates of USB that are reported in section 2.5.3.
2.5.1.1 Stratosphere
The only direct source of O3 to the troposphere with appreciable contributions to O3
concentrations is STE (other sources are indirect via precursors). STE occurs when stratospheric
air, which is relatively rich in O3, is transported across the tropopause where it enhances
tropospheric concentrations. Most STE events create enhancements that do not immediately
reach the surface. Instead, STE-enhanced O3 mixes into the free troposphere where it is
dispersed. In cases when the transported air reaches the surface before enough dispersion occurs,
it creates a localized plume of O3 referred to as a Stratospheric Ozone Intrusion (SOI). The total
stratospheric contribution includes both the well-mixed contribution from the distant stratosphere
exchanges as well as any localized SOI plume.
The total global O3 flux from the stratosphere to the troposphere is estimated at 510ฑ90
teragrams per year (Tg/y) compared to 4620ฑ600 Tg/y (post-2000 literature in Table 2 in Wu et
al., 2007) produced within the troposphere. The majority of the earth's surface is outside the U.S.
and only STE that take place over the U.S. are likely to create a large magnitude local
enhancement at a U.S. monitor. 26 A SOI that occurs outside the U.S. would likely be dispersed
26 Recently methods have been developed for identifying and estimating SOIs that have clear localized contributions
to O3 concentrations with the potential to contribute to standards' exceedances. These are described in documents
available at: https://www.epa.gov/air-quality-analysis/guidance-preparation-exceptional-events-demonstrations-
stratospheric-ozone.
April 2022
2-31 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
into the well-mixed background and reduced through chemical loss and deposition before it
reaches many monitors.
Modeling and observational studies show that SOI can episodically contribute large
amounts of O3 at a subset of U.S. monitors, but stratospheric mixing more frequently contributes
smaller quantities of O3. Modeling studies focused on seasons with frequent SOI find median
total stratospheric contributions to MDA8 are 10-22 ppb in the West and 3-13 ppb in the East
with episodic contributions up to 40 ppb mostly in the West (Table S2, Jaffe et al., 2018).
Because these studies focus on the most active season, these medians are expected to be upper
bounds for the annual average. Further, SOI are most common in the spring when MDA8 O3
concentrations above 70 ppb are less common (ISA, section 1.3.2).
2.5.1.2 Biogenic VOC
Biogenic VOCs are the quintessential "natural" source of O3 precursors. At global scales,
biogenic sources are the largest contributor to VOCs - even though local anthropogenic sources
of highly reactive VOCs can be very important in some areas. VOCs are also an important
source of carbon monoxide. Biogenic VOCs are emitted by various types of vegetation and
emissions peak in summer which is also when O3 production is fast and O3 lifetimes are short.
The large abundance of biogenic VOCs leads to NOx-limited O3 production in most of
the world. That is, concentrations of biogenic VOCs are in excess with respect to concentrations
of NOx; therefore, O3 production is controlled by the availability of NOx. The methodologies27
typically used by the air quality community estimate contribution based on sensitivity of O3
production. As a result, the sensitivity-based contribution estimate of biogenic VOC sources to
O3 shows relatively small contributions considering the large amount of emissions.
Estimates of biogenic VOC contributions in the literature are generally small compared to
NOx. For example, Lapina et al. (2014) found that North American Background (NAB)28 for
W12629 O3 was relatively insensitive to VOC (10.8% of NAB sensitivity) compared to NOx
(79.8% of NAB sensitivity). This well-known global-scale sensitivity to NOx would not exist if
concentrations of biogenic VOCs were a broadly limiting factor. Even though background O3 is
not particularly sensitive to small changes in the biogenic VOC, natural sources of VOCs are a
critical component of all background O3 estimates.
27	Source apportionment techniques and derivative-normalization techniques use sensitivity to attribute
concentrations to sources. When a concentration is insensitive to VOC sources, the contribution estimate solely
from that source of VOC will be zero.
28	North American Background is analogous to USB; but NAB is generally characterized as the O3 concentrations
that would exist in the absence of North American anthropogenic emissions.
29	W126 is a daytime weighted average concentration where higher concentrations are given greater weight based on
a sigmoidal curve (see Chapter 4).
April 2022
2-32 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
2.5.1.3	Wildland Fires
Fires emit a complex mixture of nitrogen oxides, nitrogen reservoir species (e.g., PANs),
and VOCs that are all precursors to O3. In the northern hemisphere, the fire season generally
starts in spring and extends into fall with the specific timing varying widely by region. Fires also
exhibit significant year to year variability, with emissions varying by an order of magnitude
between high and low fire years in some places (van der Werf et al., 2017). While smoke from
fires affects most of the contiguous U.S. at some point during the year, the fire season in the
western U.S. occurs primarily late in the summer. Fires across western states and parts of Canada
can contribute both to regional background and episodic surface O3 enhancements (McClure and
Jaffe, 2018).30
Ozone production in fire plumes depends on a range of factors including the type of fuel
combusted, plume age, and interactions with other air masses (e.g. urban plumes) (Jaffe and
Wigder, 2012). While some studies have estimated wildfire O3 contributions to seasonal mean
O3 of up to several ppb during high fire years in the Western U.S. (Jaffe et al., 2018), O3
production from individual fires varies substantially (Akagi et al., 2013). Several studies have
shown that locations near large fires can even experience suppressed O3 formation, perhaps due
to titration from fresh NO emissions and/or reduced solar radiation resulting from high aerosol
concentrations (McClure and Jaffe, 2018;Buysse et al., 2019). Large variability in O3 precursor
emissions from fires combined with complex in-plume dynamics and chemistry make accurately
quantifying O3 production from fires extremely difficult at both regional and local scales.31
New data from recent and upcoming field and aircraft campaigns32 are expected to
provide new insights that expand current understanding of contributions from fires to O3
concentrations in the U.S., both in the context of regional background concentrations and
production during individual fire episodes.
2.5.1.4	Lightning Nitrogen Oxides
Lightning is an indirect natural O3 precursor source. Lightning produces NOx from
molecular nitrogen and oxygen, similar to traditional combustion processes. Because NOx is the
30	Fires may occur on wildlands naturally or accidentally, or fires may be planned (prescribed) for various purposes
and set intentionally. In the USB modeling work described in section 2.5.2.1 below, emissions associated with
prescribed fires are categorized as anthropogenic emissions and are not included in estimating USB.
31	Recently methods have been developed for identifying and estimating wild or prescribed fire contributions to O3
concentrations with the potential to contribute to standards' exceedances. These are described in documents
available at https://www.epa.gov/air-quality-analysis/final-2016-exceptional-events-rule-supporting-guidance-
documents-updated-faqs.
32	Western Wildfire Experiment for Cloud Chemistry, Aerosol Absorption and Nitrogen (WE-CAN,
https://www.eol.ucar.edu/field_projects/we-can) in 2018 and Fire Influence on Regional to Global Environments
and Air Quality (FIREX-AQ, https://www.esrl.noaa.gov/csd/projects/firex-aq/) in 2019.
April 2022
2-33 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
globally limiting precursor for O3 production and lightning emits where there are few other
sources, O3 production is quite sensitive to this source. Over the U.S., lightning NOx (LNOx)
emissions peak in summer with convective activity and are characterized as having high
interannual variability (Murray, 2016). Allen et al. (2012) showed that the majority of LNOx is
emitted in the free troposphere (i.e., troposphere above the planetary boundary layer). Thus,
LNOx is produced in a NOx-limited environment where any O3 formed as a result will be
efficiently transported and loss pathways are limited.
The total NOx created by lightning is highly uncertain (Murray, 2016). Murray (2016)
discusses the uncertainty in NO yield per flash rate and the role of large spatial gradients in the
yield. The effect of such uncertainties is evident in the range of global lightning emissions
(std/mean=0.4). Murray (2016) also discusses the uncertainty in the vertical distribution of NO
production and post-production redistribution.
Jaffe et al. (2018) reviewed contributions from lightning to surface USB O3 based on
modeling studies using various flash rate yields, which shows large single day contributions to
modeled MDA8 O3 (up to 46 ppb, Murray, 2016) and smaller contributions to annual means (1-6
ppb) and seasonal means (6-10 ppb). Lapina et al. (2014) showed that, in their modeling, W126
had a 15% contribution from lightning NOx over the U.S.33 A 15% contribution is consistent
with the annual and seasonal mean contributions to MDA8 reported by Zhang et al. (2014) and
Murray (2016). Lapina et al. (2014) also noted that 40% of the lightning NOx sensitivity comes
from lightning strikes outside the U.S. The findings from these studies highlight the primary
importance of lightning NOx as a contributor to the well-mixed background concentrations
(Murray, 2016).
2.5.1.5 Natural and Agricultural Soil NOx
Nitrogen oxides from soils are a naturally occurring source that is enhanced by
anthropogenic activity. Truly natural soil NOx is created as a byproduct of nitrogen fixation in
natural environments. The fixation and byproduct release are affected by flora composition,
nitrogen availability, and environmental conditions (e.g., humidity). Human activity affects the
amount and location of soil NOx emissions by changing land cover and by increasing the
availability of nitrogen for fixation though the application of fertilizer to crop lands or additions
33 The numbers shown in this report are derived from reported values in Lapina et al. (2014) which showed
sensitivity of W126 to anthropogenic NOx sources was 58% (of that, 80% US; 9% CAN; 4% MEX) and natural
NOx sources was 25%. The remaining 17% was attributed natural isoprene (1.3%), VOCs/CO from fires (Fig 9:
~3%) and international VOC/CO (Fig 9: -14%). So non-North American anthropogenic NOx (58% * 7% non-NA
= 4%) and natural NOx (25%) create a total NAB NOx sensitivity of 29% and total NAB sensitivity of 35% (29%
/ 79.8%). Of the total sensitivity (parentheses contain percent of NAB NOx sensitivity, see Fig 12), lightning was
15% (52.9%), soil NOx was 8% (28.2%), fire NOx was 1% (4.3%) and international anthropogenic NOx was 4%
(14.5%).
April 2022
2-34 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
of nitrogen via deposition of emissions from other sources. The effect of human land cover
alteration is readily apparent in soil NOx emission measurements. Steinkamp and Lawrence
(2011), highlight that soils in pristine natural ecosystems emit more NOx compared to similar
ecosystems that have been disturbed by human activity. At the same time, human managed crop
lands emit more than natural ecosystems (pristine or disturbed) environments because of the
applied fertilizer.
Soil NOx clearly has both anthropogenic and natural sources, but these are rarely
separated in the literature. First, Hudman et al., 2012 estimate that the majority (-80%) of soil
NOx emissions are currently attributed to land surfaces without considering active fertilization or
deposition of anthropogenic nitrogen. Second, the emissions and attribution are relatively
uncertain. Finally, anthropogenic soil NOx is associated with agricultural ammonia application
that is not directly regulated in the United States. As a result, the attribution of soil NOx as a
"background" source is imperfect. In this assessment, no distinction is made between natural and
fertilizer-enhanced soil NOx and instead we include both within "natural sources."
Hudman et al. (2012) estimated the global soil NOx emissions at 10.7 TgN/y. As noted
above, soil NOx emissions are linked to nitrogen availability in the soil, which is increased by
anthropogenic activities. Hudman et al. (2012) attributed 1.8 TgN/y to anthropogenic soil
fertilization and 0.5 TgN/y to atmospheric deposition. Like lightning, most soil NOx emissions
occur outside of the U.S. Unlike lightning, soil NOx has a smaller long-range transport
component because it is emitted at the surface. For example, Lapina et al. (2014) calculated that
W126 had an 8% sensitivity to soil NOx (see footnote 26) and noted that a small fraction (only
7%) was from emissions outside the U.S. The more local sensitivity is likely due to the emission
height and spatial distribution of soil NOx.
2.5.1.6 Post-Industrial Methane
Like VOCs, CH4 is a hydrocarbon that can form O3 in the presence of NOx and sunlight.
While some atmospheric methane is emitted naturally from wetlands, wildfires, geogenic
sources, and insects, significant global methane enhancements following the industrial revolution
are clearly associated with increased emissions from anthropogenic fossil fuel combustion
(Pachauri et al., 2015). Other human activities such as livestock cultivation, landfills and land
use modification (e.g., rice paddies) also release methane. More recently, changing climate
conditions have led to increased emissions from natural sources (e.g., permafrost melting) in
some areas (Reay et al., 2018), although the exact magnitude of these effects on global methane
concentrations, and consequently O3 in the U.S., over longer time scales remains uncertain.
Due to its long atmospheric lifetime (-10 years), methane is well-mixed at seasonal and
annual time scales. As a result, isolating contributions to atmospheric methane concentrations
April 2022
2-35 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
from individual geographic areas or specific emission sectors is very difficult (Turner et al.,
2017). However, sensitivity simulations with chemical transport models can be used to assess the
overall influence of global methane concentrations on regional O3 budgets. For example, Lin et
al. (2017) used the GFDL-AM3 chemistry-climate model to estimate that increasing global
methane concentrations contributed -20% to background MDA8 O3 trends during boreal spring
and summer at several western U.S. sites during the period 1988 to 2012. In general, post-
industrial anthropogenic methane is estimated to contribute ~5 ppb to surface O3 in the U.S., an
estimate that primarily comes from modeling studies (Jaffe et al., 2018 and references therein).
A major limitation with existing model-based estimates of the influence of global
methane on current U.S. O3 concentrations is our limited understanding of historical methane
emissions. The U.S. and the rest of the world's anthropogenic methane emissions have not been
tracked quantitatively in detail until relatively recently. As a result, the pre-industrial methane
concentration is relatively unconstrained. Further, post-industrial methane can be attributed to
direct emissions and emissions from natural sources (e.g., permafrost). Many modeling studies,
including this one, do not explicitly track methane sources and sinks, further complicating
attribution in an air quality context. Therefore, the post-industrial methane contribution is
difficult to quantitatively attribute. The post-industrial enhancement of methane is clearly related
to direct anthropogenic emissions and alteration of natural emissions by human activity, which
includes both foreign and domestic contribution.
2.5.1.7 International Anthropogenic Emissions
International anthropogenic emissions are the only anthropogenic contribution to USB.
For the purposes of discussion, NOx and VOCs will be discussed separately from methane
(methane is covered in section 2.5.1.6). NOx and VOC emission estimates from outside the U.S.
are derived from international collaborative efforts like the Hemispheric Transport of Air
Pollutants (HTAP) task force of the United Nations Economic Commission for Europe
(Janssens-Maenhout et al. 2015). HTAP harmonized national emission databases from individual
countries with global estimates that cover areas without their own estimates. Collecting and
harmonizing these emission datasets requires coordination and technical expertise, which
recently occurred twice (HTAP Phase I and HTAP Phase II) and a new HTAP emission
inventory is currently underway. Global estimates that incorporate national information are
available (e.g., Community Emissions Data System and Emissions Database for Global
Atmospheric Research), but do not always have as much participation from individual countries.
This is particularly important because individual countries are most aware of regulations and
controls that have been promulgated within their borders.
April 2022
2-36 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
International anthropogenic sources of O3 include emissions within the borders of other
countries (e.g., onroad sources, power plants, etc.) as well as sources in international waters and
air space. Sources within the borders of other countries can be easily attributed to those countries
using geographical bounds based on emission source location. Some studies (e.g., Lin et al.,
2014), however, have done more complex analyses to spatially attribute emissions globally based
on the consumption of produced goods. For the purposes of this document, international
emissions are attributed based on the emission source location. Using emission source location,
maritime shipping and aircraft sources require more artificial distinctions. Typically, aircraft
takeoff and landing are assigned completely to the country where it occurs. Aircraft cruising
emissions are attributed based on geographic boundaries. This assumes that both inbound and
outbound flights change source type (domestic/international) when they cross a border.
2.5.2 Approach for Quantifying U.S. Background Ozone
Updating USB estimates is motivated by interannual variability, trends in international
anthropogenic emissions, and continual improvements in simulating processes affecting USB.
USB sources are expected to vary from year to year because natural emissions vary in response
to meteorology (e.g., temperature) and long-range transport patterns alter the efficiency of
transport from long-range USB sources (Lin et al., 2015). In addition, the scientific
characterization of background emission sources continues to evolve. As a result, we provide an
updated assessment of USB for 2016 using the latest stable version of the Community Multiscale
Air Quality (CMAQ) model applied at hemispheric to regional scales.
This assessment uses a firmly source-oriented definition of USB based on modeling. The
source composition of a model estimate can be quantified using tagging techniques or by
sensitivity analysis. By contrast, the source composition of measured O3 is difficult to isolate. In
most areas at most times, measured O3 concentrations are the result of contributions from a
variety of anthropogenic and non-anthropogenic sources. Measurements from locations
sometimes suggested to be representative of USB often have contributions from U.S.
anthropogenic sources. As a result, some researchers have filtered measurements to focus on
times when US contributions are minimized (e.g., based on wind direction or other indicators).
The measurement filtering approach is based on conceptual or quantitative models of source
contributions as a function of wind direction or another environmental variable. After correction,
the degree of contamination is minimized but not precisely known. Recently, urban
measurements have been paired with simplistic statistical models to estimate background
(Parrish et al., 2017). However, Jaffe et al. (2018) concluded that statistical adjustment cannot be
directly interpreted as "background" - even though the estimate is useful for bounding simulated
background. Due to the complications of quantifying background based on ambient air
April 2022
2-37 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
measurements, the sources that contribute to background are most clearly defined using an air
quality model. Using separate nomenclature (baseline: monitors; background: models) helps to
clearly delineate between these approaches that each have their strengths and weaknesses.
This section quantifies O3 from sources using a sensitivity approach. The multiscale
system is applied to predict total O3 and then applied multiple times to predict O3 without U.S.
anthropogenic emission sources. The difference between total O3 and O3 without the U.S.
anthropogenic emissions is used to characterize the USB.
2.5.2.1 Methodology: USB Attribution
This assessment attributes O3 to USB sources using one of several available techniques.
Jaffe et al. (2018) reviewed the methods for identifying USB contributions. The methodologies
reviewed range in complexity from simply turning off U.S. anthropogenic (or specific sources)
emissions, to normalizing derivatives from instrumented models, to complex tagging techniques
(e.g., CAMx OS AT, APCA, or Grewe, 20 1 3).34 This analysis follows the zero-out approach for
simplicity of interpretation and consistency with previous EPA analyses. In urban areas, this
approach will estimate higher natural and USB contributions than total O3 when NOx titration is
present. The estimate, therefore, is an estimate of what concentrations could be without U.S.
anthropogenic emissions and not the fraction of observed O3 that is USB.
This analysis is designed to quantify 03 specifically and separately from global natural,
international anthropogenic, and U.S. anthropogenic sources. The precursors that this analysis
focuses on are NOx and VOC because they have a response on timescales relevant to the
NAAQS planning schedules (i.e., not methane). Table 2-1 lists simulations and the sources they
exclude at the various spatial scales modeled (i.e., hemispheric - 108 km resolution, regional -
36 km resolution and regional - 12 km resolution). For international shipping and aviation, the
U.S. domain is either included (ZROW) or excluded (ZUSA). These simulations form the basis
for estimating the contributions of USB and its components. Given the long atmospheric lifetime
and attributability to U.S. sources, methane is not separately identified nor is it perturbed in any
simulations. This has the effect of attributing methane to natural processes, which are a
background source.
34 For a discussion of methods and the effect on estimates, see (Jaffe et al., 2018).
April 2022	2-38 External Review Draft-
Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
Table 2-1. Simulation names and descriptions for hemispheric-scale and regional-scale
simulations.
Simulation Description
Performed at Hemispheric A and RegionalB Scales
BASE	All emission sectors are included
ZUSA	All U.S. anthropogenic emissions are removed including prescribed fires.c
ZROW	All international anthropogenic emissions are removed including prescribed fires where
possible.
ZANTH	All anthropogenic emissions are removed including prescribed fires.
Performed at Hemispheric Scale only
ZCHN	All Chinese anthropogenic emissions are removed.
ZIND	All India anthropogenic emissions are removed.
ZSHIP	Zero all near-U.S. commercial marine vessel category 3 and all global shipping.
ZFIRE	Zero all fire emissions (agricultural, prescribed, and wild).
A Hemispheric-scale simulations use 108 km grid cells defined on a polar stereographic projection.
B Regional-scale simulations use a nested 36 km and 12km simulation on a lambert conformal projection.
c Emissions estimated to be associated with intentionally set fires ("prescribed fires") are grouped with anthropogenic fires.
Table 2-2 describes the calculations that are used to derive contributions. It is important
to note that contributions are not strictly additive. Large NOx sources can create non-linear
conditions that decrease O3 concentrations due to titration which is most relevant at night and in
the winter. In some cases, removing a source only increases the efficiency of other sources. In
that case, some anthropogenic contribution exists unless all anthropogenic sources are removed.
This residual anthropogenic contribution occurs in the model for both International and U.S.
sources. The results presented in this section focus on Base, USB, International, Natural
contributions. Some components of International and Natural were separately analyzed.
Canada/Mexico are separately quantified at both hemispheric and regional scales. The India,
China, Fire, and shipping contributions are analyzed only at the hemispheric scale and are
presented in Appendix 2B. The analyses in Appendix 2B support the interpretation in the
discussion below.
April 2022
2-39 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
Table 2-2. Expressions used to calculate contributions from specific sources.
Label
Name
Description
Expression

BASE
Total
Total Concentration
BASE

USB
USB
U.S. Background
ZUSA

USA
USA
U.S. Contribution
BASE-ZUSA

Intl
International
Rest of the World Contribution
BASE-ZROW

Natural
Natural
Natural Contribution
ZANTH

Res-Anth

Anthropogenic contribution that is
not attributed directly to either the
U.S. or International due to non-
linear chemistry
BASE - ZANTH - Ir
ltl - USA
IND
India
India Contribution
BASE-ZIND

CHN
China
China Contribution
BASE -ZCHN

Ship
Ship
Ship Contribution
BASE-ZSHIP

FIRE
Fire
Global fire contributions
BASE-ZFIRE

2.5.2.2 Methodology: Strengths, Limitations and Uncertainties
The model was evaluated to assess the accuracy of predictions and infer possible biases
in USB estimates. Evaluations included comparison to satellite retrievals, O3 sondes35,
CASTNET monitors, and AQS monitors. Results were also qualitatively compared to the
Tropospheric Ozone Assessment Report (TOAR) database, which has global O3 observations
that have been well characterized36 but Phase I, which was completed and available at the time of
analysis, only extends through 2014. The evaluation of the hemispheric simulation that provides
boundary conditions to the 36 km model simulation relies heavily upon the satellites, O3 sondes
and CASTNET monitors. Since the satellite data can be used to provide concentration estimates
in areas without surface monitors, these data are particularly useful for evaluating O3 column
totals in the hemispheric modeling. The sonde data provide a means to evaluate predictions aloft
which are important for understanding model performance of long-range transport. The regional
evaluation analysis focuses on data measured at CASTNET and AQS monitors.37 Evaluation
using the AQS monitors provides information on how the model performs at urban/suburban O3,
which may exhibit large space/time gradients in O3 concentration. CASTNET data are included
35	O3 sondes are balloon-borne instruments that ascend through the atmosphere taking O3 and meteorological
measurements. For more information, see https://www.esrl.noaa.gov/gmd/ozwv/ozsondes/.
36	The TOAR database includes O3 globally where each monitor has been consistently characterized as urban or
rural. The global observations have been processed for several metrics (MDA8, W126, etc.) and gridded to 2-
degree by 2-degree global fields for easy comparison to large-scale models.
37	In the discussion here in section 2.5, the data for CASTNET sites are referred to as "CASTNET data" and data for
all other sites in AQS are referred to as "AQS data" (even though data for many, if not all, CASTNET monitors
are stored in AQS).
April 2022
2-40 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
in the evaluation of both the hemispheric and regional models since monitoring sites in this
network are intended to represent O3 concentrations across broad areas of the U.S. Model
performance evaluation results are summarized in this chapter and provided in more detail in
Appendix 2-B.
The evaluation using sonde data shows that the hemispheric model predictions of O3 are
generally within 20% of the corresponding measurements throughout much of the free
troposphere. Near the tropopause, there is a low bias in the model that is most pronounced in the
spring. The low bias at the tropopause likely suggests an underestimate of stratospheric
exchange. Mean bias drops to below 20% in the middle troposphere (600-300 hPa). The low-bias
in the free troposphere may stem from underestimation of spring time stratospheric contribution
in some regions.
The acceptability of model performance was judged for the 2016 CMAQ O3 performance
results considering the range of performance found in recent regional O3 model applications
(NRC, 2002, Phillips et al., 2008, Simon et al„ 2012, U.S. EPA, 2009, U.S. EPA, 2018b). The
model performance results, as described in this document, demonstrate the predictions from the
2016 modeling platform closely replicate the corresponding observed concentrations in terms of
the magnitude, temporal fluctuations, and spatial differences for 8-hour daily maximum O3. At
CASTNET sites, the model performance is similarly good, but has a distinct seasonal pattern
(see Appendix 2B.3). The normalized mean bias increases from a low-bias in boreal Winter
(West: -16%; East: -14%) to relatively neutral in boreal Fall (West: 0%; East: 7%). These results
are consistent with the free troposphere bias seen in the comparison of model predictions to
sonde data. Despite the conceptual consistency, the low-bias in winter at CASTNET sites is also
influenced by local sources. For example, the Uinta Basin monitors have extremely high winter
observations that are underpredicted by the model. These are most likely due to underestimation
of O3 formed from precursors emitted by local sources as well as the need for finer resolution
meteorological inputs to capture cold pool meteorology conditions that characterize these
events.
Model predictions have historically shown poor performance for capturing the impacts
from O3 of wildfires and stratospheric intrusions. Wildfire contributions have been overpredicted
by models (Baker et al., 2016, Baker et al., 2018). Model predictions of O3 from stratospheric
intrusions have ranged from underestimated to overestimated (e.g., Emery et al., 2012). Models
are not expected to perform well in capturing the contributions from wildfires and stratospheric
38 The DIN431 CASTNET monitor, among others, is in the Uinta basin where wintertime O3 can be caused by
snow-cover enhanced photolysis combined with light VOC emissions from the oil and gas production, (see
Ahmadov et al., 2015).
April 2022
2-41 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
intrusions without a focused effort on properly characterizing the physical properties of
individual events.
This analysis uses an emission inventory with known issues in the fire inventory. The
"2016fe" inventory had double counting of some grassland fires.39 To minimize the effects of
double counting, a filter is applied to the data to remove large episodic natural influences
including fires. The filter removes days where natural contributions deviate from the mean for
that grid cell by whichever is higher: 20 ppb or twice the standard deviation for that grid cell.
Using this approach, 0.1% of grid cell days were removed — 71% of grid cells have no days
removed and fewer than 5% have more than 1% removed. Of the days that were removed, fewer
than 21% had MDA8 concentrations above 70 ppb.
This study does not directly quantify USB uncertainty. Jaffe et al. (2018) highlight that
uncertainties in USB and USB component estimates come from multi-model comparisons.
Dolwick et al., 2015) showed that multi-model estimates converged when applying bias
correction, indicating that differences in USB estimates are correlated with model performance.
No bias correction has been applied here, so in a limited manner bias in ambient predictions can
help set expectations for bias in USB. Based on hemispheric model evaluation, the stratospheric
component in spring is likely underestimated leading to a USB low bias in spring. As a single
estimate, this study relies upon the literature based ฑ10 ppb for seasonal means and higher for
individual days (Jaffe et al., 2018). Further, differences between models that share
parameterizations may not fully quantify underlying uncertainty and the year-to-year variability
complicates comparing model simulations done for different years.
2.5.3 Estimates of USB and Contributions to USB in 2016
Background O3 is known to vary seasonally, spatially, and with elevation (as discussed in
section 2.5.1, above). Seasonal variations are related to temporal changes in both sources and
sinks. Spatial variations are related to differential transport patterns and the proximity to sources
of background O3. Elevation is important in determining USB because it relates to the proximity
to the free troposphere. In addition, the seasonality and spatial relationships of USB and USA
contributions are not always aligned. As a result, USB can be highest on days with lower total
O3. For these reasons, estimates of USB and USB components (i.e., Natural and International)
contributions developed from the current modeling are summarized spatially, over time, and as a
function of total O3.
All analyses of USB and components focus on model predictions over land within the
U.S. The U.S. and adjoining areas are represented in the modeling using grid cells. Only grid
39 More information related to this issue is available on the fire working group wiki page
http://views.cira. colostate. edu/wiki/wiki/917'5# July-12-2018.
April 2022
2-42 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
cells in the U.S. are included in this analysis.40 Grid cells with water as the dominant land use
(e.g., lake or ocean) were simply excluded from analysis to acknowledge the potential bias of
total O3 over water bodies (U.S. EPA, 2018). The USB estimates provided here are all in terms
of a metric, MDA8, closely related to the form of the current O3 standards, and do not directly
apply to other metrics.
Section 2.5.3.1 characterizes the spatial variation of model-predicted MDA8 O3
concentrations and contributions using maps of seasonal averages. Section 2.5.3.2 characterizes
the time variation of the predicted MDA8 O3 and contributions using time series of spatial
averages. Section 2.5.3.3 characterizes the relationship between predicted USB components and
predicted total O3. Section 2.5.3.4 summarizes USB predictions across regions and seasons.
2.5.3.1 Spatial Characterization of O3 Contributions
Figure 2-19 and Figure 2-20 provide seasonally aggregated maps that show the spatial
distribution of total model-predicted MDA8 O3 and contributions from natural, international, and
U.S. anthropogenic sources across the U.S.
Figure 2-19 shows predicted MDA8 values for the 12 km domain averaged for spring
months (March, April, and May) for total O3 and contributions from Natural, International, and
USA. Natural is a relatively large contributor to total O3 in spring with a relatively small range of
values (ratio max:min = 2). International contributes less with a larger range (ratio max:min = 3).
There are spatial gradients primarily along parts of the Mexico border, and an overarching
general West-East gradient. The USA contribution, even in spring, has the largest variation (ratio
max:min > 20) with enhancements in some urban areas.
40 Modeling grid cells are assigned to the U.S. based on the grid cell centers. For grid cells whose area covers the
U.S. and an adjoining area, the grid cell is only assigned to the U.S. if the fraction of anthropogenic NOx
emissions contributed by the U.S. is greater than 80%. This is designed to remove grid cells from the analysis
when the model cannot differentiate the border.
April 2022	2-43 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
Figure 2-19. Predicted MDA8 total O3 concentration (top left), Natural (top right),
International (bottom left), and USA (bottom right) contributions in spring
(March, April, May). Each panel displays the simple spatial average and range
(min, max) in ppb in the lower left-hand corner of the panel.
Figure 2-20 shows the same type of information for the summer (June, July, August). The
summer total concentrations are higher than spring due to increases in USA and Natural
contributions. The international contribution spatial gradients have increased (reflecting shorter
O3 lifetimes), so that the maximum International contribution at the border is higher and the
average contribution is lower compared to spring. Similarly, the West-East gradient of Natural,
International, and USA contributions is enhanced in the summer. In addition, the USA
contributions show distinct gradients in urban areas. Figure 2-20 highlights the increasingly near-
border or high-elevation influence of international contribution during the summer when O3
concentrations are most likely to violate the NAAQS.
Natural: 21 ppb
(14, 28 ppb)
	Lฑ V
USA: 10 ppb
(1. 21 ppb)
April 2022
2-44 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
Figure 2-20. Predicted MDA8 total O3 concentration (top left), Natural (top right),
International (bottom left), and USA (bottom right) contributions in summer
(June, July, Aug). Each contribution has the spatial average and range (inin,
max) in ppb in the lower left-hand corner of the panel.
2.5.3.2 Seasonal and Geographic Variations in Ozone Contributions
Seasonal and geographic variations are an important part of background O3. The
geographic variation helps us to understand where USB contributes appreciably to O3
concentrations. The seasonal variation is particularly important as it determines whether high
USB and MDA8 concentrations above 70 ppb are likely to occur at the same time. This section
begins by characterizing the dependencies of predictions for different USB components on
season and geography to define regions for further analysis. These dependencies are used to
define regions for subsequent time series analysis.
Seasonal dependence: Comparing Figure 2-19 and Figure 2-20 highlights the seasonal
differences in the predicted contributions from Natural, International, and USA sources. Between
spring and summer, the International contribution decreases by 33%; the USA contribution
increases by 40%; and the contribution from Natural sources shows a relatively small increase of
5%. The differences in contributions between the spring and summer are due to a complex
relationship between O3 production, O3 lifetime, and therefore transport efficiency. Cooler drier
conditions increase the lifetime of O3 in winter/spring compared to summer/fall (Liu et al.,
1987). As a result, winter and spring have more efficient transport of O3 compared to summer
April 2022	2-45 External Review Draft - Do Not Quote or Cite
Natural: 22 ppb
(11. 34 ppb)
	n
USA: 14 ppb
(2, 51 ppb)

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
and fall. Summer and fall, however, have warmer weather that promotes higher local O3
production rates. Thus, summer and fall have locally fast O3 production and relatively inefficient
transport, which combined increase the relative contribution of proximate sources.
Border dependence: In the summer, model-predicted gradients of International O3 at the
borders are most obvious. As previously discussed, summer temperatures increase O3 production
rates and decrease O3 lifetimes. As a result, areas with locally high O3 are evident near the border
in southern California and the Big Bend and lower Rio Grande areas of Texas. These local
enhancements generally occur within tens of kilometers from the border due to the short O3
lifetime in summer as noted above.
Topography dependence: High elevation monitors are closer to the free troposphere; in
fact, at certain times of day and locations, the surface can sample free tropospheric air (Jaffe et
al., 2018). Complex topography can also enhance downward transport - for example, free
tropospheric air can "downwash" on the lee-side of high elevation mountains. Sites on the lee-
side can then be affected by this large-scale downwash. High elevation sites or sites influenced
by enhanced vertical transport may show higher contributions from more distant sources.
Combined Seasonal and Geographic Dependence: The simultaneous effects of
topography, proximity to international borders, and seasonal variations are highlighted by
Hovmoller diagrams (Figure 2-21). The Hovmoller diagram shows the average concentration as
a function of month (y-axis) and distance-to-border or elevation (x-axis). Due to the higher
magnitude of estimates of USB sources in the West than the East (Figure 2-19 and Figure 2-20),
the effects of distance and elevation are shown for the West. For the purposes of this analysis, we
use the 97W longitude line as a convenient way to separate the West from the East. The figures
show average estimated values and should not be used to estimate the international contribution
at any specific location. In addition, there are distinct gradients within the 100 m resolution of
the distance-to-border bins. For instance, the 0-100 km from the border grid cell values represent
a spatial average such that the locations directly adjacent to the border have Mexican
contributions higher than that average and the locations 100 km from the border have Mexican
contributions lower than that average.
Figure 2-21 shows that proximity to the border with Canada or Mexico is a good
indicator of the role of international contributions on USB predictions. In the spring, the average
international contribution can be as much as 12.4 ppb within 100 km of the border (62 miles). In
the early spring, large contributions persist further from the border because of the longer O3
lifetimes. Near the borders the contributions also have much higher variability, both from day-to-
day and between locations on the border. The contribution from international sources drops
notably in the summer months when O3 concentrations are highest. The day-to-day variability is
associated with the variations in wind direction, while the location variability is associated with
April 2022
2-46 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
the proximity to an international population center. International contributions are highest in
near-border areas of the U.S. where there are emissions sources on the other side of the border.
To isolate the effect of elevation alone, Figure 2-21 shows the predicted international
contributions as a function of elevation after excluding border areas. In the spring, higher
international contributions are seen at all elevations. The international contribution at all
elevations decreases in summer compared to spring, but to lower contributions at lower elevation
and mostly slowly for the very high elevations (> 1500 m). This is consistent with findings from
Zhang et al. (2011) who used this elevation as a threshold.
Mean 03 8HRMAX West of 97W All 12US2
L
Mean 03 8HRMAX West of 97W notnearbord 12U52
12
200 400 600 800 1000
distance from MEXCAN border (0. 1094 km)
500	1000	1500
Elevation (-12, 3660 m)
Figure 2-21. Predicted contribution of International sources as a function of distance from
Mexico/Canada (left) and at "interior" locations (excluding border areas) by
elevation (right).
Timeseries Analysis: The maps in Figure 2-19 and Figure 2-20 and the Hovmoller plots in
Figure 2-21 highlight the impact of season and location on predicted O3 and contributions. To
further characterize the temporal variations in contributions, the contribution data are averaged
over West and East regions individually using 97W as a dividing line. The coarse "all-cells"
averaging of the data from individual grid cells ignores the major features of the relationship
between the sources and receptors on a sub-regional basis. For example, there are more grid cells
with high urban density and high anthropogenic NOx in the East, so the USA contribution will
be higher in the East. Similarly, there are more high elevation areas in the West, so transported
O3 from outside the U.S. will be higher there. Within the West, however, there are also urban
areas that have both high predicted contributions from international transport and anthropogenic
emissions in the U.S. An analysis using "all-cells" will highlight the general characteristics of the
region. To highlight the within region variability in the West, we also include analyses that focus
on urban cells at high-elevation, near borders, and elsewhere. Figure 2-22 shows regions (West
April 2022
2-47 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
and East) with high-elevation and near border areas and urban areas highlighted by contours. As
can be seen, all the high-elevation areas and Mexico/U.S. border are assigned to the West, the
Canada/U.S. border extends across both East and West, and there are no high-elevation areas in
the East.
Near and High
Near Border
High Elevation
West
East
Excluded
Figure 2-22. Grid cell assignments to East (of 97W), West (of 97W), High Elevation (>
1500m), Near Border (within 100 km), and Near and High (i.e., both High
Elevation and Near Border). The purple outlines highlight grid cells with 20%
or greater urban land use. Near Border areas are in both the West and East, while
High Elevation areas are exclusively in the West. Areas matching colors denoted
East and West, are thus the Low Elevation/Interior areas.
Figure 2-23 shows the time series of regional average (C) MDA8 O3 and O3 contributions
over the year for the West and East at "all-cells," calculated using equation 2-1.
Equation 2-1
Nx
where,
Nx = number of grid cells (x) included
("• = concentration at each grid cell location (x)
The temporal pattern in the regional average clearly shows that the seasonality of MDA8
predictions for each total O3 component varies by region. The natural contribution has a single
maximum in late summer in the West, whereas, in the East there is evidence of two peaks— the
largest in late Spring and a second peak in early Fall. The somewhat lower MDA8 O3 in summer



April 2022
2-48 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
in the East requires further analysis but may be related to the lack of lightning emissions within
the regional domain. The seasonality international contribution predictions is more similar
between the two regions. The international contributions in both the West and East are greatest in
Spring, but the contribution in the West is larger both at its peak and its trough, compared to the
East. The total international contribution and the separately analyzed long-distance components
(e.g., China, India, international shipping) peak in spring when O3 lifetimes favor long-range
transport (see Appendix 2B, Figure 2B-29). However, the Canada/Mexico component of
international contributions peaks in summer because of the relative proximity to the U.S.
receptors. The predicted USA contribution increases in the summer for both the West and the
East, but the USA contribution in the West is smaller than in the East. As mentioned previously,
this "all cells" average is disproportionately rural in the West. The following analysis looks
further at the different types of land in the West, including urban areas that are more
representative of population centers that behave differently than the "all cells" analysis.
C West 97W 12km All >0 ppb
Natural
Res-Anth
Intl
USA
2016-03
2016-05
2016-07
2016-09
2016-11
2017-01
C East 97W 12km All >0 ppb
Natural
Res-Anth
Intl
USA
2016-03
2016-05
2016-07
2016-09
2016-11
2017-01
Figure 2-23. Annual time series of regional average predicted MDA8 total O3 concentration
and contributions of each source (see legend) for the West (top), and the East
April 2022
2-49 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
(bottom). Natural is global natural sources, Intl is international anthropogenic
sources, USA is U.S. anthropogenic sources, and Res-Anth is the residual
anthropogenic (see Table 2-2 for further descriptions).
Figure 2-24 shows the predicted contributions to total O3 in the West split into three
parts: the highest elevation areas, the near border areas, and Low/Interior areas with a weighted
average focusing on urban areas. Each of these subsets is illustrated in Figure 2-22, which shows
high elevation areas (exclusively in the West), near border areas (along the U.S./Mexico and
U.S./Canada borders), and dense urban areas. The Low/Interior areas are neither high elevation
nor near border. In each subset of cells, the purple outlines show the areas whose urban land use
is highest. The effect on O3 contributions of the relative amount of urban land use can be
illustrated by computing an urban area weighted average contribution (Cy), calculated using
equation 2-2.
where,
A% is the urban area in the grid cell x
The urban area weighted average gives a larger weight to data in those urban areas that have
dense emission sources (e.g., mobile). The urban area weighted average shows higher
contribution from USA while Natural and International are lower compared to Figure 2-23. The
differences between urban-weighted and non-weighted contributions are smaller in the East (not
shown) than in the West (compare Figure 2-23 top and Figure 2-24 bottom). Compared to the
West, the East has a larger fraction of land use that is urban (see Figure 2-22), which explains
this difference. Thus, the non-weighted regional average contributions in the East includes the
effects of urban areas much more so than the West. The seasonality of International is also
different between the highest elevation areas, near border areas, and urbanized areas. At
low/interior and at high-elevation sites, the simulated International contribution peaks earlier in
the year than at border sites. This earlier season peak is consistent with seasonality of O3 lifetime
necessary for long-range transport and a smaller contribution of long-distance sources (India,
China, and global shipping, see Appendix 2B, Figure 2B-30). At near-border sites, the seasonal
cycle of predicted USB contributions from Canada/Mexico and from long-range transport
combine to create a maximum later in the spring or early summer that is dominated by
Canada/Mexico contributions (see Appendix 2B, Figure 2B-30, middle panel).
April 2022	2-50 External Review Draft - Do Not Quote or Cite
Equation 2-2

-------
1
2
3
4
5
6
7
8
9
	 West 97W 12km >1500m >0 ppb ฆฆ Natural ฆ Res-Anth ฆฆ Intl ฆฆ USA
2016-01	2016-03	2016-05	2016-07	2016-09	2016-11	2017-01
	 C" West 97W 12km MX/CAN < 100km >0 ppb wm Natural ฆฆ Res-Anth ฆฆ Intl H USA
60
2016-01
2016-03
2016-05
2016-11
Figure 2-24. Annual time series of regional urban area-weighted average predicted MDA8
total Os concentration and contributions of each source (see legend) for the
High-elevation West (top), near-border West (middle), and Low/Interior West
(bottom). Natural is global natural sources, Intl is international anthropogenic
sources, USA is U.S. anthropogenic sources, and Res-Anth is the residual
anthropogenic (see Table 2-2 for further descriptions).
April 2022
2-51 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
2.5.3.3 Ozone Source Contributions as a function of Total Ozone Concentration
Background contributions are also known to vary as a function of total O3. To illustrate
the relationship, specialized scatter density plots were created to show the contributions as a
function of total O3. Unlike the rest of this section, the scatter density plots do not apply the
episodic natural filter described in section 2.5.2. Thus, episodic natural contributions including
double counted fires are included in these presentations, and the effect of large events may be
overestimated.41 In the scatter density plots (Figure 2-25 through Figure 2-27), each pixel
represents a 5 ppb O3 bin. In a traditional scatter density plot, the pixel color would represent the
proportion of all points that fall within that pixel. However, in Figure 2-25 through Figure 2-27
the color represents the fraction of grid-cell-days within each 5 ppb total O3 bin (i.e., the x-axis)
that have a particular model-predicted contribution value (i.e., the y-axis). Brighter colors show
where the most frequent model-predicted contribution (y-axis: Natural or International) lies
within each 5-ppb bin of total O3 value (x-axis). As a reference, percent contribution lines are
overlaid on the plots to help contextualize the results.
Figure 2-25 shows the simulated daily Natural contribution as a function of total MDA8
concentration in the West and East for the whole year. In both regions the majority of total O3
concentrations are under 40-50 ppb. At these low concentrations, the natural contribution
correlates well with total O3 and frequently contributes half of the total O3. At low
concentrations, natural contributions estimated by a zero-out approach can be larger than 100%
of the total prediction. This is a result of NOx-titration by local anthropogenic emissions, which
reduces O3 concentrations and is a well-known non-linearity of O3 chemistry. Thus, removing
the local NOx source increases prediction concentrations. At higher concentrations, Figure 2-25
shows that predicted natural contributions in both regions have a bimodal distribution (or a fork
in frequency of contributions). The lower mode represents a plateau of natural contributions with
increasing total O3, which represents enhancement by anthropogenic sources. The upper mode
represents instances where natural contributions are correlated with total predicted O3. In the
West, the lower mode is less dominant than the East. This suggests, at least in the modeling, that
there are more frequent model-predicted contributions from wildfires and/or stratospheric
intrusions in the West. Wildfire emissions are known to be overestimated in this emission
inventory and their contribution to O3 concentrations are also often overestimated by CMAQ
predictions. As a result, these predictions of very high natural contributions should be interpreted
41 When episodic natural events contribute to elevated O3 concentrations documented in air quality monitoring data
to such an extent that they result in a regulatorily significant exceedance or violation of the NAAQS, they can be
addressed via the Exceptional Events Rule (40 CFR 50.14).
April 2022
2-52 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
qualitatively as simply indicating that such contributions can be appreciable, rather than as
providing accurate and precise quantitative predictions.
West 97W
0	50	100	150	200
Base ppb
0	50	100	150	200
Base ppb
rr
-Q
Q.
ฃL
O
• o.i ง
u
fa
LU
ฆ 0.2
0.0
0.0
Figure 2-25. Predicted contribution of Natural as a function of predicted total (Base)
MDA8 Os concentration in the West and East. Sloped lines show percent
contribution as a quick reference. The number of cells in each column is
identified using the probability density function above the plot, which is on a
log scale that highlights infrequent high concentrations.
Figure 2-26 shows the predicted contribution in the West and East from international
anthropogenic sources. Unlike natural contributions, there is very little correlation between
international anthropogenic and total O3. There are rare large model-predicted contributions,
which are more frequent in the West than in the East and rarely contribute more than 50% total
O3 in either region. There are also negative contributions (up to -15 ppb), which arise from non-
linearities in chemistry. The largest negative contribution predictions are along the Mexico
border. These can either be NOx-titration events or cases where chemistry associated with
international NOx-sources remove precursors that would otherwise enhance O3 from U.S.
sources. Negative international contributions tend to occur at relatively low total O3
concentrations.
April 2022	2-53 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
50	100
Base ppb
50	100
Base ppb
Figure 2-26. Predicted contribution of International as a function of predicted total (Base)
MDA8 Os concentration in the West and East. Sloped lines show percent
contribution as a quick reference. The number of cells in each column is
identified using the probability density function above the plot, which is on a
log scale that highlights infrequent high concentrations.
Figure 2-27 illustrates the relationship between predictions of U.S. anthropogenic sources
and total O3. Above 50 ppb, the predicted contribution from USA increases with total O3 in both
the West and the East. The relationship is stronger in the East, than the West, where near border
contributions, fire contributions, and stratospheric exchange are smaller. Even so, the higher total
O3 in the West has a similar association of larger USA contributions at larger concentrations.
This is consistent with previous findings (Henderson et al., 2012; U.S. EPA, 2014).
•g .a 100
0.4 {

50	100
Base ppb
150	200
ฃ
- 0.6 8
50	100
Base ppb
Figure 2-27. Predicted contribution of USA as a function of predicted total (Base) MDA8
O3 concentration in the West and East. Sloped lines show percent contribution
as a quick reference. The number of cells in each column is identified using the
probability density function above the plot, which is on a log scale that
highlights infrequent high concentrations.
April 2022
2-54 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
Another way of looking at the contributions is to restrict the time series to grid cells
where the concentration is above a threshold. Restricting to grid cells with high concentrations
implicitly weights the results toward urban areas where these high concentrations occur most
frequently. Figure 2-28 shows the seasonal and regional variation of USB (International
Anthropogenic and Natural) and USA (anthropogenic only) sources on high O3 days (MDA8
>70 ppb). The largest magnitude differences between sources in the East and West come from
contributions predicted for Natural and USA sources. Recall that the West contains all the high-
elevation areas (>1500 m) and the full length of the U.S./Mexican border. Figure 2-29 includes
time series for high elevation, near Mexico border, and low-elevation interior areas separately.
Compared to the East, the low/interior sites in the West have 9 ppb larger contribution from
Natural and 2 ppb more from International. Compared to low/interior sites in the West, the high-
elevation sites have 7 ppb larger contributions from Natural and 4 ppb more from International.
For border areas, the International contribution is 13 ppb greater than in Low/Interior sites. As
previously noted, there are large gradients of predicted international contributions even within
the border areas, such that some locations within the 100 km of the border are predicted to
receive larger international contributions while others are predicted to receive substantially
smaller international contributions than noted above.
April 2022
2-55 External Review Draft - Do Not Quote or Cite

-------
04	—ill	hi—— . ^m.i 11 ii.im —i,	—ซH' —	—
2016-01	2016-03	2016-05	2016-07	2016-09	2016-11	2017-01
2
3	Figure 2-28. Annual time series of regional average predicted MDA8 Os and contributions
4	of each source to predicted MDA8 total O3 (see legend) in the West (top) and
5	East (bottom) including only those grid-cell days with MDA8 greater than 70
6	ppb. Natural is global natural sources, Intl is international anthropogenic sources,
7	USA is U.S. anthropogenic sources, and Res-Anth is the residual anthropogenic
8	(see Table 2-2 for further descriptions).
9
April 2022
2-56 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
C West 97W 12km >1500m >70 ppb
Natural
Res-Anth
Intl
USA
80
70
ฆ 60
x
< 50
a.
5 40

30
S 20
10
oJ	
2016-01
2016-03
2016-05
2016-07
2016-09
2016-11
2017-01
C West 97W 12km MX/CAN < 100km >70 ppb
Natural
Res-Anth
Intl
USA
ao
70
s50
X
< 50
ce
i 40
m1
? 30
2 20
10
0J	
2016-01
2016-03
2016-05
2016-07
2016-09
2016-11
2017-01
	 West 97W 12km Low/Interior >70 ppb
Natural

Res-Anth
Intl
USA
2016-03
2016-05
2016-07
2016-09
2016-11
2017-01
Figure 2-29. Annual time series of regional average predicted MDA8 O3 and contributions
of each source to predicted MDA8 O3 (see legend) in the high-elevation West
(top), in the near-border West (middle), and in the Low/Interior West
weighted toward urban areas (bottom) including only those grid-cell days with
MDA8 O3 greater than 70 ppb. Natural is global natural sources, Intl is
international anthropogenic sources, USA is U.S. anthropogenic sources, and Res-
Anth is the residual anthropogenic (see Table 2-2 for further descriptions).
April 2022
2-57 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
2.5.3.4 Predicted USB Seasonal Mean and USB on Peak O3 Days
The analyses above describe the contributions from the components of USB to MDA8 O3
over seasons and days. Jaffe et al. (2018) concluded that model predictions of seasonal means
have more certainty than individual daily or episodic estimates of USB. However, from a policy
perspective, it is also useful to understand the USB contributions for various regulatory-relevant
metrics. In addition to reporting predicted USB using a seasonal average metric, we also examine
predicted USB (1) on days with the highest predicted MDA8 total O3 concentrations (top 10
days); (2) on days predicted to have the 4th highest MDA8 total O3 concentrations in the year;
and (3) on days when predicted MDA8 for total O3 is above 60 ppb or above 70 ppb.
Figure 2-30 shows USB predicted by a single simulation with U.S. anthropogenic
emissions zeroed-out. Similar to what was found for the seasonal average metric, the effect of
topography and proximity to borders are readily evident for predicted MDA8 USB on the top 10
days and the 4th highest days. The differences in seasonal average contributions between the East
and West are also evident with the top 10 days metric and 4th highest day metric. The speckled
nature of the USB plot for the 4th highest day is due to the day or even season on which the 4th
high is predicted to occur, which varies from grid cell to grid cell. The season in which the 4th
highest day occurs influences the expected contribution from long-range international transport.
The average USB contributions for the top 10 days exhibit a smoother spatial pattern because
there is a tendency for high days to be grouped seasonally, even if the 4th highest is not. Because
the USB contribution varies by season, the predicted USB contribution on the predicted 4th
highest day is quite sensitive to model bias because bias may change the season on which the 4th
highest predicted day occurs.
It is also important to highlight that areas with high predicted USB contributions do not
always coincide with areas where MDA8 total O3 concentrations are predicted to be above 70
ppb. On the 10 highest predicted MDA8 O3 days, predicted USB is relatively constant over large
areas (see Figure 2-30 middle left). Within these areas of relatively constant USB, Figure 2-30
shows that the locations having model-predicted MDA8 concentrations above 70 ppb are
generally in or near urban areas (Figure 2-30 lower right).
The USB contribution predicted in urban areas on the predicted top 10 days tends to be
lower than in surrounding rural areas. This is due to the temporal anti-correlation of local
contribution with natural and international contributions. In urban areas, MDA8 total O3
concentrations above 70 ppb tend to occur in summer and fall when anthropogenic sources result
in locally high increments of O3. Also during these seasons, long-range transport is limited and
USB from intercontinental transport is at its lowest. As a result, the predicted top 10 and 4th
highest concentration days in urban areas tend to have lower predicted USB contributions than
do such days in rural parts of the region even though rural areas have lower MDA8 O3. As a
April 2022
2-58 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
result, the areas with predicted top 10 days having MDA8 total O3 above 70 ppb tend to have
lower percentage USB contributions than the surrounding areas.
Predicted USB contributions can be large on top 10 days near populated U.S./Mexico
border areas. In near-border areas with large anthropogenic emissions, international transport can
make a large contribution. For example, across the 4th highest days predicted for every grid cell
in this model simulation, the highest predicted MDA8 USB is 80 ppb (at a location immediately
adjacent to the border). Given the uncertainties associated with such single value predictions,
averaged predictions are important to consider. Compared to the maximum USB on the 4th high,
the maximum USB is 10 ppb lower for the average of top 10 days (Figure 2-30, middle left
panel) and 11 ppb lower the average of days with MDA8 above 70 ppb (Figure 2-30, lower left
panel). The very high USB values associated with international anthropogenic emissions are very
near the U.S./Mexico border and, to the extent that associated areas have been designated
nonattainment for the NAAQS, these areas may qualify under Clean Air Act section 179B, titled
"International border areas," for specified regulatory relief upon submission of a satisfactory
demonstration.
April 2022
2-59 External Review Draft - Do Not Quote or Cite

-------
1
2	Figure 2-30. Map of predicted USB contributions by O3 season for spring average (top left),
3	summer average (top right), top 10 predicted total O3 days (center left), 4th
4	highest total O3 simulated day (center right), and all days with total O3 greater
5	than 70 ppb (bottom left), along with a map of the number of days with total
6	O3 above 70 ppb (bottom right, where yellow pixels have 10+ days). Each
7	contribution has the spatial average and range (min, max) in the lower left-
8	hand corner of the panel.
9
>70 ppb: 0 days
(0, 90 days)
l\ ^
USB: 31 ppb
(17, 54 ppb)
	iiu_
USB: 37 ppb
(9. 80 ppb)
	
April 2022
2-60 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
The maps in Figure 2-30 provide a detailed spatial representation of predicted USB but
may imply more precision than can be expected from a modeling system. For example, the
maximum USB on predicted fourth highest day reaches 80 ppb near the Mexico border. The
largest USB at nearby monitoring sites was 71 ppb.42 The observed 4th highs at those monitors
occurred in late February and early March, while the predicted 4th highs occurred in summer.
After selecting the 4th highs based on the observations and applying bias correction
proportionally to contributions, the new USB at these locations is 51 and 63 ppb. The USB
values for any given grid cell may be biased due to local features of topography, meteorology,
emissions bias, or model construct.
To complement the spatially resolved data and reduce bias associated with individual
daily model predictions, we also spatially aggregate the data by NOAA climate region. The
predicted USB values by climate region are provided in Table 2-3 to Table 2-6. Similar to the
figures, the tables separately quantify all grid cells (Table 2-3), high elevation (>1500 m) areas
(Table 2-4), near border areas (Table 2-5), and low-elevation (<1500 m) interior areas (Table 2-
6). These tables show the spatial averages of USB within each climate region for the annual
average, seasonal averages, averages of days when MDA8 O3 is greater than 60 or 70 ppb,
averages of each grid cell's top 10-days, and each cell's 4th highest day. Note that top 10-day
average and 4th high day for each grid cell may be from different times of the year compared to
the neighboring grid cells. As a result, grid cells with highest O3 driven by transport in the Spring
are being mixed with grid cells with highest O3 driven by local formation. Applying these
averages to interpret observations must, therefore, be done in the full context of time, space, and
concentration range.
42 Monitor 06-025-1003 measured 4th maximum value was 74 ppb on March 1, 2016. Monitor 06-073-1011
measured 4th maximum was 75 ppb on February 28, 2016. Predicted USB on predicted 4th high at both locations
was 71 ppb without bias correction in July and August.
April 2022
2-61 External Review Draft - Do Not Quote or Cite

-------
1	Table 2-3. Predicted USB for U.S. and U.S. regions based on averages for all U.S. grid
2	cells.
Regions*
Mean MDA8 for Seasons or Year
Mean MDA8 of Values in Subset
Annual
4th highest
MDA8
DJFB
MAMC
JJAD
SONE
ANNF
>60ppb
>70ppb
Top10
U.S.
26
32
31
29
30
38
33
37
37
West
28
35
36
32
33
47
43
44
44
East
24
29
24
25
26
28
27
28
28
NW
27
33
33
32
31
43
32
41
41
W
30
34
38
34
34
47
43
46
47
WNC
24
33
36
30
31
48
44
43
44
SW
31
38
39
35
36
51
48
49
49
S
27
33
26
27
28
34
29
33
33
ENC
21
30
28
26
26
31
34
32
33
C
24
30
25
26
26
28
28
28
28
SE
25
28
20
24
24
25
22
25
25
NE
25
29
27
27
27
29
26
28
27
AU.S.=continental U.S, West= >97 degrees West longitude, East= <97 degrees West longitude, NW=Northwest, W=West,
WNC=WestNorthCentral, SW=Southwest, S=South, ENC=EastNorthCentral, C=Central, SE=Southeast, and NE=Northeast.
B Season defined as December, January and February.
c Season defined as March, April and May.
D Season defined as June, July and August.
E Season defined as September, October and November.
F Annual mean.
3
April 2022
2-62 External Review Draft - Do Not Quote or Cite

-------
1 Table 2-4. Predicted USB for high elevation locations (>1500 m).

Mean M
DA8 for Seasons or Year
Mean MDA8 of values in subset
Annual
4th highest
MDA8
Regions*
DJFB
MAMC
JJAD
SONE
ANNF
>60ppb
>70ppb
ToplO
U.S.
31
37
40
35
35
52
49
49
50
West
31
37
40
35
35
52
49
49
50
East
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
NW
29
35
38
33
34
52
42
47
48
W
32
36
42
36
36
53
47
51
52
WNC
28
35
39
34
34
52
48
48
49
SW
32
38
39
35
36
51
50
50
50
S
35
43
36
35
37
55
59
52
53
ENC
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
C
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
SE
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
NE
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
AU.S.=continental U.S, West= >97 degrees West longitude, East= <97 degrees West longitude, NW=Northwest, W=West,
WNC=WestNorthCentral, SW=Southwest, S=South, ENC=EastNorthCentral, C=Central, SE=Southeast, and NE=Northeast.
B Season defined as December, January and February.
c Season defined as March, April and May.
D Season defined as June, July and August.
E Season defined as September, October and November.
F Annual mean.
2 Table 2-5. Predicted USB for locations within 100 km of Mexico or Canada Border.
Regions*
Mean MDA8 for Seasons or Year
Mean MDA8 of values in subset
Annual
4th highest
MDA8
DJFB
MAMC
JJAD
SONE
ANNF
>60ppb
>70ppb
ToplO
U.S.
26
34
32
30
30
45
43
40
40
West
28
36
34
32
32
51
56
45
45
East
22
29
28
27
27
33
34
31
31
NW
27
32
30
31
30
46
N/A
38
38
W
30
35
41
36
36
46
51
51
51
WNC
21
33
34
29
29
49
N/A
42
42
SW
32
40
36
35
36
53
55
49
50
S
32
41
33
32
34
52
63
48
49
ENC
20
29
28
26
26
32
35
32
32
C
24
30
29
28
28
31
30
31
32
SE
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
NE
24
29
28
27
27
34
41
30
30
AU.S.=continental U.S, West= >97 degrees West longitude, East= <97 degrees West longitude, NW=Northwest, W=West,
WNC=WestNorthCentral, SW=Southwest, S=South, ENC=EastNorthCentral, C=Central, SE=Southeast, and NE=Northeast.
B Season defined as December, January and February.
c Season defined as March, April and May.
D Season defined as June, July and August.
E Season defined as September, October and November.
F Annual mean.
April 2022
2-63 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
Table 2-6. Predicted USB for low-elevation (<1500 m) that are 100 km or farther from
the border.

Mean MDA8
for Seasons or Year
Mean MDA8 of values in subset
Annual 4th
highest
MDA8
Regions*
DJFB
MAMC
JJAD
SONE
ANNF
>60ppb
>70ppb
Top 10
U.S.
25
31
28
28
28
33
30
34
34
West
27
34
34
31
31
43
39
41
41
East
24
29
24
25
26
27
27
28
28
NW
27
32
31
31
30
37
32
38
38
W
29
32
35
33
32
42
41
42
42
WNC
23
33
36
29
30
44
42
41
42
SW
29
37
38
33
34
49
43
47
47
S
26
32
26
27
28
32
26
32
32
ENC
21
30
28
26
26
31
33
32
33
C
24
30
25
26
26
28
28
28
28
SE
25
28
20
24
24
25
22
25
25
NE
25
29
26
27
27
28
25
27
26
AU.S.=continental U.S, West= >97 degrees West longitude, East= <97 degrees West longitude, NW=Northwest, W=West,
WNC=WestNorthCentral, SW=Southwest, S=South, ENC=EastNorthCentral, C=Central, SE=Southeast, and NE=Northeast.
B Season defined as December, January and February.
c Season defined as March, April and May.
D Season defined as June, July and August.
E Season defined as September, October and November.
F Annual mean.
2.5.4 Summary of USB
Background O3 results from a variety of sources, each of which has its own temporal
pattern and spatial distribution. The location and timing of these sources impacts O3 production,
dispersion and loss and thus different background O3 sources have unique seasonality and spatial
patterns. The analysis presented here provides updated model-based estimates of magnitude,
seasonality and spatial patterns of background O3 contributions. The analysis separately
characterizes the estimated magnitude and spatial/temporal patterns of MDA8 O3 from three
sources: natural, international anthropogenic, and USA anthropogenic.
The current analysis indicates that natural and USA O3 contributions peak during the
traditional O3 season (May through September), while long-range intercontinental transport of
international O3 (i.e. contributions from China, India, etc.) peaks in the spring (February through
May). The contributions from Canada/Mexico at near-border locations are associated with
relatively short-range transport and the seasonality peaks during May through September, similar
to USA anthropogenic O3. The influence of Canada/Mexico, however, is indicated by the model
predictions to have a stronger spatial gradient in summer, so Canada/Mexico contributions are
most evident near the border. Of the three categories of contributions, the USA anthropogenic is
best correlated with total O3 at concentrations above 40-50 ppb in both the West and the East
April 2022
2-64 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
suggesting that US anthropogenic emissions are usually the driving cause of high O3 events in
the US. This is largely explained by temporal patterns of background O3 influences in relation to
typical high O3 events. There can be exceptions to this rule that are generally associated natural
contributions at high-elevation, during fires events, or at near-border sites.
This modeling analysis indicates the relationship between predicted international and
USA anthropogenic contributions depend upon the international sources and the location. Long-
range transport and USA anthropogenic contributions tend peak at different times of the year, so
the contribution of international is often at its minimum when local sources are the driving factor
for high total O3 during the May through September O3 season. Even in cases where O3 formed
from international anthropogenic emissions does coincide seasonally with high O3 periods, the
impact of those sources can have large spatial variation. For example, O3 formed from
anthropogenic emissions in Canada and Mexico can peak in late spring or early summer when
total O3 is high. During this time-period, there is a strong spatial variability not shown in the
regional mean. As a result, specific days at specific locations may experience larger or smaller
contributions from cross-border transport on an episodic basis that is not well characterized by
average seasonal contributions. Another example of spatial heterogeneity is exemplified by
wintertime O3 events associated with emissions from local oil and gas production in the
Intermountain West. Even though these episodes can occur as early in the year as February,
international emissions do not contribute to them substantially. The conditions associated with
these events result in decoupling of the local air masses from the upper atmosphere, essentially
isolating air in the mountain valleys from the atmosphere above and reducing the influence of
long-range transport compared to other winter and early spring days. As a result, these unique
wintertime O3 episodes have little relative influence from international emissions despite
occurring at a time of year when long-range transport from Asia is efficient. This highlights the
need to perform location specific analysis rather than relying on regional averages.
In addition to seasonal patterns, the ISA highlights interannual patterns in background O3
as well as long-term trends (ISA, section IS.2.2.1). Natural emissions and international transport
are highly impacted by meteorological patterns which vary from year to year. One key ISA
finding is that decreasing East Asian NOx emissions starting around 2010, which would suggest
decreasing contributions from East Asia in the future if those trends continue, and therefore
decreasing spring USB.
Assessments of background O3 in the 2015 review reported regional variation in
background O3 (2013 ISA; 2014 PA). Consistent with those assessments, modeling presented
here predicts that USB is higher in the West than in the East. In this analysis, we found that on
high O3 days (greater than 70 ppb) the West-East differences are largely associated with
international contributions in near-border areas and natural contributions at high-elevation
April 2022
2-65 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
locations. The Natural component of USB exhibits the largest magnitude difference between the
West and East. International contributions from intercontinental transport (e.g., Asia) are most
important at high elevations in the West, while international contributions from Canadian and
Mexican sources are most pronounced immediately adjacent to the borders.
The modeling performed for this assessment does not differentiate between natural
sources of ozone. For this analysis we did not attempt to separately quantify the contributions
from individual Natural sources (e.g., lightning, soil, fires, stratosphere) or to address exceptional
events beyond basic screening to remove very large fire plumes. Literature-based emissions
estimates and photochemical modeling studies can help to inform the likely contributors to
natural. In the northern hemisphere, the natural NOx sources with the largest emissions estimates
are lightning (9.4 megatonN/yr), soils (5.5 megatonN/yr), and wildland fires (-2.2
megatonN/yr). Because NOx is the limiting precursor at hemispheric scales, the emissions
estimates suggest that lightning and soils are most likely the largest contributors to Natural O3,
except when impacted by specific fire episodes. As noted by Lapina et al. (2014), a large
contribution from lightning may be the result of lightning strikes outside the U.S. while the
contribution from soil NOx tends to be largest from emissions within the U.S. The distant
lightning source is likely to have its effect as part of the well-mixed background. The local soil
NOx emissions have a clear seasonal cycle and are known to have large local contributions. The
relative effect at any specific site would require further analysis, including identifying the portion
of the effect due to fertilizer.
The overall findings of this assessment are consistent with the 2014 PA, with the EPA's
Background Ozone whitepaper (U.S. EPA, 2015) and with the peer reviewed literature (e.g.,
Jaffe et al. 2018). The definition of USB is also consistent with the assessment in the 2014 PA
and includes global natural and international anthropogenic emission sources (NOx and VOC).
Specific findings from the current analysis are summarized as:
•	USB has important spatial variation that is related to geography, topography, and
international borders. The spatial variation is influenced by seasonal variation with long-
range international transport contributions peaking in the spring while US anthropogenic
contributions peak in summer.
•	The West has higher predicted USB concentrations than the East, which includes higher
contributions from International and Natural sources. Within the West, high-elevation
and near-border areas stand out as having particularly high USB. The high-elevation
areas have more International and Natural contributions than low-interior areas in the
same region. The near-border areas in the West can have substantially more international
contribution than other parts of the West.
•	The USA contributions that drive predicted MDA8 total O3 concentrations above 70 ppb
are predicted to typically peak in summer. In this typical case, the predicted USB is
overwhelmingly from Natural sources. The most notable exception to the typical case is
April 2022
2-66 External Review Draft - Do Not Quote or Cite

-------
1	reflected by predictions for an area near the Mexico border where the modeling indicates
2	that a combination of Natural and Canada/Mexico contributions can lead to predicted
3	MDA8 USB concentrations 60-80 ppb, on specific days, which is consistent with the O3
4	PA prepared for the 2015 review (2014 PA, Section 2.4).43
5	• Predicted international contributions, in most places, are lowest during the season with the
6	most frequent occurrence of MDA8 concentrations above 70 ppb. Except for the near-
7	border areas, the International contribution requires long-distance transport that is most
8	efficient in Spring.
9	• Days for which MDA8 total O3 concentrations are predicted to be above 70 ppb tend to
10	have a substantially higher model-predicted USA (anthropogenic) contribution than other
11	days in both the West and the East.
43 Uncertainties associated with such model predictions for individual days are recognized in section 2.5.3.4 above,
along with observations of how they may differ from measurements at monitoring locations in the same area. It is
also important to note that the modeling analyses presented here do not provide estimates of design values, which
are derived from monitoring data (collected over three years) and used to assess exceedances of the O3 standards.
Additionally, as noted earlier, where such exceedances occur and are shown to be caused by USB, regulations for
exceptional events may pertain.
April 2022
2-67 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
REFERENCES
Ahmadov, R, McKeen, S, Trainer, M, Banta, R, Brewer, A, Brown, S, Edwards, PM, de Gouw,
JA, Frost, GJ, Gilman, J, Helmig, D, Johnson, B, Karion, A, Koss, A, Langford, A,
Lerner, B, Olson, J, Oltmans, S, Peischl, J, Petron, G, Pichugina, Y, Roberts, JM,
Ryerson, T, Schnell, R, Senff, C, Sweeney, C, Thompson, C, Veres, PR, Warneke, C,
Wild, R, Williams, EJ, Yuan, B and Zamora, R (2015). Understanding high wintertime
ozone pollution events in an oil- and natural gas-producing region of the western US.
Atmos Chem Phys 15(1): 411-429.
Akagi, SK, Yokelson, RJ, Burling, IR, Meinardi, S, Simpson, I, Blake, DR, McMeeking, GR,
Sullivan, A, Lee, T, Kreidenweis, S, Urbanski, S, Reardon, J, Griffith, DWT, Johnson, TJ
and Weise, DR (2013). Measurements of reactive trace gases and variable 03; formation
rates in some South Carolina biomass burning plumes. Atmos Chem Phys 13(3): 1141-
1165.
Allen, DJ, Pickering, KE, Pinder, RW, Henderson, BH, Appel, KW and Prados, A (2012).
Impact of lightning-NO on eastern United States photochemistry during the summer of
2006 as determined using the CMAQ model. Atmos Chem Phys 12(4): 1737-1758.
Baker, KR, Woody, MC, Tonnesen, GS, Hutzell, W, Pye, HOT, Beaver, MR, Pouliot, G and
Pierce, T (2016). Contribution of regional-scale fire events to ozone and PM2.5 air
quality estimated by photochemical modeling approaches. Atmos Environ 140: 539-554.
Baker, KR, Woody, MC, Valin, L, Szykman, J, Yates, EL, Iraci, LT, Choi, HD, Soja, AJ,
Koplitz, SN, Zhou, L, Campuzano-Jost, P, Jimenez, JL and Hair, JW (2018).
Photochemical model evaluation of 2013 California wild fire air quality impacts using
surface, aircraft, and satellite data. Sci Total Environ 637-638: 1137-1149.
Buysse, CE, Kaulfus, A, Nair, U and Jaffe, DA (2019). Relationships between Particulate
Matter, Ozone, and Nitrogen Oxides during Urban Smoke Events in the Western US.
Environ Sci Technol 53(21): 12519-12528.
Camalier, L, Cox, W and Dolwick, P (2007). The effects of meteorology and their use in
assessing ozone trends. Atmos Environ 41: 7127-7137.
Dolwick, P, Akhtar, F, Baker, KR, Possiel, N, Simon, H and Tonnesen, G (2015). Comparison of
background ozone estimates over the western United States based on two separate model
methodologies. Atmos Environ 109: 282-296.
Emery, C, Jung, J, Downey, N, Johnson, J, Jimenez, M, Yarvvood, G and Morris, R (2012).
Regional and global modeling estimates of policy relevant background ozone over the
United States. Atmos Environ 47: 206-217.
Grewe, V (2013). A generalized tagging method. Geosci Model Dev 6(1): 247-253.
April 2022
2-68 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
Henderson, BH, Possiel, N, Akhtar, F and Simon, H. (2012). Regional and Seasonal Analysis of
North American Background Ozone Estimates from Two Studies. 08/15/2012. U.S. EPA
Research Triangle Park, NC.
Hudman, RC, Moore, NE, Mebust, AK, Martin, RV, Russell, AR, Valin, LC and Cohen, RC
(2012). Steps towards a mechanistic model of global soil nitric oxide emissions:
implementation and space based-constraints. Atmos Chem Phys 12(16): 7779-7795.
Jaffe, DA, Cooper, OR, Fiore, AM, Henderson, BH, Tonneson, GS, Russell, AG, Henze, DK,
Langford, AO, Lin, M and Moore, T (2018). Scientific assessment of background ozone
over the U.S.: Implications for air quality management. Elem Sci Anth 6(1): 56.
Jaffe, DA and Wigder, NL (2012). Ozone production from wildfires: A critical review. Atmos
Environ 51: 1-10.
Janssens-Maenhout, G., Crippa, M., Guizzardi, D., Dentener, F., Muntean, M., Pouliot, G.,
Keating, T., Zhang, Q., Kurokawa, J., Wankmiiller, R., Denier van der Gon, H., Kuenen,
J. J. P., Klimont, Z., Frost, G., Darras, S., Koffi, B., and Li, M. (2015). HTAP_v2.2: a
mosaic of regional and global emission grid maps for 2008 and 2010 to study
hemispheric transport of air pollution. Atmos Chem Phys, 15:11411-11432.
Jin, X, Fiore, AM, Murray, LT, Valin, LC, Lamsal, LN, Duncan, B, Folkert Boersma, K, De
Smedt, I, Abad, GG, Chance, K and Tonnesen, GS (2017). Evaluating a Space-Based
Indicator of Surface Ozone-NO x-VOC Sensitivity Over Midlatitude Source Regions and
Application to Decadal Trends: Space-Based Indicator of O 3 Sensitivity. Journal of
Geophysical Research: Atmospheres.
Lapina, K, Henze, DK, Milford, JB, Huang, M, Lin, M, Fiore, AM, Carmichael, G, Pfister, GG
and Bowman, K (2014). Assessment of source contributions to seasonal vegetative
exposure to ozone in the US. JOURNAL OF GEOPHYSICAL RESEARCH-
ATMOSPHERES 119(1): 324-340.
Lefohn, AS, Malley, CS, Simon, H, Wells, B, Xu, X, Zhang, L and Wang, T (2017). Responses
of human health and vegetation exposure metrics to changes in ozone concentration
distributions in the European Union, United States, and China. Atmos Environ 152: 123-
145.
Lin, J-T, Martin, RV, Boersma, KF, Sneep, M, Stammes, P, Spurr, R, Wang, P, Van Roozendael,
M, Clemer, K and Irie, H (2014). Retrieving tropospheric nitrogen dioxide from the
Ozone Monitoring Instrument: effects of aerosols, surface reflectance anisotropy, and
vertical profile of nitrogen dioxide. Atmos Chem Phys 14(3): 1441-1461.
Lin, M, Fiore, AM, Horowitz, LW, Langford, AO, Oltmans, SJ, Tarasick, D and Rieder, HE
(2015). Climate variability modulates western US ozone air quality in spring via deep
stratospheric intrusions. Nature Communications 6(1): 7105.
April 2022
2-69 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
Lin, M, Horowitz, LW, Payton, R, Fiore, AM and Tonnesen, G (2017). US surface ozone trends
and extremes from 1980 to 2014: quantifying the roles of rising Asian emissions,
domestic controls, wildfires, and climate. Atmos Chem Phys 17(4): 2943-2970.
Liu, SC, Trainer, M, Fehsenfeld, FC, Parrish, DD, Williams, EJ, Fahey, DW, Hiibler, G and
Murphy, PC (1987). Ozone production in the rural troposphere and the implications for
regional and global ozone distributions. Journal Of Geophysical Research-Atmospheres
92(D4).
McClure, CD and Jaffe, DA (2018). Investigation of high ozone events due to wildfire smoke in
an urban area. Atmos Environ 194: 146-157.
Murray, LT (2016). Lightning NO x and Impacts on Air Quality. Curr Pollut Rep 2(2): 115-133.
NRC (2002). National Research Council Committee on Estimating the Health-Risk-Reduction
Benefits of Proposed Air Pollution Regulations. National Academies Press (US).
Washington (DC).
Pachauri, RK, Mayer, L and and Intergovernment Panel on Climate Change (2015). Climate
Change 2014: Synthesis Report. Contribution of Working Groups I, II and III to the Fifth
Assessment Report of the Intergovernmental Panel on Climate Change. IPCC. Geneva,
Switzerland, https://epic.awi.de/id/eprint/37530/.
Parrish, DD, Young, LM, Newman, MH, Aikin, KC and Ryerson, TB (2017). Ozone Design
Values in Southern California's Air Basins: Temporal Evolution and U.S. Background
Contribution: Southern California Ozone Design Values. Journal of Geophysical
Research: Atmospheres 122(20): 11,166-111,182.
Phillips, S, Wang, K, Jang, C, Possiel, N, Strum, M and Fox, T (2008). Evaluation of 2002
Multi-pollutant Platform: Air Toxics, Ozone, and Particulate Matter. 7th Annual CMAS
Conference.
Reay, DS, Smith, P, Christensen, TR, James, RH and Clark, H (2018). Methane and Global
Environmental Change. Annu Rev Environ Resour 43(1): 165-192.
Simon, H, Baker, KR and Phillips, S (2012). Compilation and interpretation of photochemical
model performance statistics published between 2006 and 2012. Atmos Environ 61: 124-
139.
Simon, H, Reff, A, Wells, B, Xing, J and Frank, N (2015). Ozone trends across the United States
over a period of decreasing NOx and VOC emissions. Environ Sci Technol 49(1): 186-
195.
Steinkamp, J and Lawrence, MG (2011). Improvement and evaluation of simulated global
biogenic soil NO emissions in an AC-GCM. Atmos Chem Phys 11(12): 6063-6082.
April 2022
2-70 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
Turner, AJ, Frankenberg, C, Wennberg, PO and Jacob, DJ (2017). Ambiguity in the causes for
decadal trends in atmospheric methane and hydroxyl. Proc Natl Acad Sci USA 114(21):
5367-5372.
U.S. EPA (1978). Air Quality Criteria for Ozone and Other Photochemical Oxidants
Environmental Criteria and Assessment Office. Research Triangle Park, NC. EPA-600/8-
78-004. April 1978. Available at:
https://nepis.epa.gov/Exe/ZyPURL.cgi?Dockey=200089CW.txt.
U.S. EPA (2009). Technical Support Document for the Proposal to Designate an Emissions
Control Area for Nitrogen Oxides, Sulfur Oxides, and Particulate Matter. U.S.
Environmental Protection Agency. Research Triangle Park, NC. U.S. EPA. EPA-420-R-
007. Available at: http://www.epa.gov/otaq/regs/nonroad/marine/ci/420r09007.pdf.
U.S. EPA (2014). Policy Assessment for the Review of the Ozone National Ambient Air Quality
Standards,. U.S. Environmental Protection Agency. Research Triangle Park, NC. U.S.
EPA. EPA-452/R-14-006. Available from:
https://nepis. epa.gov/Exe/ZyPDF. cgi?Dockey=Pl 00KCZ5. txt.
U.S. EPA (2015). Implementation of the 2015 Primary Ozone NAAQS: Issues Associated with
Background Ozone White Paper for Discussion. U.S. Environmental Protection Agency.
Research Triangle Park, NC. U.S. EPA. https://www.epa.gov/sites/production/files/2016-
03/documents/whitepaper-bgo3-final.pdf.
U.S. EPA (2018a). Air Quality Modeling Technical Support Document for the Updated 2023
Projected Ozone Design Values. Office of Air Quality Planning and Standards United
States Environmental Protection Agency. RTP, NC. U.S. EPA. 83FR65878.
https://www.epa.gov/sites/production/files/2018-
06/documents/aq modelingtsd updated 2023 modeling o3 dvs.pdf.
U.S. EPA (2018b). Modeling Guidance for Demonstrating Attainment of Air Quality Goals for
Ozone, PM2.5, and Regional Haze. Office of Air Quality Planning and Standards.
Research Triangle Park, NC. U.S. EPA. EPA 454/R-18-009. Available at:
https://www3.epa.gov/ttn/scram/guidance/guide/03-PM-RH-Modeling_Guidance-
2018.pdf
U.S. EPA (2021a). Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2019. Office
of Air Quality Planning and Standards. Research Triangle Park, North Carolina. U.S.
EPA. EPA 43O-R-21-005 Available at: https://www.epa.gov/sites/default/files/2021-
04/documents/us-ghg-inventory-2021-main-text.pdf.
U.S. EPA (2021b). 2017 National Emissions Inventory: January 2021 Updated Release,
Technical Support Document. Office of Air Quality Planning and Standards. Research
Triangle Park, North Carolina. U.S. EPA. Available:
https://www. epa.gov/sites/default/files/2021-02/documents/nei2017 tsdjullJan2021.pdf
April 2022
2-71 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
van der Werf, GR, Randerson, JT, Giglio, L, van Leeuwen, TT, Chen, Y, Rogers, BM, Mu, M,
van Marie, MJE, Morton, DC, Collatz, GJ, Yokelson, RJ and Kasibhatla, PS (2017).
Global fire emissions estimates during 1997-2016. Earth Syst Sci Data 9(2): 697-720.
Wu, S, Mickley, LJ, Jacob, DJ, Logan, JA, Yantosca, RM and Rind, D (2007). Why are there
large differences between models in global budgets of tropospheric ozone? J Geophys
Res 112(D5).
Zhang, L, Jacob, DJ, Downey, NV, Wood, DA, Blewitt, D, Carouge, CC, van Donkelaar, A,
Jones, DBA, Murray, LT and Wang, Y (2011). Improved estimate of the policy-relevant
background ozone in the United States using the GEOS-Chem global model with 1/2ฐ x
2/3ฐ horizontal resolution over North America. Atmos Environ 45(37): 6769-6776.
Zhang, L, Jacob, DJ, Yue, X, Downey, NV, Wood, DA and Blewitt, D (2014). Sources
contributing to background surface ozone in the US Intermountain West. Atmos Chem
Phys 14(11): 5295-5309.
April 2022
2-72 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
3 RECONSIDERATION OF THE PRIMARY STANDARD
This chapter presents and evaluates the policy implications of the key aspects of the
scientific and technical information pertaining to this reconsideration of the 2020 decision on the
O3 primary standard. Specifically, the chapter presents key aspects of the available evidence of
the health effects of O3, as documented in the 2020 ISA, with support from the prior ISA and
AQCDs, and associated public health implications.1 It also presents key aspects of the
quantitative risk and exposure analyses conducted for the 2020 review (and originally presented
in the 2020 PA), with the details provided in Appendices 3C and 3D. Together this information
provides the basis for our evaluation of the scientific information regarding health effects of O3
in ambient air and the potential for effects to occur under air quality conditions associated with
the existing standard (or any alternatives considered), as well as the associated implications for
public health.
Our evaluation in this chapter is framed around key policy-relevant questions derived
from the IRP (IRP, section 3.1.1), and also takes into account, as relevant, assessments of the
evidence and quantitative exposure/risk analyses in prior reviews. In this way we identify key
policy-relevant considerations and summary conclusions regarding the public health protection
provided by the current standard for the Administrator's consideration in this reconsideration of
the 2020 decision on the primary O3 standard.
Within this chapter, background information on the current standard is summarized in
section 3.1. The general approach for considering the available information, including policy-
relevant questions identified to frame our policy evaluation, is summarized in section 3.2. Key
aspects of the available health effects evidence and associated public health implications and
uncertainties are addressed in section 3.3, and the quantitative exposure and risk information,
with associated uncertainties, is addressed in section 3.4. Section 3.5 summarizes the key
evidence- and exposure/risk-based considerations identified in our evaluation, and also presents
associated preliminary conclusions of this analysis. Key remaining uncertainties and areas for
future research are identified in section 3.6.
1 The ISA builds on evidence and conclusions from previous assessments, focusing on synthesizing and integrating
the newly available evidence (ISA, section IS. 1.1). Past assessments are generally cited when providing further,
still relevant, details that informed the current assessment but are not repeated in the latest assessment.
April 2022
3-1
External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
3.1 BACKGROUND ON THE CURRENT STANDARD
The current primary O3 standard of 0.070 ppm,2 as the annual fourth-highest daily
maximum 8-hour average concentration, averaged across three consecutive years, was set in
2015 and retained without revision in 2020 (80 FR 65292, October 26, 2015; 85 FR 87256,
December 31, 2020). Establishment of this standard, and its retention in 2020, were based on the
extensive body of evidence spanning several decades documenting the causal relationship
between O3 exposure and a broad range of respiratory effects, that had been augmented by
evidence available since the 2008 review (80 FR 65292, October 26, 2015; 2013 ISA, p. 1-14).
A key consideration driving the 2015 decision was the newly available evidence of adverse
respiratory effects from controlled human exposure studies in healthy adults at an exposure
concentration lower than had been previously studied (80 FR 65342-47 and 65362-66, October
26, 2015). While the study subjects in the vast majority of the controlled human exposure studies
(and in all of these studies conducted at the lowest exposures) are healthy adults, the EPA's
establishment of the standard in 2015, and its retention in 2020, focused particularly on
implications of these studies to insure protection of much less well studied at-risk populations,3
such as people with asthma, and particularly children with asthma (80 FR 65343, October 26,
2015; 85 FR 87305, December 31, 2020).
The 2020 review of the 2015 standard also considered differences in the health effects
evidence since 2015 for effects other than respiratory effects. Specifically, the newly available
evidence supported updated conclusions regarding metabolic effects, cardiovascular effects, and
mortality (ISA, Table ES-1). For example, while the evidence available in the 2015 review was
sufficient to conclude that the relationships for short-term O3 exposure with cardiovascular
health effects and mortality were likely to be causal, that conclusion was no longer supported by
the more expansive evidence base which the 2020 ISA determines to be suggestive of, but not
sufficient to infer, a causal relationship for these health effect categories (ISA, Appendix 4,
section 4.1.17; Appendix 6, section 6.1.8). Further, newly available evidence since 2015 supports
a new determination that the relationship between short-term O3 exposure and metabolic effects
2	Although ppm are the units in which the level of the standard is defined, the units, ppb, are more commonly used
throughout this PA for greater consistency with their use in the more recent literature. The level of the current
primary standard, 0.070 ppm, is equivalent to 70 ppb.
3	As used here and similarly throughout the document, the term population refers to persons having a quality or
characteristic in common, such as, and including, a specific pre-existing illness or a specific age or lifestage. A
lifestage refers to a distinguishable time frame in an individual's life characterized by unique and relatively stable
behavioral and/or physiological characteristics that are associated with development and growth. Identifying at-
risk populations includes consideration of intrinsic (e.g., genetic or developmental aspects) or acquired (e.g.,
disease or smoking status) factors that increase the risk of health effects occurring with exposure to O3 as well as
extrinsic, nonbiological factors, such as those related to socioeconomic status, reduced access to health care, or
exposure.
April 2022
3-2 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
is likely to be causal (ISA, section IS.4.3.3). The basis for this conclusion is largely experimental
animal studies in which the exposure concentrations are well above those in the controlled
human exposure studies for respiratory effects as well as above those likely to occur in areas of
the U.S. that meet the current standard (85 FR 87270, December 31, 2020). Thus, while new
conclusions were reached in the 2020 review for these non-respiratory effect categories, they did
not lead to a change in focus for the standard, which continued to be protection of at-risk
populations from respiratory effects, as the effects causally related to O3 at the lowest exposure
levels.
With regard to respiratory effects, the health effects evidence base available in the 2015
and 2020 reviews documents a broad range of effects associated with O3 exposure (2013 ISA, p.
1-14; 2020 ISA, p. ES4-10). Such effects range from small, transient and/or reversible changes in
pulmonary function and pulmonary inflammation (documented in controlled human exposure
studies involving exposures ranging from 1 to 8 hours) to more serious health outcomes such as
emergency department visits and hospital admissions, which have been associated with ambient
air concentrations of O3 in epidemiologic studies (2013 ISA, section 6.2; 2020 ISA, Appendix 3,
sections 3.1.5.1 and 3.1.5.2).4
Across the different study types, the controlled human exposure studies, which were
recognized to provide the most certain evidence indicating the occurrence of health effects in
humans following specific O3 exposures, additionally document the roles of ventilation rate,5
exposure duration, and exposure concentration, in eliciting responses to O3 exposure (80 FR
65343, October 26, 2015; 2014 PA, section 3.4). For example, the exposure concentrations
eliciting a given level of response in subjects at rest are higher than those eliciting a response in
subjects exposed while at elevated ventilation, such as while exercising (2013 ISA, section
6.2.1.1).6 Accordingly, of particular interest is the extent and magnitude of exposures during
4	In addition to extensive controlled human exposure and epidemiologic studies, the evidence base includes
experimental animal studies that provide insight into potential modes of action for these effects, contributing to
the coherence and robust nature of the evidence.
5	Ventilation rate (VE) is a specific technical term referring to breathing rate in terms of volume of air taken into the
body per unit of time. A person engaged in different activities will exert themselves at different levels and
experience different ventilation rates.
6	In the controlled human exposure studies, the magnitude or severity of the respiratory effects induced by O3 is
influenced by ventilation rate (in addition to exposure duration and exposure concentration), with physical
activity increasing ventilation and potential for effects. In studies of generally healthy young adults exposed while
at rest for 2 hours, 500 ppb is the lowest concentration eliciting a statistically significant Ch-induced reduction in
group mean lung function measures, while a much lower concentration produces a statistically significant
response in lung function when the ventilation rate of the group of study subjects is sufficiently increased with
exercise (2013 ISA, section 6.2.1.1). For example, the lowest exposure concentration examined that elicited a
statistically significant Ch-induced reduction in group mean lung function in an exposure of 2 hours or less was
120 ppb in a 1-hour exposure of trained cyclists who maintained a high exertion level throughout the exposure
April 2022
3-3 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
periods of elevated ventilation, such as while exercising, under air quality conditions of interest.
Thus, key considerations in the establishment of the standard in 2015 and in its review in 2020
were the population exposure and risk assessments performed for air quality conditions
associated with just meeting the standard (and with alternative air quality scenarios). These
assessments, which included a focus on the at-risk populations of children and children with
asthma, analyzed the occurrence of exposures to O3 concentrations of interest by individuals
breathing at elevated rates and characterized the associated risk.
The Administrator's judgment in establishing the standard in 2015 was based primarily
on the extensive evidence of respiratory effects health effects evidence for O3 with a focus on the
public health implications of the exposure and risk analyses conducted in that review. In the
review concluded in 2020, the Agency considered the health effects evidence base, including that
newly available since the 2015 decision, and the updated exposure/risk analyses. In 2020, the
Administrator reaffirmed judgments of the 2015 decision associated with establishment of the
different elements of the standard and made additional judgments reflecting the information
current to the review, concluding that the existing standard, set in 2015, continued to provide the
requisite public health protection with an adequate margin of safety (85 FR 87300-87306,
December 31, 2020). Key aspects of the health effects evidence and exposure and risk
information available in the 2020 review, as well as the associated judgments reflecting
consideration of associated limitations and uncertainties, are summarized below for each of the
four basic elements of the NAAQS (indicator, averaging time, form, and level), in turn.
In 1979, O3 was established as the indicator for a standard meant to provide protection
against photochemical oxidants in ambient air (44 FR 8202, February 8, 1979). In setting the
current standard in 2015 and reviewing it in 2020, the Administrator considered the available
information presented in the ISA and PA, along with advice from the CASAC and public
comment. Both the 2013 and 2020 IS As specifically noted that O3 is the only photochemical
oxidant (other than nitrogen dioxide) that is routinely monitored and for which a comprehensive
database exists (2013 ISA, section 3.6; 80 FR 65347, October 26, 2015; 2020 ISA, p. IS-3; 85
FR 87301, December 31, 2020). The 2020 ISA further noted that "the primary literature
evaluating the health and ecological effects of photochemical oxidants includes ozone almost
exclusively as an indicator of photochemical oxidants" (2020 ISA, p. IS-3). In both reviews, the
CASAC indicated its support for O3 as the appropriate indicator. Based on these considerations
and public comments, the Administrators in both reviews concluded that O3 remains the most
appropriate indicator for a standard meant to provide protection against photochemical oxidants
period (2013 ISA, section 6.2.1.1; Gong et al., 1986) or after 2-hour exposure (heavy intermittent exercise) of
young healthy adults (2013 ISA, section 6.2.1.1; McDonnell et al., 1983).
April 2022
3-4 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
in ambient air, and they retained O3 as the indicator for the primary standard (80 FR 65347,
October 26, 2015; 85 FR 87306; December 31, 2020).
The 8-hour averaging time for the primary O3 standard was established in 1997 with the
decision to replace the then-existing 1-hour standard with an 8-hour standard (62 FR 38856, July
18, 1997). The decision in that review was based on newly available evidence from numerous
controlled human exposure studies in healthy adults of adverse respiratory effects resulting from
6- to 8-hour exposures, as well as quantitative analyses indicating the control provided by an 8-
hour averaging time of both 8-hour and 1-hour peak exposures and associated health risk (62 FR
38861, July 18, 1997; U.S. EPA, 1996). The 1997 decision was also consistent with advice from
the CASAC (62 FR 38861, July 18, 1997; 61 FR 65727, December 13, 1996). This averaging
time has been retained in each of the three NAAQS reviews since then (73 FR 16436, March 27,
2008; 80 FR 65292, October 26, 2015; 85 FR 87256, December 31, 2020). In the establishment
of the existing standard in 2015 and its review in 2020, the averaging time was retained in light
of both the strong evidence for 03-associated respiratory effects following short-term exposures
and the available evidence related to effects following longer-term exposures (80 FR 65347-50,
October 26, 2015). The 2015 decision on a revised standard recognized that an 8-hour averaging
time is similar to the exposure periods evaluated in the more recent controlled human exposure
studies conducted at the lowest concentrations, and that other evidence, including that from
epidemiologic studies did not provide a strong basis of support for alternative averaging times
(80 FR 65348, October 26, 2015). Further, in 2015 the considerations on a revised standard also
included consideration of the extent to which the available evidence and exposure/risk
information suggested that a standard with an 8-hour averaging time can provide protection
against respiratory effects associated with longer-term exposures to ambient air O3. Based on the
then-available evidence and information discussed in detail in the 2013 ISA, 2014 Health Risk
and Exposure Assessment (HREA), and 2014 PA, along with CASAC advice and public
comments, the Administrator concluded that a standard with an 8-hour averaging time (and
revised level) could effectively limit health effects attributable to both short- and long-term O3
exposures and that it was appropriate to retain the 8-hour averaging time (80 FR 65350, October
26, 2015). The EPA reached similar conclusions in the 2020 review and retained the 8-hour
averaging time (85 FR 87306; December 31, 2020).
While giving foremost consideration to the adequacy of public health protection provided
by the combination of all elements of the standard, including the form, in 2015 the Administrator
placed considerable weight on the findings from prior reviews with regard to the use of the ซth-
high metric, as described below (80 FR 65350-65352, October 26, 2015). Based on these
findings and consideration of CASAC advice, the Administrator judged it appropriate to retain
the fourth-high form, more specifically the annual fourth-highest daily maximum 8-hour O3
April 2022
3-5 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
average concentration, averaged over 3 years (80 FR 65352, October 26, 2015). The EPA
reached similar conclusions in the 2020 review and retained the form of the annual fourth-
highest daily maximum 8-hour O3 average concentration, averaged over 3 years (85 FR 87306;
December 31, 2020).
The concentration-based form (e.g., the //th-high metric) of the existing standard was
established in the 1997 review when it was recognized that such a form better reflects the
continuum of health effects associated with increasing O3 concentrations than an expected
exceedance form,7 which had been the form of the standard prior to 1997. Unlike an expected
exceedance form, a concentration-based form gives proportionally more weight to years when 8-
hour O3 concentrations are well above the level of the standard than years when 8-hour O3
concentrations are just above the level of the standard. With regard to a specific concentration-
based form, the fourth-highest daily maximum was selected in 1997, recognizing that a less
restrictive form (e.g., fifth highest) would allow a larger percentage of sites to experience O3
peaks above the level of the standard, and would allow more days on which the level of the
standard may be exceeded when the site attains the standard (62 FR 38868-38873, July 18,
1997), and there was not a basis identified for selection of a more restrictive form (62 FR 38856,
July 18, 1997). In subsequent reviews, the EPA also considered the potential value of a
percentile-based form, recognizing that such a statistic is useful for comparing datasets of
varying length because it samples approximately the same place in the distribution of air quality
values, whether the dataset is several months or several years long (73 FR 16474-75, March 27,
2008). However, the EPA concluded that, because of the differing lengths of the monitoring
season for O3 across the U.S., a percentile-based statistic would not be effective in ensuring the
same degree of public health protection across the country.8 The importance of a form that
provides stability to ongoing control programs was also recognized.9 Advice from the CASAC in
the 2015 review supported this, stating that this concentration-based form that is averaged over
three years "provides health protection while allowing for atypical meteorological conditions that
can lead to abnormally high ambient ozone concentrations which, in turn, provides programmatic
7	The first O3 standard, set in 1979 as an hourly standard, had an expected exceedance form, such that attainment
was defined as when the expected number of days per calendar year, with maximum hourly average concentration
greater than 0.12 ppm, was equal to or less than 1 (44 FR 8202, February 8, 1979).
8	Specifically, a percentile-based form would allow more days with higher air quality values (i.e., higher 03
concentrations) in locations with longer O3 seasons relative to locations with shorter 03 seasons.
9	In the case of O3, for example, it was noted that it was important to have a form that provides stability and
insulation from the impacts of extreme meteorological events that are conducive to O3 occurrence. Such events
could have the effect of reducing public health protection, to the extent they result in frequent shifts in and out of
attainment due to meteorological conditions because such frequent shifting could disrupt an area's ongoing
implementation plans and associated control programs (73 FR 16475, March 27, 2008).
April 2022
3-6 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
stability" (Frey, 2014, p. 6; 80 FR 65352, October 26, 2015). Advice from the CASAC did not
raise objections with the indicator, averaging time and form of the existing standard (Cox, 2020).
In establishing the level of the standard in 2015 and in the decision to retain it in 2020,
the Administrator at each time carefully considered: (1) the assessment of the health effects
evidence and conclusions reached in the ISA; (2) the available quantitative exposure/risk
analyses, including associated limitations and uncertainties, described in detail in the HREA (in
the 2015 review) or appendices of the 2020 PA (in 2020); (3) considerations and staff
conclusions and associated rationales in the PA; (4) advice and comments from the CASAC;
and, (5) public comments (80 FR 65362, October 26, 2015; 85 FR 37300, December 31, 2020).
In weighing the health effects evidence and making judgments regarding the public health
significance of the quantitative estimates of exposures and risks allowed by the existing standard
and potential alternative standards considered, as well as judgments regarding margin of safety,
both of the decisions, in 2015 and 2020, considered the currently available information,
including EPA judgments in prior reviews, advice from the CASAC, statements of the American
Thoracic Society (ATS), an organization of respiratory disease specialists, and public comments.
In so doing, each decision recognized that the determination of what constitutes an adequate
margin of safety is expressly left to the judgment of the EPA Administrator. See Lead Industries
Ass'n v. EPA, 647 F.2d 1130, 1161-62 (D.C. Cir 1980);Mississippi v. EPA, 744 F.3d 1334, 1353
(D.C. Cir. 2013). In NAAQS reviews generally, evaluations of how particular primary standards
address the requirement to provide an adequate margin of safety include consideration of such
factors as the nature and severity of the health effects, the size of the sensitive population(s) at
risk, and the kind and degree of the uncertainties present. Consistent with past practice and long-
standing judicial precedent, in both the 2015 and 2020 decisions, the Administrator took into
account the need for an adequate margin of safety as an integral part of their decision-making.
The 2015 decision to set the level of the revised primary O3 standard at 70 ppb placed the
greatest weight on the results of controlled human exposure studies and on quantitative analyses
based on information from these studies, particularly analyses comparing exposure estimates for
study area populations of children at elevated exertion to exposure benchmark concentrations
(exposures of concern), consistent with CASAC advice and interpretation of the scientific
evidence (80 FR 65362, October 26, 2015; Frey, 2014b).10 This weighting reflected the
recognition that controlled human exposure studies provide the most certain evidence indicating
the occurrence of health effects in humans following specific O3 exposures, and, in particular,
10 The Administrator viewed the results of other quantitative analyses in this review - the lung function risk
assessment, analyses of O3 air quality in locations of epidemiologic studies, and epidemiologic-study-based
quantitative health risk assessment - as being of less utility for selecting a particular standard level among a range
of options (80 FR 65362, October 26, 2015).
April 2022
3-7 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
that the effects reported in the controlled human exposure studies are due solely to O3 exposures,
and are not complicated by the presence of co-occurring pollutants or pollutant mixtures (as is
the case in epidemiologic studies) (80 FR 65362-65363, October 26, 2015).n. With regard to this
evidence, the Administrator at that time recognized that: (1) the largest respiratory effects, and
the broadest range of effects, have been studied and reported following exposures to 80 ppb O3
or higher (i.e., decreased lung function, increased airway inflammation, increased respiratory
symptoms, airway hyperresponsiveness, and decreased lung host defense12); (2) exposures to O3
concentrations somewhat above 70 ppb13 have been shown to both decrease lung function and to
result in respiratory symptoms; and (3) exposures to O3 concentrations as low as 60 ppb have
been shown to decrease lung function and to increase airway inflammation (80 FR 65363,
October 26, 2015). The Administrator also noted that 70 ppb was well below the O3 exposure
concentration documented to result in the widest range of respiratory effects (i.e., 80 ppb), and
also below the lowest O3 exposure concentration shown in 6.6-hour exposures with quasi-
continuous exercise to result in the combination of lung function decrements and respiratory
symptoms (80 FR 65363, October 26, 2015).
Consideration of the controlled human exposure study results and quantitative analyses
based on information from those studies focused primarily, both in 2015 and 2020, on the
exposure-based comparison-to-benchmarks analysis. This analysis characterizes the extent to
which individuals in at-risk populations could experience O3 exposures, while engaging in their
daily activities, with the potential to elicit the effects reported in controlled human exposure
studies for concentrations at or above specific benchmark concentrations. The analysis conducted
for the 2020 review reflected a number of updates and improvements and provided estimates
with reduced uncertainty compared to those from the 2015 review (see section 3.4.1 below for
details). The results for analyses in both reviews are characterized through comparison of
exposure concentration estimates to three benchmark concentrations of O3: 60, 70, and 80 ppb.
These are based on the three lowest concentrations targeted in studies of 6- to 6.6-hour exposures
of generally healthy adults engaging in quasi-continuous exercise (at a moderate level of
exertion), and that yielded different occurrences, of statistical significance, and severity of
11	Other quantitative exposure/risk analyses (e.g., the lung function risk assessment, analyses of O3 air quality in
locations of epidemiologic studies, and epidemiologic-study-based quantitative health risk assessment) were
viewed as providing information in support of the 2015 decision to revise the then-current standard level of 75
ppb, but of less utility for selecting a particular standard level among a range of options (80 FR 65362, October
26, 2015).
12	Host defense refers to a decreased ability to repel pathogens and resist infection.
13	For the 70 ppb target exposure, the time weighted average concentration across the full 6.6-hour exposure was 73
ppb and the mean O3 concentration during the exercise portion of the study protocol was 72ppb, based on O3
measurements during the six 50-minute exercise periods (Schelegle et al., 2009).
April 2022
3-8 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
respiratory effects (80 FR 65312, October 26, 2015; 85 FR 87277; December 31, 2020; 2020 PA,
section 3.3.3).14 A second exposure-based analysis provided population risk estimates of the
occurrence of days with Cb-attributable lung function reductions of varying magnitudes by using
the exposure-response (E-R) information in the form of E-R functions or other quantitative
descriptions of biological processes.15 These latter estimates were given less weight in the
Administrator's decisions in both the 2015 and 2020 reviews due to a recognition of relatively
greater uncertainty in interpretation of the results. Analyses in the 2020 PA quantitatively
illustrated this greater uncertainty associated with the lung function risk estimates related to their
greater reliance on estimation of responses at exposure levels below those that have been studied
(80 FR 65464, October 26, 2015; 85 FR 87277, December 31, 2020; 2020 PA, section 3.4.4).
In the 2015 decision to revise the standard level to 70 ppb (while retaining the existing
indicator, averaging time and form) and also the 2020 decision to retain that level (and all other
standard elements), without revision, the exposure analysis results for each of the three
benchmarks were considered in the context of the Administrator judgments concerning each
benchmark. Such judgments of the Administrator in setting the standard level of 70 ppb in 2015
are briefly summarized below. These are followed by a description of key aspects of the
considerations and judgments associated with the decision to retain this standard in 2020.
In the 2015 considerations of the degree of protection to be provided by a revised
standard, and the extent to which that standard would be expected to limit population exposures
to the broad range of O3 exposures shown to result in health effects, the Administrator focused
particularly on the exposure analysis estimates of two or more exposures of concern. Placing the
most emphasis on a standard that limits repeated occurrences of exposures at or above the 70 and
80 ppb benchmarks, while at elevated ventilation, the Administrator noted that a standard of the
existing form and averaging time with a revised level of 70 ppb was estimated to eliminate the
occurrence of two or more days with exposures at or above 80 ppb and to virtually eliminate the
occurrence of two or more days with exposures at or above 70 ppb for all children and children
with asthma, even in the worst-case year and location evaluated (80 FR 65363-65364, October
26, 2015).16 The Administrator's consideration of exposure estimates at or above the 60 ppb
benchmark (focused most particularly on multiple occurrences), an estimated exposure to which
14	The studies given primary focus were those for which 03 exposures occurred over the course of 6.6 hours during
which the subjects engaged in six 50-minute exercise periods separated by 10-minute rest periods, with a 35-
minute lunch period occurring after the third hour (e.g., Folinsbee et al., 1988 and Schelegle et al., 2009).
Responses after O3 exposure were compared to those after filtered air exposure.
15	The E-R information and quantitative models derived from it are based on controlled human exposure studies.
16	Under conditions just meeting an alternative standard with a level of 70 ppb across the 15 urban study areas, the
estimate for two or more days with exposures at or above 70 ppb was 0.4% of children, in the worst year and
worst area (80 FR 65313, Table 1, October 26, 2015).
April 2022
3-9 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
the Administrator was less confident would result in adverse effects,17 was primarily in the
context of considering the extent to which the health protection provided by a revised standard
included a margin of safety against the occurrence of adverse Cb-induced effects (80 FR 65364,
October 26, 2015). In this context, the Administrator noted that a revised standard with a level of
70 ppb was estimated to protect the vast majority of children in urban study areas (i.e., about
96% to more than 99% of children in individual areas) from experiencing two or more days with
exposures at or above 60 ppb (while at moderate or greater exertion).18
Given the considerable protection provided against repeated exposures of concern for all
three benchmarks, including the 60 ppb benchmark, the Administrator in 2015 judged that a
standard with a level of 70 ppb would incorporate a margin of safety against the adverse O3-
induced effects shown to occur in the controlled human exposure studies following exposures
(while at moderate or greater exertion) to a concentration somewhat higher than 70 ppb (80 FR
65364, October 26, 2015).19 The Administrator also judged the estimates of one or more
exposures (while at moderate or greater exertion) at or above 60 ppb to also provide support for
her somewhat broader conclusion that "a standard with a level of 70 ppb would incorporate an
adequate margin of safety against the occurrence of O3 exposures that can result in effects that
are adverse to public health" (80 FR 65364, October 26, 2015).20
17	The 2015 decision noted that "the Administrator is notably less confident in the adversity to public health of the
respiratory effects that have been observed following exposures to O3 concentrations as low as 60 ppb," citing,
among other considerations, "uncertainty in the extent to which short-term, transient population-level decrease in
FEVi would increase the risk of other, more serious respiratory effects in that population" (80 FR 54363, October
26, 2015). Note: FEVi (a measure of lung function response) is the forced expiratory volume in one second.
18	The 2015 decision also noted the Administrator's consideration of the extent to which she judged that adverse
effects could occur following specific O3 exposures related to each of the three benchmarks. The Administrator
recognized the interindividual variability in responsiveness in her interpretation of the exposure analysis results
noting noted "that not everyone who experiences an exposure of concern, including for the 70 ppb benchmark, is
expected to experience an adverse response," further judging "that the likelihood of adverse effects increases as
the number of occurrences of O3 exposures of concern increases." And "[i]n making this judgment, she note[d]
that the types of respiratory effects that can occur following exposures of concern, particularly if experienced
repeatedly, provide a plausible mode of action by which O3 may cause other more serious effects. Therefore, her
decisions on the primary standard emphasize [d] the public health importance of limiting the occurrence of
repeated exposures to O3 concentrations at or above those shown to cause adverse effects in controlled human
exposure studies" (80 FR 65331, October 26, 2015).
19	In so judging, she noted that the CAS AC had recognized the choice of a standard level within the range it
recommended based on the scientific evidence (which was inclusive of 70 ppb) to be a policy judgment (80 FR
65355, October 26, 2015; Frey, 2014b).
20	While the Administrator was less concerned about single exposures, especially for the 60 ppb benchmark, she
judged the HREA of one-or-more estimates informative to margin of safety considerations. In this regard, she
noted that "a standard with a level of 70 ppb is estimated to (1) virtually eliminate all occurrences of exposures of
concern at or above 80 ppb; (2) protect the vast majority of children in urban study areas from experiencing any
exposures of concern at or above 70 ppb (i.e., > about 99%, based on mean estimates; Table 1); and (3) to achieve
substantial reductions, compared to the [then-]current standard, in the occurrence of one or more exposures of
concern at or above 60 ppb (i.e., about a 50% reduction; Table 1)" (80 FR 65364, October 26, 2015).
April 2022
3-10 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
The 2020 review of the 2015 standard also focused on the exposure-based analyses in the
context of results from the controlled human exposure studies of exposures from 60 to 80 ppb,
recognizing this information on exposure concentrations found to elicit respiratory effects in
exercising study subjects to be unchanged from what was available in the 2015 review (2020 PA,
section 3.3.1; 85 FR 87302, December 31, 2020).21 In considering the significance of responses
documented in these studies and in the full evidence base for the purposes of judging
implications of the available information on public health protection provided by the current
standard, several aspects, limitations and uncertainties of the evidence base were noted. For
example, as also recognized in 2015, the responses reported from exposures ranging from 60 to
80 ppb are transient and reversible in the study subjects who are largely healthy, adult subjects.
Such study data are lacking at these exposure levels for children and people with asthma, and the
evidence indicates that such responses, if repeated or sustained, particularly in people with
asthma, pose risks of effects of greater concern, including asthma exacerbation, as cautioned by
the CASAC (85 FR 87302, December 31, 2020).22
As in 2015, the Administrator in 2020 also considered statements from the ATS, as well
as judgments made by the EPA in considering similar effects in previous NAAQS reviews (85
FR 87270-72, 87302-87305, December 31, 2020; 80 FR 65343, October 26, 2015). The ATS
statements included one newly available in the 2020 review (Thurston et al., 2017), which is
generally consistent with the prior statement (that was considered in the 2015 review) including
the attention that the prior statement gives to at-risk or vulnerable population groups, while also
broadening the discussion of effects, responses, and biomarkers to reflect the expansion of
scientific research in these areas (ATS, 2000; Thurston et al., 2017). The Administrator
recognized the role of such statements, as described by the ATS, as proposing principles or
considerations for weighing the evidence rather than offering "strict rules or numerical criteria"
(ATS, 2000, Thurston et al., 2017). In keeping with this intent of these statements (to avoid
21	With regard to the epidemiologic studies of respiratory effects, the Administrator recognized that, as a whole,
these investigations of associations between O3 and respiratory effects and health outcomes (e.g., asthma-related
hospital admission and emergency department visits) provided strong support for the conclusions of causality but
the studies were less informative regarding exposure concentrations associated with O3 air quality conditions that
meet the current standard. He noted that the evidence base in the 2020 review did not include new evidence of
respiratory effects associated with appreciably different exposure circumstances than the evidence available in the
2015 review, including particularly any circumstances that would also be expected to be associated with air
quality conditions likely to occur under the current standard.
22	The CASAC noted that "[a]rguably the most important potential adverse effect of acute ozone exposure in a child
with asthma is not whether it causes a transient decrement in lung function, but whether it causes an asthma
exacerbation'' and that O3 "has respiratory effects beyond its well-described effects on lung function," including
increases in airway inflammation which also have the potential to increase the risk for an asthma exacerbation.
The CASAC further cautioned with regard to repeated episodes of airway inflammation, indicating that they have
the potential to contribute to irreversible reductions in lung function (Cox, 2020, Consensus Responses to Charge
Questions pp. 7-8).
April 2022
3-11 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
specific criteria), the statements, in discussing what constitutes an adverse health effect, do not
comprehensively describe all the biological responses raised, e.g., with regard to magnitude,
duration or frequency of small pollutant-related changes in lung function.
The Administrator also recognized the limitations in the available evidence base with
regard to our understanding of these aspects of such changes that may be associated with
exposure concentrations of interest (e.g., as estimated in the exposure analysis). Notwithstanding
these limitations and associated uncertainties, he took note of the emphasis of the earlier ATS
statement on consideration of individuals with preexisting compromised function, such as that
resulting from asthma (an emphasis which is reiterated and strengthened in the current
statement), agreeing that these were important considerations in his judgment on the adequacy of
protection provided by the current standard for at-risk populations.
Among such important considerations, it was recognized that the controlled human
exposure studies, primarily conducted in healthy adults, on which the depth of our understanding
of 03-related health effects is based, in combination with the larger evidence base, informs our
conceptual understanding of O3 responses in people with asthma and in children. Aspects of the
EPA's understanding continue to be limited, however, including with regard to the risk of
particular effects and associated severity for these less studied population groups that may be
posed by 7-hour exposures with exercise to concentrations as low as 60 ppb that are estimated in
the exposure analyses for the 2020 review (85 FR 87303, December 31, 2020).
Collectively, these aspects of the evidence and associated uncertainties contributed to the
recognition that for O3 in the 2020 review, as for other pollutants and other reviews, the available
evidence base in a NAAQS review generally reflects a continuum, consisting of levels at which
scientists generally agree that health effects are likely to occur, through lower levels at which the
likelihood and magnitude of the response become increasingly uncertain. As is the case in
NAAQS reviews in general, the 2020 decision regarding the primary O3 standard depended on a
variety of factors, including science policy judgments and public health policy judgments. These
factors included judgments regarding aspects of the evidence and exposure/risk estimates, such
as judgments concerning the Administrator's interpretation of the different benchmark
concentrations, in light of the available evidence and of associated uncertainties, as well as
judgments on the public health significance of the effects that have been observed at the
exposures evaluated in the health effects evidence. These judgments are rooted in interpretation
of the evidence, which reflects a continuum of health-relevant exposures, with less confidence
and greater uncertainty in the existence of adverse health effects as one considers lower O3
exposures. The factors relevant to judging the adequacy of the standards also included the
interpretation of, and decisions as to the relative weight to place on, different aspects of the
results of the exposure and risk assessment for the areas studied and the associated uncertainties.
April 2022
3-12 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
Together, factors identified here informed the Administrator's judgment about the degree of
protection that is requisite to protect public health with an adequate margin of safety, including
the health of sensitive groups, and, accordingly, his conclusion of the requisiteness of the
existing standard to protect public health with an adequate margin of safety (85 FR 87303,
December 31, 2020).
In placing greater weight and giving primary attention to the comparison-to-benchmarks
analysis, the Administrator recognized that, as noted in the 2020 PA, the comparison-to-
benchmarks analysis (newly updated in the 2020 review with a number of improvements over
the 2014 analysis, as described in section 3.4.1 below) provides for characterization of risk for
the broad array of respiratory effects documented in the controlled human exposure studies,
facilitating consideration of an array of respiratory effects, including but not limited to lung
function decrements (85 FR 87294, December 31, 2020). The Administrator recognized the three
benchmark concentrations (60, 70 and 80 ppb) to represent exposure conditions (during quasi-
continuous exercise) associated with different levels of respiratory response (both with regard to
the array of effects and severity of individual effects) in the subjects studied and to inform his
judgments on different levels of risk that might be posed to unstudied members of at-risk
populations. The highest benchmark concentration (80 ppb) represented an exposure where
multiple controlled human exposure studies involving 6.6-hour exposures during quasi-
continuous exercise demonstrate a range of Cb-related respiratory effects including inflammation
and airway responsiveness, as well as respiratory symptoms and lung function decrements in
healthy adult subjects. The second benchmark (70 ppb) represented an exposure level below the
lowest exposures that have reported both statistically significant lung function decrements23 and
increased respiratory symptoms (reported at 73 ppb, Schelegle et al 2009) or statistically
significant increases in airway resistance and responsiveness (reported at 80 ppb, Horstman et
al., 1990). The lowest benchmark (60 ppb) represents still lower exposure, and a level for which
findings from controlled human exposure studies of largely healthy subjects have included:
statistically significant decrements in lung function (with mean decrements ranging from 1.7% to
3.5% across the four studies with average exposures of 60 to 63 ppb), but not respiratory
symptoms; and, a statistically significant increase in a biomarker of airway inflammatory
response relative to filtered air exposures in one study (Kim et al, 2011).
23 The study group mean lung function decrement for the 73 ppb exposure was 6%, with individual decrements of
15% or greater (moderate or greater) in about 10% of subjects and decrements of 10% or greater in 19% of
subjects. Decrements of 20% or greater were reported in 6.5% of subjects (Schelegle et al., 2009; 2020 PA, Table
3-2 and Appendix 3D, Table 3D-20). In studies of 80 ppb exposure, the percent of study subjects with individual
FEVi decrements of this size ranged up to nearly double this (2020 PA, Appendix 3D, Table 3D-20).
April 2022
3-13 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
In turning to the exposure/risk analysis results, the Administrator considered the
controlled human exposure evidence represented by these benchmarks noting that due to
differences among individuals in responsiveness, not all people experiencing exposures (e.g., to
73 ppb), experience a response, such as a lung function decrement, and among those
experiencing a response, not all will experience an adverse effect (85 FR 87304, December 31,
2020). Accordingly, the Administrator noted that not all people estimated to experience an
exposure of 7-hour duration while at elevated exertion above even the highest benchmark would
be expected to experience an adverse effect, even members of at-risk populations. With these
considerations in mind, he noted that while single occurrences could be adverse for some people,
particularly for the higher benchmark concentration where the evidence base is stronger, the
potential for adverse response and greater severity increased with repeated occurrences (as
cautioned by the CASAC). The Administrator also noted that while the exposure/risk analyses
provide estimates of exposures of the at-risk population to concentrations of potential concern,
they do not provide information on how many of such populations will have an adverse health
outcome. Accordingly, in considering the exposure/risk analysis results, while giving due
consideration to occurrences of one or more days with an exposure at or above a benchmark,
particularly the higher benchmarks, he judged multiple occurrences to be of greater concern than
single occurrences.
In this context, the Administrator considered the exposure risk estimates, focusing first on
the results for the highest benchmark concentration (80 ppb), which represents an exposure well
established to elicit an array of responses in sensitive individuals among study groups of largely
healthy adult subjects, exposed while at elevated exertion. Similar to judgments of past
Administrators, the Administrator in 2020 judged these effects in combination and severity to
represent adverse effects for individuals in the population group studied, and to pose a risk of
adverse effects for individuals in at-risk populations, most particularly people with asthma, as
noted above. Accordingly, he judged that the primary standard should provide protection from
such exposures. In considering the exposure/risk estimates, he focused on the results for children,
and children with asthma, given the higher frequency of exposures of potential concern for
children compared to adults, in terms of percent of the population groups. The exposure/risk
estimates indicated more than 99.9% to 100% of children and children with asthma, on average
across the three years, to be protected from one or more occasions of exposure at or above this
level; the estimate is 99.9% of children with asthma and of all children for the highest year and
study area (85 FR 87279, Table 2, December 31, 2020). Further, no children in the simulated
populations (zero percent) were estimated to be exposed more than once (two or more occasions)
in the 3-year simulation to 7-hr concentrations, while at elevated exertion, at or above 80 ppb (85
FR 87279, Table 2, December 31, 2020). These estimates indicated strong protection against
April 2022
3-14 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
exposures of at-risk populations that have been demonstrated to elicit a wide array of respiratory
responses in multiple studies (85 FR 87304, December 31, 2020).
The Administrator next considered the results for the second benchmark concentration
(70 ppb), which is just below the lowest exposure concentration (73 ppb) for which a study has
reported a combination of a statistically significant increase in respiratory symptoms and
statistically significant lung function decrements in sensitive individuals in a study group of
largely healthy adult subjects, exposed while at elevated exertion (Schelegle et al., 2009).
Recognizing the lack of evidence for people with asthma from studies at 80 ppb and 73 ppb, as
well as the emphasis in the ATS statement on the vulnerability of people with compromised
respiratory function, such as people with asthma, the Administrator judged it appropriate that the
standard protect against exposure, particularly multiple occurrences of exposure, to somewhat
lower levels. In so doing, he noted that the exposure/risk estimates indicate more than 99% of
children with asthma, and of all children, to be protected from one or more occasions in a year,
on average, of 7-hour exposures to concentrations at or above 70 ppb, while at elevated exertion
(85 FR 87279, Table 2, December 31, 2020). The estimate is 99% of children with asthma for
the highest year and study area (85 FR 87279, Table 2, December 31, 2020). Further, he noted
that 99.9% of these groups were estimated to be protected from two or more such occasions, and
100%) from still more occasions. These estimates also indicated strong protection of at-risk
populations against exposures similar to those demonstrated to elicit lung function decrements
and increased respiratory symptoms in healthy subjects, a response described as adverse by the
ATS (85 FR 87304, December 31, 2020).
In consideration of the exposure/risk results for the lowest benchmark (60 ppb), the
Administrator noted that the lung function decrements in controlled human exposure studies of
largely healthy adult subjects exposed while at elevated exertion to concentrations of 60 ppb,
although statistically significant, were much reduced from that observed in the next higher
studied concentration (73 ppb), both at the mean and individual level, and were not reported to
be associated with increased respiratory symptoms in healthy subjects.24 In light of these results
and the transient nature of the responses, the Administrator did not judge these responses to
represent adverse effects for generally healthy individuals. However, he further considered these
findings specifically with regard to protection of at-risk populations, such as people with asthma.
In this regard, he noted that such data are lacking for at-risk groups, such as people with asthma,
and considered the evidence and comments from the CASAC regarding the need to consider
endpoints of particular importance for this population group, such as risk of asthma exacerbation
24 The response for the 60 ppb studies is also somewhat lower than that for the 63 ppb study (Table 1; 2020 PA,
Appendix 3D, Table 3D-20).
April 2022
3-15 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
and prolonged inflammation. He took note of comments from the CASAC (and also noted in the
ATS statement) that small lung function decrements in this at-risk group may contribute to a risk
of asthma exacerbation, an outcome described by the CASAC as "arguably the most important
potential adverse effect'' of O3 exposure for a child with asthma. Thus, he judged it important for
the standard to provide protection that reduces such risks. With regard to the inflammatory
response, he noted the evidence indicating the role of repeated occurrences of inflammation in
contributing to severity of response. Thus, he found repeated occurrences of exposure events of
potential concern to pose greater risk than single events, leading him to place greater weight on
exposure/risk estimates for multiple occurrences (85 FR 87304-87305, December 31, 2020).
Thus, in this context, and given that the 70 ppb benchmark represents an exposure level
somewhat below the lowest exposure concentration for which both statistically significant lung
function decrements and increased respiratory symptoms have been reported in largely healthy
adult subjects, the Administrator considered the exposure/risk estimates for the third benchmark
of 60 ppb to be informative most particularly to his judgments on an adequate margin of safety.
In so doing, he took note that these estimates indicate more than 96% to more than 99% of
children with asthma to be protected from more than one occasion in a year (two or more), on
average, of 7-hour exposures to concentrations at or above this level (60 ppb), while at elevated
exertion (85 FR 87279, Table 2, December 31, 2020). Additionally, the analysis estimates more
than 90% of all children, on average across the three years, to be protected from one or more
occasions of exposure at or above this level. The Administrator found this to indicate an
appropriate degree of protection from such exposures (85 FR 87305, December 31, 2020).
The Administrator additionally considered whether it was appropriate to consider a more
stringent standard that might be expected to result in reduced O3 exposures. As an initial matter,
he considered the advice from the CASAC. With regard to the CASAC advice, while part of the
Committee concluded the evidence supported retaining the current standard without revision,
another part of the Committee reiterated advice from the prior CASAC, which while including
the current standard level among the range of recommended standard levels, also provided policy
advice to set the standard at a lower level. In considering this advice in the 2020 review, as it was
raised by part of the then-current CASAC, the Administrator noted the slight differences of the
current exposure and risk estimates from the corresponding 2014 estimates for the lowest
benchmark, which were those considered by the CASAC in 2014 (85 FR 87280, Table 3,
December 31, 2020). For example, while the 2014 HREA estimated 3.3 to 10.2% of children, on
average, to experience one or more days with exposures at or above 60 ppb (and as many as
18.9%) in a single year), the comparable estimates for the current analyses are lower (3.2 to 8.2%>
on average and 10.6%> in a single year), particularly with regard to the upper end of the range of
averages and the highest in a single year. While the estimates for two or more days with
April 2022
3-16 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
occurrences at or above 60 ppb, on average across the assessment period, were more similar
between the two assessments, the 2020 estimate for the single highest year was much lower (9.2
versus 4.3%). The Administrator additionally recognized the 2020 PA finding that the factors
contributing to these differences, which includes the use of air quality data reflecting
concentrations much closer to the now-current standard than was the case in the 2015 review,
also contribute to a reduced uncertainty in the current estimates (85 FR 87275-87279, December
31, 2020; 2020 PA, sections 3.4 and 3.5). Thus, he noted that the exposure analysis estimates in
the 2020 review indicate the current standard to provide appreciable protection against multiple
days with a maximum exposure at or above 60 ppb. In the context of his consideration of the
adequacy of protection provided by the standard and of the CAA requirement that the standard
protect public health, including the health of at-risk populations, with an adequate margin of
safety, the Administrator concluded, "in light of all of the considerations raised here, that the
current standard provides appropriate protection, and that a more stringent standard would be
more than requisite to protect public health" (85 FR 87306; December 31, 2020).
Therefore, based on his consideration of the evidence and exposure/risk information,
including that related to the lowest exposures studied in controlled human exposure studies, and
the associated uncertainties, the Administrator judged that the current standard provides the
requisite protection of public health, including an adequate margin of safety, and thus should be
retained, without revision. Accordingly, he also concluded that a more stringent standard was not
needed to provide requisite protection and that the current standard provides the requisite
protection of public health under the Act (85 FR 87306, December 31, 2020).
3.2 GENERAL APPROACH AND KEY ISSUES
As is the case for primary NAAQS reviews, this reconsideration of the 2020 decision on
the primary O3 standard is fundamentally based on using the Agency's assessment of the
scientific evidence and associated quantitative analyses to inform the Administrator's judgments
regarding a primary standard that is requisite to protect public health with an adequate margin of
safety. This approach builds on the substantial assessments and evaluations performed over the
course of O3 NAAQS reviews to inform our understanding of the key-policy relevant issues in
this reconsideration of the 2020 decision.
The evaluations in the PA of the scientific assessments in the ISA (building on prior such
assessments), augmented by the quantitative risk and exposure analyses,25 are intended to inform
25 The overarching purpose of the quantitative exposure and risk analyses is to inform the Administrator's
conclusions on the public health protection afforded by the current primary standard. An important focus is the
assessment, based on current tools and information, of the potential for exposures and risks beyond those
indicated by the information available at the time the standard was established.
April 2022
3-17 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
the Administrator's public health policy judgments and conclusions, including his decisions
regarding the O3 standards. The PA considers the potential implications of various aspects of the
scientific evidence, the exposure/risk-based information, and the associated uncertainties and
limitations. Thus, the approach for this PA is to draw on the evaluation of the scientific and
technical information available in the 2020 review to address a series of key policy-relevant
questions using both evidence- and exposure/risk-based considerations. Together, consideration
of the available evidence and information will inform the answer to the following initial
overarching question:
• Do the available scientific evidence and exposure-/risk-based information support or
call into question the adequacy of the public health protection afforded by the current
primary O3 standard?
In reflecting on this question, we will consider the body of scientific evidence, assessed
in the 2020 ISA and used as a basis for developing or interpreting exposure/risk analyses,
including whether it supports or calls into question the scientific conclusions reached in the 2020
review regarding health effects related to exposure to ambient air-related O3. Information that
may be informative to public health judgments regarding significance or adversity of key effects
is also be considered. Additionally, the available exposure and risk information will be
considered, including with regard to the extent to which it may continue to support judgments
made in the 2020 review. Further, in considering this question with regard to the primary O3
standard, as in all NAAQS reviews, we give particular attention to exposures and health risks to
at-risk populations.26 Evaluation of the available scientific evidence and exposure/risk
information with regard to consideration of the current standard and the overarching question
above focuses on key policy-relevant issues by addressing a series of questions on specific
topics. Figure 3-1 summarizes, in general terms, the approach to considering the available
information in the context of policy-relevant questions pertaining to the primary standard.
26 As used here and similarly throughout this document, the term population refers to persons having a quality or
characteristic in common, such as a specific pre-existing illness or a specific age or lifestage. Identifying at-risk
populations involves consideration of susceptibility and vulnerability. Susceptibility refers to innate (e.g., genetic
or developmental aspects) or acquired (e.g., disease or smoking status) sensitivity that increases the risk of health
effects occurring with exposure to O3. Vulnerability refers to an increased risk of Ch-related health effects due to
factors such as those related to socioeconomic status, reduced access to health care or exposure.
April 2022
3-18 External Review Draft - Do Not Quote or Cite

-------
Potential Alternative Standards for Consideration
1
2	Figure 3-1. Overview of general approach for the primary Os standard.
April 2022
3-19 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
The Agency's approach with regard to the O3 primary standard is consistent with
requirements of the provisions of the CAA related to the review of the NAAQS and with how the
EPA and the courts have historically interpreted these provisions. As discussed in section 1.2
above, these provisions require the Administrator to establish primary standards that, in the
Administrator's judgment, are requisite (i.e., neither more nor less stringent than necessary) to
protect public health with an adequate margin of safety. Consistent with the Agency's approach
across NAAQS reviews, the approach of the PA to informing these judgments is based on a
recognition that the available health effects evidence generally reflects continuums that include
ambient air exposures for which scientists generally agree that health effects are likely to occur
through lower levels at which the likelihood and magnitude of response become increasingly
uncertain. The CAA does not require the Administrator to establish a primary standard at a zero-
risk level or at background concentration levels, but rather at a level that reduces risk sufficiently
so as to protect public health, including the health of sensitive groups,27 with an adequate margin
of safety.
The Agency's decisions on the adequacy of the current primary standard and, as
appropriate, on any potential alternative standards considered in a review are largely public
health policy judgments made by the Administrator. The four basic elements of the NAAQS (i.e.,
indicator, averaging time, form, and level) are considered collectively in evaluating the health
protection afforded by the current standard, and by any alternatives considered. Thus, the
Administrator's final decisions in such reviews draw upon the scientific evidence for health
effects, quantitative analyses of population exposures and/or health risks, as available, and
judgments about how to consider the uncertainties and limitations that are inherent in the
scientific evidence and quantitative analyses.
3.3 HEALTH EFFECTS EVIDENCE
The health effects evidence on which this PA for the reconsideration of the 2020 decision
on the O3 primary standard will focus is the evidence as assessed and described in the 2020 ISA
and prior ISAs or AQCDs. As described in section 1.5 above, the EPA has provisionally
considered more recently available studies that were raised in public comments in the 2020
review or were identified in a literature search that the EPA conducted for this reconsideration of
more recently available controlled human exposure studies (Luben et al., 2020; Duffney et al.
27 More than one population group may be identified as sensitive or at-risk in a NAAQS review. Decisions on
NAAQS reflect consideration of the degree to which protection is provided for these sensitive population groups.
To the extent that any particular population group is not among the identified sensitive groups, a decision that
provides protection for the sensitive groups would be expected to also provide protection for other population
groups.
April 2022
3-20 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
2022). The provisional consideration of these studies concluded that, taken in context, the
associated information and findings did not materially change any of the broad scientific
conclusions of the ISA regarding the health and welfare effects of O3 in ambient air or warrant
reopening the air quality criteria for this review. Thus, the discussion below focuses on the health
effects evidence assessment, with associated conclusions, as described in the 2020 ISA.
3.3.1 Nature of Effects
The health effects evidence base for O3 includes decades of extensive evidence that
clearly describes the role of O3 in eliciting an array of respiratory effects and the more recent
evidence suggests the potential for relationships between O3 exposure and other effects. As was
established in prior O3 NAAQS reviews, the most commonly observed effects, and those for
which the evidence is strongest are transient decrements in pulmonary function and respiratory
symptoms, such as coughing and pain on deep inspiration, as a result of short-term exposures
particularly when breathing at elevated rates (ISA, section IS.4.3.1; 2013 ISA, p. 2-26). These
effects are demonstrated in the large, long-standing evidence base of controlled human exposure
studies28 (1978 AQCD, 1986 AQCD, 1996 AQCD, 2006 AQCD, 2013 ISA, ISA). Lung function
effects are also positively associated with ambient air O3 concentrations in epidemiologic panel
studies, available in past reviews, that describe these associations for outdoor workers and
children attending summer camps in the 1980s and 1990s (2013 ISA, section 6.2.1.2; ISA,
Appendix 3, section 3.1.4.1.3). Collectively, the epidemiologic evidence base documents
consistent, positive associations of O3 concentrations in ambient air with lung function effects in
epidemiologic panel studies29 and with more severe health outcomes in other epidemiologic
studies, including asthma-related emergency department visits and hospital admissions (2013
ISA, sections 6.2.1.2 and 6.2.7; ISA, Appendix 3, sections 3.1.4.1.3, 3.1.5.1 and 3.1.5.2).
Extensive animal toxicological evidence informs a detailed understanding of mechanisms
underlying the respiratory effects of short-term exposures, and studies in animal models also
provide evidence for effects of longer-term O3 exposure on the developing lung (ISA, Appendix
3, sections 3.1.11 and 3.2.6).
28	The vast majority of the controlled human exposure studies (and all of the studies conducted at the lowest
exposures) involved young healthy adults (typically 18-35 years old) as study subjects (ISA, section 3.1.4; 2013
ISA, section 6.2.1.1). There are also some 1-8 hr controlled human exposure studies in older adults and adults
with asthma, and there are still fewer controlled human exposure studies in healthy children (i.e., individuals aged
younger than 18 years) or children with asthma (See, for example, Appendix 3 A, Table 3 A-3).
29	Panel studies are a type of longitudinal epidemiologic study. The studies referenced here include a number of such
past studies investigating O3 and lung function measures in groups of children attending summer camp and
respiratory symptoms in groups of children with asthma (ISA, sections 3.1.4.1.3 and 3.1.5.3; 2013 ISA, sections
6.2.1.2 and 6.2.4.1).
April 2022
3-21 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
• Does the available scientific evidence alter prior conclusions regarding the health
effects attributable to exposure to O3?
The available scientific evidence, as assessed in the ISA, continues to support the prior
conclusion that short-term O3 exposure causes respiratory effects. Specifically, the full body of
evidence continues to support the conclusions of a causal relationship of respiratory effects with
short-term O3 exposures and a likely causal relationship of respiratory effects with longer-term
exposures (ISA, sections IS.4.3.1 and IS.4.3.2). The evidence base described in the 2020 ISA
which is expanded from the evidence available in the 2015 review (and described in the 2013
ISA), also indicates a likely causal relationship between short-term O3 exposure and metabolic
effects,30 which were not evaluated as a separate category of effects in the 2015 review when less
evidence was available (ISA, section IS.4.3.3). The more recent evidence is primarily from
experimental animal research. For other types of health effects, recent evidence has led to
different conclusions from those reached previously. Specifically, the evidence base described in
the 2020 ISA, particularly in light of the additional controlled human exposure studies, is less
consistent than what was previously available and less indicative of 03-induced cardiovascular
effects.31 This recent evidence has altered conclusions from the 2015 review with regard to
relationships between short-term O3 exposures and cardiovascular effects and mortality, such
that likely causal relationships are no longer concluded.32 Thus, as discussed in the ISA,
conclusions have changed for some effects based on the recent evidence, and conclusions are
newly reached for an additional category of health effects. The prior conclusions on respiratory
effects, however, continue to be supported.
3.3.1.1 Respiratory Effects
The available evidence, as described in the 2020 ISA, continues to support the conclusion
of a causal relationship between short-term O3 exposure and respiratory effects (ISA, section
IS. 1.3.1). The strongest evidence for this comes from controlled human exposure studies
30	The term "metabolic effects" is used in the ISA to refer metabolic syndrome (a collection of risk factors including
high blood pressure, elevated triglycerides and low high density lipoprotein cholesterol), diabetes, metabolic
disease mortality, and indicators of metabolic syndrome that include alterations in glucose and insulin
homeostasis, peripheral inflammation, liver function, neuroendocrine signaling, and serum lipids (ISA, section
IS.4.3.3).
31	As described in the ISA, "[t]he number of controlled human exposure studies showing little evidence of ozone
induced cardiovascular effects has grown substantially" and "the plausibility for a relationship between short-
term ozone exposure to cardiovascular health effects is weaker than it was in the previous review, leading to the
revised causality determination" (ISA, p. IS-43).
32	The evidence for cardiovascular, reproductive and nervous system effects, as well as mortality, is "suggestive of,
but not sufficient to infer" a causal relationship with short- or long-term O3 exposures (ISA, Table IS-1). The
evidence is inadequate to infer the presence or absence of a causal relationship between long-term O3 exposure
and cancer (ISA, section IS4.3.6.6).
April 2022
3-22 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
demonstrating Cte-related respiratory effects in generally healthy adults.33 The key evidence
comes from the body of controlled human exposure studies that document respiratory effects in
people exposed for short periods (6.6 to 8 hours) during quasi-continuous exercise.34 The
potential for O3 exposure to elicit health outcomes more serious than those assessed in the
experimental studies, particularly for children with asthma, continues to be indicated by the
epidemiologic evidence of associations of O3 concentrations in ambient air with increased
incidence of hospital admissions and emergency department visits for an array of health
outcomes, including asthma exacerbation, COPD exacerbation, respiratory infection, and
combinations of respiratory diseases (ISA, Appendix 3, sections 3.1.5 and 3.1.6). The strongest
such evidence is for asthma-related outcomes and specifically asthma-related outcomes for
children, indicating an increased risk for people with asthma and particularly children with
asthma (ISA, Appendix 3, section 3.1.5.7).
Respiratory responses observed in human subjects exposed to O3 for periods of 8 hours or
less, while intermittently or quasi-continuously exercising, include reduced lung function
decrements (e.g., based on forced expiratory volume in one second [FEVi] measurements),35
respiratory symptoms, increased airway responsiveness, mild bronchoconstriction (measured as a
change in specific airway resistance [sRaw]), and pulmonary inflammation, with associated
injury and oxidative stress (ISA, Appendix 3, section 3.1.4; 2013 ISA, sections 6.2.1 through
6.2.4). The available mechanistic evidence, discussed in greater detail in the ISA, describes
pathways involving the respiratory and nervous systems by which O3 results in pain-related
respiratory symptoms and reflex inhibition of maximal inspiration (inhaling a full, deep breath),
commonly quantified by decreases in forced vital capacity (FVC) and total lung capacity. This
reflex inhibition of inspiration combined with mild bronchoconstriction contributes to the
33	The phrases "healthy adults" or "healthy subjects" are used to distinguish from subjects with asthma or other
respiratory diseases, because "the study design generally precludes inclusion of subjects with serious health
conditions," such as individuals with severe respiratory diseases (2013 ISA, p. lx).
34	A quasi-continuous exercise protocol is common to these controlled exposure studies where, in the case of a 6.6-
hour study, subjects complete six 50-minute periods of exercise, each followed by 10-minute periods of rest, in
addition to a 30-minute lunch exposure period at rest (e.g., ISA, Appendix 3, section 3.1.4.1.1, and p. 3-11; 2013
ISA, section 6.2.1.1).
35	The measure of lung function response most commonly considered across O3 NAAQS reviews is changes in
FEVi. In considering controlled human exposure studies, an Ch-induced change in FEVi is typically the
difference between the decrement observed with O3 exposure ([post-exposure FEVi minus pre-exposure FEVi]
divided by pre-exposure FEVi) and what is generally an improvement observed with filtered air (FA) exposure
([postexposure FEVi minus pre-exposure FEVi] divided by pre-exposure FEVi). As explained in the 2013 ISA,
"[n]oting that some healthy individuals experience small improvements while others have small decrements in
FEVi following FA exposure, investigators have used the randomized, crossover design with each subject serving
as their own control (exposure to FA) to discern relatively small effects with certainty since alternative
explanations for these effects are controlled for by the nature of the experimental design" (2013 ISA, pp. 6-4 to 6-
5).
April 2022
3-23 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
observed decrease in forced expiratory volume in one second (FEVi), the most common metric
used to assess Cb-related pulmonary function effects. The evidence also indicates that the
additionally observed inflammatory response is correlated with mild airway obstruction,
generally measured as an increase in sRaw (ISA, Appendix 3, section 3.1.3). As described in
section 3.3.3 below, the prevalence and severity of respiratory effects in controlled human
exposure studies, including symptoms (e.g., pain on deep inspiration, shortness of breath, and
cough) increases, with increasing O3 concentration, exposure duration, and ventilation rate of
exposed subjects (ISA, Appendix 3, sections 3.1.4.1 and 3.1.4.2).
Within the evidence base from controlled human exposure studies, the majority of studies
involve healthy adult subjects (generally 18 to 35 years old), although there are studies involving
subjects with asthma, and a limited number of studies, generally of durations shorter than four
hours, involving adolescents and adults older than 50 years. A summary of salient observations
of O3 effects on lung function, based on the controlled human exposure study evidence reviewed
in the 1996 and 2006 AQCDs, and recognized in the 2013 ISA, continues to pertain to this
evidence base as it exists today "(1) young healthy adults exposed to >80 ppb O3 develop
significant reversible, transient decrements in pulmonary function and symptoms of breathing
discomfort if minute ventilation (Ve) or duration of exposure is increased sufficiently [i.e., as
measured by FEVi and/or FVC]; (2) relative to young adults, children experience similar
spirometric responses but lower incidence of symptoms from O3 exposure; (3) relative to young
adults, ozone-induced spirometric responses are decreased in older individuals; (4) there is a
large degree of inter-subject variability in physiologic and symptomatic responses to O3, but
responses tend to be reproducible within a given individual over a period of several months; and
(5) subjects exposed repeatedly to O3 for several days experience an attenuation of spirometric
and symptomatic responses on successive exposures, which is lost after about a week without
exposure" (ISA, Appendix 3, section 3.1.4.1.1, p. 3-11).36
The evidence is most well established with regard to the effects, reversible with the
cessation of exposure, that are associated with short-term exposures of several hours. For
example, the evidence indicates a rapid recovery from Cb-induced lung function decrements
(e.g., reduced FEVi) and respiratory symptoms (2013 ISA, section 6.2.1.1). However, in some
cases, such as after exposure to higher concentrations such as 300 ppb, the recovery phase may
be slower and involve a longer time period (e.g., at least 24 hours [hrs]). Repeated daily exposure
studies at such higher concentrations also have found FEVi response to be enhanced on the
second day of exposure. This enhanced response is absent, however, with repeated exposure at
36 A spirometric response refers to a change in the amount of air breathed out of the body (forced expiratory
volumes) and the associated time to do so (e.g., FEVi).
April 2022
3-24 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
lower concentrations, perhaps as a result of a more complete recovery or less damage to
pulmonary tissues (2013 ISA, section pp. 6-13 to 6-14; Folinsbee et al., 1994).
With regard to airway inflammation and the potential for repeated occurrences to
contribute to further effects, Cb-induced respiratory tract inflammation "can have several
potential outcomes: (1) inflammation induced by a single exposure (or several exposures over
the course of a summer) can resolve entirely; (2) continued acute inflammation can evolve into a
chronic inflammatory state; (3) continued inflammation can alter the structure and function of
other pulmonary tissue, leading to diseases such as fibrosis; (4) inflammation can alter the
body's host defense response to inhaled microorganisms, particularly in potentially at-risk
populations such as the very young and old; and (5) inflammation can alter the lung's response to
other agents such as allergens or toxins" (2013 ISA, p. 6-76; ISA Appendix 3, section 3.1.5.6).
With regard to Cb-induced increases in airway responsiveness, the controlled human exposure
study evidence for healthy adults generally indicates a resolution within 18 to 24 hours after
exposure, with slightly longer persistence in some individuals (ISA, Appendix 3, section
3.1.4.3.1; 2013 ISA, p. 6-74; Folinsbee and Hazucha, 2000).
The extensive evidence base for Cb-related health effects, compiled over several decades,
continues to indicate respiratory responses to short exposures as the most sensitive effects of O3.
This array of respiratory effects, including reduced lung function, respiratory symptoms,
increased airway responsiveness, and inflammation are of increased significance to people with
asthma given aspects of the disease that contribute to a baseline status that includes chronic
airway inflammation and greater airway responsiveness than people without asthma (ISA,
section 3.1.5). For example, O3 exposure of a magnitude that increases airway responsiveness
may put such people at potential increased risk for prolonged bronchoconstriction in response to
asthma triggers (ISA, Appendix 3, p. 3-7, 3-28; 2013 ISA, section 6.2.9; 2006 AQCD, section
8.4.2). The increased significance of effects in people with asthma and risk of increased exposure
for children (from greater frequency of outdoor exercise as described in Section 3.3.2) is
illustrated by the epidemiological findings of positive associations between O3 exposure and
asthma-related emergency department visits and hospital admissions for children with asthma.
Thus, the evidence indicates O3 exposure to increase the risk of asthma exacerbation, and
associated outcomes, in children with asthma.
With regard to an increased susceptibility to infectious diseases, the experimental animal
evidence continues to indicate, as described in the 2013 ISA and past AQCDs, a potential role
for O3 exposures through effects on defense mechanisms of the respiratory tract (2013 ISA,
section 6.2.5). Evidence regarding respiratory infections and associated effects has been
augmented by a number of epidemiologic studies reporting positive associations between short-
April 2022
3-25 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
term O3 concentrations and emergency department visits for a variety of respiratory infection
endpoints (ISA, Appendix 3, section 3.1.7; 2013 ISA, section 6.2.5).
Although the long-term exposure conditions that may contribute to further respiratory
effects are less well understood, the evidence-based conclusion remains that there is likely to be
a causal relationship for such exposure conditions with respiratory effects (ISA, section IS.4.3.2).
Most notably, experimental studies, including with nonhuman infant primates, have provided
evidence relating O3 exposure to allergic asthma-like effects, and epidemiologic cohort studies
have reported associations of O3 concentrations in ambient air with asthma development in
children (ISA, Appendix 3, section 3.2.4.1.3 and 3.2.6). The biological plausibility of such a role
for O3 has been indicated by animal toxicological evidence on biological mechanisms (ISA,
Appendix 3, sections 3.2.3 and 3.2.4.1.2). Specifically, the animal evidence, including the
nonhuman primate studies of early life O3 exposure, indicates that such exposures can cause
"structural and functional changes that could potentially contribute to airway obstruction and
increased airway responsiveness," which are hallmarks of asthma (ISA, Appendix 3, section
3.2.6, p. 3-113).
Overall, the recent respiratory effects evidence is generally consistent with the evidence
base in the 2015 review (ISA, Appendix 3, section 3.1.4). A few recent studies provide insights
in previously unexamined areas, both with regard to human study groups and animal models for
different effects, while other studies confirm and provide depth to prior findings with updated
protocols and techniques (ISA, Appendix 3, sections 3.1.11 and 3.2.6). Thus, our current
understanding of the respiratory effects of O3 is similar to that in the 2015 review.
One aspect of the evidence, augmented in the 2020 review as compared with the 2015
review, concerns pulmonary function in adults older than 50 years of age. Previously available
evidence in this age group indicated smaller 03-related decrements in middle-aged adults (35 to
60 years) than in adults 35 years of age and younger (2006 AQCD, p. 6-23; 2013 ISA, p. 6-22;
ISA, Appendix 3, section 3.1.4.1.1.2). A recent multicenter study of 55- to 70-year old subjects
(average of 60 years), conducted for a 3-hour duration involving alternating 15-minute rest and
exercise periods and a 120 ppb exposure concentration, reported a statistically significant O3
FEVi response (ISA, Appendix 3, section 3.1.4.1.1.2; Arjomandi et al., 2018). While there is not
a precisely comparable study in younger adults, the mean response for the 55- to 70-year olds,
1.2% 03-related FEVi decrement, is lower than results for somewhat comparable exposures in
adults aged 35 or younger, suggesting somewhat reduced responses to O3 exposure in this older
age group (ISA, Appendix 3, section 3.1.4.1.1.2; Arjomandi et al., 2018; Adams, 2000; Adams,
April 2022
3-26 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
2006a).37 Such a reduced response in middle-aged and older adults compared to young adults is
consistent with conclusions in the past (2013 ISA, section 6.2.1.1; 2006 AQCD, section 6.4).
The strongest evidence of Cb-related health effects continues to document the respiratory
effects of O3 (ISA, section ES.4.1). There are no new studies, however, of 6.6-hour exposures
(with exercise) to O3 concentrations below those previously studied.38 Among the newly
available studies in the 2020 ISA, are several controlled human exposure studies that
investigated lung function effects of higher exposure concentrations (e.g., 100 to 300 ppb) in
healthy individuals younger than 35 years old, with findings generally consistent with previous
studies (ISA, Appendix 3, section 3.4.1.1.2, p. 3-17). The newly available animal toxicological
studies augment the previously available information concerning mechanisms underlying the
effects documented in experimental studies. Lastly, newly available epidemiologic studies of
hospital admissions and emergency department visits for a variety of respiratory outcomes
supplement the previously available evidence with additional findings of consistent associations
with O3 concentrations across a number of study locations (ISA, Appendix 3, sections 3.1.4.1.3,
3.1.5, 3.1.6.1.1, 3.1.7.1 and 3.1.8). These studies include a number that report positive
associations for asthma-related outcomes, as well as a few for COPD-related outcomes. Together
these epidemiologic studies continue to indicate the potential for O3 exposures to contribute to
such serious health outcomes, particularly for people with asthma.
3.3.1.2 Other Effects
As was the case for the evidence available previously, the evidence for health effects
other than those on the respiratory system is more uncertain than that for respiratory effects. For
some of these other categories of effects, the more recent evidence as described in the 2020 ISA
has contributed to changes to conclusions reached in the 2015 review. For example,
cardiovascular effects and mortality are no longer concluded to be likely causally related to O3
exposures based on newly available evidence in combination with the uncertainties that had been
recognized for the previously available evidence. Additionally, newly available evidence also led
37	For the same exposure concentration of 120 ppb, Adams (2006a) observed an average 3.2%, statistically
significant, Ch-related FEVi decrement in young adults (average age 23 years) at the end of the third hour of an 8-
hour protocol that alternated 30 minutes of exercise and rest, with the equivalent ventilation rate (EVR) averaging
20 L/min-m2 during the exercise periods (versus 15 to 17 L/min-m2 in Aijomandi et al., 2018]). For the same
concentration with a lower EVR during exercise (17 L/min-m2), although with more exercise, Adams (2000)
observed a 4%, statistically significant, 03-related FEVi decrement in young adults (average age 22 years) after
the third hour of a 6.6-hour protocol (alternating 50 minutes exercise and 10 minutes rest).
38	The 2020 ISA includes a newly available 3-hr study of subjects aged 55 years of age or older that involves a
slightly lower target ventilation rate for the exercise periods. The exposure concentrations were 120 ppb and 70
ppb, only the former of which elicited a statistically significant FEVi decrement in this age group of subjects
(ISA, Appendix 3, section 3.1.4.1.1.2).
April 2022
3-27 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
to conclusions for another category, metabolic effects, for which formal causal determinations
were previously not articulated.
The ISA finds the evidence for metabolic effects sufficient to conclude that there is likely
to be a causal relationship with short-term O3 exposures (ISA, section IS.4.3.3). The evidence of
metabolic effects of O3 comes primarily from experimental animal study findings that short-term
O3 exposure can impair glucose tolerance, increase triglyceride levels and elicit fasting
hyperglycemia and increase hepatic gluconeogenesis (ISA, Appendix 5, section 5.1.8, and Table
5-3). The exposure conditions from these studies generally involve much higher O3
concentrations than those commonly occurring in areas of the U.S. where the current standard is
met. For example, the animal studies include 4-hour concentrations of 400 to 800 ppb (ISA,
Appendix 5, Tables 5-8 and 5-10). In addition, an epidemiologic study of a Taiwanese cohort
and 2002 air quality that was available in the 2015 review has reported positive associations of
multiday average O3 concentrations in ambient air with changes in two indicators of glucose and
insulin homeostasis (ISA, Appendix 5, sections 5.1.3.1.1 and 5.1.8).
The ISA additionally concludes that the evidence is suggestive of, but not sufficient to
infer, a causal relationship between long-term O3 exposures and metabolic effects (ISA, section
IS.4.3.6.2). As with metabolic effects and short-term O3, the primary evidence is from
experimental animal studies in which the exposure concentrations are appreciably higher than
those commonly occurring in the U.S. For example, the animal studies include exposures over
several weeks to concentrations of 250 ppb and higher (ISA, Appendix 5, section 5.2.3.1.1). The
somewhat limited epidemiologic evidence related to long-term O3 concentrations and metabolic
effects includes several studies reporting increased odds of being overweight or obese or having
metabolic syndrome and increased hazard ratios for diabetes incidence with increased O3
concentrations (ISA, Appendix 5, sections 5.2.3.4.1, 5.2.5 and 5.2.9, Tables 5-12 and 5-15).
With regard to cardiovascular effects and total (nonaccidental) mortality and short-term
O3 exposures, the conclusions in the ISA regarding the potential for a causal relationship have
changed from what they were in the 2015 review after integrating the previously available
evidence with the more recently available evidence. The relationships are now characterized as
suggestive of, but not sufficient to infer, a causal relationship (ISA, Appendix 4, section 4.1.17;
Appendix 6, section 6.1.8). This reflects several aspects of the evidence base: (1) a now-larger
body of controlled human exposure studies providing evidence that is not consistent with a
cardiovascular effect in response to short-term O3 exposure; (2) a paucity of epidemiologic
evidence indicating more severe cardiovascular morbidity endpoints,39 that would be expected if
39 These include emergency department visits and hospital admission visits for cardiovascular endpoints including
myocardial infarctions, heart failure or stroke (ISA, Appendix 6, section 6.1.8).
April 2022
3-28 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
the impaired vascular and cardiac function (observed in animal toxicological studies) was the
underlying basis for cardiovascular mortality (for which epidemiologic studies have reported
some positive associations with O3); and (3) the remaining uncertainties and limitations
recognized in the 2013 ISA (e.g., lack of control for potential confounding by copollutants in
epidemiologic studies) that still remain. Although there exists consistent or generally consistent
evidence for a limited number of Cb-induced cardiovascular endpoints in animal toxicological
studies and cardiovascular mortality in epidemiologic studies, there is a general lack of
coherence between these results and findings in controlled human exposure and epidemiologic
studies of cardiovascular health outcomes (ISA, section IS. 1.3.1). Related to this updated
conclusion for cardiovascular effects, the evidence for short-term O3 and mortality is also
updated (ISA, Appendix 6, section 6.1.8). While there remain consistent, positive associations
between short-term O3 and total (nonaccidental), respiratory, and cardiovascular mortality (and
there are some studies reporting associations to remain after controlling for PM10 and NO2), the
full evidence base does not describe a continuum of effects that could lead to cardiovascular
mortality.40 Therefore, because cardiovascular mortality is the largest contributor to total
mortality, the relatively limited biological plausibility and coherence within and across
disciplines for cardiovascular effects (including mortality) contributes to an accompanying
change in the causality determination for total mortality (ISA, section IS.4.3.5). Thus, the
evidence for cardiovascular effects and total mortality, as evaluated in the ISA, is concluded to
be suggestive of, but not sufficient to infer, a causal relationship with short-term (as well as long-
term) O3 exposures (ISA, section IS. 1.3.1).
For other health effect categories, EPA's conclusions, as described in the ISA, are largely
unchanged from those in the 2015 review. For example, the available evidence for reproductive
effects, as well as for effects on the nervous system, continue to be suggestive of, but not
sufficient to infer, a causal relationship (ISA, section IS.4.3.6). Additionally, the evidence is
inadequate to determine if a causal relationship exists between O3 exposure and cancer (ISA,
section IS.4.3.6.6).
3.3.2 Public Health Implications and At-risk Populations
The public health implications of the evidence regarding 03-related health effects, as for
other effects, are dependent on the type and severity of the effects, as well as the size of the
population affected. Such factors are discussed here in the context of our consideration of the
40 Due to findings from controlled human exposure studies examining clinical endpoints (e.g., blood pressure) that
do not indicate an O3 effect and from epidemiologic studies examining cardiovascular-related hospital admissions
and emergency department visits that do not find positive associations, a continuum of effects that could lead to
cardiovascular mortality is not apparent (ISA, Appendices 4 and 6).
April 2022
3-29 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
health effects evidence related to O3 in ambient air. Additionally, we summarize the available
information related to judgments or interpretative statements developed by public health experts,
including particularly experts in respiratory health. This section also summarizes the current
information on population groups at increased risk of the effects of O3 in ambient air.
With regard to O3 in ambient air, the potential public health impacts relate most
importantly to the role of O3 in eliciting respiratory effects, the category of effects that the ISA
concludes to be causally related to O3 exposure. Controlled human exposure studies have
documented reduced lung function, respiratory symptoms, increased airway responsiveness, and
inflammation, among other effects, in largely healthy adults exposed while at elevated
ventilation, such as while exercising. Such effects, if of sufficient severity and in individuals
with compromised respiratory function, such as individuals with asthma, are plausibly related to
emergency department visits and hospital admissions for asthma which have been associated
with ambient air concentrations of O3 in epidemiologic studies (as summarized in section 3.3.1
above; 2013 ISA, section 6.2.7; ISA, Appendix 3, sections 3.1.5.1 and 3.1.5.2).
The clinical significance of individual responses to O3 exposure depends on the health
status of the individual, the magnitude of the changes in pulmonary function, the severity of
respiratory symptoms, and the duration of the response among other factors. While a particular
reduction in FEVi or increase in inflammation or airway responsiveness may not be of concern
for a healthy group,41 it may increase the risk of a more severe effect in a group with asthma. As
a more specific example, the same increase in inflammation or airway responsiveness in
individuals with asthma could predispose them to an asthma exacerbation event triggered by an
allergen to which they may be sensitized (e.g., ISA, Appendix 3, section 3.1.5.6.1; 2013 ISA,
sections 6.2.3 and 6.2.6). Duration and frequency of documented effects is also reasonably
expected to influence potential adversity and interference with normal activity. In summary,
consideration of differences in magnitude or severity, and also the relative transience or
persistence of the responses (e.g., FEVi changes) and respiratory symptoms, as well as pre-
existing sensitivity to effects on the respiratory system, and other factors, are important to
characterizing implications for public health effects of an air pollutant such as O3 (ATS, 2000;
Thurston et al., 2017).
Decisions made in past reviews of the O3 primary standard and associated judgments
regarding adversity or health significance of measurable physiological responses to air pollutants
41 For example, for most healthy individuals, moderate effects on pulmonary function, such as transient FEVi
decrements smaller than 20% or transient respiratory symptoms, such as cough or discomfort on exercise or deep
breath, would not be expected to interfere with normal activity, while larger pulmonary function effects (e.g.,
FEVi decrements of 20% or larger lasting longer than 24 hours) and/or more severe respiratory symptoms are
more likely to interfere with normal activity for more of such individuals (e.g., 2014 PA, p. 3-53; 2006 AQCD,
Table 8-2).
April 2022
3-30 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
have been informed by guidance, criteria or interpretative statements developed within the public
health community, including the ATS, an organization of respiratory disease specialists, as well
as the CASAC. The ATS released its initial statement (titled Guidelines as to What Constitutes
an Adverse Respiratory Health Effect, with Special Reference to Epidemiologic Studies of Air
Pollution) in 1985 and updated it in 2000 (ATS, 1985; ATS, 2000). The ATS described its 2000
statement as being intended to "provide guidance to policy makers and others who interpret the
scientific evidence on the health effects of air pollution for the purposes of risk management"
(ATS, 2000). The statement further asserts that "principles to be used in weighing the evidence
and setting boundaries" and "the placement of dividing lines should be a societal judgment"
(ATS, 2000). The ATS explicitly states that it does "not attempt to provide an exact definition or
fixed list of health impacts that are, or are not, adverse," providing instead "a number of
generalizable 'considerations'" and that there "cannot be precise numerical criteria, as broad
clinical knowledge and scientific judgments, which can change over time, must be factors in
determining adversity" (ATS, 2000). A more recent ATS statement, while generally consistent
with the 2000 statement in the attention that statement gives to at-risk or vulnerable population
groups, broadens the discussion of effects, responses and biomarkers to reflect the expansion of
scientific research in these areas (Thurston, et al., 2017). The more recent statement additionally
notes that it does not offer "strict rules or numerical criteria, but rather proposes considerations to
be weighed in setting boundaries between adverse and nonadverse health effects," providing a
general framework for interpreting evidence that proposes a "set of considerations that can be
applied in forming judgments" for this context (Thurston et al., 2017). Thus, the most recent
statement expands upon (with some specificity) and updates the prior statement by retaining
previously identified considerations, including, for example, its emphasis on consideration of
vulnerable populations, while retaining core consistency with the earlier ATS statement
(Thurston et al., 2017; ATS, 2000).
With regard to pulmonary function decrements, the earlier ATS statement concluded that
"small transient changes in forced expiratory volume in 1 s[econd] (FEVi) alone were not
necessarily adverse in healthy individuals, but should be considered adverse when accompanied
by symptoms" (ATS, 2000). The more recent ATS statement continues to support this
conclusion and also gives weight to findings of such lung function changes in the absence of
respiratory symptoms in individuals with pre-existing compromised function, such as that
resulting from asthma (Thurston et al., 2017). More specifically, the recent ATS statement
expresses the view that when occurring in individuals with pre-existing compromised function,
such as asthma, the occurrence of "small lung function changes" "should be considered adverse
... even without accompanying respiratory symptoms" (Thurston et al., 2017). In keeping with
the intent of these statements to avoid specific criteria, neither statement provides more specific
April 2022
3-31 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
descriptions of such responses, such as with regard to magnitude, duration or frequency of small
pollutant-related lung function changes, for consideration of such conclusions. The earlier ATS
statement, in addition to emphasizing clinically relevant effects, also emphasized both the need
to consider changes in "the risk profile of the exposed population," and effects on the portion of
the population that may have a diminished reserve that puts its members at potentially increased
risk if affected by another agent (ATS, 2000). In a similar vein, the more recent statement
emphasizes the distinction between population changes and individual changes in lung function
measures noting that for an exposed group of study subjects, while the mean change or reduction
may be small, some individual study group members will have larger reductions which in some
cases may have passed a threshold for clinical importance (Thurston et al., 2017). These
concepts, including the consideration of the magnitude of effects occurring in just a subset of
study subjects, continue to be recognized as important in the more recent ATS statement
(Thurston et al., 2017) and continue to be relevant to the evidence base for O3.
• Does the available evidence alter our prior understanding of populations that are
particularly at risk from O3 exposures? What are important uncertainties in that
evidence?
The newly available information regarding O3 exposures and health effects among
sensitive populations, as thoroughly evaluated in the ISA, has not altered our understanding of
human populations at particular risk of health effects from O3 exposures (ISA, section IS.4.4).
For example, the respiratory effects evidence, extending decades into the past and augmented by
new studies in this review, supports the conclusion that "individuals with pre-existing asthma are
at greater risk of ozone-related health effects based on the substantial and consistent evidence
within epidemiologic studies and the coherence with toxicological studies" (ISA, p. IS-57).
Numerous epidemiological studies document associations of O3 with asthma exacerbation. Such
studies indicate the associations to be strongest for populations of children which is consistent
with their generally greater time outdoors while at elevated exertion. Together, these
considerations indicate people with asthma, including particularly children with asthma, to be at
relatively greater risk of 03-related effects than other members of the general population (ISA,
sections IS.4.3.1 and IS.4.4.2, Appendix 3).42
With respect to people with asthma, the limited evidence from controlled human
exposure studies (which are primarily in adult subjects) indicates similar magnitude of FEVi
decrements as in people without asthma (ISA, Appendix 3, section 3.1.5.4.1). Across studies of
other respiratory effects of O3 (e.g., increased respiratory symptoms, increased airway
42 Populations or lifestages can be at increased risk of an air pollutant-related health effect due to one or more
factors. These factors can be intrinsic, such as physiological factors that may influence the internal dose or
toxicity of a pollutant, or extrinsic, such as sociodemographic, or behavioral factors.
April 2022
3-32 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
responsiveness and increased lung inflammation), the responses observed in study subjects
generally do not differ due to the presence of asthma, although the evidence base is more limited
with regard to study subjects with asthma (ISA, Appendix 3, section 3.1.5.7). However, the
features of asthma (e.g., increased airway responsiveness) contribute to a risk of asthma-related
responses, such as asthma exacerbation in response to asthma triggers, which may increase the
risk of more severe health outcomes (ISA, section 3.1.5). For example, a particularly strong and
consistent component of the epidemiologic evidence is the appreciable number of epidemiologic
studies that demonstrate associations between ambient air O3 concentrations and hospital
admissions and emergency department visits for asthma (ISA, section IS.4.4.3.1).43 We
additionally recognize that in these studies, the strongest associations (e.g., highest effect
estimates) or associations more likely to be statistically significant are those for childhood age
groups, which are, as recognized in section 3.4, the age groups most likely to spend time
outdoors during afternoon periods (when O3 may be highest) and at activity levels corresponding
to those that have been associated with respiratory effects in the human exposure studies (ISA,
Appendix 3, sections 3.1.4.1 and 3.1.4.2).44 The epidemiologic studies of hospital admissions
and emergency department visits are augmented by a large body of individual-level
epidemiologic panel studies that demonstrated associations of short-term ozone concentrations
with respiratory symptoms in children with asthma. Additional support comes from
epidemiologic studies that observed Cb-associated increases in indicators of airway inflammation
and oxidative stress in children with asthma (ISA, section IS.4.3.1). Together, this evidence
continues to indicate the increased risk of population groups with asthma (ISA, Appendix 3,
section 3.1.5.7).
Children and outdoor adult workers, are at increased risk largely due to their generally
greater time spent outdoors while at elevated exertion rates (including in summer afternoons and
43	In addition to asthma exacerbation, the epidemiologic evidence also includes findings of positive associations of
increased O3 concentrations with hospital admissions or emergency department visits for COPD exacerbation and
other respiratory diseases (ISA, Appendix 3, sections 3.1.6.1.3 and 3.1.8).
44	Evaluations of activity pattern data indicate children to more frequently spend time outdoors during afternoon and
early evening hours, while at moderate or greater exertion level, than other age groups (Appendix 3D, section
3D.2.5.3, including Figure 3D-9; 2014 HREA, section 5.4.1.5 and Appendix 5G, section 5G-1.4). For example,
for days with some time spent outdoors, children spend, on average, approximately 2'/i hours of afternoon time
outdoors, 80% of which is at a moderate or greater exertion level, regardless of their asthma status (Appendix 3D,
section 3D.2.5.3). Adults, for days having some time spent outdoors, also spend approximately 2'/i hours of
afternoon time outdoors regardless of their asthma status but the percent of afternoon time at moderate or greater
exertion levels for adults (about 55%) is lower than that observed for children. Such analyses also note greater
participation in outdoor events during the afternoon, compared to other times of day, for children ages 6 through
19 years old during the warm season months (ISA, Appendix 2, section 2.4.1, Table 2-1). Analyses of the limited
activity pattern data by health status do not indicate asthma status to have appreciable impact (Appendix 3D,
section 3D.2.5.3; 2014 HREA, section 5.4.1.5).
April 2022
3-33 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
early evenings when O3 levels may be higher)45 This behavior makes them more likely to be
exposed to O3 in ambient air under conditions contributing to increased dose, e.g., elevated
ventilation taking greater air volumes into the lungs46 (ISA, section IS.4.4.2; 2013 ISA, section
5.2.2.7). Thus, in light of the evidence summarized in the prior paragraphs, children and outdoor
workers with asthma may be at increased risk of more severe outcomes, such as asthma
exacerbation. Further, with regard to children, there is experimental evidence from early life
exposures of nonhuman primates that indicates the potential for effects in childhood (through
adolescence) when human respiratory systems are under development (ISA, sections IS.4.4.2 and
IS.4.4.4.1). As noted in the ISA, "these experimental studies indicate that early-life ozone
exposure can cause structural and functional changes that could potentially contribute to airway
obstruction and increased airway responsiveness" (ISA, p. IS-52). Overall, the available
evidence, while not increasing our knowledge about susceptibility or at-risk status of these
population groups, is consistent with that in the 2015 review (ISA, section IS.4.4).
Evidence available in the 2020 ISA for older adults, a population identified as at risk in
the 2015 review, adds little to the evidence previously available (ISA, sections IS.4.4.2 and
IS.4.4.4.2; Table IS-10). The ISA notes, however, that "[t]he majority of evidence for older
adults being at increased risk of health effects related to ozone exposure comes from studies of
short-term ozone exposure and mortality evaluated in the 2013 Ozone ISA" (ISA, p. IS-52).
Such studies are part of the larger evidence base that is now concluded to be suggestive, but not
sufficient to infer a causal relationship of O3 with mortality (ISA, sections IS.4.3.5 and
IS.4.4.4.2, Appendix 4, section 4.1.16.1 and 4.1.17).
The ISA also expressly considered the evidence regarding O3 exposure and health effects
among populations with several other potential risk factors. As in the 2015 review, there is
suggestive evidence of low socioeconomic status (SES) as a factor associated with potentially
increased risk of 03-related health effects (2013 ISA, section 8.3.3 and p. 8-37; ISA, section
IS.4.4). The 2013 ISA concluded that "[ojverall, evidence is suggestive of SES as a factor
affecting risk of 03-related health outcomes based on collective evidence from epidemiologic
studies of respiratory hospital admissions but inconsistency among epidemiologic studies of
45	More specifically regarding outdoor workers, in 2020 about 4% of civilian workers were required to spend more
than two-thirds of their workday outdoors. Among construction, landscaping and groundskeeping workers, about
80-90% were required to spend more than two-thirds of their working day outside. Other employment sectors,
including highway maintenance, protection services, extraction and other construction trades like engineers and
equipment operators also had a high percentage of employees who spent most of their workday outdoors (Bureau
of Labor Statistics, 2020). Such jobs often include physically demanding tasks and involve increased ventilation
rates, increasing the potential for exposure to O3.
46	Additionally, compared to adults, children have higher ventilation rates relative to their lung volume which tends
to increase the dose normalized to lung surface area (ISA, p. IS-60).
April 2022
3-34 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
mortality and reproductive outcomes," additionally stating that "[f]urther studies are needed to
confirm this relationship, especially in populations within the U.S." (2013 ISA, p. 8-28). The
evidence in the 2020 ISA adds little to the evidence previously available in this area (ISA,
section IS.4.4.2 and Table IS-10). Regarding populations identified by race or ethnicity,
including American Indians or Native Americans, the evidence continued to be inadequate to
make a determination regarding a potential for increased risk (ISA, section IS.4.4, Table IS-10).
The ISA in the 2015 review additionally identified a role for dietary anti-oxidants such as
vitamins C and E in influencing risk of Cb-related effects, such as inflammation, as well as a role
for genetic factors to also confer either an increased or decreased risk (2013 ISA, sections 8.1
and 8.4.1). No recently available evidence was evaluated in the ISA that would inform or change
these prior conclusions (ISA, section IS.4.4 and Table IS-10).
• What does the available information indicate with regard to the size of at-risk
populations and their distribution in the U.S.?
The magnitude and characterization of a public health impact is dependent upon the size
and characteristics of the populations affected, as well as the type or severity of the effects. As
summarized above, children are an at-risk population and children under the age of 18 account
for 22.3% of the total U.S. population, with 6.0% of the total population being children under 5
years of age (U.S. Census Bureau, 2019). Further, as summarized above, a key population most
at risk of health effects associated with O3 in ambient air is people with asthma. The National
Center for Health Statistics data for 2019 indicate that approximately 7.8% of the U.S.
population has asthma (Table 3-1; CDC, 2019). This is one of the principal populations that the
primary O3 NAAQS is designed to protect (80 FR 65294, October 26, 2015). Table 3-1 below
considers the currently available information that helps to characterize key features of this
population.47
The age group for which asthma prevalence documented by these data is greatest is
children aged five to 19, with 9.1% of children aged five to 14 and 7.4% of children aged 15-19
having asthma. In 2012 (the most recent year for which such an evaluation is available), asthma
was the leading chronic illness affecting children (Bloom et al., 2013). The prevalence is greater
for boys than girls (for those less than 18 years of age). Among populations of different races or
ethnicities, black non-Hispanic children have the highest prevalence, at 13.5%. Asthma
prevalence is also increased among populations in poverty. For example, 11.8% of people living
in households below the poverty level have asthma, compared to 7.2%, on average, of those
47 Additionally, as part of the 2019 National Health Interview Survey, about 41% of people with asthma reported
having had an asthma attack or asthma episode within the prior 12 months, with this percentage being slightly
greater among children with asthma (44%) compared to adults with asthma (40%). A summary is available in
Tables 5-1 and 6-1 of the survey (https://www.cdc.gov/asthma/most_recent_national_asthma_data.htm).
April 2022
3-35 External Review Draft - Do Not Quote or Cite

-------
1	living above it. Populations groups with relatively greater asthma prevalence, such as
2	populations in poverty and children, might be expected to have a relatively greater potential for
3	03-related health impacts.48
4
5	Table 3-1. National prevalence of asthma, 2019.
Characteristic A
Number with Current Asthma
Percent with Current
(in thousands)6
Asthma
Total
25,131
7.8
Child (Age <18)
5,104
7.0
Adult (Age 18+)
20,026
8.0
All Age Groups
0-4 years
517
2.6
5-14 years
3,725
9.1
15-19 years
1,529
7.4
20-24 years
2,092
9.9
25-34 years
3,574
8.0
35-64 years
9,594
7.8
65+ years
4,069
7.7
Child Age Group
0-4 years
517
2.6
5-11 years
2,345
8.3
12-17 years
2,241
8.9
12-14 years
1,379
10.8
15-17 years
861
7.0
Sex
Males
10,487
6.6
Boys (Age <18)
3,122
8.4
Men (Age 18+)
7,364
6.1
Females
14,643
8.9
Girls (Age <18)
1,981
5.5
Women (Age 18+)
12,662
9.8
Race/Ethnicity
White NHc
15,094
7.7
Child (Age <18)
2,385
6.4
Adult (Age 18+)
12,701
8.1
Black NH
4,105
10.6
Child (Age <18)
1,289
13.5
Adult (Age 18+)
2,814
9.7
AI/ANE NH
349
10.7
Child (Age <18)
67
8.2
48 As summarized in section 3.1 above, the current standard was set to protect at-risk populations, which include
people with asthma. Accordingly, populations with asthma living in areas not meeting the standard would be
expected to be at increased risk of effects.
April 2022
3-36 External Review Draft - Do Not Quote or Cite

-------
Characteristic A
Number with Current Asthma Percent with Current
(in thousands)B Asthma
Adult (Age 18+)
281
11.6
Asian NH
697
3.8
Child (Age <18)
130
3.7
Adult (Age 18+)
567
3.8
Multiple0 NH
867
12.6
Child (Age <18)
339
11.2
Adult (Age 18+)
527
13.7
Hispanic, all
3,874
6.6
Child (Age <18)
1,387
7.5
Adult (Age 18+)
2,486
6.1
Hispanic, MexicanF
1,933
5.3
Child (Age<18)
725
6.1
Adult (Age 18+)
1,207
5.0
Hispanic, OtherF
1,929
8.5
Child (Age<18)
656
10.0
Adult (Age 18+)
1,273
7.9
Federal Poverty Threshold
Below 100% of poverty level
100% to less than 250% of poverty level
250% to less than 450% of poverty level
450% of poverty level or higher
4,814
7,837
6,345
6,138
11.8
8.5
7.3
5.9
A Numbers within selected characteristics may not sum to total due to rounding
B Includes persons who answered "yes" to the questions "Have you EVER been told by a doctor or other health
professional that you had asthma" and "Do you still have asthma?"
c NH = non-Hispanic
D Subcategory includes 'Other single and multiple races' for 2019
E AI/AN = American Indian/ Alaska Native
F As a subset of Hispanic
Adapted from 2019 National Health Interview Survey, Tables 3-1 and 4-1
(,https://www. cdc. gov/asthma/most_recent_national_asthma_data. htm).
1
2	3.3.3 Exposure Concentrations Associated with Effects
3	The extensive evidence base for O3 health effects, compiled over several decades and
4	evaluated in the ISA, continues to indicate respiratory responses to short-term exposures as the
5	most sensitive effects. As at the time of the 2015 review, the EPA's conclusions regarding
6	exposure concentrations of O3 associated with respiratory effects reflect the extensive
7	longstanding evidence base of controlled human exposure studies of short-term O3 exposures of
8	people with and without asthma.49 These studies have documented an array of respiratory effects,
9	including reduced lung function, respiratory symptoms, increased airway responsiveness, and
49 As recognized elsewhere, the studies are largely conducted with adult subjects.
April 2022
3-37 External Review Draft - Do Not Quote or Cite

-------
1	inflammation, in study subjects following 1- to 8-hour exposures, primarily while exercising.
2	The severity of observed responses, the percentage of individuals responding, and strength of
3	statistical significance at the study group level have been found to increase with increasing
4	exposure (ISA; 2013 ISA; 2006 AQCD). Factors influencing exposure include activity level or
5	ventilation rate, exposure concentration, and exposure duration (ISA; 2013 ISA; 2006 AQCD).
6	For example, evidence from studies with similar duration and exercise aspects (6.6-hour duration
7	with six 50-minute exercise periods) demonstrates an exposure-response relationship for O3-
8	induced reduction in lung function (Figure 3-2).50'51 This specific evidence was integral to the
9	Administrator's judgments and decisions in 2015 and 2020 (80 FR 65292, October 26, 2015; 85
50	For a subset of the studies included in Figure 3-2 (those with face mask rather than chamber exposures), there is
no O3 exposure during some of the 6.6-hr experiment (e.g., during the lunch break). Thus, while the exposure
concentration during the exercise periods is the same for the two types of studies, the time-weighted average
(TWA) concentration across the full 6.6-hr period differs slightly. For example, in the facemask studies of 120
ppb, the TWA across the full 6.6-hour experiment is 109 ppb (Appendix 3A, Table 3A-2).
51	The relationship also exists for size of FEVi decrement with alternative exposure or dose metrics, including total
inhaled O3 and intake volume averaged concentration.
April 2022
3-38 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
FR 87256, December 31, 2020).
Qv
A Adams (2006)	,
X Adams (2003)	^
~	Adams (2002)
~ Harstman et al (1&90)
*	Kim et al. (2011). McDonnell et al. (2012)
O McDonnell etal. (1991)
O Scbelegle et al, (2009)
~	Folinsbee etal. (1988)
ฆ Folinsbeeetal, (1994)
A Adams (2000)
•	Adams and Ollison (1997)
ov
F = Face mask
A Av	V = Varying
+ F
_ASL
IF
~f.v
~
30 40 50 60 70 80 90 100 110 120 130
Ozone (ppb)
Figure 3-2. Group mean 03-induced reduction in FEVi from controlled human exposure
studies of healthy adults exposed for 6.6 hours with quasi-continuous exercise.
FEVi values plotted reflect group mean 03-induced percent change in FEVi, based
on subtraction of the group mean filtered air percent change (post-pre exposure)
from the group mean O3 percent change in FEVi (adapted from Appendix 3A; ISA,
Appendix 3, Figure 3-1). Concentrations are the time-weighted averages of target
concentrations across full 6.6-hour period in chamber studies (or the average of
target concentrations across the six exposures in face mask studies).
• Does the available evidence alter prior conclusions regarding the exposure duration
and concentrations associated with health effects? Does the available scientific
evidence indicate health effects attributable to exposures to O3 concentrations lower
than previously reported?
The available evidence, as documented in the ISA, including that newly available in the
2020 review, does not alter our conclusions from the 2015 review on exposure duration and
April 2022
3-39 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
concentrations associated with Cte-related health effects. These conclusions were largely based
on the body of evidence from the controlled human exposure studies. A limited number of newly
available controlled human exposure studies are described in the ISA, although none involve
lower exposure concentrations than those previously studied (e.g., Figure 3-2) or find effects not
previously reported (ISA, Appendix 3, section 3.1.4).52
The extensive evidence base for O3 health effects, compiled over several decades,
continues to indicate respiratory responses to short-term exposures as the most sensitive effects
of O3. As summarized in section 3.3.1.1 above, an array of respiratory effects is well documented
in controlled human exposure studies of subjects exposed for 1 to 8 hours, primarily while
exercising. The risk of more severe health outcomes associated with such effects is increased in
people with asthma as illustrated by the epidemiological findings of positive associations
between O3 exposure and asthma-related emergency department visits and hospital admissions.
The magnitude of respiratory response (e.g., size of lung function decrements and
magnitude of symptom scores) documented in the controlled human exposure studies is
influenced by ventilation rate, exposure duration, and exposure concentration. When performing
physical activities requiring elevated exertion, ventilation rate is increased, leading to greater
potential for health effects due to an increased internal dose (2013 ISA, section 6.2.1.1, pp. 6-5 to
6-11). Accordingly, the exposure concentrations eliciting a given level of response after a given
exposure duration is lower for subjects exposed while at elevated ventilation, such as while
exercising (2013 ISA, pp. 6-5 to 6-6). For example, in studies of generally healthy young adults
exposed while at rest for 2 hours, 500 ppb is the lowest concentration eliciting a statistically
significant 03-induced group mean lung function decrement, while a 1- to 2-hour exposure to
120 ppb produces a statistically significant response in lung function when the ventilation rate of
the group of study subjects is sufficiently increased with exercise (2013 ISA, pp. 6-5 to 6-6).
The exposure conditions (e.g., duration and exercise) given primary focus in the past
several reviews are those of the 6.6-hour study design, which involves six 50-minute exercise
periods during which subjects maintain a moderate level of exertion to achieve a ventilation rate
of approximately 20 L/min per m2 body surface area while exercising. The 6.6 hours of exposure
in these studies has generally occurred in an enclosed chamber and the study design includes
three hours in each of which is a 50-minute exercise period and a 10-minute rest period, followed
by a 35-minute lunch (rest) period, which is followed by three more hours of exercise and rest, as
52 No 6.6-hour studies are newly available (ISA, Appendix 3, section 3.1.4.1.1). The newly available studies are
generally for exposures of three hours or less, and in nearly all instances involve exposure (while at elevated
exertion) to concentrations above 100 ppb (ISA, Appendix 3, section 3.1.4).
April 2022
3-40 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
before lunch.53 Most of these studies performed to date involve exposure maintained at a
constant (unchanging) concentration for the full duration, although a subset of studies have
concentrations that vary (generally in a stepwise manner) across the exposure period and are
selected so as to achieve a specific target concentration as the exposure average (Appendix 3 A,
Table 3A-2).54
No studies of the 6.6-hour quasi-continuous exercise design are newly available since the
2015 review. The previously available studies of this design document statistically significant
03-induced reduction in lung function (FEVi) and increased pulmonary inflammation in young
healthy adults exposed to O3 concentrations as low as 60 ppb. Statistically significant group
mean changes in FEVi, also often accompanied by statistically significant increases in
respiratory symptoms, become more consistent across such studies of exposures to higher O3
concentrations, such as 70 ppb and 80 ppb (Table 3-2; Appendix 3A, Table 3A-1). The lowest
exposures concentration for which these studies document a statistically significant increase in
respiratory symptoms is somewhat above 70 ppb, at 73 ppb55 (Schelegle et al., 2009; Appendix
3 A, Table 3A-1). In the 6.6-hour studies, the group means of 03-induced56 FEVi reductions for
target exposure concentrations at or below 70 ppb are approximately 6% or lower (Figure 3-2,
Table 3-2). For example, the group means of 03-induced FEVi decrements reported in these
studies that are statistically significantly different from the responses in filtered air are 6.1% for
the 70 ppb target (73 ppb time weighted average based on measurements) and 1.7% to 3.5% for
the 60 ppb target (Figure 3-2, Table 3-2).
The group mean 03-induced FEVi decrements generally increase with increasing O3
exposures, reflecting increases in both the number of the individuals affected and the magnitude
53	A few studies have involved exposures by facemask rather than in a chamber. To date, there is little research
differentiating between exposures conducted with a facemask and in a chamber since the pulmonary responses of
interest do not seem to be influenced by the exposure mechanism. However, similar responses have been seen in
studies using both exposure methods at higher O3 concentrations (Adams, 2002; Adams, 2003). In the facemask
designs, there is a short period of zero exposure, such that the total period of exposure is closer to 6 hours than 6.6
(Adams, 2000; Adams, 2002; Adams, 2003).
54	In these studies, the exposure concentration changes for each of the six hours in which there is exercise and the
concentration during the 35-minute lunch is the same as in the prior (third) hour with exercise. For example, in
the study by Adams (2006b), the protocol for the 6.6-hour period is as follows: 60 minutes at 0.04 ppm, 60
minutes at 0.07 ppm, 95 minutes at 0.09 ppm, 60 minutes at 0.07 ppm, 60 minutes at 0.05 ppm and 60 minutes at
0.04 ppm.
55	Measurements are reported in this study for each of the six 50-minute exercise periods, for which the mean is 72
ppb (Schelegle et al., 2009). Based on these data, the time-weighted average concentration across the full 6.6-
hour duration was 73 ppb (Schelegle et al., 2009). The study design includes a 35-minute lunch period following
the third exposure hour during which the exposure concentration remains the same as in the third hour.
56	Consistent with the ISA and 2013 ISA, the phrase "03-induced" decrement or reduction in lung function or FEVi
refers to the percent change from pre-exposure measurement of the O3 exposure minus the percent change from
pre-exposure measurement of the filtered air exposure (2013 ISA, p. 6-4).
April 2022
3-41 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
of the FEVi reduction (Figure 3-2). For example, following 6.6-hour exposures to a lower
concentration (40 ppb), for which decrements were not statistically significant at the group mean
level, none of 60 subjects across two separate studies experienced an Cb-induced FEVi reduction
as large as 15% or more (Appendix 3D, Table 3D-19). Across the four experiments (with
number of subjects ranging from 30 to 59 subjects) that have reported results for 60 ppb target
exposure,57 the number of subjects experiencing this magnitude of FEVi reduction (at or above
15%) varied (zero of 30, one of 59, two of 31 and two of 30 exposed subjects), while, together,
they represent 3% of all 150 subjects. The percentage of subjects (with reductions of 15% or
more) increased to 10% (three of 31 subjects) for the study at 73 ppb (70 ppb target
concentration) and is higher still (16%) in a variable exposure study at 80 ppb (Appendix 3D,
Tables 3D-19 and 3D-30; Schelegle et al., 2009). In addition to illustrating the E-R relationship,
these findings also illustrate the considerable variability in magnitude of responses observed
among study subjects (Table 3-2, Figure 3-2; ISA, Appendix 3, section 3.1.4.1.1; 2013 ISA, p. 6-
13).
57 For these four experiments, the average concentration across the 6.6-hour period ranged from 60 to 63 ppb
(Appendix 3A, Table 3A-2).
April 2022
3-42 External Review Draft - Do Not Quote or Cite

-------
1	Table 3-2. Summary of 6.6-hour controlled human exposure study-findings, healthy
2	adults.
Endpoint
O3 Target
Exposure
Concentration*
Statistically
Significant
EffectB
03-lnduced Group
Mean ResponseB
Study
FEVi
Reduction
120 ppb
Yes
-10.3% to -15.9% c
Horstman et al. 1990; Adams 2002;
Folinsbee et al. (1988); Folinsbee et al.
(1994); Adams, 2002; Adams 2000; Adams
and Ollison 1997ฐ
100 ppb
Yes
-8.5% to -13.9% c
Horstman et al., 1990; McDonnell et al.,
1991ฐ
87 ppb
Yes
-12.2%
Schelegle et al., 2009
80 ppb
Yes
-7.5%
Horstman et al., 1990
-7.7%
McDonnell et al., 1991
-6.5%
Adams, 2002
-6.2% to -5.5% c
Adams, 2003
-7.0% to -6.1% c
Adams, 2006b
-7.8%
Schelegle et al., 2009
ND E
-3.5%
Kim et al., 2011 F
70 ppb
Yes
-6.1%
Schelegle et al., 2009
60 ppb
Yes
G
-2.9%
-2.8%
Adams, 2006b; Brown et al., 2008
Yes
-1.7%
Kim et al., 2011
No
-3.5%
Schelegle et al., 2009
40 ppb
No
-1.2%
Adams, 2002
No
-0.2%
Adams, 2006b
Increased
Respiratory
Symptoms
120 ppb
Yes
Increased symptom
scores
Horstman et al. 1990; Adams 2002;
Folinsbee et al. 1988; Folinsbee et al. 1994;
Adams, 2002; Adams 2000; Adams and
Ollison 1997; Horstman et al., 1990;
McDonnell et al., 1991; Schelegle et al.,
2009; Adams, 2003; Adams, 2006b H
100 ppb
Yes
87 ppb
Yes
80 ppb
Yes
70 ppb
Yes
60 ppb
No
Adams, 2006b; Kim et al., 2011; Schelegle
et al., 2009; Adams, 2002 H
40 ppb
No
Airway
Inflammation
80 ppb
Yes
Multiple indicators1
Devlin et al., 1991; Alexis et al., 2010
60 ppb
Yes
Increased neutrophils
Kim et al., 2011
Increased
Airway
Resistance and
Responsiveness
120 ppb
Yes
Increased
Horstman et al., 1990; Folinsbee et al.,
1994 (O3 induced sRaw not reported)
100 ppb
Yes
Horstman et al., 1990
80 ppb
Yes
Horstman et al., 1990
This refers to the average concentration across the six exercise periods as targeted by authors. This differs from the time-
weighted average concentration for the full exposure periods (targeted or actual). For example, as shown in Appendix 3A, Table
3A-2, in chamber studies implementing a varying concentration protocol with targets of 0.03, 0.07, 0.10, 0.15, 0.08 and 0.05
apm, the exercise period average concentration is 0.08 ppm while the time weighted average for the full exposure period (based
on targets) is 0.082 ppm due to the 0.6 hour lunchtime exposure between periods 3 and 4. In some cases this also differs from
the exposure period average based on study measurements. For example, based on measurements reported in Schelegle et
April 2022
3-43 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
al., (2009), the full exposure period average concentration for the 70 ppb target exposure is 73 ppb, and the average
concentration during exercise is 72 ppb.
3 Statistical significance based on the O3 compared to filtered air response at the study group mean (rounded here to decimal).
3 Ranges reflect the minimum to maximum FEV1 decrements across multiple exposure designs and studies. Study-specific
values and exposure details provided in the PA, Appendix 3A, Tables 3A-1 and 3A-2, respectively.
3 Citations for specific FEV1 findings for exposures above 70 ppb are provided in PA, Appendix 3A, Table 3A-1.
E ND (not determined) indicates these data have not been subjected to statistical testing.
F The data for 30 subjects exposed to 80 ppb by Kim et al. (2011) are presented in Figure 5 of McDonnell et al. (2012).
3 Adams (2006) reported FEV1 data for 60 ppb exposure by both constant and varying concentration designs. Subsequent
analysis of the FEV1 data from the former found the group mean O3 response to be statistically significant (p < 0.002) (Brown et
al., 2008; 2013 ISA, section 6.2.1.1). The varying-concentration design data were not analyzed by Brown et al., 2008.
H Citations for study-specific respiratory symptoms findings are provided in the PA, Appendix 3A, Table 3A-1.
Increased numbers of bronchoalveolar neutrophils, permeability of respiratory tract epithelial lining, cell damage, production of
proinflammatory cytokines and prostaglandins (ISA, Appendix 3, section 3.1.4.4.1; 2013 ISA, section 6.2.3.1).	
For shorter exposure periods (e.g., from one to two hours), with heavy intermittent or
very heavy continuous exercise, higher exposure concentrations, ranging from 80 ppb to 400
ppb, have been studied (ISA, Appendix 3A, section 3.1, Table 3A-3; 2013 ISA, section 6.2.1.1;
2006 AQCD, Chapter 6). Across these shorter-duration studies (which involved ventilation rates
2-3 times greater than in the prolonged [6.6- or 8-hour] exposure studies),58 the lowest exposure
concentration for which statistically significant respiratory effects were reported is 120 ppb, for a
1-hour exposure combined with continuous very heavy exercise and a 2-hour exposure with
intermittent heavy exercise. As recognized above the increased ventilation rate associated with
increased exertion increases the amount of O3 entering the lung, where depending on dose and
the individual's susceptibility, it may cause respiratory effects (2013 ISA, section 6.2.1.1). Thus,
for exposures involving a lower exertion level, a comparable response would not be expected to
occur without a longer duration at this concentration (120 ppb), as is illustrated by the 6.6-hour
study results for this concentration (Appendix 3A, Table 3A-1).
With regard to epidemiologic studies reporting positive associations between O3 exposure
concentrations and respiratory health outcomes such as asthma-related emergency department
visits and hospitalizations, these studies are generally primarily focused on investigating the
existence of a relationship between O3 occurring in ambient air and specific health outcomes,
(versus detailing the specific exposure circumstances eliciting such effects). Accordingly, while
as a whole, this evidence base of epidemiologic studies provides strong support for the
conclusions of causality as summarized in section 3.3.1 above,59 these studies provide less
information on details of the specific O3 exposure circumstances that may be eliciting health
effects associated with such outcomes, and whether these occur under air quality conditions that
58	A quasi-continuous exercise protocol is common to the prolonged exposure studies where study subjects complete
six 50-minute periods of exercise, each followed by 10-minute periods of rest (2013 ISA, section 6.2.1.1).
59	Combined with the coherent evidence from experimental studies, the epidemiologic studies "can support and
strengthen determinations of the causal nature of the relationship between health effects and exposure to ozone at
relevant ambient air concentrations" (ISA, p. ES-17).
April 2022
3-44 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
meet the current standard.60 For example, these studies generally do not measure personal
exposures of the study population or track individuals in the population with a defined exposure
to O3 alone. Further, the vast majority of these studies were conducted in locations and during
time periods that would not have met the current standard. The extent to which reported
associations with health outcomes in the resident populations in these studies are influenced by
the periods of higher concentrations during times that did not meet the current standard is
unknown. While this does not lessen their importance in the evidence base documenting the
causal relationship between O3 and respiratory effects, it means they are less informative in
considering O3 exposure concentrations occurring under air quality conditions allowed by the
current standard. Notwithstanding this, we have considered the epidemiologic studies identified
in the ISA as to what they might indicate regarding O3 exposure concentrations in this regard.
Consistent with the evaluation of the epidemiologic evidence of associations between O3
exposure and respiratory health effects in the ISA, we focus on those studies conducted in the
U.S. and Canada as including populations and air quality characteristics that may be most
relevant to circumstances in the U.S. (ISA, Appendix 3, section 3.1.2). Among the epidemiologic
studies finding a statistically significant positive relationship of short- or long-term O3
concentrations with respiratory effects, there are no single-city studies conducted in the U.S. in
locations with ambient air O3 concentrations that would have met the current standard for the
entire duration of the study (see Appendix 3B, Table 3B-1; ISA, Appendix 3, Tables 3-13, 3-14,
3-39, 3-41, 3-42 and Appendix 6, Tables 6-5 and 6-6;). There are (among this large group of
studies) two single city studies conducted in western Canada that include locations for which the
highest-monitor design values61 fell just below 70 ppb, at 65 and 69 ppb (Appendix 3B, Table
3B-1; Kousha and Rowe, 2014; Villeneuve et al., 2007). These studies did not, however, include
analysis of correlations with other co-occurring pollutants or of the strength of the associations
when accounting for effects of copollutants in copollutant models (ISA, Tables 3-14 and 3-39).
Thus, these studies pose significant limitations with regard to informing conclusions regarding
specific O3 exposure concentrations and elicitation of such effects. There are also a handful of
multicity studies conducted in the U.S. or Canada in which the O3 concentrations in a subset of
the study locations and for a portion of the study period appear to have met the current standard
(Appendix 3B). Concentrations in other portions of the study area or study period, however, do
not meet the standard, or data were not available in some cities for the earlier years of the study
60	For example, these studies generally do not measure personal exposures of the study population or track
individuals in the population with a defined exposure to O3 alone.
61	As described in chapter 2, a design value is the metric used to describe air quality in a given area relative to the
level of the standard, taking the averaging time and form into account. For example, a design value of 70 ppb just
meets the current primary standard.
April 2022
3-45 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
period when design values for other cities in the study were well above 70 ppb. The extent to
which reported associations with health outcomes in the resident populations in these studies are
influenced by the periods of higher concentrations during times that did not meet the current
standard is unknown. Additionally, with regard to multicity studies, the reported associations
were based on the combined dataset from all cities, complicating interpretations regarding the
contribution of concentrations in the small subset of locations that would have met the current
standard compared to that from the larger number of locations that would have violated the
standard (Appendix 3B, Table 3B-1 and Table 3B-2).62 Further, given that populations in such
studies may have also experienced longer-term, variable and uncharacterized exposure to O3 (as
well as to other ambient air pollutants), "disentangling the effects of short-term ozone exposure
from those of long-term ozone exposure (and vice-versa) is an inherent uncertainty in the
evidence base" (ISA, p. IS-87 [section IS.6.1]). While given the depth and breadth of the
evidence base for O3 respiratory effects, such uncertainties do not change our conclusions
regarding the causal relationship between O3 and respiratory effects.
With regard to the experimental animal evidence (largely rodent studies) and exposure
conditions associated with respiratory effects, the exposure concentrations in the animal studies
are generally much greater than those examined in the controlled human exposure studies
(summarized above) and higher than concentrations commonly occurring in ambient air in areas
of the U.S. where the current standard is met. This is also true for the small number of early life
studies in nonhuman primates (recognized in section 3.3.1.1 above) that reported O3 to contribute
to allergic asthma-like effects in infant primates.63 The exposures eliciting the effects in these
studies included multiple 5-day periods with O3 concentrations of 500 ppb over 8-hours per day,
exposure conditions appreciably greater than occur in areas of the U.S. where the current
standard is met (ISA, Appendix 3, section 3.2.4.1.2).
With regard to short-term O3 and metabolic effects, the category of nonrespiratory effects
for which the ISA concludes there to be a likely causal relationship with O3, the evidence base is
comprised primarily of experimental animal studies, as summarized in section 3.3.1.2 above
(ISA, Appendix 5, section 5.1). The exposure conditions from these studies, however, generally
involve much higher O3 concentrations than those examined in the controlled human exposure
studies for respiratory effects (and much higher than concentrations occurring in ambient air in
62	As recognized in the 2015 review, "multicity studies do not provide a basis for considering the extent to which
reported O3 health effects associations are influenced by individual locations with ambient [air] 03 concentrations
low enough to meet the current O3 standard versus locations with O3 concentrations that violate this standard" (80
FR 64344, October 26, 2015).
63	These studies indicate that sufficient early-life O3 exposure can cause structural and functional changes that could
potentially contribute to airway obstruction and increased airway responsiveness (ISA, Table IS-10, p. 3-92 and
p.3-113).
April 2022
3-46 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
areas of the U.S. where the current standard is met). For example, the animal studies include 4-
hour concentrations of 400 to 800 ppb (ISA, Appendix 5, Table 5-8).64 The two epidemiologic
studies reporting statistically significant positive associations of O3 with metabolic effects (e.g.,
changes in glucose, insulin, metabolic clearance) are based in Taiwan and South Korea,
respectively.65 Given the potential for appreciable differences in air quality patterns between
Taiwan and South Korea and the U.S., as well as differences in other factors that might affect
exposure (e.g., activity patterns), those studies are of limited usefulness for informing our
understanding of exposure concentrations and conditions eliciting such effects in the U.S. (ISA,
Appendix 5, section 5.1).
Thus, as in the 2015 review, the exposure to which we give greatest attention, particularly
with regard to considering O3 exposures expected under air quality conditions that meet the
current standard, are those informed by the controlled human exposure studies. The full body of
evidence described in the current ISA continues to indicate respiratory effects as the effects
associated with lowest exposures, with conditions of exposure (e.g., duration, ventilation rate,
and concentration) influencing dose and associated response. Evidence for other categories of
effects does not indicate effects at comparably low exposures.
3.3.4 Uncertainties in the Health Effects Evidence
• To what extent have previously identified uncertainties in the health effects evidence
been reduced or do important uncertainties remain?
We have not identified any new uncertainties in the evidence since the 2015 review.
However, we continue to recognize important uncertainties that also existed at that time. This
array of important areas of uncertainty relates to the available health evidence, including that
newly available in the 2020 review, and is summarized below.
Although the evidence clearly demonstrates that short-term O3 exposures cause
respiratory effects, as was the case in the last review, we continue to recognize uncertainties that
remain in several aspects of our understanding of these effects. Such uncertainties include those
associated with the severity and prevalence of responses to short (e.g., 6.6- to 8-hour) O3
exposures at and below 60 ppb and responses of some population groups not well represented in
the evidence base of controlled human exposure studies (e.g., children and people with asthma).
64	The exposure concentration in the single controlled human exposure study of metabolic effects (e.g., 300 ppb) are
also well above those examined in the respiratory effect studies (ISA, Appendix 5, Table 5-7).
65	Of the five epidemiologic studies discussed in the ISA that investigate associations between short-term 03
exposure and metabolic effects, three are conducted in Asia or South America and two are conducted in the U.S.
The two U.S. studies report either a null or negative association of metabolic markers with O3 concentration, and
while the South American study (focused on hospital admissions associated with diabetes complications) reported
positive associations with 24-hr average concentrations for some subgroups, no associations were statistically
significant (ISA, Appendix 5, Tables 5-6 and 5-9).
April 2022
3-47 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
There are also uncertainties concerning the potential influence of exposure history and co-
exposure to other pollutants on the relationship between short-term O3 exposure and respiratory
effects. With regard to the full health effects evidence base, we also recognize as an important
uncertainty the extent to which O3 exposures are related to health effects other than respiratory
effects. The following discussion touches on each of these types of uncertainty.
The majority of the available studies have generally involved healthy young adult
subjects, although there are some studies involving subjects with asthma, and a limited number
of studies, generally of very short durations (i.e., less than four hours), involving adolescents and
adults older than 50. While there is evidence from short (6.6- to 8-hour) controlled exposure
studies of healthy adult subjects to concentrations as low as 40 ppb, the only controlled human
exposure study of such a duration (7.6 hours with quasi-continuous light exercise) conducted in
people with asthma was for an exposure concentration of 160 ppb (Appendix 3 A, Table 3A-2).
Given the paucity of studies using subjects that have asthma, particularly those at exposure
concentrations likely to occur under conditions meeting the current standard, uncertainties
remain with regard to characterizing the response in people with asthma while at elevated
ventilation to lower exposure concentrations, e.g., below 80 ppb. The extent to which the
epidemiologic evidence, including that recently available, can inform this area of uncertainty
also may be limited.66 As discussed in section 3.3.2 above, given the effects of asthma on the
respiratory system, exposures associated with relatively mild respiratory responses in largely
healthy people may pose an increased risk of more severe responses, including asthma
exacerbation, in people with asthma. Such considerations remain areas of uncertainty at this
time. Thus, uncertainty remains with regard to the extent to which the controlled human
exposure study evidence describes the responses of the populations, such as children with
asthma, that may be most at risk of Cb-related respiratory effects (e.g., through an increased
likelihood of severe responses, or greatest likelihood of response).
Other areas of uncertainty concerning the potential influence of O3 exposure history and
co-exposure to other pollutants on the relationship between short-term O3 exposures and
respiratory effects also remain in the evidence base. As in the epidemiologic evidence in the
2015 review, there is a limited number of studies that include copollutant analyses for a small set
of pollutants (e.g., PM or NO2). Recent studies with such analyses suggest that observed
associations between O3 concentrations and respiratory effects are independent of co-exposures
66 Associations of health effects with O3 that are reported in the epidemiologic analyses are based on air quality
concentration metrics used as surrogates for the actual pattern of O3 exposures experienced by study population
individuals over the period of a particular study. Therefore, the studies are limited in what they can convey
regarding the specific patterns of exposure circumstances (e.g., magnitude of concentrations over specific
duration and frequency) that might be eliciting reported health outcomes.
April 2022
3-48 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
to correlated pollutants or aeroallergens (ISA, sections IS.4.3.1 and IS.6.1; Appendix 3, sections
3.1.10.1 and 3.1.10.2). Despite the increased prevalence of copollutant modeling in recent
epidemiologic studies, however, uncertainty still exists with regard to the independent effect of
O3 given the high correlations observed for some copollutants in some studies and the small
fraction of all atmospheric pollutants included in these analyses (ISA, section IS.4.3.1; Appendix
2, section 2.5). We also note that neither of the two epidemiologic studies of respiratory
outcomes conducting in Canadian areas that would have met the current standard included
copollutant modeling (as recognized in section 3.3.3 above).
Further, although there remains uncertainty in the evidence with regard to the potential
role of exposures to O3 in eliciting health effects other than respiratory effects, the evidence has
been strengthened since the 2015 review with regard to metabolic effects. As noted in section
3.3.1.2 above, the ISA newly identifies metabolic effects as likely to be causally related to short-
term O3 exposures. The evidence supporting this relationship is limited and not without its own
uncertainties. For example, as noted in section 3.3.1.2 above, the conclusion is based primarily
on animal toxicological studies conducted at much higher O3 concentrations than those common
in ambient air in the U.S. A limited number of epidemiologic studies of short-term O3
concentrations and metabolic effects are available, many of which did not control for
copollutants confounding; just two studies, both in Asia, report significant positive associations
with changes in markers of glucose homeostasis (ISA, Appendix 5; sections 5.1.8 and 5.3).
Uncertainty is increased with regard to a relationship between O3 exposure and
cardiovascular effects and mortality, as discussed in section 3.3.1.2 above, including regarding a
now-larger body of controlled human exposure studies providing evidence that is not consistent
with a cardiovascular effect in response to short-term O3 exposure; and a paucity of
epidemiologic evidence indicating more severe cardiovascular morbidity endpoints, that would
be expected if the impaired vascular and cardiac function (observed in animal toxicological
studies) was the underlying basis for cardiovascular mortality (for which epidemiologic studies
have reported some positive associations with O3). Additionally, uncertainties and limitations
recognized in the 2013 ISA (e.g., lack of control for potential confounding by copollutants in
epidemiologic studies) still remain (ISA, section IS. 1.3.1). As discussed in section 3.3.1.2, these
uncertainties also pertain to conclusions regarding short-term O3 and mortality (ISA, Appendix
6, section 6.1.8). Uncertainties are unchanged with regard to other nonrespiratory categories of
effects (described in section 3.3.1.2 above) for which the evidence is either suggestive of, but not
sufficient to infer, a causal relationship or is inadequate to determine if a causal relationship
exists with O3 (ISA, section IS.4.3).
In summary, while there are some changes with regard to limitations and uncertainties of
the health effects evidence base, some key uncertainties associated with the evidence for
April 2022
3-49 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
respiratory effects that were identified in the 2015 review remain, including those related to the
extent of effects at concentrations below those evaluated in controlled human exposure studies,
and the potential for more severe impacts in individuals with asthma, including particularly
children, and in other at-risk populations.
3.4 EXPOSURE AND RISK INFORMATION
Our consideration of the scientific evidence, as in each review of the O3NAAQS, is
informed by results from quantitative analyses of estimated population exposure and consequent
risk. Estimates from the exposure-based analyses, particularly the comparison of daily maximum
exposures to benchmark concentrations, were most informative to the Administrator's decision
in the 2015 review (as summarized in section 3.1 above). This largely reflected the EPA
conclusion that "controlled human exposure studies provide the most certain evidence indicating
the occurrence of health effects in humans following specific O3 exposures," and recognition that
"effects reported in controlled human exposure studies are due solely to O3 exposures, and
interpretation of study results is not complicated by the presence of co-occurring pollutants or
pollutant mixtures (as is the case in epidemiologic studies)" (80 FR 65343, October 26, 2015).67
Therefore, the quantitative analyses developed in the 2020 review focused on exposure-based
risk analyses, in reflection of the emphasis given to these types of analyses and the
characterization of their uncertainties in the 2015 review, along with the availability of new or
updated information, models, and tools that address those uncertainties (IRP, Appendix 5A).
This reconsideration of the 2020 decision will rely on the exposure-based risk analyses
performed in the 2020 review, which were first presented in the 2020 PA and considered in the
2020 decision. These analyses are also presented here and described in detail in the associated
Appendices 3C and 3D. In section 3.4.1, we summarize the conceptual model for the assessment,
as well as key aspects of the assessment design, including the study areas, populations simulated,
67 In the 2015 review, the Administrator placed relatively less weight on the air quality epidemiologic-based risk
estimates, in recognition of an array of uncertainties, including, for example, those related to exposure
measurement error (80 FR 65346, October 26, 2015). In so doing, she recognized key uncertainties in utilizing
the estimated air concentrations and epidemiologic study relationships (often called epidemiologic-based risk
estimates) (80 FR 65316; 79 FR 75277-75279; 2014 HREA, sections 3.2.3.2 and 9.6). These included the
heterogeneity in effect estimates between locations, the potential for exposure measurement errors, and
uncertainty in the interpretation of the shape of concentration-response functions at lower O3 concentrations, as
well as uncertainties related to the public health importance of increases in relatively low O3 concentrations
following air quality adjustment. Lower confidence was also placed in the results of the epidemiologic-based risk
assessment of respiratory mortality risks associated with long-term O3 exposures in consideration of several
factors. Importantly since that time, the causal determinations for short-term O3 exposure with mortality in the
current ISA differ from the 2013 ISA. The current determinations for both short-term and long-term O3 exposure
(as summarized in section 3.1 above) are that the evidence is "suggestive" but not sufficient to infer causal
relationships for O3 with mortality (ISA, Table IS-1).
April 2022
3-50 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
modeling tools, and exposure and risk metrics derived. Sections 3.4.2 and 3.4.3 summarize the
assessment results. Key limitations and uncertainties associated with the assessment estimates
are identified in section 3.4.4. and potential public health implications are discussed in section
3.4.5. An overarching consideration is whether the current exposure and risk information alters
overall conclusions reached in the 2015 review regarding the health risk associated with
exposure to O3 in ambient air which formed an important foundation in the establishment at that
time of the existing standard.
3.4.1 Conceptual Model and Assessment Approach
The long-standing evidence base for Cb-related health effects is comprised of a large
assemblage of controlled human exposure studies, laboratory animal research studies, and air
quality epidemiologic studies. Together, these health effect studies lead to the strongly supported
conclusion that O3 exposure causes respiratory effects (as summarized in section 3.3 above).
This conclusion is strongest with regard to short-term O3 exposures, for which the ISA and
science assessments in prior reviews have determined there to be a causal relationship. The ISA
additionally determines the relationship between long-term exposure and respiratory effects, as
well as between short-term exposures and metabolic effects to be likely causal, recognizing that
associated uncertainties remain in the evidence. Given the relatively greater strength of the
evidence and understanding of the relevant exposure conditions, as well as availability of
appropriate data and modeling tools, the exposure and risk analysis is focused on respiratory
risks associated with short-term O3 exposures.
The controlled human exposure studies document the occurrence of an array of
respiratory effects in humans in a variety of short-term exposure circumstances. These studies, in
combination with the laboratory animal studies, inform our understanding of the mode of action
for 03-attributable effects, including those health outcomes associated with ambient air
concentrations in air quality epidemiologic studies (ISA, Appendix 3, section 3.1.3). Figure 3-3
below illustrates the conceptual model for O3 in ambient air and respiratory effects, with a
particular focus on short-term exposures and including linkages with the risk metrics assessed in
the quantitative analyses described here.
April 2022
3-51 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
a.
=>
to
O
CL
X
LU
CO
o
UJ
vi
P
o >-
CO
*
C/>
a:
Emissions of 03 precursors to ambient air
Indoor Air 4	
03 originating
indoors
Children and adults (all and subgroups with asthma)


Inhalation while at moderate or greater exertion


Respiratory system

r
Lung function decrements (FEV.,, sRaw), inflammation, respiratory
symptoms, etc.


1
Exposures of Concern:
Number & percent of
people experiencing a
day with an exposure (at
elevated breathing rates)
above benchmarks
Lung Function Risk:
Number & percent of
people experiencing a
day with an 03-induced
FEV-1 reduction
(>10%, 15%, 20%)
Incidence of Respiratory
Health Outcomes
(e.g., hospital emergency
department [ED] visits,
hospital admissions [HA],
mortality)
Figure 3-3. Conceptual model for exposure-based risk assessment. Solid lines indicate
processes explicitly modeled in the assessment. Dashed lines indicate relationships
that are not explicitly modeled.
The exposure-based analyses, described in detail in Appendix 3D, were developed based
on this conceptual model, in consideration of the information newly available in the 2020 review.
In these analyses, we have estimated O3 exposures and resulting risk for air quality conditions of
interest, most particularly air quality conditions that just meet the current primary O3 standard.
These analyses inform our understanding of the protection provided by the current primary
standard from effects that the health effects evidence indicates to be elicited in some portion of
exercising people exposed for several hours to elevated O3 concentrations.
The analysis approach employed is summarized in Figure 3-4 below and described in
detail in Appendices 3C and 3D. This approach incorporates the use of an array of models and
data to develop population exposure and risk estimates for a set of eight urban study areas.
Ambient air O3 concentrations were estimated in each study area using an approach that relies on
a combination of ambient air monitoring data, atmospheric photochemical modeling and
April 2022
3-52 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
statistical methods (described in detail in Appendix 3C). Population exposure and risk modeling
is employed to characterize exposures and related lung function risk associated with the ambient
air concentration estimates (described in detail in Appendix 3D). While the lung function risk
analysis focuses only on the specific O3 effect of FEVi reduction, the comparison-to-benchmark
approach, with its use of multiple benchmark concentrations, provides for characterization of the
risk of other respiratory effects, the type and severity of which increase with increased exposure
concentration.
CO
US
a
^5
Ambient Air Monitoring Data (hourly concentrations)
	T	
Hourly concentrations at monitoring sites
1 Adjustments 1

Photochemical
	
Air Quality

Modeling
a>
=3
CD
O
CL
X
LLi
scenarios (just meeting the current standard and other design values)
[ Voronoi Neighbor Averaging (VNA) Interpolation I
		
Hourly concentrations at census tracts
Exposure Modeling (APEX)
(exposure concentrations and ventilation rate for each individual's exposure events)
CO
bd
Population counts
of 7-hour daily
maximum 03
exposures at
elevated ventilation
Time series of 03
exposure events
(concentrations and
ventilation rates) for
each individual
Health-Based
Benchmark
Concentrations
Controlled Human
Exposure Data
(exposures involving
moderate or greater
exertion)
Population counts
of 7-hour daily
maximum 03
exposures at
elevated ventilation
MSS-FEV,
Lung Function
Risk Model
Exposure to Benchmark Comparison
Output: Number and percent of simulated at-risk
populations estimated to experience 1 or more days
with daily maximum exposures, at moderate or
greater exertion, at or above benchmark
concentrations (60 ppb, 70 ppb, 80 ppb)
Exposure-
Response
(E-R)
Function
Lung Function Risk
Output: Number and percent of simulated at-risk
populations estimated to experience 1 or more days
with specified 03-related lung function responses
(FEV, >10%, 15% and 20%)
Figure 3-4. Analysis approach for exposure-based risk analyses. Dashed lines and gray box
indicate the sole lung function risk approach used prior to 2014 HREA.
The analyses estimate exposure and risk for simulated populations in eight study areas in
Atlanta, Boston, Dallas, Detroit, Philadelphia, Phoenix, Sacramento and St. Louis. The eight
study areas represent a variety of circumstances with regard to population exposure to short-term
April 2022
3-53 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
concentrations of O3 in ambient air. The eight study areas range in total population size from
approximately two to eight million and are distributed across the U.S. in seven of the nine
different NOAA climate regions: the Northeast, Southeast, Central, East North Central, South,
Southwest and West (Karl and Koss, 1984). Assessment of this set of study areas and the
associated exposed populations is intended to be informative to the EPA's consideration of
potential exposures and risks that may be associated with the air quality conditions that meet the
current primary standard.
This set of eight study areas represents a streamlined set as compared to the 15 study
areas in the 2015 review, with the areas chosen to ensure they reflect the full range of air quality
and exposure variation expected across major urban areas in the U.S. (2014 HREA, section 3.5).
As a specific example, while seven of the eight study areas were also included in the 2014
HREA, the eighth study area was not, and has been included in the more recent assessment to
insure representation of a large city in the southwest. Additionally, the years simulated reflect
more recent emissions and atmospheric conditions subsequent to data used in the 2014 HREA,
and therefore represent O3 concentrations somewhat nearer the current standard than was the
case for study areas included in the HREA of the 2015 review (Appendix 3C, Table 3C and 2014
HREA, Table 4-1). Thus, the urban study areas (e.g., combined statistical areas that include
urban and suburban populations) the exposure and risk analyses discussed here reflect an array of
air quality, meteorological, and population exposure conditions.
Consistent with the health effects evidence (summarized in section 3.3 above), the focus
of the assessment is on short-term exposures of individuals in the population during times when
they are breathing at an elevated rate. Exposure and risk are characterized for four population
groups that include representation of key at-risk populations (children and people with asthma),
as described in section 3.3.2 above. Two of the four groups are populations of school-aged
children, aged 5 to 18 years:68 all children and children with asthma. Two are populations of
adults: all adults and adults with asthma. Another population identified as at risk for O3, outdoor
workers, was not included due to appreciable data limitations, a decision also made for past
exposure assessments.69
68	The child population group focuses on ages 5 to 18 in recognition of data limitations and uncertainties, including
those related to accurately simulating activities performed, estimating physiological attributes, and also
challenges in asthma diagnoses for children younger than 5 years old.
69	Outdoor workers, due to the requirements of their job spend more time outdoors at elevated exertion. For a
number of reasons, including the appreciable data limitations (e.g., related to specific durations of time spent
outdoors and activity data), and associated uncertainties summarized in Table 3D-64 of Appendix 3D, this group
was not simulated in this assessment. Limited exploratory analyses of a hypothetical outdoor worker population
in the 2014 HREA (single study area, single year) for the 75 ppb air quality scenario estimated an appreciably
greater portion of this population to experience exposures at or above benchmark concentrations than the full
April 2022
3-54 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
Asthma prevalence estimates for each of the entire populations in the eight study areas
ranges from 7.7 to 11.2%; the rates for children in these study areas range from 9.2 to 12.3%
(Appendix 3D, section 3D.3.1). Spatial variation within each study area related to the population
distribution of age, sex, and family income was also taken into account.70 For children, this
variation is greatest in the Detroit study area, with census tract level, age-specific asthma
prevalence estimates ranging from 6.4 to 13.2% for girls and from 7.7 to 25.5% for boys
(Appendix 3D, Table 3D-3).
Ambient air O3 concentrations were estimated in each study area for the air quality
conditions of interest by adjusting hourly ambient air concentrations, from monitoring data for
the years 2015-2017, using a photochemical model-based approach and then applying a spatial
interpolation technique to produce air quality surfaces with high spatial and temporal resolution
(Appendix 3C).71 The photochemical modeling outputs included both modeled O3 concentrations
and sensitivities of O3 concentrations to changes in NOx emissions for each hour in a single year
at all ambient air monitor locations (Appendix 3C, sections 3C.4 and 3C.5). Linear regression
was used with these single-year model outputs to create relationships between the sensitivities
and O3 concentrations at each monitoring location for each hour of the day during each of the
four seasons. The relationships between hourly sensitivities and hourly O3 for each season were
then used with three years of ambient air monitoring data at each location to predict hourly
sensitivities for the complete 3-year record at each monitoring location. From these, we
calculated hourly O3 concentrations at each monitor location based on iteratively increasing NOx
reductions to determine the adjustments necessary for the monitor location with the highest
design value in each study area to just meet the target value, e.g., 70 ppb for the current standard
scenario (Appendix 3C, section 3C.5). Hourly O3 concentrations for all census tracts comprising
each study area were then derived from the model adjusted hourly concentrations at the ambient
air monitor locations using the Voronoi Neighbor Averaging (VNA) spatial interpolation
technique (Appendix 3C, section 3C.6). The final products were datasets of ambient air O3
concentration estimates with high temporal and spatial resolution (hourly concentrations in 500
adult or child populations simulated, although there are a number of uncertainties associated with the estimates
due to appreciable limitations in the data underlying the analyses (2014 HREA, section 5.4.3.2). It is expected
that if an approach similar to that used in the 2014 HREA had been used for this assessment a generally similar
pattern might be observed, although with somewhat lower overall percentages based on the comparison of current
estimates with estimates from the 2014 HREA (Appendix 3D, section 3D.3.2.4).
70	As described in Appendix 3D, section 3D.2.2.2, asthma prevalence in each study area is estimated based on
combining regional national prevalence information from NHIS with U.S census tract level population data by
linking demographic information related to age, sex, and family income. Then, further adjustments were made
using state-level prevalence obtained from the U.S. Behavioral Risk Factor Surveillance System. See Appendix
3D, Attachment 1 for details.
71	A similar approach was used to develop the air quality scenarios for the 2014 HREA.
April 2022
3-55 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
to 1700 census tracts) for each of the eight study areas (Appendix 3C, section 3C.7) representing
each of the three air quality scenarios assessed.72
The photochemical modeling approach involved use of the Comprehensive Air Quality
Model with Extensions (CAMx), version 6.5, instrumented with the higher order decoupled
direct method (HDDM)73 The CAMx-HDDM was run with emissions estimates and
meteorology data for calendar year 2016 to estimate the O3 sensitivities,74 and the linear
regressions of the modeled O3 concentrations to their respective sensitivities were applied to
hourly O3 concentrations reported at ambient air monitors for the 2015-2017 period to determine
the adjustments needed for each air quality scenario (Appendix 3C, sections 3C.4 and 3C.5). We
maximized the spatial representation of the monitoring data by using all available monitors
within each study area (between 12 and 30) in addition to those within 50 km of the study area
boundaries (yielding between 5 and 31 additional monitors per area). Because we selected study
areas having design values close to the level of the current standard, the levels of NOx emissions
adjustments needed to meet the air quality scenarios of interest were generally lower than those
used in the 2014 HREA, thus reducing one of the important sources of uncertainty associated
with these air quality estimates.
Population exposures were estimated using the EPA's Air Pollutant Exposure model
(APEX) version 5, which probabilistically generates a large sample of hypothetical individuals
from demographic and activity pattern databases and simulates each individual's movements
through time and space to estimate their time-series of O3 exposures occurring within indoor,
outdoor, and in-vehicle microenvironments (Appendix 3D, section 3D.2).75 The APEX model
accounts for the most important factors that contribute to human exposure to O3 from ambient
air, including the temporal and spatial distributions of people and ambient air O3 concentrations
throughout a study area, the variation of ambient air-related O3 concentrations within various
microenvironments in which people conduct their daily activities, and the effects of activities
72	For this assessment, high spatial and temporal resolution O3 concentration datasets were created for conditions
representing each area meeting the current standard of 70 ppb and two alternative air quality scenarios characterized
by ozone concentrations that would result in design values of 75 and 65 ppb representing a level slightly above and a
level slightly below the current standard.
73	Details on the models, methods and input data used to estimate ambient air concentrations for the eight study
areas are provided in Appendix 3C. The "higher ordef' aspect of the HDDM tool refers to the capability of
capturing nonlinear response curves (Appendix 3C, section 3C.5.1).
74	Sensitivities of 03 refer to predicted incremental changes in 03 concentrations in response to incremental changes
in precursor emissions (e.g., NOx emissions).
75	The APEX model is a probabilistic model that estimates population exposure using a stochastic, event-based
microenvironmental approach. This model has a history of application, evaluation, and progressive model
development in estimating human exposure, dose, and risk for reviews of NAAQS for gaseous pollutants,
including the 2015 review of the 03 NAAQS (U.S. EPA, 2008; U.S. EPA, 2009; U.S. EPA, 2010; U.S. EPA,
2014; U.S. EPA, 2018).
April 2022
3-56 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
involving different levels of exertion on breathing rate (or ventilation rate) for the exposed
individuals of different sex, age, and body mass in the study area (Appendix 3D, section 3D.2).
The APEX model generates each simulated person or profile by probabilistically selecting values
for a set of profile variables, including demographic variables, health status and physical
attributes (e.g., residence with air conditioning, height, weight, body surface area) and activity-
specific ventilation rate (Appendix 3D, section 3D.2).
By incorporating individual activity patterns76 and estimating physical exertion for each
exposure event,77 the model addresses an important determinant of individual's exposure (2013
ISA, section 4.4.1). This aspect of the exposure modeling is critical in estimating exposure,
ventilation rate, O3 intake (dose), and health risk resulting from ambient air concentrations of
O3 78 Because of variation in O3 concentrations among the different microenvironments in which
individuals are active, the amount of time spent in each location, as well as the exertion level of
the activity performed, will influence an individual's exposure to O3 from ambient air and
potential for adverse health effects. Activity patterns vary both among and within individuals,
resulting in corresponding variations in exposure across a population and over time (2013 ISA,
section 4.4.1). For each exposure event, APEX tracks activity performed, ventilation rate,
exposure concentration, and duration for all simulated individuals throughout the assessment
period. This time-series of exposure events serves as the basis for calculating exposure and risk
metrics of interest.
The APEX model estimates of population exposures for simulated individuals breathing
at elevated rates79 are used to characterize health risk based on information from the controlled
human exposure studies on the incidence of lung function decrements in study subjects who are
exposed over multiple hours while intermittently or quasi-continuously exercising (Appendix
3D, section 3D.2.8). In drawing on this evidence base for this purpose, the assessment gives
76	To represent personal time-location-activity patterns of simulated individuals, the APEX model draws from the
CHAD developed and maintained by the EPA (McCurdy, 2000; U.S. EPA, 2019). The CHAD is comprised of
data from several surveys that collected activity pattern data at city, state, and national levels. Included are
personal attributes of survey participants (e.g., age, sex), the locations visited, and activities performed by survey
participants throughout a day, and the time-of-day activities occurred and their duration (Appendix 3D, section
3D.2.5.1).
77	An exposure event occurs when a simulated individual inhabits a microenvironment for a specified time, while
engaged at a constant exertion level and experiencing a particular pollutant concentration. If the
microenvironmental concentration and/or activity/activity level changes, a new exposure event occurs (McCurdy
and Graham, 2003).
78	Indoor sources are generally minor in comparison to O3 from ambient air (ISA, Appendix 2, section 2.4.3) and are
not accounted for by the exposure modeling in this assessment.
79	Based on minute-by-minute activity levels, and physiological characteristics of the simulated person, APEX
estimates an equivalent ventilation rate (EVR), by normalizing the simulated individuals' activity-specific
ventilation rate to their body surface area (Appendix 3D, section 3D.2.2.3.3).
April 2022
3-57 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
primary focus to the well-documented controlled human exposure studies summarized in
Appendix 3A, Table 3A-1 for 6.6-hour average exposure concentrations ranging from 40 ppb to
120 ppb (Figure 3-2; ISA, Appendix 3, Figure 3-3). Health risk is characterized in two ways,
producing two types of risk metrics: one involving comparison of population exposures
involving elevated exertion to benchmark concentrations (that are specific to elevated exertion
exposures), and the second involving estimated population occurrences of ambient air Cb-related
lung function decrements (Figure 3-2). The first risk metric estimates population occurrences of
daily maximum 7-hour average exposure concentrations (during periods of elevated breathing
rates) at or above concentrations of potential concern (benchmark concentrations). The second
metric (lung function risk) uses E-R information for O3 exposures and FEVi decrements to
estimate the portion of the simulated at-risk population expected to experience one or more days
with an Cb-related FEVi decrement of at least 10%, 15% and 20%. Both of these metrics are
used to characterize health risk associated with O3 exposures among the simulated population
during periods of elevated breathing rates. Similar risk metrics were also derived in the HREA
for the 2015 review and the associated estimates informed the Administrator's 2015 decision on
the current standard (80 FR 65292, October 26, 2015).
The general approach and methodology for the exposure-based assessment is similar to
that used in the 2015 review although a number of updates and improvements, related to the air
quality, exposure and risk aspects of the assessment, have been implemented (Appendices 3C
and 3D). These are summarized here.
•	The ambient air monitoring data used is from a more recent period (e.g., 2015-2017)
during which O3 concentrations in the eight study areas are at or near the current standard
(Appendix 3C, Table 3C-1). This contrasts with the 2014 HREA use of 2006-2010 air
monitoring data, that for many study areas included design values (for unadjusted
concentrations) well above (e.g., by more than 10 ppb) the level of the then-existing
standard (2014 HREA, section 4.3.1.1, Table 4-1). The use of more recent ambient air
monitoring data in the current analysis allows for smaller adjustments to develop the air
quality conditions of interest, thus contributing to generally lesser uncertainty in the
concentrations estimated in each air quality scenario.
•	The most recent CAMx model, with updates to the treatment of atmospheric chemistry
and physics within the model, is used to derive spatially and temporally varying
relationships between changes to emissions and modeled O3 concentrations, which are
then used in adjusting ambient air concentrations to just meet the air quality scenarios.
Model inputs represent recent year emissions, meteorology, and international transport
(e.g., 2016). The 2016-based inputs were derived using updated methods for calculating
emissions, as well as updated meteorological and hemispheric photochemical models
(described in more detail in Appendix 3C).
•	A significantly expanded CHAD, with now nearly 180,000 diaries, including over 25,000
for school-aged children is drawn on in the exposure modeling (Appendix 3D, section
April 2022
3-58 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
3D.2.5.1), as are updated National Health and Nutrition Examination Survey data (2009-
2014), which are the basis for the age- and sex-specific body weight distributions used to
specify the individuals in the modeled populations (Appendix 3D, section 3D.2.2.3.1).
•	Population exposure modeling inputs include the most recent U.S. Census demographics
and commuting data (i.e., 2010), meteorological data to reflect the assessment years
studied (e.g., 2015-2017), and updated estimates of asthma prevalence for all census
tracts in all study areas (e.g., 2013-2017). Regarding asthma prevalence, the more recent
information includes increased prevalence reported for adults and for children aged 10-17
years (Akinbami et al., 2016; CDC, 2016).80
•	The APEX equations used to estimate of ventilation rate (Ve) and resting metabolic rate
have been updated such that the overall statistical model fit and predictability has been
improved (U.S. EPA, 2018, Appendix H).
•	The approach for deriving population exposure estimates, both for comparison to
benchmark concentrations and for use in deriving lung function risk using the E-R
function, has been modified to provide for a better match of the simulated population
exposure estimates with the 6.6-hour duration of the controlled human exposure studies
and with the study subject ventilation rates (Appendix 3D, section 3D.2.8.1). The
modifications include deriving estimates for exposures of a duration and ventilation rate
more closely corresponding to the duration and average ventilation rate across the 6.6-
hour duration in the controlled human exposure studies (Appendix 3D, section 3D.2.8.1).
81
•	In addition to the E-R function, as updated in the 2014 HREA, an updated version of the
McDonnell Stewart Smith model (MSS-FEVi model, McDonnell et al., 2013) is used to
estimate individual-based lung function risk. Although the impact on risk estimates is
unclear, the updated MSS model has been described as better accounting for intra-subject
variability, yielding an improved model fit (McDonnell et al., 2013; Appendix 3D,
section 3D.2.8.2.2).
The comparison-to-benchmarks analysis characterizes the extent to which individuals in
at-risk populations could experience O3 exposures, while engaging in their daily activities, with
the potential to elicit the effects reported in controlled human exposure studies for concentrations
at or above specific benchmark concentrations. Results are characterized through comparison of
exposure concentrations to three benchmark concentrations of O3: 60, 70, and 80 ppb. These are
based on the three lowest concentrations targeted in studies of 6- to 6.6-hour exposures, with
quasi-continuous exercise (at moderate level of exertion), and that yielded different occurrences
80	For more information, see https://www.cdc.gov/nchs/products/databriefs/db239.htm.
81	Estimated exposures for a 7-hour duration are used in the comparison to benchmark concentrations (that are based
on the 6.6-hour exposure studies). The use of 7-hour exposure duration provides for a closer match of the duration
for the benchmark concentrations to the duration of population exposure concentration estimates than the 8-hour
exposure concentrations used in the last review. Additionally, an equivalent ventilation rate (EVR) of at least 17.3
L/min-m2 is used to more closely correspond to the average across the 6.6 hours of the controlled human
exposure studies (Appendix 3D, section 3D.2.8.1).
April 2022
3-59 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
of statistical significance, and severity of respiratory effects (section 3.3.3 above; Appendix 3 A,
section 3A.1; Appendix 3D, section 3D.2.8.1). The lowest benchmark, 60 ppb, represents the
lowest exposure concentration for which controlled human exposure studies have reported
statistically significant respiratory effects (as summarized in section 3.3.3 above). Exposure to
approximately 70 ppb82 averaged over a similar time resulted in a larger group mean lung
function decrement, as well as a statistically significant increase in prevalence of respiratory
symptoms over what was observed for 60 ppb (Figure 3-3; ISA, Appendix 3, section 3.1.4.1.1;
Schelegle et al., 2009). Studies of exposures to approximately 80 ppb have reported larger lung
function decrements at the study group mean than following exposures to 60 or 70 ppb, in
addition to an increase in airway inflammation, increased respiratory symptoms, increased
airway responsiveness, and decreased resistance to other respiratory effects (Figure 3-3 and
section 3.3.3, above; ISA, Appendix 3, sections 3.1.4.1-3.1.4.4).
The APEX-generated exposure concentrations for comparison to these benchmark
concentrations is the average of concentrations encountered by an individual while at an activity
level that elicits the specified elevated ventilation rate.83 The incidence of such exposures at or
above the benchmark concentrations are summarized for each simulated population, study area,
and air quality scenario as discussed in sections 3.4.2 and 3.4.3 below (Appendix 3D).
The lung function risk analysis estimates (in two different ways) the extent to which
individuals in exposed populations could experience different sizes of Cb-induced lung function
decrements. The two different approaches utilize the evidence from the 6.6-hour controlled
human exposure studies in different ways.84 One, the population-based E-R function, uses
quantitative descriptions of the E-R relationships for study group incidence of different
82	The design for the study on which the 70 ppb benchmark concentration is based, Schelegle et al. (2009), involved
varying concentrations across the full exposure period. The study reported the average O3 concentration measured
during each of the six exercise periods. The mean concentration across these six values is 72 ppb. The 6.6-hr time
weighted average based on the six reported measurements and the study design is 73 ppb (Schelegle et al., 2009).
Other 6.6-hr studies generally report an exposure concentration precision at or below 3 ppb (e.g., Adams, 2006b).
83	The model averages the ventilation rate (Ve) for the exposed individual (based on the activities performed) over 7-
hour periods. This is done based on the APEX estimates of Ve and exposure concentration for every individual's
time-series of exposure events. For the exposure duration of interest (e.g., 7 hours), the model derives and outputs
the daily maximum average Ve (and hence an equivalent ventilation rate or EVR) and simultaneously occurring
exposure concentration for the specified duration for each simulated individual. To reasonably extrapolate the
ventilation rate of the controlled human study subjects (i.e., adults having a specified body size and related lung
capacity), who were engaging in quasi-continuous exercise during the study period, to individuals having varying
body sizes (e.g., children with smaller size and related lung capacity), an equivalent ventilation rate (EVR) was
calculated by normalizing the ventilation rate (L/min) by body surface area (m2). Seven-hour exposure
concentrations associated with 7-hour average EVR at or above the target of 17.3 ฑ1.2 L/min-m2 (i.e., the value
corresponding to average EVR across the 6.6-hour study duration in the controlled human exposure studies) are
compared to the benchmark concentrations (Appendix 3D, section 3D.2.8.1).
84	The two approaches also estimate responses associated with unstudied exposure circumstances and population
groups in different ways.
April 2022
3-60 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
magnitudes of lung function decrements based on the individual study subject observations. The
second, the individual-based MSS model, uses quantitative estimations of biological processes
identified as important in eliciting the different sizes of decrements at the individual level, with a
factor that also provides a representation of intra- and inter-individual response variability
(Appendix 3D, section 3D.2.8.2.2). The two approaches, described in detail in Appendix 3D,
utilize evidence from the 6.6-hour controlled human exposure studies in different ways, and
accordingly, differ in their strengths, limitations, and uncertainties.
The E-R function used for estimating the risk of lung function decrements was developed
from the individual study subject measurements of Cb-related FEVi decrements from the 6.6-
hour controlled human exposure studies targeting mean exposure concentrations from 120 ppb
down to 40 ppb (Appendix 3D, Table 3D-19; Appendix 3A, Figure 3A-1). The FEVi responses
reported in these studies have been summarized in terms of percent of study subjects
experiencing Cb-related decrements equal to at least 10%, 15% or 20%. Across the exposure
range from 40 to 120 ppb, the percentage of exercising study subjects with asthma estimated to
have at least a 10% O3 related FEVi decrement increases from 0 to 7% (a statistically non-
significant response at exposures of 40 ppb) up to approximately 50 to 70% (at exposures of 120
ppb) (Appendix 3D, Section 3D.2.8.2.1, Table 3D-19). The E-R function relies on equations that
describe the fraction of the population experiencing a particular size decrement as a function of
the exposure concentration experienced while at the target ventilation rate.85 This type of risk
model has been used in risk assessments since the 1997 O3 NAAQS review. As used here, the
functions (fraction of the population having of a day or more per simulation period with at least
one decrement of one of the specified sizes) are applied to the APEX estimates of 7-hour average
exposure concentrations concomitant with the target ventilation level estimated by APEX, with
the results presented in terms of number of individuals in the simulated populations (and percent
of the population) estimated to experience a day (or more) with a lung function decrement at or
above 10%, 15% and 20%.
The MSS model, also used for estimating the risk of lung function decrements, was
developed using the extensive database from controlled human exposure studies that has been
compiled over the past several decades, and biological concepts based on that evidence
(McDonnell et al., 2012; McDonnell et al., 2013). The model mathematically estimates the
magnitude of FEVI decrement as a function of inhaled O3 dose (based on concentration &
ventilation rate) over the time period of interest (Appendix 3D, section 3D.2.8.2.2). The
simulation of decrements is dynamic, based on a balance between predicted development of the
85 This risk model was updated in the 2015 review to include the more recently available study data at that time
(Appendix 3D, section 3D.2.8.2.1).
April 2022
3-61 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
decrement in response to inhaled dose and predicted recovery (using a decay factor). Each
occurrence of decrements of interest (e.g., at or above 10%, 15% and 20%) is tallied. This model
was first applied in combination with the APEX model to generate lung function risk estimates
in the 2015 03 NAAQS review (80 FR 65314, October 26, 2015).86
To generate risk estimates for lung function decrements, the model is applied to the
APEX estimates of exposure concentration and ventilation for every exposure event experienced
by a simulated individual. The model then utilizes its mathematical descriptions of dose
accumulation and decay, and relationship of dose to response, to estimate the magnitude of O3
response associated with the sequence of exposure events in each individual's day. We report the
MSS model risk results using the same metrics as for the E-R function, i.e., number of
individuals in the simulated populations (and percent of the population) estimated to experience
a day (or more) per simulation period with a lung function decrement at or above 10%, 15% and
20%.
The comparison-to-benchmark analysis (involving comparison of 7-hour average
exposure concentrations that coincide with a 7-hour average elevated ventilation rates) provides
perspective on the extent to which the air quality being assessed could be associated with
discrete exposures to O3 concentrations reported to result in an array of respiratory effects. For
example, estimates of such exposures can indicate the potential for Cb-related effects in the
exposed population, including effects for which we do not have E-R functions that could be used
in quantitative risk analyses (e.g., airway inflammation). Thus, the comparison-to-benchmark
analysis differs from the two lung function risk analyses with their specific focus on lung
function decrements and provides for a broader risk characterization with consideration of the
array of Cb-related respiratory effects.
3.4.2 Population Exposure and Risk Estimates for Air Quality Just Meeting the Current
Standard
In this section, we consider the exposure and risk estimates in the context of the
following questions.
• What are the nature and magnitude of O3 exposures and associated health risks for
air quality conditions just meeting the current standard? What portions of the
exposed populations are estimated to experience exposures of concern or lung
function decrements?
To address these questions, we consider the estimates provided by the exposure and risk
simulations for the eight urban study areas with air quality conditions adjusted to just meet the
86 As noted below, the MSS model used in the current assessment has been updated since the 2015 review based on
the most recent study by its developers (McDonnell et al., 2013).
April 2022
3-62 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
current standard (Appendix 3D, sections 3D.3.2 through 3D.3.3). In considering these estimates
here and their associated limitations, uncertainties and implications in greater depth in sections
3.4.5 and 3.5 below, we particularly focus on the extent of protection provided by the standard
from O3 exposures of potential concern. As described in the prior section, the exposure and risk
analyses present two types of risk estimates for the 3-year simulation in each study area: (1) the
number and percent of simulated people experiencing exposures at or above the particular
benchmark concentrations of interest in a year, while breathing at elevated rates; and (2) the
number and percent of people estimated to experience at least one Cb-related lung function
decrement (specifically, FEVi reductions of a magnitude at or above 10%, 15% or 20%) in a
year and the number and percent of people estimated to experience multiple lung function
decrements.
As an initial matter regarding the objectives for the analysis approach, we note that the
analyses and the use of an urban case study approach (summarized in section 3.4.1 above) are
intended to provide assessments of air quality scenarios, including in particular one just meeting
the current standard, for a diverse set of areas and associated exposed populations. These
analyses are not intended to provide a comprehensive national assessment. Nor is the objective to
present an exhaustive analysis of exposure and risk in the areas that currently just meet the
current standard and/or of exposure and risk associated with air quality adjusted to just meet the
current standard in areas that currently do not meet the standard. Rather, the purpose is to assess,
based on current tools and information, the potential for exposures and risks beyond those
indicated by the information available at the time the standard was established. Accordingly, use
of this approach recognizes that capturing an appropriate diversity in study areas and air quality
conditions (that reflect the current standard scenario)87 is an important aspect of the role of the
exposure and risk analyses in informing the Administrator's conclusions on the public health
protection afforded by the current standard.
Of the two types of risk metrics derived in the exposure and risk analyses, we turn first to
the results for the benchmark-based risk metric, which are summarized in terms of the percent of
the simulated populations of all children and children with asthma estimated to experience at
87 A broad variety of spatial and temporal patterns of 03 concentrations can exist when ambient air concentrations
just meet the current standard. These patterns will vary due to many factors including the types, magnitude, and
timing of emissions in a study area, as well as local factors, such as meteorology and topography. We focused our
current assessment on specific study areas having ambient air concentrations close to conditions that reflect air
quality that just meets the current standard. Accordingly, assessment of these study areas is more informative to
evaluating the health protection provided by the current standard than would be an assessment that included areas
with much higher and much lower concentrations.
April 2022
3-63 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
least one day per year88 with a 7-hour average exposure concentration at or above the different
benchmark concentrations while breathing at elevated rates under air quality conditions just
meeting the current standard (Table 3-3). The estimates for the adult populations, in terms of
percentages, are generally lower, due to the lesser amount and frequency of time spent outdoors
at elevated exertion (Appendix 3D, section 3D.3.2). Given the recognition of people with asthma
as an at-risk population and the relatively greater amount and frequency of time spent outdoors at
elevated exertion of children, we focus here on the estimates for children, including children with
asthma.
Under air quality conditions just meeting the current standard, less than 0.1% of any
study area's children with asthma, on average, were estimated to experience any days per year
with a 7-hour average exposure at or above 80 ppb, while breathing at elevated rates (Table 3-3).
With regard to the 70 ppb benchmark, the study areas' estimates for children with asthma range
up to 0.7 percent (0.6% for all children), on average across the 3-year period, and range up to
1.0% in a single year (Table 3-3). Approximately 3% to nearly 9% of each study area's
simulated children with asthma, on average across the 3-year period, are estimated to experience
one or more days per year with a 7-hour average exposure at or above 60 ppb (Table 3-3). This
range is very similar for the populations of all children (Table 3-3).
Regarding multiday occurrences, we see that no children are estimated to experience
more than a single day with a 7-hour average exposure at or above 80 ppb in any year simulated
in any study location (Table 3-3). For the 70 ppb benchmark, the estimate is less than 0.1% of
any area's children (on average across 3-year period), both those with asthma and all children
(Table 3-3, Figure 3-4). The estimates for the 60 ppb benchmark are slightly higher, with up to
3% of children estimated to experience more than a single day with a 7-hour average exposure at
or above 60 ppb, on average (and more than 4% in the highest year across all eight study area
locations) (Table 3-3).
These estimates are based on analyses that, while based on conceptually similar
approaches to those used in the 2014 HREA, reflect the updates and revisions to those
approaches that have been implemented since that time. Taking that into consideration, the
estimates for the 3-year period from the current assessment for air quality conditions simulated to
just meet the current standard are of a magnitude roughly similar, although slightly lower at the
upper end of the ranges, to the estimates for these same populations in the 2014 HREA. For
88 The three years of ambient air O3 concentrations analyzed in the exposure assessment analyses include
concentrations during the O3 seasons for that area. These seasons capture the times during the year when
concentrations are elevated (80 FR 65419-65420, October 26, 2015). While the duration of an O3 season for each
year may vary across the study areas, for the purposes of the exposure and risk analyses, the O3 season in each
study area is considered synonymous with a year.
April 2022
3-64 External Review Draft - Do Not Quote or Cite

-------
1	example, for air quality conditions just meeting the standard with a level of 70 ppb, the 2014
2	HREA estimated 0.1 to 1.2% of children to experience at least one day with exposure at or above
3	70 ppb, while at elevated ventilation (Appendix 3D, section 3D.3.2.4, Table 3D-38). There are a
4	number of differences between the quantitative modeling and analyses performed in the current
5	assessment and the 2014 HREA that likely contribute to the small differences in estimates
6	between the two assessments (e.g., 2015-2017 vs. 2006-2010 distribution of ambient air
7	concentrations, full statistical distribution of ventilation rates vs. a 5th percentile point estimate,
8	7-hour vs. 8-hour exposure durations).
9	Table 3-3. Percent and number of simulated children and children with asthma
10	estimated to experience at least one or more days per year with a 7-hour
11	average exposure at or above indicated concentration while breathing at an
12	elevated rate in areas just meeting the current standard.
Exposure
Concentration
(ppb)
One or more days
Two or more days
Four or more days
Average per
year
Highest in a
single year
Average
per year
Highest in a
single year
Average
per year
Highest in a
single year
Children with asthma - percent of simulated population A
>80
0B — <0.1 c
0.1
0
0
0
0
>70
0.2-0.7
1.0
<0.1
0.1
0
0
>60
3.3-8.8
11.2
0.6-3.2
4.9
<0.1 -0.8
1.3
-numl
b er of individuals A
>80
0-67
202
0
0
0
0
>70
93-1145
1616
3-39
118
0
0
>60
1517-8544
11776
282-2609
3977
23-637
1033
All children - percent of simulated population A
>80
0B - <0.1
0.1
0
0
0
0
>70
0.2-0.6
0.9
<0.1
0.1
0 - <0.1
<0.1
>60
3.2-8.2
10.6
0.6-2.9
4.3
<0.1 -0.7
1.1
-numl
b er of individuals A
>80
0-464
1211
0
0
0
0
>70
727-8305
11923
16-341
757
0-5
14
>60
14928-
69794
96261
2601 -
24952
36643
158-5997
9554
A Estimates for each study area were averaged across the 3-year assessment period. Ranges reflect the ranges of averages.
B A value of zero (0) means that there were no individuals estimated to have the selected exposure in any year.
cAn entry of <0.1 is used to represent small, non-zero values that do not round upwards to 0.1 (i.e., <0.05).
13
14	In framing these same exposure estimates from the perspective of estimated protection
15	provided by the current standard, these results indicate that, in the single year with the highest
16	concentrations across the 3-year period, 99% of the population of children with asthma would
17	not be expected to experience such a day with an exposure at or above the 70 ppb benchmark;
April 2022
3-65 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
99.9% would not be expected to experience such a day with exposure at or above the 80 ppb
benchmark. The estimates, on average across the 3-year period, indicate that over 99.9%, 99.3%
and 91.2%) of the population of children with asthma would not be expected to experience a day
with a 7-hour average exposure while at elevated ventilation that is at or above 80 ppb, 70 ppb
and 60 ppb, respectively (Table 3-3 above). Further, with regard to multiple days, more than
approximately 97% of all children or children with asthma (on average across a 3-year period),
are estimated to be protected against multiple days of exposures at or above 60 ppb. These
estimates indicate generally similar protection to that described in establishing the current
standard in 2015 (as summarized in section 3.1 above), with slightly greater level of protection
for occurrences at 70 ppb (see section 3.5.2 below, refer to Table 3-8).
With regard to lung function risk, the estimates for all children and for children with
asthma are again roughly similar, with the higher end of the ranges for the eight study areas
being just slightly higher in some cases for the children with asthma (Table 3-4). The lung
function risk estimates from the MSS model are appreciably higher than those based on the E-R
function (full results in Appendix 3D, section 3D.3.3). This difference relates to the fact, noted in
section 3.4.1 above, that the two lung function risk approaches are based on different aspects of
the controlled human exposure study evidence and differ in how they extrapolate beyond the
exposure study conditions and observations. Accordingly, uncertainties associated with the two
modeling approaches also differ (as discussed in section 3.4.4 below). The E-R function risk
approach conforms more closely to the circumstances of the 6.6-hour controlled human exposure
studies, such that the 7-hour duration and moderate or greater exertion level are necessary for
nonzero risk. This approach additionally, however, uses a continuous function which predicts
responses for exposure concentrations below those studied down to zero. As a result, exposures
below those studied in the controlled human exposures will result in a fraction of the population
being estimated by the E-R function to experience a lung function decrement (albeit to an
increasingly small degree with decreasing exposures). The MSS model, which has been
developed based on a conceptualization intended to reflect a broader set of controlled human
exposure studies (e.g., including studies of exposures to higher concentrations for shorter
durations), does not require a 7-hour exposure period for the model to generate an estimated
response, and lung function decrements are estimated for exertion below moderate or greater
levels, as well as for exposure concentrations lower than those that have been studied (Appendix
3D, section 3D.3.4.2; 2014 HREA section 6.3.3). These differences in the models, accordingly,
result in differences in the extent to which they produce estimates that reflect the particular
April 2022
3-66 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
conditions of the available controlled human exposure studies and the frequency and magnitude
of the measured responses in those studies.89
For example, the 6.6-hour controlled human exposure studies have reported
approximately 3% of subjects exposed to an average concentration of 60 ppb and 10% of
subjects exposed to 70 ppb to have at least a 15% FEVi decrement (Appendix 3D, Table 3D-20
and Figure 3D-11). Table 3-3 above shows that, at a maximum, approximately 11% and 1% of
children with asthma are estimated in a single year to have a day with daily maximum 7-hour
exposure at or above the 60 ppb and 70 ppb benchmarks, respectively, indicating that perhaps
10%) (11%) minus 1 %>) might be expected to have a day with an exposure at or above 60 ppb but
less than 70 ppb. If the simulated children had the same sensitivity as the controlled human
exposure study subjects, it might be expected that 0.3% (3% times 10%>) of this group could have
a 15%o (or larger) FEVi decrement resulting from concentrations at or above 60 ppb and less than
70 ppb and 0.1%> (10%> times 1%>) of this group could have a 15% (or larger) decrement resulting
from concentrations at or above 70 ppb. Accordingly, this would yield an estimated lung
function risk for the simulated population of 0.4% for decrements of 15% or larger. This
contrasts with the estimates based on the E-R function, that are at most a 1% risk (Table 3-4),
and the MSS model estimates, that are at most an 8.7% risk (Table 3-4).
89 The two models, their bases in the evidence and associated limitations and uncertainties are discussed in detail in
Appendix 3D, sections 3D.2.8.2 and 3D.3.4.
April 2022
3-67 External Review Draft - Do Not Quote or Cite

-------
1	Table 3-4. Percent of simulated children and children with asthma estimated to
2	experience at least one or more days per year with a lung function decrement
3	at or above 10,15 or 20% while breathing at an elevated rate in areas just
4	meeting the current standard.
Lung Function
DecrementA
One or more days
Two or more days
Four or more days
Average
per year
Highest in a
single year
Average
per year
Highest in a
single year
Average per
year
Highest in a
single year
E-R Function

percent of simulatec
children with asthma A
> 20%
CO
O
I
CM
O
0.4
CM
O
I
O
0.2
O
I
CD
O
V
0.1
> 15%
0.5-0.9
1.0
CO
O
I
CO
0
0.6
O
I
CM
O
0.4
> 10%
CO
CO
I
CO
csj
3.6
1.5-2.4
2.6
0.9-1.7
1.8

percent of all simulated children A
> 20%
CO
O
I
CM
O
0.4
CM
O
I
O
0.2
A
O
I
O
0.1
> 15%
OO
0
I
LO
0
0.9
LO
O
I
CO
0
0.6
O
I
CM
O
0.4
> 10%
CO
I
CM
CM
3.3
CM
CM
I
CO
2.4
CD
I
OO
O
1.7
MSS Model

percent of simulatec
children with asthma A
> 20%
LO
CO
I
OO
3.9
CM
I
OO
O
2.5
0.3-1.1
1.3
> 15%
4.5-8.2
8.7
CT>
I
CM
CM
5.3
1.1 -2.9
3.3
> 10%
13.9-22
23.3
8.0-14.9
16
CO
I
CO
CO
10.5

percent of all simulated children A
> 20%
1.7-3.1
3.6
O
CO
I
2.0
CD
O
I
CO
0
1.1
> 15%
4.1-7.1
7.8
CO
I
CM
4.9
LO
CM
I
O
2.9
> 10%
13.2-20.4
21.8
7.4-13.6
14.8
3.9-8.8
9.7
A Estimates for each urban case study area were averaged across the 3-year assessment period. Ranges reflect the ranges
across urban study area averages.
B An entry of <0.1 is used to represent small, non-zero values that do not round upwards to 0.1 (i.e., <0.05).
5
6	3.4.3 Population Exposure and Risk Estimates for Additional Air Quality Scenarios
7	In addition to estimating population exposure and risk for O3 concentrations simulated to
8	occur under air quality conditions when the current standard is just met, the exposure and risk
9	analyses also estimated population exposure and risk in the eight study areas for two additional
10	air quality scenarios. In these scenarios, the air quality conditions were adjusted such that the
11	monitor location with the highest concentrations in each area had a design value just equal to
12	either 75 ppb or 65 ppb.
13	The results for the comparison-to-benchmarks analysis for these additional air quality
14	scenarios are summarized in Table 3-5 below for all three benchmark concentrations. The
15	estimates for these two additional scenarios differ markedly from the results for air quality just
16	meeting the current standard (summarized in Table 3-3 above). For simplicity, the summary of
17	the comparison discussed here focuses on the 70 ppb benchmark concentration, which falls just
April 2022
3-68 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
below the time-weighted exposure concentration for which there was a statistically significant
lung function decrement and also a statistically significant increase in respiratory symptom score
in one of the controlled human exposure studies, as noted in section 3.3.3 (ISA, Appendix 3,
section 3.1.4.1.1; Schelegle et al., 2009). The pattern is similar for the other two benchmarks,
although in general, the differences of the results for the additional scenarios from the results for
the current standard (presented in section 3.4.2) are somewhat greater for the higher benchmark
and slightly smaller for the lower benchmark.
Under air quality conditions in the 75 ppb scenario, estimated percentages of children
with asthma expected to experience at least one day per year with exposures at or above the
benchmark concentrations are two or more times higher than the estimates discussed in section
3.4.2 above for air quality conditions just meeting the current standard. For example, the
minimum and maximum percentages, on average per year across the study areas, of children
with asthma estimated to experience one or more days with exposures at or above the 70 ppb
benchmark are five and three times, respectively, greater than the corresponding percentages for
conditions associated with the current standard (Table 3-3 and Table 3-5). The highest estimated
percentage in a single year for the 70 ppb benchmark is more than twice as high for the 75 ppb
scenario compared to conditions associated with the current standard. The corresponding
estimate for two or more days per year is even greater for the 75 ppb scenario versus the current
standard scenario (Table 3-3 and Table 3-5).
In contrast, under air quality conditions in the 65 ppb scenario, the estimated percentages
of children with asthma expected to experience at least one day per year with exposures above
the benchmark concentrations are at most one third the estimates discussed in section 3.4.2 above
for air quality conditions just meeting the current standard (Table 3-3 and Table 3-5). The
highest estimated percentage of children expected to experience two or more days a year at or
above the 70 ppb benchmark drops to zero for the 65 ppb scenario compared to <0.1% for air
quality conditions just meeting the current standard (Table 3-3, Table 3-5).
As with the estimates for air quality just meeting the current standard, and as expected
given the various exposure and risk analysis updates implemented, the estimates discussed here
for the additional air quality scenarios are also slightly different from the estimates for such
scenarios that were derived in the 2015 review. However, the differences are not of such a
magnitude that the estimates for one air quality scenario in the current analyses are similar to
results for a different scenario in the 2015 review. For example, while the current estimates for
the 75 ppb air quality scenario are somewhat lower for some benchmarks than those for that
scenario in the 2015 review, they are still higher than the estimates from the 2015 review for the
air quality scenario just meeting the current standard.
April 2022
3-69 External Review Draft - Do Not Quote or Cite

-------
1	Table 3-5. Percent and number of simulated children and children with asthma
2	estimated to experience one or more days per year with a daily maximum 7-
3	hour average exposure at or above indicated concentration while breathing at
4	an elevated rate - additional air quality scenarios.
Exposure
One or more days
Two or more days
Four or more days
Concentration
(ppb)
Average per
year
Highest in a
single year
Average per
year
Highest in a
single year
Average per
year
Highest in a
single year
Air quality scenario for 75 ppb
Children with asthma
- percent of simulated population A


>80
<0.1B-0.3
0.6
O
0
1
A
O
<0.1
0
0
>70
1.1 -2.1
3.9
O
I
o
0.8
0
1
A
O
0.1
>60
7.6-17.1
19.2
CD
CO
I
o
CM
11.0
CO
CO
I
o
4.4
- number of individuals A
>80
23-410
888
1^
i
o
20
0
0
>70
502 - 2480
4544
36-316
637
0-33
99
>60
3538-14054
17673
1188-7232
8931
204-2708
3595
All child
ren
- percent of simulated population A


>80
<0.1 b-0.3
0.6
O
0
1
A
O
<0.1
0
0
>70
O
cm
I
3.4
CO
o
I
o
0.7
<0.1
<0.1
>60
6.6-15.7
17.9
I
00
o
9.9
O
CO
I
o
4.1
- number of individuals A
>80
129-3127
6658
0-54
121
0
0
>70
4915-19794
34981
414-2750
5775
3-141
368
>60
34918-133400
162894
11087-67747
83660
1813 - 25773
34902
Air quality scenario for 65 ppb
Children with asthma
- percent of simulated population A


>80
O
I
A
O
<0.1
0
0
0
0
>70
CM
O
I
o
0.3
0
0
0
0
>60
0.5-2.5
4.3
CO
o
I
O
V
0.6
O
I
A
O
0.1
- number of individuals A
>80
0-23
68
0
0
0
0
>70
0-311
455
0
0
0
0
>60
212-3542
5165
13-386
709
0-14
42
All child
ren
- percent of simulated population A


>80
CD
V
I
o
<0.1
0
0
0
0
>70
CM
o
I
o
0.2
O
I
A
O
<0.1
0
0
>60
CO
CM
I
o
3.7
CO
O
I
O
V
0.5
0
1
A
O
<0.1
- number of individuals A
>80
0-38
114
0
0
0
0
>70
0-2495
3140
0-13
23
0
0
>60
1832 - 29486
39772
83-3681
7188
0-179
354
A Estimates for each study area were averaged across the 3-year assessment period. Ranges reflect the ranges of averages.
B An entry of <0.1 is used to represent small, non-zero values that do not round upwards to 0.1 (i.e., <0.05).
c A value of zero (0) means that there were no individuals estimated to have the selected exposure in any year.
April 2022
3-70 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
Lung function risk estimated for children and children with asthma in air quality
scenarios with design values just above and below the current standard are presented in detail in
Appendix 3D, section 3D.3.3. The patterns of the estimates are, as expected, higher for the 75
ppb air quality scenario and lower for the 65 ppb scenario. For each scenario, the differences in
risk estimates between the two models is similar to that which occurs with the risk estimates for
air quality just meeting the current standard (as discussed in section 3.4.2 above). These
estimates (for both lung function risk approaches) are less different from those for the current
standard air quality scenario than are differences noted above for the comparison-to-benchmarks
estimates. This is due to the greater influence on the risk results of exposures associated with the
low O3 concentrations that are less affected by air quality adjustments used to develop air
concentration surfaces for which the highest-concentration location has a design value just
meeting the different targets.
3.4.4 Key Uncertainties
In this section, we consider the uncertainties associated with the quantitative estimates of
exposure and risk, including those recognized by the characterization of uncertainty in Appendix
3D (section 3D.3.4). This characterization is based on an approach intended to identify and
compare the relative impact that important sources of uncertainty may have on the exposure and
risk estimates. The approach utilized is largely qualitative and is adapted from the World Health
Organization (WHO) approach for characterizing uncertainty in exposure assessment (WHO,
2008) augmented by several quantitative sensitivity analyses of key aspects of the assessment
approach (described in detail in Appendix 3D, section 3D.3.4). This characterization and
associated analyses build upon information generated from a previously conducted quantitative
sensitivity analysis of population-based O3 exposure modeling (Langstaff, 2007), considering the
various types of data, algorithms, and models that together yield exposure and risk estimates for
the eight study areas. In this way, we considered the limitations and uncertainties underlying
these data, algorithms and models and the extent of their influence on the resultant exposure/risk
estimates using the general approach applied in past risk and exposure assessments for O3,
nitrogen oxides, carbon monoxide and SOx (U.S. EPA, 2008; U.S. EPA, 2010; U.S. EPA, 2014;
U.S. EPA, 2018).
The exposure and risk uncertainty characterization and quantitative sensitivity analyses,
presented in Appendix 3D, section 3D.3.4, involve consideration of the various types of inputs
and approaches that together result in the exposure and risk estimates for the eight study areas. In
this way the limitations and uncertainties underlying these inputs and approaches and the extent
of their influence on the resultant exposure/risk estimates are considered. Consistent with the
WHO (2008) guidance, the overall impact of the uncertainty is scaled by considering the extent
April 2022
3-71 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
or magnitude of the impact of the uncertainty as implied by the relationship between the source
of the uncertainty and the exposure and risk output. The characterization in Appendix 3D also
evaluated the direction of influence, indicating how the source of uncertainty was judged, or
found, to quantitatively affect the exposure and risk estimates, e.g., likely to over- or under-
estimate (Appendix 3D, section 3D.3.4.1).
• What are the important uncertainties associated with the exposure and risk
estimates?
Based on the uncertainty characterization and associated analyses in Appendix 3D and
consideration of associated policy implications, we recognize several areas of uncertainty as
particularly important in our consideration of the exposure and risk estimates, while also
recognizing several areas where new or updated information reduced uncertainties in the
exposure and risk estimates compared to those in the 2015 review. In so doing, we note areas
that pertain to estimates for both types of risk metrics, as well as areas that pertain more to one
type of estimate versus the other. We also note differences in the uncertainties that pertain to
each of the two approaches used for the lung function risk metric.
An overarching and important area of uncertainty, remaining from the 2015 review and
important to our consideration of the exposure and risk analysis results, relates to the underlying
health effects evidence base. The quantitative analysis focuses on the evidence providing the
"strongest evidence" of O3 respiratory effects (ISA, p. IS-1), the controlled human exposure
studies, and on the array of respiratory responses documented in those studies (e.g., lung function
decrements, respiratory symptoms, increased airway responsiveness and inflammation). The
comparison-to-benchmarks analysis is particularly focused on consideration of the potential for
exposures that pose a risk of experiencing this array of effects. We note, however, evidence is
lacking from controlled human exposure studies of 6.6-hour duration at the lower concentrations
(e.g., 60, 70 and 80 ppb) for children and for people of any age with asthma. While the limited
evidence informing our understanding of potential risk to people with asthma is uncertain, it
indicates the potential for this group, given their disease status, to be at risk (e.g., of asthma
exacerbation), as summarized in section 3.3.4 above. Such a conclusion is consistent with the
epidemiological study findings of positive associations of O3 concentrations with asthma-related
emergency department visits and hospital admissions (and the higher effect estimates from these
studies), as referenced in section 3.3.1 above and presented in detail in the ISA. Thus, we
recognize uncertainty in interpretation of the exposure and risk estimates in the broader context
(e.g., as discussed in section 3.4.5 below).
Key uncertainties and limitations in data and tools that affect the quantitative estimates of
exposure and risk, particularly in their interpretation in the context of considering the current
standard, relate to each step in the assessment. These include uncertainty related to estimation of
April 2022
3-72 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
the concentrations in ambient air for the current standard and the additional air quality scenarios;
lung function risk approaches that rely, to varying extents, on extrapolating from controlled
human exposure study conditions to lower exposure concentrations, lower ventilation rates, and
shorter durations; and characterization of risk for particular population groups that may be at
greatest risk, particularly for people with asthma, and particularly children with asthma. Areas in
which uncertainty has been reduced by new or updated information or methods include the use
of updated air quality modeling, with a more recent model version and model inputs, applied to
study areas with design values near the current standard, as well as updates to several inputs to
the exposure model, including changes to the exposure duration to better match those in the
controlled human exposure studies and an alternate approach to characterizing periods of activity
while at moderate or greater exertion for simulated individuals.
With regard to the analysis approach overall, two updates since the 2014 HREA reduce
uncertainty in the results. The first relates to identifying when simulated individuals may be at
moderate or greater exertion, with the new approach reducing the potential for overestimation of
the number of people achieving the associated ventilation rate, which was an important
uncertainty in the 2014 HREA. Additionally, the current analysis focus on exposures of 7 hours
duration better represents the 6.6-hour exposures from the controlled human exposure studies
(than the 8-hour exposure durations used for the 2014 HREA and prior assessments).
Additional aspects of the analytical design pertaining to both exposure-based risk metrics
include the estimation of ambient air O3 concentrations for the air quality scenarios, and main
components of the exposure modeling. Uncertainties include the modeling approach used to
adjust ambient air concentrations to meet the air quality scenarios of interest and the method
used to interpolate monitor concentrations to census tracts. While the adjustment to conditions
near, just above, or just below the current standard is an important area of uncertainty, the size of
the adjustment needed to meet a given air quality scenario is minimized with the selection of
study areas for which recent O3 design values were near the level of the current standard. Also,
more recent data are used as inputs for the air quality modeling, such as more recent O3
concentration data (2015-2017), meteorological data (2016) and emissions data (2016), as well
as a recently updated air quality photochemical model which includes state-of-the-science
atmospheric chemistry and physics (Appendix 3C). Further, the number of ambient monitors
sited in each of the eight study areas provides a reasonable representation of spatial and temporal
variability for the air quality conditions simulated in those areas.
Among other key aspects, there is uncertainty associated with the simulation of study
area populations (and at-risk populations), including those with particular physical and personal
attributes. As also recognized in the 2014 HREA, exposures could be underestimated for some
population groups that are frequently and routinely outdoors during the summer (e.g., outdoor
April 2022
3-73 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
workers, children.90 In addition, longitudinal activity patterns do not exist for these and other
potentially important population groups (e.g., those having respiratory conditions other than
asthma), limiting the extent to which the exposure model outputs reflect information that may be
particular to these groups. Important uncertainties in the approach used to estimate energy
expenditure (i.e., metabolic equivalents of work or METs used to estimate ventilation rates),
include the use of longer-term average MET distributions to derive short-term estimates, along
with extrapolating adult observations to children. Both of these approaches are reasonable based
on the availability of relevant data and appropriate evaluations conducted to date, and
uncertainties associated with these steps are somewhat reduced in the current analyses (compared
to the 2014 HREA) because of the added specificity and use of redeveloped METs distributions
(based on newly available information), which is expected to more realistically estimate activity-
specific energy expenditure.
With regard to the exposure and risk modeling aspects of the two risk metrics, we
recognize that there are some uncertainties that apply to the estimation of lung function risk and
not the comparison-to-benchmarks analysis. For example, both lung function risk approaches
utilized in the risk analyses incorporate some degree of extrapolation beyond the exposure
circumstances evaluated in the controlled human exposure studies in recognition of the potential
for lung function decrements to be greater in unstudied population groups than is evident from
the available studies. For example, both models generate nonzero predictions for 7-hour
concentrations below the 6.6-hour concentrations investigated in the controlled human exposure
studies. In considering these risk estimates, we recognize that the uncertainty in the lung function
risk estimates increases with decreasing exposure concentration, and is particularly increased for
concentrations below those evaluated in controlled exposure studies (section 3.4.4 and Appendix
3D, section 3D.3.4). Further, the two lung function risk approaches differ in how they
extrapolate beyond the controlled human exposure study conditions and in the impact on the
estimates. The E-R function risk approach generates nonzero predictions from the full range of
potential nonzero concentrations for 7-hour average durations in which the average exertion
levels meets or exceeds the target. The MSS model, which draws on evidence-based concepts of
how human physiological processes respond to O3, extrapolates beyond the controlled
experimental conditions, with regard to exposure concentration, exposure duration, and also,
ventilation rate (both magnitude and duration). The impact of this extrapolation, and the
difference between the two models in its extent beyond the studied exposure circumstances, is
illustrated by differences in the percent of the risk estimates derived on days for which the
90 As described in section 3.4.1 above, the child populations modeled were school ages (ages 5 to 18), in recognition
of limitations and uncertainties in the data for children younger than five years.
April 2022
3-74 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
highest 7-hour average concentration is below the lowest 6.6-hour exposure concentration tested
(Table 3-6 and Table 3-7). For example, while 3 to 6% of the risk to children (based on single-
year estimates for three study areas) of experiencing at least one day with decrements greater
than 20% estimated by the E-R model is associated with exposure concentrations below 40 ppb
(the lowest exposure concentration studied, and at which no decrements of this severity occurred
in any study subjects), 25% to nearly 40% of MSS model estimates of decrements greater than
20% derive from exposures below 40 ppb (Table 3-6 and Table 3-7). Further, using ventilation
rates lower than those used for the E-R function risk approach (which are based on the controlled
human exposure study conditions) also contribute to relatively greater risks estimated by the
MSS model. Limiting the MSS model results to estimates for individuals with at least the same
exertion level achieved by study subjects (>17.3 L/min-m2), reduces the risks of experiencing at
least one lung function decrement by an amount between 24 to 42% (Appendix 3D, Table 3D-
69).
The difference between the two models for risk contribution from low concentrations is
smaller for risk estimates for two or more days than the estimates for one or more days. This is
largely because the percent contribution to low-concentration risk for two or more decrement
days predicted by the E-R approach is, by design, greater than the corresponding contribution to
low-concentration risk for one or more days.91 This also occurs because the MSS model
estimates risk from a larger variety of exposure and ventilation conditions (Table 3-6, Table 3-7).
Further, many of the uncertainties previously identified as part of the 2014 HREA unique to the
MSS model remain as important uncertainties in the current assessment. For example, the
extrapolation of the MSS model age parameter down to age 5 (from the age range of 18- to 35-
year old study subjects to which the model was fit) is an important uncertainty given that
children are an at-risk population of particular interest in this assessment. Also, there is
uncertainty in estimating the frequency and magnitude of lung function decrements as a result of
the statistical form and parameters used for the MSS model inter- and intra-individual variability
terms. Each of these, among other newly identified MSS model uncertainties, are evaluated and
discussed in the current uncertainty characterization (Appendix 3D, section 3D.3.4). As a whole,
the differences between the two lung function risk approaches described above and the estimates
generated by these approaches indicate appreciably greater uncertainty associated with the MSS
91 The E-R function approach uses the daily maximum exposure concentration for the simulated population. By
design, every individual would more than likely have a lower exposure on the second day than that experienced
on the first day, and so on for each progressive day throughout the simulation period. Therefore, if any risk is
estimated, the distribution of exposures would be shifted more so to lower concentrations for a greater proportion
of the population.
April 2022
3-75 External Review Draft - Do Not Quote or Cite

-------
1	model estimates than the E-R function estimates due to the significantly greater portion of
2	relatively low concentrations contributing to risk.
3	Table 3-6. Percent of risk estimated for air quality just meeting the current standard in
4	three study areas using the E-R function approach on days where the daily
5	maximum 7-hour average concentration is below specified values.
Size of
Lung
Function
Decrement
Percent of child population at risk of decrement from specific 7-hour concentrations A
Percent of one-or-more-days risk
Percent of two-or-more-days risk
< 30 ppb
< 40 ppb
< 50 ppb
< 60 ppb
< 30 ppb
< 40 ppb
< 50 ppb
< 60 ppb
> 20%
0.7-1%
3-6%
12-25%
39-70%
2-3%
7-12%
24-44%
67-93%
> 15%
2-3%
6-11%
19-34%
48-78%
4-5%
12-18%
34-54%
75-95%
> 10%
4-5%
11-16%
29-45%
61 -86%
7-9%
18-25%
45-63%
83-97%
A The ranges presented are based on 1-year simulations in three study areas (Atlanta, Dallas, and St Louis); the values
presented here are rounded to whole numbers or at least one significant digit (full results are in Appendix 3D, section
3D.3.4.2, Table 3D-62).
6	Table 3-7. Percent of risk estimated for air quality just meeting the current standard in
7	three study areas using the MSS model approach on days where the daily
8	maximum 7-hour average concentration is below specified values.
Size of
Lung
Function
Decrement
Percent of child population at risk of decrement from specified 7-hour concentrations A
Percent of one-or-more-days risk
Percent of two-or-more-days risk
< 30 ppb
< 40 ppb
< 50 ppb
< 60 ppb
< 30 ppb
< 40 ppb
< 50 ppb
< 60 ppb
> 20%
5-9%
25-38%
63-78%
88-96%
5-10%
28-42%
66-81%
90-98%
> 15%
11-18%
36-51%
72 - 84%
92-98%
11-19%
38-54%
74-87%
93-99%
> 10%
25-32%
57-67%
84-91%
96-99%
26-33%
57-68%
84-91%
96-99%
A The ranges presented are based on 1-year simulations in three study areas (Atlanta, Dallas, and St Louis); the values
presented here are rounded to whole numbers or at least one significant digit (full results are in Appendix 3D, section
3D.3.4.2, Table 3D-63).
9
10	An additional area in which uncertainty has been reduced for the exposure estimates is
11	related to the approach to identifying when simulated individuals may be at moderate or greater
12	exertion. The approach used in the current assessment reduces the potential for overestimation of
13	the number of people achieving the associated ventilation rate, an important uncertainty
14	identified in the 2014 HREA. We also note that the exposure duration in the assessment was a 7-
15	hour averaging time, which was selected to better represent the 6.6-hour exposures from the
16	controlled human exposure studies, compared to the 8-hour exposure durations used in the model
17	in the 2014 HREA and prior assessments.
18	In summary, among the multiple uncertainties and limitations in data and tools that affect
19	the quantitative estimates of exposure and risk and their interpretation in the context of
April 2022
3-76 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
considering the current standard, we recognize several here as particularly important, noting that
some of these uncertainties are similar to those recognized in the 2015 review. These include
uncertainty related to estimation of the concentrations in ambient air for the current standard and
the additional air quality scenarios; lung function risk approaches that rely, to varying extents, on
extrapolating from controlled human exposure study conditions to lower exposure
concentrations, lower ventilation rates, and shorter durations; and, characterization of risk for
particular population groups that may be at greatest risk, particularly for people with asthma,
particularly children. We also recognize several areas in which uncertainty has been reduced by
new or updated information or methods, including more refined air quality modeling based on
selection of study areas with design values near the current standard and more recent model
inputs, as well as updates to several inputs to the exposure model including changes to the
exposure duration to better match those in the controlled human exposure studies and an
alternate approach to characterizing periods of activity while moderate or greater exertion for
simulated individuals.
3.4.5 Public Health Implications
In considering public health implications of the quantitative exposure and risk estimates
that may inform the Administrator's judgments in this area, this section discusses the information
pertaining to the following question.
• To what extent are the estimates of exposures and risks to at-risk populations
associated with air quality conditions just meeting the current standard reasonably
judged important from a public health perspective?
Several factors are important to the consideration of public health implications. These
include the magnitude or severity of the effects associated with the estimated exposures, as well
as their adversity at the individual and population scale. Other important considerations include
the size of the population estimated to experience such effects or to experience exposures
associated with such effects. Thus, the discussion here reflects consideration of the health
evidence, and exposure and risk estimates, as well as the consideration of potential public health
implications in previous NAAQS decisions and ATS policy statements (as also discussed in
section 3.3.2).
In considering the severity of responses associated with the exposure and risk estimates,
we take note of the health effects evidence for the different benchmark concentrations and
judgments made with regard to the severity of these effects in the 2015 review. We recognize the
greater prevalence of more severe lung function decrements among study subjects exposed to 80
ppb or higher concentrations (compared to the study findings for lower exposure concentrations),
as well as the prevalence of other effects such as respiratory symptoms; thus, such exposures (of
April 2022
3-77 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
80 ppb and greater) are appropriately considered to be associated with adverse respiratory effects
consistent with past and recent ATS position statements and with EPA's judgments in
establishing the current standard in 2015.92 Further, in the controlled human exposure study of an
average exposure level somewhat above 70 ppb (73 ppb), statistically significant increases in
transient lung function decrements (specifically reduced FEVi) and respiratory symptoms have
been reported, leading EPA to also characterize these exposure conditions as being associated
with adverse responses, consistent with ATS statements as summarized in section 3.1 above
(e.g., 80 FR 65343, 65345, October 26, 2015; 85 FR 87304, December 31, 2020). Studies of
controlled human exposures to the lowest benchmark concentration of 60 ppb have found small
but statistically significant Cb-related decrements in lung function and airway inflammation
(without increased incidence of respiratory symptoms).
We additionally take note of the greater significance of estimates for multiple
occurrences of exposures at or above these benchmarks consistent with the evidence. This is
consistent with past O3 NAAQS reviews in which it was recognized, using the example of effects
such as inflammation, that while isolated occurrences can resolve entirely, repeated occurrences
from repeated exposure could potentially result in more severe effects (2013 ISA, section 6.2.3
and p. 6-76). The ascribing of greater significance to repeated occurrences of exposures of
potential concern is also consistent with public health judgments in NAAQS reviews for other
pollutants, such as SOx and carbon monoxide (84 FR 9900, March 18, 2019; 76 FR 54307,
August 31, 2011).
The exposure-based analyses include two types of metrics, one involving comparison-to-
benchmark concentrations corresponding to 6.6-hour exposure concentrations to which
exposures while at elevated ventilation have elicited lung function decrements, and the second
involving estimates of lung function risk with regard to such decrements of magnitudes at or
above 10%, 15% or 20%. Based on evidence base described in the 2020 ISA, which is largely
consistent with that available in the 2015 review (as summarized in section 3.3.1 above), the
quantitative exposure and risk analyses results in which we have the greatest confidence are
estimates from the comparison-to-benchmarks analysis, as discussed in section 3.4.4 above.
In light of the conclusions that people with asthma and children are at-risk populations
for 03-related health effects (summarized in section 3.3.2 above) and the exposure and risk
analysis findings of higher exposures and risks for children (in terms of percent of that
population), we have focused the discussion here on children, and specifically children with
92 The ATS statements indicate that consideration of differences in magnitude or severity, and also the relative
transience or persistence of the adverse responses (e.g., FEVi changes) and respiratory symptoms, as well as pre-
existing sensitivity to effects on the respiratory system, and other factors, is important to characterizing
implications for public health effects of an air pollutant such as O3 (ATS, 2000; Thurston et al., 2017).
April 2022
3-78 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
asthma. We recognize that the exposure and risk estimates indicate that in some areas of the U.S.
where O3 concentrations just meet the current standard, on average across the 3-year period
simulated, just over 0.5%, and less than 0.1% of the simulated population of children with
asthma might be expected to experience a single day per year with a 7-hour exposure at or above
70 ppb and 80 ppb, respectively, while breathing at an elevated rate. With regard to the lowest
benchmark considered (60 ppb), the corresponding percentage is just over 8%, with higher
percentages in some individual years (Table 3-6). The corresponding estimates for the air quality
scenario with higher O3 concentrations are notably higher (Table 3-5). For example, for the 75
ppb air quality scenario, 1.1% to 2.1% of children with asthma, on average across the 3-year
design period, are estimated to experience at least one day with exposure concentrations at or
above 70 ppb, while at moderate or greater exertion, with as many as 3.9% in a single year
(Table 3-5). For the 60 ppb benchmark, the single-day occurrence estimates for the 75 ppb
scenario range up to nearly 16%. Estimates for the 65 ppb scenario are appreciably lower.
With regard to estimates of lung function decrements, we focus on the E-R model
estimates as having less associated uncertainty, as discussed in section 3.4.4 above. The exposure
and risk analysis estimates 0.2 to 0.3% of children with asthma, on average across the 3-year
design period to experience one or more days with a lung function decrement at or above 20%,
and 0.5 to 0.9 % to experience one or more days with a decrement at or above 15% (Table 3-4
above). In a single year, the highest estimate is 1.0% of this at-risk population expected to
experience one or more days with a decrement at or above 15%. The corresponding estimate for
two or more days is 0.6% (Table 3-4 above). As discussed in section 3.4.3 above, the estimates
for the 75 ppb air quality scenario are notably higher, while the estimates for the 65 ppb scenario
are notably lower (Table 3-5). In reviewing the lung function risk estimates, we note the
uncertainties discussed in section 3.4.4 above, including the appreciable portion of these
estimates that are based on quantifying risk for exposure concentrations below those studied.
The size of the at-risk population (people with asthma, particularly children) in the U.S.
is substantial. As summarized in section 3.3.2, nearly 8% of the total U.S. population93 and 7.0%
of U.S. children have asthma. The asthma prevalence in U.S. child populations (younger than 18
years) of different races or ethnicities ranges from 7.5% for all Hispanic children to 13.5% for
black non-Hispanic children (Table 3-1 above). This is well reflected in the exposure and risk
analysis study areas in which the asthma prevalence ranged from 7.7% to 11.2% of the total
populations and 9.2% to 12.3% of the children. In each study area, the prevalence varies among
93 The number of people in the US with asthma is estimated to be about 25 million. As shown in Table 3-1 the
estimated number of people with asthma was 25,131,000 in 2019.
April 2022
3-79 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
census tracts, with the highest tract having a prevalence in boys of 25.5% and a prevalence in
girls of 17.1% (Appendix 3D, Table 3D-3).
The exposure and risk analyses inherently recognize that variability in human activity
patterns (where people go and what they do) is key to understanding the magnitude, duration,
pattern, and frequency of population exposures. For O3 in particular, the amount and frequency
of afternoon time outdoors at moderate or greater exertion is an important factor for
understanding the fraction of the population that might experience O3 exposures that have
elicited respiratory effects in controlled human exposure studies (2014 HREA, section 5.4.2). In
considering the available information regarding prevalence of behavior (time outdoors and
exertion levels) and daily temporal pattern of O3 concentrations, we take note of the findings of
evaluations of the data in the CHAD. Based on these evaluations of human activity pattern data,
it appears that children and adults both, on average, spend about 2 hours of afternoon time
outdoors per day, but differ substantially in their participation in these events at elevated exertion
levels (rates of about 80% versus 60%, respectively) (2014 HREA, section 5.4.1.5), indicating
children are more likely to experience exposures that may be of concern. This is one basis for
their identification as an at-risk population for Cb-related health effects. The human activity
pattern evaluations have also shown there is little to no difference in the amount or frequency of
afternoon time outdoors at moderate or greater exertion for people with asthma compared with
those who do not have asthma (2014 HREA, section 5.4.1.5). Further, recent CHAD analyses
indicate that while 46 - 73% of people do not spend any afternoon time outdoors at moderate or
greater exertion, a fraction of the population (i.e., between 5.5 - 6.8% of children) spend more
than 4 hours per day outdoors at moderate or greater exertion and may have greater potential to
experience exposure events of concern than adults (Appendix 3D, section 3D.2.5.3 and Figure
3D-9). It is this potential that contributes importance to consideration of the exposure and risk
estimates.
In considering the public health implications of the exposure and risk estimates across the
eight study areas, we note the purpose for the study areas is to illustrate exposure circumstances
that may occur in areas that just meet the current standard, and not to estimate exposure and risk
associated with conditions occurring in those specific locations today. To the extent that
concentrations in the specific areas simulated may differ from others across the U.S., the
exposure and risk estimates for these areas are informative to consideration of potential
exposures and risks in areas existing across the U.S. that have air quality and population
characteristics similar to the study areas assessed, and that have ambient concentrations of O3
that just meet the current standard today or that will be reduced to do so at some period in the
future. We note that numerous areas across the U.S. have air quality for O3 that is near or above
April 2022
3-80 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
the existing standard.94 Thus, the air quality and exposure circumstances assessed in the eight
study areas are of particular importance in considering whether the available information calls
into question the adequacy of public health protection afforded by the current standard.
The exposure and risk estimates for the eight study areas reflect differences in exposure
circumstances among those areas and illustrate the exposures and risks that might be expected to
occur in other areas with such circumstances under air quality conditions that just meet the
current standard (or the alternate conditions assessed). Thus, the exposure and risk estimates
indicate the magnitude of exposure and risk that might be expected in many areas of the U.S.
with O3 concentrations at or near the current standard. Although the methodologies and data used
to estimate population exposure and lung function risk in this assessment differ in several ways
from what was used in the 2015 review, the findings and considerations summarized here present
a pattern of exposure and risk that is generally similar to that considered in the last review (as
described in section 3.4.2 above), and indicate a level of protection generally consistent with that
described in the 2015 decision.
In summary, the considerations raised here are important to conclusions regarding the
public health significance of the exposure and risk results. We recognize that such conclusions
also depend in part on public health policy judgments that weigh in the Administrator's decision
regarding the protection afforded by the current standard. Such judgments that are common to
NAAQS decisions include those related to public health implications of effects of differing
severity (75 FR 355260 and 35536, June 22, 2010; 76 FR 54308, August 31, 2011; 80 FR 65292,
October 26, 2015). Such judgments also include those concerning the public health significance
of effects at exposures for which evidence is limited or lacking, as discussed in section 3.4.4
above, such as effects at the lower benchmark concentrations considered and lung function risk
estimates associated with exposure concentrations lower than those tested or for population
groups not included in the controlled exposure studies.
3.5 KEY CONSIDERATIONS REGARDING THE CURRENT PRIMARY
STANDARD
In considering what the available evidence and exposure/risk information indicate with
regard to the current primary O3 standard, the overarching question we consider is:
94 Based on data from 2016-2018, 142 counties have 03 concentrations that exceed the current standard. Population
size in these counties ranges from approximately 20,000 to more than ten million, with a total population of over
112 million living in counties that exceed the current standard. Air quality data are from Table 4. Monitor Status
in the Excel file labeled ozone_designvalues_20162018_final_06_28_19.xlsx downloaded from
https://www.epa.gov/air-trends/air-quality-design-values. Population sizes are based on 2017 estimates from the
U.S. Census Bureau (https://www.census.gov/programs-surveys/popest.html).
April 2022
3-81 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
•	Does the available scientific evidence- and exposure/risk-based information support
or call into question the adequacy of the protection afforded by the current primary
O3 standard?
To assist us in interpreting the available scientific evidence and the results of recent
quantitative exposure/risk analyses to address this question, we have focused on a series of more
specific questions, as detailed in sections 3.5.1 and 3.5.2 below. In considering the scientific and
technical information, we take into account the information available at the time of the 2015
review and information newly available in the 2020 review, which have been critically analyzed
and characterized in the 2013 ISA for the 2015 review and the ISA for the 2020 review,
respectively. In this context, a primary consideration is whether the available information alters
overall prior conclusions regarding health effects associated with photochemical oxidants,
including O3, in ambient air.
3.5.1 Evidence-based Considerations
In considering the evidence with regard to the overarching question posed above
regarding the adequacy of the current standard, we address a series of more specific questions
that focus on policy-relevant aspects of the evidence. These questions begin with consideration
of the available evidence on health effects associated with exposure to photochemical oxidants,
and particularly O3.
•	Is there evidence that indicates the importance of photochemical oxidants other than
O3 with regard to abundance in ambient air, and potential for human exposures and
health effects?
The 2020 ISA did not identify any newly available evidence regarding the importance of
photochemical oxidants other than O3 with regard to abundance in ambient air, and potential for
health effects.95 As summarized in section 2.1 above, O3 is one of a group of photochemical
oxidants formed by atmospheric photochemical reactions of hydrocarbons with nitrogen oxides
in the presence of sunlight, with O3 being the only photochemical oxidant other than nitrogen
dioxide that is routinely monitored in ambient air. Data for other photochemical oxidants are
generally derived from a few special field studies such that national scale data for these other
oxidants are scarce (ISA, Appendix 1, section 1.1; 2013 ISA, sections 3.1 and 3.6). Moreover,
few studies of the health impacts of other photochemical oxidants beyond O3 have been
identified by literature searches conducted for other recent O3 assessments (ISA, Appendix 1,
section 1.1). As stated in the ISA, "the primary literature evaluating the health... effects of
95 Close agreement between past O3 measurements and the photochemical oxidant measurements upon which the
early photochemical oxidants NAAQS was based indicated the very minor contribution of other oxidant species
in comparison to O3 (U.S. DHEW, 1970).
April 2022
3-82 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
photochemical oxidants includes ozone almost exclusively as an indicator of photochemical
oxidants" (ISA, section IS. 1.1, p. IS-3). Thus, the evidence base for health effects of
photochemical oxidants does not indicate an importance of any other photochemical oxidants.
For these reasons, discussion of photochemical oxidants in this document focuses on O3.
•	Does the available scientific evidence alter prior conclusions regarding the nature of
health effects attributable to human exposure to O3 from ambient air?
The evidence, as evaluated in the 2020 ISA, is largely consistent with the conclusion in
the last ISA (in the 2015 review) regarding the health effects causally related to O3 exposures,
and most specifically regarding respiratory effects, which, as in the past, are concluded to be
causally related to short-term exposures to O3. Also, as in the 2015 and prior reviews, respiratory
effects are concluded to be likely causally related to longer-term O3 exposures (ISA, section
IS. 1.3.1, Appendix 3). Further, while a causal determination was not made in the 2015 review
regarding metabolic effects, the 2020 ISA finds there to be sufficient evidence to conclude there
to likely be a causal relationship of short-term O3 exposures and metabolic effects and finds the
evidence to be suggestive of, but not sufficient to infer, such a relationship between long-term O3
exposure and metabolic effects (ISA, section IS. 1.3.1). This is based on more recently available
evidence, largely from experimental animal studies, on these effects (ISA, Appendix 5).
Additionally, the EPA's causal determinations regarding cardiovascular effects and mortality
have been updated from what they were in 2013 ISA based on more recently available evidence
in combination with uncertainties that had been identified in the previously available evidence
(ISA, Appendix 4, section 4.1.17 and Appendix 6, section 6.1.8). The EPA has concluded that
the evidence base is suggestive of, but not sufficient to infer, causal relationships between O3
exposures (short- and long-term) and cardiovascular effects, mortality, reproductive and
developmental effects, and nervous system effects (ISA, section IS. 1.3.1). As in the 2015 and
prior O3 NAAQS reviews, the strongest evidence, including with regard to characterization of
relationships between O3 exposure and occurrence and magnitude of effects, is for respiratory
effects, and particularly for effects such as lung function decrements, respiratory symptoms,
airway responsiveness, and respiratory inflammation.
•	Does the available evidence alter our prior understanding of populations that are
particularly at risk from O3 exposures?
The evidence, as evaluated in the 2020 ISA, does not alter our prior understanding of
populations at risk from health effects of O3 exposures. As in the past, people with asthma, and
particularly children, are the at-risk population groups for which the evidence is strongest. In
addition to populations with asthma, groups with relatively greater exposures, particularly those
who spend more time outdoors during times when ambient air concentrations of O3 are highest
April 2022
3-83 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
and while engaged in activities that result in elevated ventilation, are recognized as at increased
risk. Such groups include outdoor workers and children. Other groups for which the evidence is
less clear include older adults, individuals with reduced intake of certain nutrients and
individuals with certain genetic variants. Recent evidence does not provide additional
information for these groups beyond the evidence available at the time of the 2015 review (ISA,
section IS.4.4).
• Does the available evidence alter past conclusions regarding the exposure duration
and concentrations associated with health effects? To what extent does the scientific
evidence indicate health effects attributable to exposures to O3 concentrations lower
than previously reported and what are important uncertainties in that evidence?
The available evidence documented in the 2020 ISA regarding O3 exposures associated
with health effects is largely similar to that available at the time of the 2015 review and does not
indicate effects attributable to exposures of shorter duration or lower concentrations than
previously understood. Respiratory effects continue to be the effects for which the experimental
information regarding exposure concentrations eliciting effects is well established, as
summarized in section 3.3.3 above. Such information allows for characterization of potential
population risk associated with O3 in ambient air under conditions allowed by the current
standard. The more recently available controlled human exposure studies, as discussed in section
3.3.3 above, are conducted over shorter durations while at much higher concentrations than the
key set of 6.6-hour studies that have been the focus of the last several reviews. The respiratory
effects evidence includes support from a large number of epidemiologic studies. The positive
associations of O3 with respiratory health outcomes (e.g., asthma-related hospital admissions and
emergency department visits) reported in these studies are coherent with findings from the
controlled human exposure and experimental animal studies. All but a few of these studies,
however, are conducted in areas during periods when the current standard is not met, making
them less useful with regard to indication of health effects of exposures allowed by the current
standard.
Within the evidence base for the recently identified category of metabolic effects, the
evidence derives largely from experimental animal studies of exposures appreciably higher than
those for the 6.6-hour human exposure studies along with a small number of epidemiologic
studies. As discussed in section 3.3.3 above, these studies do not prove to be informative to our
consideration of exposure circumstances likely to elicit health effects.
Thus, the 6.6-hour controlled human exposure studies of respiratory effects remain the
focus for our consideration of exposure circumstances associated with O3 health effects. Based
on these studies, the exposure concentrations investigated range from as low as approximately 40
ppb to 120 ppb. This information on concentrations that have been found to elicit effects for 6.6-
April 2022
3-84 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
hour exposures while exercising is unchanged from what was available in the 2015 review. The
lowest concentration for which lung function decrements have been found to be statistically
significantly increased over responses to filtered air remains approximately 60 ppb, at which
group mean decrements on the order of 2% to 4% have been reported (Table 3-2, Figure 3-2).
Respiratory symptoms were not increased with this exposure level.96 Exposure to concentrations
slightly above 70 ppb, with quasi-continuous exercise, has been reported to elicit statistically
significant increases in both lung function decrements and respiratory symptom scores, as
summarized in section 3.3.3 above. Still greater group mean and individual responses in lung
function decrements and respiratory symptom scores, as well as inflammatory response and
airway responsiveness, are reported for higher exposure concentrations.
• To what extent have previously identified uncertainties in the health effects evidence
been reduced or do important uncertainties remain?
Uncertainties identified in the health effects evidence at the time of the 2015 review
generally remain. These include uncertainties related to the susceptibility of population groups
not studied, the potential for effects to result from exposures to concentrations below those
included in controlled human exposure studies, and the potential for increased susceptibility as a
result of prior exposures. We additionally recognize uncertainties associated with the
epidemiologic studies (e.g., the potential for copollutant confounding and exposure measurement
error). In this context, however, we note the appreciably greater strength in the epidemiologic
evidence in its support for determination of a causal relationship for respiratory effects than the
epidemiologic evidence related to other categories, such as metabolic effects, more recently
determined to have a likely causal relationship with short-term O3 exposures (as summarized in
section 3.3.1 above).
3.5.2 Exposure/risk-based Considerations
Our consideration of the scientific evidence is informed by results from a quantitative
analysis of estimated population exposure and associated risk, as at the time of the 2015 review.
The overarching consideration in this section is whether the current exposure/risk information
alters overall conclusions of the 2015 review regarding health risk associated with exposure to
O3 in ambient air. As in our consideration of the evidence in section 3.3.1 above, we have
focused the discussion regarding the exposure/risk information around key questions to assist us
in considering the exposure/risk analyses of at-risk populations living in a set of urban areas
96 A statistically significant increase in sputum neutrophils (a marker of increased airway inflammation) was
observed in one controlled human exposure study following 6.6-hour exposures to 60 ppb (Table3-2, Figure 3-2;
Appendix 3 A).
April 2022
3-85 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
under air quality conditions simulated to just meet the existing primary O3 standard. These
questions are as follows.
• To what extent are the estimates of exposures and risks to at-risk populations
associated with air quality conditions just meeting the current standard reasonably
judged important from a public health perspective? What are the important
uncertainties associated with any exposure/risk estimates?
The exposure and risk analyses conducted for the 2020 review, as described in section
3.4, provide exposure and risk estimates associated with air quality that might occur in an area
under conditions that just meet the current standard. These estimates illustrate the differences
likely to occur across various locations with such air quality as a result of area-specific
differences in emissions, meteorological and population characteristics. In understanding these
results, we note that the eight study areas provide a variety of circumstances with regard to
population exposure to concentrations of O3 in ambient air. These study areas reflect different
combinations of different types of sources of O3 precursor emissions, and also illustrate different
patterns of exposure to O3 concentrations in a populated area in the U.S. (Appendix 3C, section
3C.2). In this way, the eight areas provide a variety of examples of exposure patterns that can be
informative to the EPA's consideration of potential exposures and risks that may be associated
with air quality conditions occurring under the current O3 standard. While the same conceptual
air quality scenario is simulated in all eight study areas (i.e., conditions that just meet the existing
standard), variability in emissions patterns of O3 precursors, meteorological conditions, and
population characteristics in the study areas contribute to variability in the estimated magnitude
of exposure and associated risk across study areas.
In considering the exposure and risk results, we focus first on estimates for the eight
study areas from the comparison-to-benchmarks analysis, the results in which we have the
greatest confidence, as discussed in section 3.4.4 above. These results for urban areas with air
quality that just meets the current standard indicate that up to 0.7% of children with asthma, on
average across the 3-year period, and up to 1.0% in a single year might be expected to
experience, while at elevated exertion, at least one day with a 7-hour average O3 exposure
concentration at or above 70 ppb (Table 3-3). As noted earlier, this benchmark concentration
reflects the finding of statistically significant 03-related decrements and increased respiratory
symptoms in a controlled human exposure study of individuals at elevated exertion. Less than
0.1% of this population group is estimated to have multiple days with an occurrence of this
exposure level (Table 3-3). For the benchmark concentration of 80 ppb (which reflects the
potential for more severe effects), a much lower percentage of children with asthma, <0.1% on
average across the 3-year period, with 0.1% in the highest single year, might be expected to
experience, while at elevated exertion, at least one day with such a concentration (Table 3-3).
April 2022
3-86 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
There are no children with asthma estimated to experience more than a single day per year with a
7-hour average O3 concentration at or above 80 ppb (Table 3-3). With regard to the lowest
benchmark concentration of 60 ppb, 8.8% of children with asthma, on average across the 3-year
period, might be expected to experience one or more days with a 7-hour average O3 exposure
concentration at or above 60 ppb (the concentration associated with less severe effects), and just
over 11% in the highest single year (Table 3-3). Regarding multiple day occurrences, the
percentages for more than a single day occurrence are 3%, on average across the three years, and
just below 5% in the highest single year period (Table 3-3).
The estimates for the additional air quality scenarios differ as would be expected. For the
75 ppb air quality scenario, the percent of children with asthma that might be expected to
experience at least one day with a 7-hour average O3 exposure concentration, while at elevated
exertion, at or above 70 ppb, is a factor of three or more higher than for the current standard
(Table 3-3, Table 3-5). The corresponding estimates for multiple days are a factor of four or
more higher than those for air quality just meeting the current standard. By comparison,
corresponding estimates for the 65 ppb scenario are approximately a third those for the current
standard scenario, with a correspondingly smaller incremental difference in absolute number of
children (Table 3-3, Table 3-5). With regard to the 80 ppb benchmark, the difference of the 75
ppb scenario from the current standard is a factor of three (for average across the 3-year period)
to six (for the highest in a single year) (Table 3-3, Table 3-5). In contrast, the estimates for the 80
ppb benchmark (which is associated with the more severe effects) in the 65 ppb air quality
scenario are nearly identical to those for the current standard (Table 3-3, Table 3-5).
With regard to the estimates of lung function risk, as an initial matter we note the
uncertainty associated with these estimates, as discussed in section 3.4.4 above. In this context,
we also recognize the lesser uncertainty associated with estimates derived using the E-R function
(in comparison to estimates based on MSS model). Accordingly, it is those estimates which we
consider here for air quality conditions just meeting the current standard. The E-R lung function
risk analysis for the eight study areas indicates that the percent of children with asthma in an
urban area that just meets the current standard that might be expected to experience one or more
days with a lung function decrement of at least 15% or 20% might range up 0.9% or 0.3%,
respectively, on average across the three years, and 1.0% or 0.4%, respectively, in a single-year
period (Table 3-4). The estimates for a day with a decrement of at least 10% might range up to
3.3%), on average across the three years, and just over 3.5% in a single-year period (Table 3-4).
With regard to multiple day occurrences, the percent of children with asthma that might be
expected to experience two or more days with a lung function decrement of at least 10% may be
as high as 2.4%, on average across the three years, and 2.6% in a single year (Table 3-4), with
much smaller percentages for larger decrements. For multiple days with a decrement of at least
April 2022
3-87 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
15% or 20%, the corresponding percentages are much lower, 0.6% or 0.2%, respectively, on
average across the three years, and 0.6% or 0.2%, respectively, in a single year period (Table 3-
4).
We also consider the estimates from this assessment in light of the estimates from the
2014 HREA that were a focus of the decision on the standard in 2015. The estimates across all
study areas from this assessment are generally similar to those reported in the 2015 review across
all study areas included in that HREA, particularly for the two or more occurrences and for the
80 ppb benchmark (Table 3-8).97 In our consideration here, we focus on the full array of study
areas (e.g., rather than limiting to areas common to the two assessments) given the purpose of the
assessments in providing estimates across a range of study areas to inform decision making with
regard to the exposures and risks that may occur across the U.S. in areas that just meet the
current standard. We note only slight differences, particularly for the lower benchmarks, and
most particularly in the estimates for the highest year. For example, for the 70 ppb benchmark,
the lower and higher end of the range of average per year percent of children with at least one
day above the benchmark from the 2014 HREA are both twice the corresponding values from the
current assessment (Table 3-8). Consideration of the percentage of children estimated to
experience a day or more with an exposure at or above 70 ppb across the three air quality
conditions in the two assessments, however, indicates that differences between air quality
scenarios in the current assessment remain appreciably larger than the slight differences in
estimates between the two assessments for a given scenario. The factors likely contributing to the
slight differences between the two assessments, such as for the lowest benchmark, include
greater variation in ambient air concentrations in some of the study areas in the 2014 HREA, as
well as the lesser air quality adjustments required in study areas for the current assessment due to
closer proximity of conditions to meeting the current standard (70 ppb).98 Other important
differences between the two assessments are the updates made to the ventilation rates used for
identifying when a simulated individual is at moderate or greater exertion and the use of 7 hours
for the exposure duration. Both of these changes were made to provide closer linkages to the
conditions of the controlled human exposure studies which are the basis for the benchmark
concentrations. Thus, we recognize there to be reduced uncertainty associated with the current
estimates. Overall, particularly in light of differences in the assessments, we conclude the current
97	For consistency with the estimates highlighted in the 2015 review, Table 3-8 focuses on the simulated population
of all children (versus the simulated population for children with asthma that are the focus in section 3.4).
98	The 2014 HREA air quality scenarios involved adjusting 2006-2010 ambient air concentrations, and some study
areas had design values in that time period that were well above the then-existing standard (and more so for the
current standard). Study areas included the current exposure analysis had 2015-2017 design values close to the
current standard, requiring less of an adjustment for the current standard (70 ppb) air quality scenario.
April 2022
3-88 External Review Draft - Do Not Quote or Cite

-------
1	estimates to be generally similar to those which were the focus in the 2015 decision on
2	establishing the current standard.
3
4	Table 3-8. Comparison of current assessment and 2014 HREA (all study areas) for
5	percent of children estimated to experience at least one, or two, days with an
6	exposure at or above benchmarks while at moderate or greater exertion.
Air Quality
Scenario
(DVฐ, ppb)
Estimated average % of simulated children
with at least one day per year
at or above benchmark
(highest in single season)
Estimated average % of simulated children
with at least two days per year
at or above benchmark
(highest in single season)
Current PA A
2014 HREA B
Current PA A
2014 HREA B
Benchmark Exposure Concentration of 80 ppb
75
<0.1A-0.3 (0.6)
0-0.3(1.1)
0 - <0.1 (<0.1)
0(0.1)
70
0 - <0.1 (0.1)
0-0.1 (0.2)
0(0)
0(0)
65
0 - <0.1 (<0.1)
0(0)
0(0)
0(0)
Benchmark Exposure Concentration of 70 ppb
75
1.1-2.0 (3.4)
0.6-3.3(8.1)
0.1-0.3(0.7)
0.1-0.6 (2.2)
70
0.2-0.6 (0.9)
0.1-1.2(3.2)
<0.1 (0.1)
0-0.1 (0.4)
65
0-0.2(0.2)
0 - 0.2 (0.5)
0 - <0.1 (<0.1)
0(0)
Benchmark Exposure Concentration of 60 ppb
75
6.6-15.7(17.9)
9.5-17.0(25.8)
1.7-8.0 (9.9)
3.1 -7.6 (14.4)
70
3.2-8.2 (10.6)
3.3-10.2 (18.9)
0.6-2.9 (4.3)
0.5-3.5(9.2)
65
0.4-2.3(3.7)
0 - 4.2 (9.5)
<0.1-0.3(0.5)
0-0.8(2.8)
A For the current analysis, calculated percent is rounded to the nearest tenth decimal using conventional rounding. Values
equal to zero are designated by "0" (there are no individuals exposed at that level). Small, non-zero values that do not round
upwards to 0.1 (i.e., <0.05) are given a value of "<0.1"
B For the 2014 HREA. calculated percent was rounded to the nearest tenth decimal using conventional rounding. Values that
did not round upwards to 0.1 (i.e., <0.05) were given a value of "0".
c The monitor location with the highest concentrations in each area had a design value just equal to the indicated value.
7
8	3.5.3 Preliminary Conclusions on the Primary Standard
9	This section describes our preliminary conclusions for the Administrator's consideration
10	with regard to the current primary O3 standard. These preliminary conclusions are based on
11	considerations described in the sections above, and in the discussion below regarding the
12	available scientific evidence (as summarized in the 2020 ISA, and the ISA and AQCDs from
13	prior reviews), and the risk and exposure information developed in the 2020 review and
14	summarized in section 3.4 above. Taking into consideration the discussions above in this chapter,
15	this section addresses the following overarching policy question.
April 2022
3-89 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
• Do the available scientific evidence- and exposure/risk-based information support or
call into question the adequacy of the protection afforded by the current primary O3
standard?
In considering this question, we recognize that, as is the case in NAAQS reviews in
general, the extent to which the protection provided by the current primary O3 standard is judged
to be adequate will depend on a variety of factors, including science policy judgments and public
health policy judgments. These factors include public health policy judgments concerning the
appropriate benchmark concentrations on which to place weight, as well as judgments on the
public health significance of the effects that have been observed at the exposures evaluated in the
health effects evidence. The factors relevant to judging the adequacy of the standards also
include the interpretation of, and decisions as to the weight to place on, different aspects of the
results of the quantitative exposure risk analyses and the associated uncertainties. Thus, we
recognize that the Administrator's conclusions regarding the adequacy of the current standard
will depend in part on public health policy judgments, science policy judgments, including those
regarding aspects of the evidence and exposure/risk estimates, and judgments about the degree of
protection that is requisite to protect public health with an adequate margin of safety.
Our response to the overarching question above takes into consideration the discussions
that address the specific policy-relevant questions in prior sections of this document (see section
3.2) and builds on the approach from previous reviews. We focus first on consideration of the
evidence, including that newly available in the 2020 ISA, including the extent to which it alters
prior key conclusions supporting the current standard. We then turn to consideration of the
quantitative exposure and risk estimates developed for the 2020 review, including associated
limitations and uncertainties. We consider what they indicate regarding the level of protection
from adverse effects provided by the current standard, as well as the extent to which
exposure/risk estimates may indicate differing conclusions regarding air quality conditions
associated with the current standard from those based on past assessments. We additionally
consider the key aspects of the evidence and exposure/risk estimates emphasized in establishing
the current standard, and the associated public health policy judgments and judgments about the
uncertainties inherent in the scientific evidence and quantitative analyses that are integral to
decisions on the adequacy of the current primary O3 standard.
As an initial matter, we recognize the continued support in the available evidence for O3
as the indicator for photochemical oxidants, as recognized in section 3.5.1 above. Of the
photochemical oxidants, O3 is the only one other than nitrogen dioxide (for which there are
separate NAAQS) that is routinely monitored in ambient air. Further, as stated in the ISA, "the
primary literature evaluating the health and ecological effects of photochemical oxidants
includes ozone almost exclusively as an indicator of photochemical oxidants" (ISA, section
April 2022
3-90 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
IS. 1.1, p. IS-3). In summary, the evidence base for health effects of photochemical oxidants does
not indicate an importance of any other photochemical oxidants as it includes O3 almost
exclusively as an indicator of photochemical oxidants, thus continuing to support the
appropriateness of O3 as the indicator for photochemical oxidants.
In considering the extensive evidence base for health effects of O3, we give particular
attention to the longstanding evidence of respiratory effects causally related to O3 exposures.
This array of effects, and the underlying evidence base, was integral to the basis for setting the
current standard in 2015. As summarized in section 3.3.1 above and addressed in detail in the
ISA, the available evidence base does not include new evidence of respiratory effects associated
with appreciably different exposure circumstances, including any that would be expected to
occur under air quality conditions associated with the current standard. Thus, in considering the
information available at this time, we continue to focus on exposure circumstances associated
with the current standard as those of importance in this reconsideration.
Further, while the evidence base has been augmented somewhat since the 2015 review,
we note that the newly available evidence does not lead to different conclusions regarding the
respiratory effects of O3 in ambient air or regarding exposure concentrations associated with
those effects; nor does it identify different populations at risk of 03-related effects. For example,
as in the 2015 review, people of all ages with asthma, children, and outdoor workers, are
populations at increased risk of respiratory effects related to O3 in ambient air. Children with
asthma, which number approximately five million in the U.S., may be particularly at risk (section
3.3.2 and Table 3.1)." In these ways, the health effects evidence is consistent with evidence
available in the 2015 review when the current standard was established. This strong evidence
base continues to demonstrate a causal relationship between short-term O3 exposures and
respiratory effects, including in people with asthma. This conclusion is primarily based on
evidence from controlled human exposure studies that was available at the time the standard was
set that reported lung function decrements and respiratory symptoms in people exposed to O3 for
6.6 hours during which they engage in five hours of exercise. Support is also provided by the
experimental animal and epidemiologic evidence that is coherent with the controlled exposure
studies. The epidemiologic evidence, including that recently available, includes studies reporting
positive associations for asthma-related hospital admissions and emergency department visits,
which are strongest for children, with short-term O3 exposures. Based collectively on this
evidence, populations identified as at risk of such effects include people with asthma and
children.
99 The size of the U.S. population with asthma is approximately 25 million.
April 2022
3-91 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
As in the 2015 review, the most certain evidence of health effects in humans elicited by
exposures to specific O3 exposure concentrations is provided by controlled human exposure
studies. This category of short-term studies includes an extensive evidence base of 1- to 3-hour
studies, conducted with continuous or intermittent exercise and generally involving relatively
higher exposure concentrations (e.g., greater than 120 ppb).100 Given the lack of ambient air
concentrations of this magnitude in areas meeting the current standard (see section 2.4.1 above
and Appendix 2A), we continue to focus primarily on a second group of somewhat longer-
duration studies of much lower exposure concentrations. These studies employ a 6.6-hour
protocol that includes six 50-minute periods of exercise at moderate or greater exertion. There
are no new such studies with exercise available since the 2015 review. Thus, the newly available
evidence does not extend our understanding of the range of exposure concentrations that elicit
effects in such studies beyond what was understood previously.
Similarly, as in the 2015 review, 60 ppb remains the lowest exposure concentration
(target concentration, as average across exercise periods) at which statistically significant lung
function decrements have been reported in the 6.6-hour exposure studies. Two studies have
assessed exposure concentrations at the lower concentration of 40 ppb, with no statistically
significant finding of Cb-related FEVi decrements for the group mean in either study (which is
just above 1% in one study, and well below in the second). At 60 ppb, the group mean Cb-related
decrement in FEVi ranges from approximately 2 to 4%, with associated individual study subject
variability in decrement size. In the single study assessing the next highest exposure
concentration (just above 70 ppb, at 73 ppb),101 the group mean FEVi decrement (6%) was also
statistically significant, as were respiratory symptom scores. At higher exposure concentrations,
the incidence of both respiratory symptom scores and Cb-related lung function decrements in the
study subjects is increased. Other respiratory effects, such as inflammatory response and airway
resistance are also increased at higher exposures (ISA; 2013 ISA; 2006 AQCD).
In considering what may be indicated by the epidemiologic evidence with regard to
exposure concentrations eliciting effects, we recognize that of the numerous epidemiologic
studies of respiratory outcome associations with O3 in ambient air, none were conducted in U.S.
100	Table 3A-3 in Appendix 3 summarizes controlled human exposures to O3 for 1 to 2 hours during continuous or
intermittent exercise in contrast to similar exposure durations at rest. This table was adapted from Table 7-1 in the
1996 AQCD and Table AX6-1 in the 2006 CD, with additional studies from Table AX6-13 in the 2006 AQCD, as
well as more recent studies from the 2013 ISA and the ISA.
101	As noted in sections 3.1.1 and 3.3.3 above, the 70 ppb target exposure comes from Schelegle et al. (2009). That
study reported, based on O3 measurements during the six 50-minute exercise periods, that the mean O3
concentration during the exercise portion of the study protocol was 72 ppb. Based on the measurements for the
six exercise periods, the time weighted average concentration across the full 6.6-hour exposure was 73 ppb
(Schelegle et al., 2009).
April 2022
3-92 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
locations during time periods when the current standard was met. In fact, the vast majority of
these studies were conducted in locations and during time periods that would not have met the
current standard, thus making them less useful for considering the potential for O3 concentrations
allowed by the current standard to contribute to health effects. While there were a handful of
multi-city studies in which the O3 concentrations in a subset of the study locations and for a
portion of the study period appear to have met the current standard, data were not available in
some cities for the earlier years of the study period when design values for other cities were well
above 70 ppb (as discussed in section 3.3.3). We recognize that the study analyses and
associations reported were based on the combined dataset across the full time period (and, for
multicity studies, from all cities), and the extent to which risk associated with exposures derived
from the concentrations in the subset of years (and locations) that would have met the current
standard compared to that from the years (and locations) that would have violated the standard
influenced the study findings is not clear. There were no studies conducted in U.S. locations with
ambient air O3 concentrations that would meet the current standard for the entire duration of the
study (i.e., with design values102 at or below 70 ppb). Thus, the epidemiologic studies provide
limited insight regarding exposure concentrations associated with health outcomes that might be
expected under air quality conditions that meet the current standard (section 3.3.3 above). Thus,
the studies of 6.6-hour exposures with quasi-continuous exercise, and particularly those for
concentrations ranging from 60 to 80 ppb continue to provide an appropriate focus in this
reconsideration.
As in the 2015 review, we recognize some uncertainty, reflecting limitations in the
evidence base, with regard to the exposure levels eliciting effects as well as the severity of the
effects in some population groups not included in the available controlled human exposure
studies, such as children and individuals with asthma. Further, we note uncertainty in the extent
or characterization of effects at exposure levels below those studied. In this context, we
recognize that the controlled human exposure studies, primarily conducted in healthy adults, on
which the depth of our understanding of 03-related health effects is based, provide limited, but
nonetheless important information with regard to responses in people with asthma or in children.
We also note that the evidence indicates that responses such as those observed in the controlled
human exposure studies, if repeated or sustained, particularly in people with asthma, can pose
risks of effects of greater concern, including asthma exacerbation. We also take note of
statements from the ATS, and judgments made by the EPA in considering similar effects in past
NAAQS reviews (80 FR 65343, October 26, 2015; 85 FR 87302, December 31, 2020). In so
102 As described in chapter 2, a design value is the metric used to describe air quality in a given area relative to the
level of the standard, taking the averaging time and form into account. For example, a design value of 70 ppb just
meets the current primary standard.
April 2022
3-93 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
doing, we recognize the role of such statements in proposing principles or considerations for
weighing the evidence rather than offering "strict rules or numerical criteria" (ATS, 2000;
Thurston et al., 2017).
The more recent statement is generally consistent with the prior (2000) statement, that
was considered in the 2015 O3 NAAQS review, including the attention that statement gives to at-
risk or vulnerable population groups, while also broadening the discussion of effects, responses,
and biomarkers to reflect the expansion of scientific research in these areas. One example of this
increased specificity is in the discussion of small changes in lung function (in terms of FEVi) in
people with compromised function, such as people with asthma (Thurston et al., 2017). We note
that, in keeping with the intent of these statements to avoid specific criteria, the statements, in
discussing what constitutes an adverse health effect, do not comprehensively describe all the
biological responses raised, e.g., with regard to magnitude, duration or frequency of small
pollutant-related changes in lung function. These concepts, including the consideration of the
magnitude of effects occurring in just a subset of study subjects, continue to be recognized as
important in the more recent ATS statement (Thurston et al., 2017) and continue to be relevant to
the evidence base for O3. In this context, we also recognize the limitations in the available
evidence base with regard to our understanding of these aspects (e.g. magnitude, duration and
frequency) of such changes (e.g., in lung function) that may be associated with exposure
concentrations of interest, including with regard to the exposure levels eliciting effects (as well
as the severity or magnitude of the effects) in some population groups not included in the
available controlled human exposure studies, such as children and individuals with asthma.
Notwithstanding these limitations, we recognize that the controlled human exposure studies,
primarily conducted in healthy adults, on which the depth of our understanding of Cb-related
health effects is based, in combination with the larger evidence base, inform our conceptual
understanding of O3 responses in people with asthma and in children. Aspects of our
understanding continue to be limited, however, including with regard to the risk of particular
effects and associated severity for these less studied population groups that may be posed by 7-
hour exposures with exercise to concentrations as low as 60 ppb that are estimated in the
exposure analyses. Notwithstanding these limitations and associated uncertainties, we take note
of the emphasis of the ATS statement on consideration of effects in individuals with pre-existing
compromised function, such as that resulting from asthma (an emphasis which is reiterated and
strengthened in the current statement) Such considerations are important to the judgments on the
adequacy of protection provided by the current standard for at-risk populations. Collectively,
these aspects of the evidence and associated uncertainties contribute to a recognition that for O3,
as for other pollutants, the available evidence base in a NAAQS review generally reflects a
continuum, consisting of exposure levels at which scientists generally agree that health effects
April 2022
3-94 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
are likely to occur, through lower levels at which the likelihood and magnitude of the response
become increasingly uncertain.
As at the time the current standard was set in 2015, the exposure and risk estimates
developed from modeling exposures to O3 derived from precursors emitted into ambient air are
critically important to consideration of the potential for exposures and risks of concern under air
quality conditions of interest, and consequently are critically important to judgments on the
adequacy of public health protection provided by the current standard. In turning to consideration
of the public health implications of estimated occurrences of exposures (while at increased
exertion) to the three benchmark concentrations (60, 70 and 80 ppb), we note the respiratory
effects reported for this range of concentrations in controlled human exposure studies during
quasi-continuous exercise. In this context, we recognize that the three benchmarks represent
exposure conditions associated with different levels of respiratory responses in the subjects
studied and can inform judgments on different levels of risk that might be posed to unstudied
members of at-risk populations. The highest benchmark concentration (80 ppb) represents an
exposure where multiple controlled human exposure studies, involving 6.6-hour exposures
during quasi-continuous exercise, demonstrate a range of Cb-related respiratory effects. These
respiratory effects include a statistically significant increase in multiple types of respiratory
inflammation indicators in multiple studies; statistically significantly increased airway resistance
and responsiveness; statistically significant FEVi decrements; and statistically significant
increases in respiratory symptoms (Table 3.2). In one variable exposure study for which 80 ppb
was the exposure period average concentration, the study subject group mean FEVi decrement
was nearly 8%, with individual decrements of 15% or greater (of moderate or greater size) in
16% of subjects and decrements of 10% or greater in 32% of subjects (Schelegle et., al 2009;
Table 3.2; Appendix 3D, Figure 3D-11 and Table 3D-20); the percentages of individual subjects
with decrements greater than 10 or 15% were lower in other studies for this exposure (Appendix
3D, Figure 3D-11 and Table 3D-20). The second benchmark (70 ppb) represents an exposure
level below the lowest exposures that have reported both statistically significant FEVi
decrements103 and increased respiratory symptoms (reported at 73 ppb, Schelegle et al., 2009) or
statistically significant increases in airway resistance and responsiveness (reported at 80 ppb,
Horstman et al., 1990). The lowest benchmark (60 ppb) represents still lower exposure, and a
level for which findings from controlled human exposure studies of largely healthy subjects have
included: statistically significant decrements in lung function (with group mean decrements
103 The study group mean lung function decrement for the 73 ppb exposure was 6%, with individual decrements of
15% or greater (of moderate or greater size) in about 10% of subjects and decrements of 10% or greater in 19% of
subjects. Decrements of 20% or greater were reported in 6.5% of subjects (Schelegle et al., 2009; Table 3-2;
Appendix 3D, Figure 3D-11 and Table 3D-20).
April 2022
3-95 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
ranging from 1.7% to 3.5% across the four studies with average exposures of 60 to 63 ppb104),
but not respiratory symptoms; and, a statistically significant increase in a biomarker of airway
inflammatory response relative to filtered air exposures in one study (Kim et al, 2011; Table 3.2).
In this context, we additionally note that while not all people experiencing such
exposures experience a response (e.g. lung function decrement), as illustrated by the percentages
cited above, and among those individuals that experience a response, not all will experience an
adverse effect, the likelihood of adverse effects increase as the number of occurrences of O3
exposures of concern increases (as recognized in the 2015 decision establishing the current
standard).105 Thus, while single occurrences can be adverse for some people, particularly for the
higher benchmark concentration where the evidence base is stronger, the potential for adverse
response increases with repeated occurrences (particularly for people with asthma). Accordingly,
we recognize that the exposure/risk analyses provide estimates of exposures of the at-risk
population to concentrations of potential concern but are not yet able to provide information on
how many of such populations will have an adverse health outcome. Thus, in considering the
exposure/risk analysis results, while taking note of the extent of occurrences of one or more days
with an exposure at or above a benchmark, particularly the higher benchmarks, we additionally
recognize the potential for multiple occurrences to be of greater concern than single occurrences
(as was judged in establishing the current standard in 2015).
In the 2015 decision establishing the current standard, the controlled human exposure
study evidence as a whole provided context for consideration of the 2014 HREA results for the
exposures of concern (i.e., the comparison-to-benchmarks analysis) (80 FR 65363, October 26,
2015).106 Similarly, in this reconsideration of the 2020 decision to retain the standard, the
104	Among subjects in all four of these studies, individual FEVi decrements of at least 15% were reported in 3% of
subjects, with 7% of subjects reported to have decrements at or above a lower value of 10% (Appendix 3D,
Figure 3D-11 and Table 3D-20).
105	The 2015 decision establishing the current standard stated for example, "the Administrator acknowledge[d] such
interindividual variability in responsiveness in her interpretation of estimated exposures of concern." In this 2015
decision context, the Administrator noted "that not everyone who experiences an exposure of concern, including
for the 70 ppb benchmark, is expected to experience an adverse response," judging "that the likelihood of adverse
effects increases as the number of occurrences of O3 exposures of concern increases." In making this judgment,
the Administrator noted that "the types of respiratory effects that can occur following exposures of concern,
particularly if experienced repeatedly, provide a plausible mode of action by which O3 may cause other more
serious effects." Therefore, the 2015 decision included her emphasis on "the public health importance of limiting
the occurrence of repeated exposures to O3 concentrations at or above those shown to cause adverse" (80 FR
65331, October 26, 2015).
106	As summarized in section 3.1 above, the decision in the 2015 review considered the breadth of the O3 respiratory
effects evidence, recognizing the relatively greater significance of effects reported for exposures at and above 80
ppb as well as the greater array of effects elicited. The decision additionally emphasized consideration of the
much less severe effects associated with lower exposures, such as 60 ppb, in light of the need for a margin of
safety in setting the standard.
April 2022
3-96 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
evidence base of 6.6-hour controlled human exposure studies, particularly those of exposures
from 60 to 80 ppb, which is little changed from the 2015 review, provides context for our
consideration of the public health implications of the results from the updated exposure/risk
analyses. In our consideration of these analyses, we first note several ways in which they differ
from and improve upon those available in the 2015 review. For example, we note the number of
improvements to input data and modeling approaches summarized in section 3.4.1 above. As in
past reviews, exposure and risk are estimated from air quality scenarios designed to just meet an
O3 standard in all its elements. That is, the air quality scenarios are defined by the highest design
value in the study area, which is the location with the highest 3-year average of annual fourth
highest daily maximum 8-hour O3 concentrations (e.g., equal to 70 ppb for the current standard
scenario). The risk and exposure analyses include air quality simulations based on more recent
ambient air quality data that include O3 concentrations closer to the current standard than was the
case for the analyses in the 2015 review. As a result, much smaller reductions in precursor
emissions were needed in the photochemical modeling than was the case with the 2014 HREA.
Further, this modeling was updated to reflect the current state of the science. Additionally, the
approach for deriving population exposure estimates, both for comparison to benchmark
concentrations and for use in deriving lung function risk using the E-R function approach, has
been modified to provide for a better match of the simulated population exposure estimates with
the 6.6-hour duration of the controlled human exposure studies and with the study subject
ventilation rates. Together, these differences, as well as a variety of updates to model inputs, are
believed to reduce uncertainty associated with our interpretation of the analysis results.
As we consider the exposure and risk estimates, we also take note of the array of air
quality and exposure circumstances represented by the eight study areas. As summarized in
section 3.2.2 above, the areas fall into seven of the nine climate regions in the continental U.S.
The population sizes of the associated metropolitan areas range in size from approximately 2.4 to
8 million and vary in population demographic characteristics. While there are uncertainties and
limitations associated with the exposure and risk estimates, as noted in section 3.4.4 above, the
factors recognized here contribute to their usefulness in informing judgments relevant to the
Administrator's consideration of the current standard.
While there are more adults in the U.S. with asthma than children with asthma, the
exposure and risk analysis results in terms of percent of the simulated at-risk populations,
indicate higher frequency of exposures of potential concern and risks for children as compared to
adults. This finding relates to children's greater frequency and duration of outdoor activity, as
well as their greater activity level while outdoors (section 3.4.3 above). In light of these
conclusions and findings, we have focused our consideration of the exposure and risk analyses
here on children.
April 2022
3-97 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
As can be seen by variation in exposure estimates across the study areas, the eight study
areas represent an array of exposure circumstances, including those contributing to relatively
higher and relatively lower exposures and associated risk. As recognized in Appendix 3D and in
section 3.4.3 above, the risk and exposure analyses are not intended to provide a comprehensive
national assessment. Rather, the analyses for this array of study areas and air quality patterns are
intended to indicate the magnitude of exposures and risks that may be expected in areas of the
U.S. that just meet the current standard but that may differ in ways affecting population
exposures of interest. In that way, the exposure and risk estimates are intended to be informative
to the EPA's consideration of potential exposures and risks associated with the current standard
and the Administrator's decision on the adequacy of protection provided by the current standard.
While we note reduced uncertainty in several aspects of the exposure and risk analysis
approach (as summarized above), we continue to recognize the relatively greater uncertainty
associated with the lung function risk estimates compared to the results of the comparison-to-
benchmarks analysis (and the greater uncertainty with the estimates derived using the MSS
model approach than the E-R approach). Thus, we focus primarily on the estimates of exposures
at or above different benchmark concentrations that represent different levels of significance of
03-related effects, both with regard to the array of effects and severity of individual effects.
Based on all of the above, and taking into consideration related information, limitations
and uncertainties, such as those recognized above, we address the extent to which the recently
available information supports or calls into question the adequacy of protection afforded by the
current standard. Focusing on the air quality scenario for the current standard, we note that
across all eight study areas, which provide an array of exposure situations, less than 1% of
children with asthma are estimated to experience, while breathing at an elevated rate, a daily
maximum 7-hour exposure per year at or above 70 ppb, on average across the 3-year period, with
a maximum of 1% for the study area with the highest estimates in the highest single year (as
summarized in section 3.4.2 above). Further, the percentage for at least one day with such an
exposure above 80 ppb is less than 0.1%, as an average across the 3-year period (and 0.1% or
less in each of the three years simulated across the eight study areas). No simulated individuals
were estimated to experience more than a single such day with an exposure at or above the 80
ppb benchmark. Although the exposure and risk analysis approaches have been updated since the
2015 review as summarized in section 3.4.1 above, these estimates are generally similar to the
comparable estimates for these benchmarks from the 2014 HREA considered at the time the
current standard was set,107 with only slight differences observed, e.g., for the lowest benchmark.
107 For example, in the 2015 decision to set the standard level at 70 ppb, the Administrator took note of several
findings for the air quality scenarios for this level, noting that "a revised standard with a level of 70 ppb is
April 2022
3-98 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
We take note, however, of the differences across air quality scenarios for both sets of estimates
which remain appreciably larger than the slight differences between the current and 2014
estimates. Thus, we observe that the current estimates of children and children with asthma that
might be expected to experience a day with an exposure while exercising at or above the three
benchmark concentrations are generally similar to those that were a primary focus of the decision
in establishing the current standard in 2015.
We additionally consider the estimates of 7-hour exposures, at elevated ventilation, at or
above 60 ppb. In so doing, we recognize that the role of this consideration in the 2015 decision
was in the context of the judgment of the Administrator at the time regarding an adequate margin
of safety for the new standard. We additionally recognize the greater significance of risk for
multiple occurrences of days at or above this benchmark, given the associated greater potential
for more lasting effects. The exposure analysis estimates indicate fewer than 1% to just over 3%
of children with asthma, on average across the 3-year period, to be expected to experience two or
more days with an exposure at or above 60 ppb, while at elevated ventilation. This finding of
about 97% to more than 99% of children protected from experiencing two or more days with
exposures at or above 60 ppb while at elevated exertion is quite similar to the characterization of
such estimates at the time of the 2015 decision establishing the current standard (as summarized
in section 3.1.2.4 above),108 and half that indicated by the comparable estimates for air quality
just meeting the slightly higher design value of 75 ppb. In addition to this level of protection at
the lower exposure level (of 60 ppb), the current information also indicates more than 99% of
children with asthma, on average per year, to be protected from a day or more with an exposure
at or above 70 ppb. In light of public health judgments by the EPA in prior NAAQS reviews, and
related considerations, as well as ATS guidance, we recognize a greater concern for 7-hour
exposures generally at or above 70 and 80 ppb (while at elevated exertion) than such exposures
to O3 concentrations below 70 ppb, and a greater concern for repeated (versus single)
occurrences of such exposures at concentrations at or above 60 ppb up to 70 ppb. With this in
mind, we find the current exposure and risk estimates to indicate that the current standard is
estimated to eliminate the occurrence of two or more exposures of concern to O3 concentrations at or above 80
ppb and to virtually eliminate the occurrence of two or more exposures of concern to O3 concentrations at or
above 70 ppb for all children and children with asthma, even in the worst-case year and location evaluated" (80
FR 65363, October 26, 2015). This statement remains true for the results of the current assessment (Table 3-8).
108 For example, with regard to the 60 ppb benchmark, for which the 2015 decision placed relatively greater weight
on multiple (versus single) occurrences of exposures at or above it, the Administrator at that time noted the 2014
HREA estimates for the 70 ppb air quality scenario that estimated 0.5-3.5% of children to experience multiple
such occurrences on average across the study areas, stating that the now-current standard "is estimated to protect
the vast majority of children in urban study areas ... from experiencing two or more exposures of concern at or
above 60 ppb" (80 FR 65364, October 26, 2015). The corresponding estimates, on average across the 3-year
period in the current assessments, are remarkedly similar at 0.6 -2.9% (Table 3-8).
April 2022
3-99 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
likely to provide a high level of protection from Cte-related health effects to at-risk populations of
all children and children with asthma. We additionally recognize such protection to be generally
similar to what was estimated when the standard was set in 2015.
As recognized above, the protection afforded by the current standard stems from its
elements collectively, including the level of 70 ppb, the averaging time of eight hours and the
form of the annual fourth-highest daily maximum concentration averaged across three years. The
current evidence as considered in the ISA, the current air quality information as analyzed in
chapter 2 of this document, and the current risk and exposure information (presented in
Appendix 3D and summarized in section 3.4 above) provide continued support to these elements,
as well as to the current indicator, as discussed earlier in this section.
In summarizing the information discussed thus far, we reflect on the key aspects of the
2015 decision that established the current standard. As an initial matter, effects associated with
6.6-hour exposures with quasi-continuous exercise (in controlled human exposure studies) to 73
ppb O3 (as a time-weighted average) included both lung function decrements and respiratory
symptoms, which the EPA recognized to be adverse; this judgment was based on consideration
of the EPA decisions in prior NAAQS reviews and CASAC advice, as well as ATS guidance (80
FR 65343, October 26, 2015). We note that the newly available information since the 2015
review includes an additional statement from ATS on assessing adverse effects of air pollution
which is generally consistent with the earlier statement (available at the time of the 2015
decision), e.g., continuing to emphasize potentially at-risk groups, including specific
consideration of effects in people with compromised lung function. While recognizing the
differences between the current and past exposure and risk analyses, as well as uncertainties
associated with such analyses, we note a rough consistency of the associated estimates when
considering the array of study areas in both reviews. Overall, the recent quantitative analyses
appear to comport with the conclusions reached in the 2015 review regarding control expected to
be exerted by the current standard on exposures of concern.
We additionally recognize that decisions regarding the adequacy of the current standard
depend in part on public health policy judgments, such as those identified above, and judgments
about when a standard is requisite to protect the public health, including the health of at-risk
populations, allowing for an adequate margin of safety. In this context, we take note of the long-
standing health effects evidence that documents the effects of 6.6-hour O3 exposures on people
exposed while breathing at elevated rates and recognize that these effects have been reported in a
few individuals for the lowest concentration studied in exposure chambers (40 ppb). Thus, in
considering the exposure analysis estimates for 7-hour exposures at and above 60 ppb, we also
take note of the variability in the responses at low concentrations, including, for example, the
variation in average response to a 7-hour 60 ppb exposure with exercise (group mean FEVi
April 2022
3-100 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
decrement of 1.7 to 3.5% change), as well as the lack of statistically significant decrements in
lung function from such exposures at concentrations below 60 ppb. Consistent with the EPA's
judgments in previous reviews, we also recognize the greater potential for health risk from
repeated (versus isolated single) occurrences In light of this, we note that the exposure estimates
indicate the current standard may be expected to protect more than 97% of populations of
children with asthma residing in areas just meeting the current standard from experiencing more
than a single day per year with an exposure at or above 60 ppb, on average over a 3-year period.
We additionally note the estimates that indicate protection of more than 99.9% of children with
asthma living in such areas from experiencing any days with a 7-hour exposure while at elevated
exertion of 80 ppb or higher in a 3-year period, on average. In light of ATS guidance, CASAC
advice and EPA judgments and considerations in pastNAAQS reviews, these results indicate a
high level of protection of key at-risk populations from Cb-related health effects that is a
generally similar level of protection to what was articulated when the standard was set in 2015
and retained in 2020. Thus, the evidence and exposure/risk information, including that related to
the lowest exposures studied, lead us to conclude that the combined consideration of the body of
evidence and the quantitative exposure estimates including the associated uncertainties, do not
call into question the adequacy of the protection provided by the current standard. Rather, this
information continues to provide support for the current standard, and thus supports
consideration of retaining the current standard, without revision.
In reaching these conclusions, we recognize that the Administrator's decisions in primary
standard reviews, in general, are largely public health judgments, as described above. We further
note that different public health policy judgments (e.g., from those made in both 2020 and 2015)
could lead to different conclusions regarding the extent to which the current standard provides
protection of public health with an adequate margin of safety. Such public health judgments
include those related to the appropriate degree of public health protection that should be afforded
to protect against risk of respiratory effects in at-risk populations, such as asthma exacerbation
and associated health outcomes in people with asthma, as well as with regard to the appropriate
weight to be given to differing aspects of the evidence and exposure/risk information, and how to
consider their associated uncertainties. For example, different judgments might give greater
weight to more uncertain aspects of the evidence or reflect a differing view with regard to margin
of safety. Such judgments are left to the discretion of the Administrator. In this context, we note
that the scientific evidence and quantitative exposure and risk information in the record on which
this reconsideration is based are largely unchanged. Staff conclusions regarding the adequacy of
the current standards thus remain unchanged from those reached in the 2020 PA.
In summary, the newly available health effects evidence, critically assessed in the 2020
ISA as part of the full body of evidence, reaffirms conclusions on the respiratory effects
April 2022
3-101 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
recognized for O3 in prior reviews. Further, we observe the general consistency of the more
recent evidence with the evidence that was available in the 2015 review with regard to key
aspects on which the current standard is based. We additionally note the quantitative exposure
and risk estimates for conditions just meeting the current standard that indicate a generally
similar level of protection for at-risk populations from respiratory effects, as that described in the
2015 review for the now-current standard. We also recognize limitations and uncertainties
associated with the available information, similar to those at the time of the 2015 review.
Collectively, these considerations (including those discussed above) provide the basis for the
preliminary conclusion that the available evidence and exposure/risk information does not call
into question the adequacy of protection provided by the existing standard or the scientific and
public health judgments that informed the 2020 decision to retain the current standard, which
was established in the 2015 review. Accordingly, we conclude it is appropriate in this
reconsideration of the 2020 decision that consideration be given to retaining the current primary
standard of 0.070 ppm O3, as the fourth-highest daily maximum 8-hour concentration averaged
across three years, without revision. In light of this conclusion, we have not identified any
potential alternative standards for consideration.
3.6 KEY UNCERTAINTIES AND AREAS FOR FUTURE RESEARCH
In this section, we highlight key uncertainties associated with reviewing and establishing
the primary O3 standard, while additionally recognizing that research in these areas may be
informative to the development of more efficient and effective control strategies. The list in this
section includes key uncertainties and data gaps thus far highlighted in this review of the primary
standard. A critical aspect of our consideration of the evidence and the quantitative risk/exposure
estimates is our understanding of O3 effects below the lowest concentrations studied in
controlled human exposure studies, for longer exposures and for different population groups,
particularly including people with asthma. Additional information in several areas would reduce
uncertainty in our interpretation of the available information for purposes of risk characterization
and, accordingly, reduce uncertainty in characterization of 03-related health effects. In this
section, we highlight areas for future health-related research, model development, and data
collection activities to address these uncertainties and limitations in the current scientific
evidence. These areas are similar to those highlighted in past reviews.
Exposure and Risk Assessment Data and Tools:
• An important aspect of risk assessment and characterization to inform decisions regarding
the primary standard is our understanding of the exposure-response relationship for 03-
related health effects in at-risk populations. Additional research is needed to more
April 2022
3 -102 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
comprehensively assess risk of respiratory effects in at-risk individuals exposed to O3 in
the range of 40 to 80 ppb, and lower, for 6.6 hours while engaged in moderate exertion.
•	Population- or cohort-based information on human exposure and associated health effects
for healthy adults and children and at-risk populations, including people with asthma, to
relevant levels and durations of O3 concentrations in ambient air, including exposure
information in various microenvironments and at varying activity levels, is needed to
better evaluate current and future O3 exposure and lung function risk models. Such
information across extended periods would facilitate evaluation of exposure models for
the O3 season.
•	Collection of time-activity data over longer time periods, and particularly for children
(including under the age of five), is needed to reduce uncertainty in the modeled exposure
distributions that form an important part of the basis for decisions regarding NAAQS for
O3 and other air pollutants. Research addressing energy expenditure and associated
breathing rates in various population groups, particularly healthy children and children
with asthma, in various locations, across the spectrum of physical activity, including
sleep to vigorous exertion, is needed.
Health Effects Evidence Base:
•	Epidemiologic studies assessing the influence of "long-term" or "short-term" O3
exposures is complicated by a lack of knowledge regarding the exposure history of study
populations. Further, existing studies generally focus on either long-term or short-term
exposure separately, thereby making it difficult to assess whether a single short-term
high-level exposure versus a repeated long-term low-level exposure, or a combination of
both short-term high-level and repeated long-term low-level exposures, influence health
outcomes of the study subjects. Epidemiologic studies that include exposure
measurements across a longer-term assessment period and can simultaneously assess the
impact of these various elements of exposure (i.e., magnitude, frequency, durations, and
pattern) are needed.
•	The extent to which the broad mix of photochemical oxidants as well as other copollutants
in the ambient air (e.g., PM, NO2, SO2, etc.) may play a role in modifying or contributing
to the observed associations between ambient air O3 concentrations and reported health
outcomes continues to be an important research question. A better understanding of the
broader mixture of photochemical oxidants other than O3 in ambient air, the associated
human exposures, and of the extent to which effects of the mixture may differ from those
of O3, would be informative to future NAAQS reviews. Studies that examine and
improve analytical approaches to better understand the role of copollutants, as well as
temperature, in contributing to potential confounding or effect modification in
epidemiologic models would be helpful.
•	Most epidemiologic study designs remain subject to uncertainty due to use of fixed-site
ambient air monitors serving as a surrogate for exposure measurements. The accuracy
with which measurements made at stationary outdoor monitors actually reflect subjects'
exposure is not yet fully understood. The degree to which discrepancies between
April 2022
3-103 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
stationary monitor measurements and actual pollutant exposures introduces error into
statistical estimates of pollutant effects in epidemiologic studies needs to be investigated.
•	For health endpoints reported in epidemiologic studies, such as respiratory hospital
admissions, emergency department visits, and premature mortality, a more
comprehensive characterization of the exposure circumstances (including ambient air
concentrations, as well as duration of exposure and activity levels of individuals) eliciting
such effects is needed
•	Further research investigating additional uncertainties and factors that modify
epidemiologic associations, particularly for different population groups would improve
our understanding in these areas.
•	The evidence base, expanded by evidence newly available for the 2020 review, indicates a
likely causal relationship between short-term O3 exposure and metabolic effects. Further
research characterizing perturbations of glucose and insulin homeostasis by O3 in
controlled human exposure studies at exertion and in animal toxicology studies at
concentrations closer to the current standard are needed inform decisions regarding the
primary standard. The collection of population-based information on clinical health
outcomes such as metabolic syndrome, diabetes, etc., as well as intermediate indicators
like insulin resistance is also needed for an array of populations and lifestages. Such
studies would provide an improved understanding of relationships between O3 exposure
and metabolic-related health outcomes.
Air Quality:
•	Advances in photochemical modeling representations of the atmosphere and in high
spatial and temporal resolution estimates of ozone precursor emissions will further reduce
uncertainties in photochemical modeling used in estimating O3 concentrations for
different air quality scenarios.
A more robust ambient monitoring network is needed to better understand ozone concentration
gradients in urban areas. With the recent development of low-cost ozone sensors, this could be
achieved in the near future.
April 2022
3 -104 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
REFERENCES
Adams, WC (2000). Ozone dose-response effects of varied equivalent minute ventilation rates. J
Expo Anal Environ Epidemiol 10(3): 217-226.
Adams, WC (2002). Comparison of chamber and face-mask 6.6-hour exposures to ozone on
pulmonary function and symptoms responses. Inhal Toxicol 14(7): 745-764.
Adams, WC (2003). Comparison of chamber and face mask 6.6-hour exposure to 0.08 ppm
ozone via square-wave and triangular profiles on pulmonary responses. Inhal Toxicol
15(3): 265-281.
Adams, WC (2006a). Human pulmonary responses with 30-minute time intervals of exercise and
rest when exposed for 8 hours to 0.12 ppm ozone via square-wave and acute triangular
profiles. Inhal Toxicol 18(6): 413-422.
Adams, WC (2006b). Comparison of chamber 6.6-h exposures to 0.04-0.08 PPM ozone via
square-wave and triangular profiles on pulmonary responses. Inhal Toxicol 18(2): 127-
136.
Akinbami, LJ, Simon, AE and Schoendorf, KC (2016). Trends in allergy prevalence among
children aged 0-17 years by asthma status, United States, 2001-2013. J Asthma 53(4):
356-362.
Alexis, NE, Lay, JC, Hazucha, M, Harris, B, Hernandez, ML, Bromberg, PA, Kehrl, H, Diaz-
Sanchez, D, Kim, C, Devlin, RB and Peden, DB (2010). Low-level ozone exposure
induces airways inflammation and modifies cell surface phenotypes in healthy humans.
Inhal Toxicol 22(7): 593-600.
Arjomandi, M, Balmes, JR, Frampton, MW, Bromberg, P, Rich, DQ, Stark, P, Alexis, NE,
Costantini, M, Hollenbeck-Pringle, D, Dagincourt, N and Hazucha, MJ (2018).
Respiratory Responses to Ozone Exposure. MOSES (The Multicenter Ozone Study in
Older Subjects). Am J Respir Crit Care Med 197(10): 1319-1327.
ATS (1985). Guidelines as to what constitutes an adverse respiratory health effect, with special
reference to epidemiologic studies of air pollution. Am Rev Respir Dis 131(4): 666-668.
ATS (2000). What constitutes an adverse health effect of air pollution? Am J Respir Crit Care
Med 161(2): 665-673.
Brown, JS, Bateson, TF and McDonnell, WF (2008). Effects of exposure to 0.06 ppm ozone on
FEV1 in humans: a secondary analysis of existing data. Environ Health Perspect 116(8):
1023-1026.
Bureau of Labor Statistics (2020). U.S. Department of Labor, The Economics Daily, Civilian
occupations required to spend the most time outdoors in 2020 on the Internet at
https://www.bls.gov/opub/ted/2021/civilian-occupations-required-to-spend-the-most-
time-outdoors-in-2020.htm (visited January 26, 2022).
April 2022
3-105 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
CDC (2016). Current Asthma Prevalence by Weight Status Among Adults: United States, 2001-
2014. NCHS Data Brief No. 239, March 2016. Available at:
https://www. cdc.gov/nchs/products/databriefs/db239. htm.
CDC (2019). National Health Interview Survey, 2019. National Center for Health Statistics,
CDC. Washington, DC. Most Recent National Asthma Data. Available at:
www.cdc.gov/asthma/most recent national asthma data.htm. Accessed January 26,
2022.
Cox, LA. (2020). Letter from Louis Anthony Cox, Jr., Chair, Clean Air Scientific Advisory
Committee, to Administrator Andrew R. Wheeler. Re:CASAC Review of the EPA's
Policy Assessment for the Review of the Ozone National Ambient Air Quality Standards
(ExternalReview Draft - October 2019). February 19, 2020. EPA-CASAC-20-003.
Available at:
https://yosemite. epa.gov/sab/sabproduct.nsf/264cb122 7d55e02c85257402007446a4/4 713
D217BC07103485258515006359BA/$File/EPA-CASAC-20-003.pdf
Devlin, RB, McDonnell, WF, Mann, R, Becker, S, House, DE, Schreinemachers, D and Koren,
HS (1991). Exposure of humans to ambient levels of ozone for 6.6 hours causes cellular
and biochemical changes in the lung. Am J Respir Cell Mol Biol 4(1): 72-81.
Duffney, PF, Brown, JS, and Stone, SL (2022). Memorandum to the Review of the Ozone
National Ambient Air Quality Standards (NAAQS) Docket (EPA-HQ-ORD-2018-
0279). Re: Provisional Evaluation of Newly Identified Controlled Human Exposure
Studies in the context of the 2020 Integrated Science Assessment for Ozone and Related
Photochemical Oxidants. April 15, 2020.
Folinsbee, LJ and Hazucha, MJ (2000). Time course of response to ozone exposure in healthy
adult females. Inhal Toxicol 12: 151-167.
Folinsbee, LJ, Horstman, DH, Kehrl, HR, Harder, S, Abdul-Salaam, S and Ives, PJ (1994).
Respiratory responses to repeated prolonged exposure to 0.12 ppm ozone. Am J Respir
CritCare Med 149(1): 98-105.
Folinsbee, LJ, McDonnell, WF and Horstman, DH (1988). Pulmonary function and symptom
responses after 6.6-hour exposure to 0.12 ppm ozone with moderate exercise. JAPCA
38(1): 28-35.
Frey, HC. (2014). Letter from Dr. H. Christopher Frey, Chair, Clean Air Scientific Advisory
Committee to Honorable Gina McCarthy, Administrator, US EPA. Re: CASAC Review
of the EPA's Second Draft Policy Assessment for the Review of the Ozone National
Ambient Air Quality Standards. June 26, 2014. EPA-CASAC-14-004. Available at:
https: //nepis. epa.gov/Exe/ZyPURL. cgi?Dockey=P 100JR6F.txt.
Frey, HC and Samet, JM. (2012). Letter from Dr. H. Christopher Frey, Chair, Clean Air
Scientific Advisory Committee Ambient Air Monitoring & Methods Committee and
Jonathan Samet, Immediate Past Chair, Clean Air Scientific Advisory Committee, to
Administrator Lisa Jackson. Re: CASAC Review of the EPA's Policy Assessment for the
April 2022
3-106 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
Review of the Ozone National Ambient Air Quality Standards (First External Review
Draft - August 2012). November 26, 2012. EPA-CASAC-13-003. Available at:
https://nepis.epa.gov/Exe/ZyPURL.cgi?Dockey=P100J7PQ. txt.
Gong, H, Jr., Bradley, PW, Simmons, MS and Tashkin, DP (1986). Impaired exercise
performance and pulmonary function in elite cyclists during low-level ozone exposure in
a hot environment. Am J Respir Crit Care Med 134(4): 726-733.
Horstman, DH, Folinsbee, LJ, Ives, PJ, Abdul-Salaam, S and McDonnell, WF (1990). Ozone
concentration and pulmonary response relationships for 6.6-hour exposures with five
hours of moderate exercise to 0.08, 0.10, and 0.12 ppm. Am Rev Respir Dis 142(5):
1158-1163.
Karl, T and Koss, WJ (1984). Regional and national monthly, seasonal, and annual temperature
weighted by area, 1895-1983. 4-3. National Environmental Satellite and Data
Information Service (NESDIS). Asheville, NC.
Kim, CS, Alexis, NE, Rappold, AG, Kehrl, H, Hazucha, MJ, Lay, JC, Schmitt, MT, Case, M,
Devlin, RB, Peden, DB and Diaz-Sanchez, D (2011). Lung function and inflammatory
responses in healthy young adults exposed to 0.06 ppm ozone for 6.6 hours. Am J Respir
Crit Care Med 183(9): 1215-1221.
Kousha, T and Rowe, BH (2014). Ambient ozone and emergency department visits due to lower
respiratory condition. Int J Occup Med Environ Health 27(1): 50-59.
Langstaff, J (2007). Memorandum to Ozone NAAQS Review Docket (EPA-HQ-OAR-2005-
0172). Analysis of Uncertainty in Ozone Population Exposure Modeling. Docket
Document ID: EPA-HQ-OAR-2005-0172-0174.
Luben, T, Lassiter, M and Herrick, J (2020). Memorandum to Ozone NAAQS Review Docket
(EPA-HQ-ORD-2018-0279). RE: List of Studies Identified by Public Commenters That
Have Been Provisionally Considered in the Context of the Conclusions of the 2020
Integrated Science Assessment for Ozone and Related Photochemical Oxidants.
December 2020. Docket Document ID: EPA-HQ- OAR-2018-0279-0560.
Mar, TF and Koenig, JQ (2009). Relationship between visits to emergency departments for
asthma and ozone exposure in greater Seattle, Washington. Ann Allergy, Asthma
Immunol 103(6): 474-479.
McCurdy, T (2000). Conceptual basis for multi-route intake dose modeling using an energy
expenditure approach. J Expo Anal Environ Epidemiol 10(1): 86-97.
McCurdy, T and Graham, SE (2003). Using human activity data in exposure models: Analysis of
discriminating factors. J Expo Anal Environ Epidemiol 13(4): 294-317.
McDonnell, WF, Horstman, DH, Hazucha, MJ, Seal, E, Jr., Haak, ED, Salaam, SA and House,
DE (1983). Pulmonary effects of ozone exposure during exercise: Dose-response
characteristics. J Appl Physiol (1985) 54(5): 1345-1352.
April 2022
3-107 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
McDonnell, WF, Kehrl, HR, Abdul-Salaam, S, Ives, PJ, Folinsbee, LJ, Devlin, RB, O'Neil, JJ
and Horstman, DH (1991). Respiratory response of humans exposed to low levels of
ozone for 6.6 hours. Arch Environ Health 46(3): 145-150.
McDonnell, WF, Stewart, PW and Smith, MV (2013). Ozone exposure-response model for lung
function changes: an alternate variability structure. Inhal Toxicol 25(6): 348-353.
McDonnell, WF, Stewart, PW, Smith, MV, Kim, CS and Schelegle, ES (2012). Prediction of
lung function response for populations exposed to a wide range of ozone conditions.
Inhal Toxicol 24(10): 619-633.
Schelegle, ES, Morales, CA, Walby, WF, Marion, S and Allen, RP (2009). 6.6-hour inhalation of
ozone concentrations from 60 to 87 parts per billion in healthy humans. Am J Respir Crit
Care Med 180(3): 265-272.
Thurston, GD, Kipen, H, Annesi-Maesano, I, Balmes, J, Brook, RD, Cromar, K, De Matteis, S,
Forastiere, F, Forsberg, B, Frampton, MW, Grigg, J, Heederik, D, Kelly, FJ, Kuenzli, N,
Laumbach, R, Peters, A, Rajagopalan, ST, Rich, D, Ritz, B, Samet, JM, Sandstrom, T,
Sigsgaard, T, Sunyer, J and Brunekreef, B (2017). A joint ERS/ATS policy statement:
what constitutes an adverse health effect of air pollution? An analytical framework. Eur
Respir J 49(1).
U.S. Census Bureau (2021). Quick Facts: Population estimates in the United States as of July 1,
2021.	https://www.census.gov/quickfacts/fact/table/US/PST045221 Accessed Febuary 2,
2022.
U.S. DHEW (Department of Health, Education, and Welfare) (1970). Air Quality Criteria for
Photochemical Oxidants. Washington, D.C.: National Air Pollution Control
Administration; publication no. AP-63. Available from: NTIS, Springfield, VA; PB-
190262/BA
U.S. EPA (1996). Review of national ambient air quality standards for ozone: Assessment of
scientific and technical information: OAQPS staff paper . Office of Air Quality Planning
and Standards. Research Triangle Park, NC. U.S. EPA. EPA-452/R-96-007. June 1996.
Available at: http://nepis.epa.gov/Exe/ZyPDF.cgi?Dockey=2000DKJT.PDF.
U.S. EPA (2008). Risk and Exposure Assessment to Support the Review of the NO2 Primary
National Ambient Air Quality Standard. EPA-452/R-08-008a. Office of Air Quality
Planning and Standards. Research Triangle Park, NC. Available at:
https://www 3. epa.gov/ttn/naaqs/standards/nox/snoxcrrea. html.
U.S. EPA (2009). Risk and Exposure Assessment to Support the Review of the SO2 Primary
National Ambient Air Quality Standard. Office of Air Quality Planning and Standards.
Research Triangle Park, NC. US EPA. EPA-452/R-09-007. Available at:
https://www3.epa.gov/ttn/naaqs/standards/so2/data/200908S02REAFinalReport.pdf
U.S. EPA (2010). Quantitative Risk and Exposure Assessment for Carbon Monoxide - Amended.
Office of Air Quality Planning and Standards. Research Triangle Park, NC. U.S. EPA.
April 2022
3-108 External Review Draft-Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
EPA-452/R-10-006. Available at: https://www.epa.gov/naaqs/carbon-monoxide-co-
standards-risk-and-exposure-assessments-current-review.
U.S. EPA (2014). Health Risk and Exposure Assessment for Ozone. (Final Report). Office of Air
Quality Planning and Standards. Research Triangle Park, NC. U.S. EPA. EPA-452/R-14-
004a. August 2014. Available at:
https://nepis. epa.gov/Exe/ZyPURL. cgi?Dockey=P 100KBUF. txt.
U.S. EPA (2018). Risk and Exposure Assessment for the Review of the Primary National
Ambient Air Quality Standard for Sulfur Oxides. Office of Air Quality Planning and
Standards. Research Triangle Park, NC. U.S. EPA. EPA-452/R-18-003. Available at:
https://www.epa.gov/sites/production/files/2018-05/documents/primary so2 naaqs -
final rea - may 20I8.pdf.
U.S. EPA (2019). The Consolidated Human Activity Database (CHAD). Documentation and
User's Guide. Research Triangle Park, NC. US EPA. EPA-452/B-19-001. Available at:
https://www.epa.gov/healthresearch/consolidated-human-activity-database-chad-use-
human-exposure-and-health-studies-and.
U.S. EPA (2020a). Integrated Science Assessment for Ozone and Related Photochemical
Oxidants. U.S. Environmental Protection Agency. Washington, DC. Office of Research
and Development. EPA/600/R-20/012. Available at: https://www.epa.gov/isa/ integrated-
science-assessment-isa-ozoneand- related-photochemical-oxidants.
U.S. EPA (2020b). Policy Assessment for the Review of the Ozone National Ambient Air
Quality Standards. U.S. Environmental Protection Agency, Office of Air Quality
Planning and Standards, Heath and Environmental Impacts Division. Research Triangle
Park, NC. U.S. EPA. EPA-452/R-20-001. 2020 Available at: https://www.epa.gov/
naaqs/ozone-o3-standards-policy assessments-current-review
Villeneuve, PJ, Chen, L, Rowe, BH and Coates, F (2007). Outdoor air pollution and emergency
department visits for asthma among children and adults: A case-crossover study in
northern Alberta, Canada. Environ Health 6: 40.
WHO (2008). Uncertainty and Data Quality in Exposure Assessment. The Internaltional
Programme on Chemical Safety. Geneva. WHO.
https://www.who.int/ipcs/publications/methods/harmonization/exposure assessment.pdf.
April 2022
3-109 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
4 RECONSIDERATION OF THE SECONDARY
STANDARD
This chapter presents and evaluates the policy implications of the available scientific and
technical information pertaining to this reconsideration of the 2020 decision on the O3 secondary
standard. Specifically, the chapter presents key aspects of the available evidence of the welfare
effects of O3, as documented in the 2020 ISA, with support from the prior ISA and AQCDs, and
associated public welfare implications, as well as key aspects of quantitative analyses, including
air quality and environmental exposure-related information that has been updated for this
reconsideration using more recent air quality monitoring data, and is presented in detail in
appendices 4D and 4F associated with this chapter. Together all of this information provides the
foundation for our evaluation of the scientific information regarding welfare effects of O3 in
ambient air and the potential for welfare effects to occur under air quality conditions associated
with the current standard (or any alternatives considered), as well as the associated public
welfare implications. Our evaluation is framed around key policy-relevant questions derived
from the questions included in the IRP (IRP, section 3.2.1) and also takes into account, as
relevant, prior assessments of the evidence and quantitative exposure/risk analyses. In light of all
of these considerations, we will identify key policy-relevant considerations and summary
conclusions regarding the public welfare protection provided by the current standard for the
Administrator's consideration in this reconsideration.
Within this chapter, background information on the current standard, including
considerations in its establishment in the 2015 review, is summarized in section 4.1. The general
approach for considering the available information, including policy-relevant questions identified
to frame our policy evaluation, is summarized in section 4.2. Key aspects of the available welfare
effects evidence and associated public welfare implications and uncertainties are addressed in
section 4.3, and the current air quality and exposure information, with associated uncertainties, is
addressed in section 4.4. Section 4.5 summarizes the key evidence- and air quality or exposure-
based considerations identified in our evaluation, and also presents associated preliminary
conclusions of this analysis. Key remaining uncertainties and areas for future research are
identified in section 4.6.
4.1 BACKGROUND ON THE CURRENT STANDARD
As a result of the O3 NAAQS review completed in 2015, the level of the secondary
standard was revised to 0.070 ppm, in conjunction with retaining the indicator (O3), averaging
time (8 hours) and form (fourth-highest annual daily maximum 8-hour average concentration,
April 2022
4-1
External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
averaged across three years). The establishment of this standard in 2015, and its retention in
2020, is based primarily on consideration of the extensive welfare effects evidence base
compiled from more than fifty years of extensive research on the phytotoxic effects of O3,
conducted both in and outside of the U.S., that documents the impacts of O3 on plants and their
associated ecosystems (U.S. EPA, 1978, 1986, 1996, 2006, 2013). Key considerations in the
2015 decision were the scientific evidence and technical analyses available at that time, as well
as the Administrator's judgments regarding the available welfare effects evidence, the
appropriate degree of public welfare protection for the revised standard, and available air quality
information on seasonal cumulative exposures (in terms of the W126-based exposure index1) that
may be allowed by such a standard (80 FR 65292, October 26, 2015).
The 2020 decision to retain the standard, without revision, additionally took into account
updates to the evidence base since the 2015 review, and associated conclusions regarding welfare
effects; updated and expanded quantitative analyses of air quality data, including the frequency
of cumulative exposures of potential concern and of elevated hourly concentrations in areas with
air quality meeting the standard; and also the August 2019 decision of the D.C. Circuit
remanding the 2015 secondary standard to the EPA for further justification or reconsideration, as
mentioned earlier in Section 1.3 {Murray Energy Corp. v. EPA, 936 F.3d 597 [D.C. Cir. 2019]).
In the August 2019 decision, the court held that EPA had not adequately explained its decision to
focus on a 3-year average for consideration of the cumulative exposure, in terms of W126,
identified as providing requisite public welfare protection, or its decision to not identify a
specific level of air quality related to visible foliar injury. The EPA's decision not to use a
seasonal W126 index as the form and averaging time of the secondary standard was also
challenged, but the court did not reach a decision on that issue, concluding that it lacked a basis
to assess the EPA's rationale because the EPA had not yet fully explained its focus on a 3-year
average W126 in its consideration of the standard. Accordingly, the 2020 decision included
discussion of these areas to address these aspects of the court's decision.
Among the updates to the welfare effects evidence considered in the 2020 decision was
the welfare effects evidence for two insect-related categories of effects with new determinations
in the 2020 ISA. Specifically, the 2020 ISA concluded the evidence sufficient to infer likely
causal relationships of O3 with alterations of plant-insect signaling and insect herbivore growth
and reproduction. Uncertainties in the evidence for the effects, however, precluded a full
understanding of the effects, the air quality conditions that might elicit them, and the potential
1 The W126 index is a cumulative seasonal metric described as the sigmoidally weighted sum of all hourly O3
concentrations during a specified daily and seasonal time window, with each hourly O3 concentration given a
weight that increases from zero to one with increasing concentration (80 FR 65373-74, October 26, 2015). The
units for W126 index values are ppm-hours (ppm-hrs). More detail is provided in section 4.3.3.1.1 below.
April 2022
4-2 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
for impacts in a natural ecosystem. Together this resulted in a lack of clarity in the
characterization of these effects, and a lack of important quantitative information to consider
such effects in the context of reviewing the standard, such as in judging how particular ambient
air concentrations of O3 relate to the degree of impacts on public welfare related to these effects.
With regard to the more well-established vegetation-related effects of O3 in ambient air,
the extensive evidence base considered in the 2015 and 2020 decisions documents an array of
effects, ranging from the organism scale to larger-scale impacts, such as those on populations,
communities, and ecosystems. These categories of effects which the 2013 and 2020 IS As
identified as causally or likely causally related to O3 in ambient air include: reduced vegetation
growth, reproduction, crop yield, productivity and carbon sequestration in terrestrial systems;
alteration of terrestrial community composition, belowground biogeochemical cycles and
ecosystem water cycling; and visible foliar injury (2013 ISA, Appendix 9; 2020 ISA, Appendix
8).2 Across the different types of studies, the strongest quantitative evidence available in both the
2015 and 2020 reviews for effects from O3 exposure on vegetation comes from controlled
exposure studies of growth effects in a number of species (2013 ISA, p. 1-15). Of primary
importance in considering the appropriate level of protection for the standard, both in the 2015
decision establishing it and in its 2020 retention, were the studies of O3 exposures that reduced
growth in tree seedlings from which E-R functions of seasonal relative biomass loss (RBL)3 have
been established (80 FR 65385-86, 65389-90, October 26, 2015). Consistent with advice from
the CASAC in both reviews, the Administrators considered the effects of O3 on tree seedling
growth as a surrogate or proxy for the broader array of vegetation-related effects of O3, ranging
from effects on sensitive species to broader ecosystem-level effects (80 FR 65369, 65406,
October 26, 2015; 85 FR 87319, 87399, December 31, 2020).
In their consideration of O3 effects on tree seedling growth, the Administrators in both
the 2015 and 2020 decisions ascribed importance to the intended use of the natural resources and
ecosystems potentially affected. For example, the 2015 decision considered the available
evidence and quantitative analyses in the context of an approach for considering and identifying
public welfare objectives for the revised standard (80 FR 65403-65408, October 26, 2015). In
light of the extensive evidence base of O3 effects on vegetation and associated terrestrial
2	The 2020 ISA also newly determined the evidence sufficient to infer likely causal relationships of 03 with
increased tree mortality, although it does not indicate a potential for O3 concentrations that occur in locations that
meet the current standard to cause this effect (85 FR 87319, December 31, 2020; 2020 PA, section 4.3.1).
3	These functions were developed to quantify Ch-related reduced growth in tree seedlings relative to control
treatments (without O3). In this way, RBL is the percentage by which the O3 treatment growth in a growing
season differs from the control seedlings over the same period, and the functions provide a quantitative estimate
of the reduction in a year's growth as a percentage of that expected in the absence of O3 (2013 ISA, section 9.6.2;
2020 PA, Appendix 4A).
April 2022
4-3 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
ecosystems, the Administrator, in both decisions, focused on protection against adverse public
welfare effects of Cb-related effects on vegetation, giving particular attention to such effects in
natural ecosystems, such as those in areas with protection designated by Congress, and areas
similarly set aside by states, tribes and public interest groups, with the intention of providing
benefits to the public welfare for current and future generations (80 FR 65405, October 26, 2015;
85 FR 87344, December 31, 2020).
Climate-related effects were also considered in both reviews (2013 ISA, Appendix 10,
Section 10.3; 2020 ISA, Appendix 9, Section 9.2 and 9.3). In 2020, as was the case when the
standard was set in 2015, the evidence documents tropospheric O3 as a greenhouse gas causally
related to radiative forcing, and likely causally related to subsequent effects on variables such as
temperature and precipitation. In 2020, as in 2015, limitations and uncertainties in the evidence
base affected characterization of the extent of any relationships between ground-level O3
concentrations in ambient air in the U.S. and climate-related effects and preclude quantitative
characterization of climate responses to changes in ground-level O3 concentrations in ambient air
at regional or national (vs global) scales. As a result, the EPA recognized the lack of important
quantitative tools with which to consider such effects in its review of the standard. For example,
it was not feasible to relate different patterns of ground-level O3 concentrations at the regional
(or national) scale in the U.S. with specific risks of alterations in temperature, precipitation, and
other climate-related variables. Thus, the available information did not provide a sufficient basis
for use in considering the adequacy of the secondary standard in either review (80 FR 65370,
October 26, 2015; 85 FR 87337-87339, December 31, 2020).
For quantifying effects on tree seedling growth as a surrogate or proxy for a broader array
of vegetation-related effects using the RBL metric, in 2015 and 2020 the evidence base provided
established E-R functions for seedlings of 11 tree species (80 FR 65391-92, October 26, 2015;
2014 PA, Appendix 5C; 85 FR 87307-9, 87313-4, December 31, 2020; 2020 PA, Appendix 4A).
Cumulative O3 exposure was evaluated in terms of the W126 cumulative seasonal exposure
index, an index supported by the evidence in the 2013 and 2020 IS As for this purpose and that
was consistent with advice from the CASAC in both reviews (2013 ISA, section 9.5.3, p. 9-99;
80 FR 65375, October 26, 2015; 2020 ISA, section 8.13; 85 FR 87307-8, December 31, 2020).
In judgments regarding effects that are adverse to the public welfare, the decision setting the
standard in 2015, and that retaining it in 2020, both utilized the RBL as a quantitative tool within
a larger framework of considerations pertaining to the public welfare significance of O3 effects
(80 FR 65389, October 26, 2015; 73 FR 16496, March 27, 2008; 85 FR 87339-41, December 31,
2020).
Accordingly, in both the 2015 and 2020 decisions, consideration of the appropriate public
welfare protection objective for the secondary standard gave prominence to the estimates of tree
April 2022
4-4 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
seedling growth impacts (in terms of RBL) for a range of W126 index values, developed from
the E-R functions for 11 tree species (80 FR 65391-92, Table 4, October 26, 2015; 85 FR 87339-
41, December 31, 2020). The Administrators also incorporated into their considerations the
broader evidence base associated with forest tree seedling biomass loss, including other less
quantifiable effects of potentially greater public welfare significance. That is, in drawing on
these RBL estimates, the Administrators noted they were not simply making judgments about a
specific magnitude of growth effect in seedlings that would be acceptable or unacceptable in the
natural environment. Rather, though mindful of associated uncertainties, the RBL estimates were
used as a surrogate or proxy for consideration of the broader array of related vegetation-related
effects of potential public welfare significance, which included effects on individual species and
extending to ecosystem-level effects (80 FR 65406, October 26, 2015; 85 FR 87304, December
31, 2020). This broader array of vegetation-related effects included those for which public
welfare implications are more significant but for which the tools for quantitative estimates were
more uncertain.
In the 2015 decision to revise the standard level to 70 ppb, and also the 2020 decision to
retain that standard, without revision, air quality analyses played an important role in the
Administrator's judgments. Such judgments of the Administrator in setting the standard in 2015
are briefly summarized below. These are followed by a summary of additional key aspects of the
considerations and judgments associated with the decision to retain this standard in 2020.
In using the RBL estimates as a proxy, the Administrator in 2015 focused her attention on
a revised standard that would generally limit cumulative exposures to those for which the median
RBL estimate for seedlings of the 11 species with established E-R functions would be somewhat
below 6% (80 FR 65406-07, October 26, 2015).4 She noted that the median RBL estimate was
6% for a cumulative seasonal W126 exposure index of 19 ppm-hrs (80 FR 65391-92, Table 4,
October 26, 2015). Given the information on median RBL at different W126 exposure levels,
using a 3-year cumulative exposure index for assessing vegetation effects,5 the potential for
4	The Administrator noted the CASAC view regarding 6%, most particularly the CASAC's characterization of this
level of effect in the median studied species as "unacceptably high" (Frey, 2014, pp. iii, 13, 14). These
comments were provided in the context of CASAC's considering the significance of effects associated with a
range of alternatives for the secondary standard (80 FR 65406, October 26, 2015).
5	Based on a number of considerations, the Administrator recognized greater confidence in judgments related to
public welfare impacts based on a 3-year average metric than a single-year metric, and consequently concluded it
to be appropriate to use a seasonal W126 index averaged across three years forjudging public welfare protection
afforded by a revised secondary standard. For example, she recognized uncertainties associated with
interpretation of the public welfare significance of effects resulting from a single-year exposure, and that the
public welfare significance of effects associated with multiple years of critical exposures are potentially greater
than those associated with a single year of such exposure. She additionally concluded that use of a 3-year average
metric could address the potential for adverse effects to public welfare that may relate to shorter exposure periods,
including a single year (80 FR 65404, October 26, 2015).
April 2022
4-5 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
single-season effects of concern, and CASAC comments on the appropriateness of a lower value
for a 3-year average W126 index, the Administrator concluded it was appropriate to identify a
standard that would restrict cumulative seasonal exposures to 17 ppm-hrs or lower, in terms of a
3-year W126 index, in nearly all instances (80 FR 65407, October 26, 2015). Based on such
information, available at that time, to inform consideration of vegetation effects and their
potential adversity to public welfare, the Administrator additionally judged that the RBL
estimates associated with marginally higher exposures in isolated, rare instances were not
indicative of effects that would be adverse to the public welfare, particularly in light of
variability in the array of environmental factors that can influence O3 effects in different systems
and uncertainties associated with estimates of effects associated with this magnitude of
cumulative exposure in the natural environment (80 FR 65407, October 26, 2015).
Using these objectives, the 2015 decision regarding a standard revised from the then-
existing (2008) standard was based on extensive air quality analyses that included the most
recently available data as well as air monitoring data that extended back more than a decade (80
FR 65408, October 26, 2015; Wells, 2015). These analyses evaluated the cumulative seasonal
exposure levels in locations meeting different alternative levels for a standard of the existing
form and averaging time. These analyses supported the Administrator's judgment that a standard
with a revised level in combination with the existing form and averaging time could achieve the
desired level of public welfare protection, considered in terms of cumulative exposure, quantified
as the W126 index (80 FR 65408, October 26, 2015). Based on the extensive air quality analyses
and consideration of the W126 index value associated with a median RBL of 6%, and the W126
index values at monitoring sites that met different levels for a revised standard of the existing
form and averaging time, the Administrator additionally judged that a standard level of 70 ppb
would provide the requisite protection. The Administrator noted that such a standard would be
expected to limit cumulative exposures, in terms of a 3-year average W126 exposure index, to
values at or below 17 ppm-hrs, in nearly all instances, and accordingly, to eliminate or virtually
eliminate cumulative exposures associated with a median RBL of 6% or greater (80 FR 65409,
October 26, 2015).
The 2015 decision also took note of the well-recognized evidence for visible foliar injury
and crop yield effects. However, the RBL information available for seedlings of a set of 11 tree
species was judged to be more useful (particularly in a role as surrogate for the broader array of
vegetation-related effects) in informing judgments regarding the nature and severity of effects
associated with different air quality conditions and associated public welfare significance than
the available information on visible foliar injury and crop yield effects (80 FR 65405-06,
October 26, 2015). With regard to visible foliar injury, while the Administrator recognized the
potential for this effect to affect the public welfare in the context of affecting value ascribed to
April 2022
4-6 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
natural forests, particularly those afforded special government protection, she also recognized
limitations in the available information that might inform consideration of potential public
welfare impacts related to this vegetation effect noting the significant challenges in judging the
specific extent and severity at which such effects should be considered adverse to public welfare
(80 FR 65407, October 26, 2015).6 Similarly, while Cb-related growth effects on agricultural and
commodity crops had been extensively studied and robust E-R functions developed for a number
of species, the Administrator found this information less useful in informing judgments
regarding an appropriate level of public welfare protection (80 FR 65405, October 26, 2015).7
In summary, the 2015 decision focused primarily on the information related to trees and
growth impacts in identifying the public welfare objectives for the revised secondary standard
(80 FR 65409-65410, October 26, 2015). In this context, the Administrator in 2015 judged that
the 70 ppb standard would protect natural forests in Class I and other similarly protected areas
against an array of adverse vegetation effects, most notably including those related to effects on
growth and productivity in sensitive tree species. She additionally judged that the new standard
would be sufficient to protect public welfare from known or anticipated adverse effects. These
judgments by the Administrator at that time appropriately recognized that the CAA does not
require that standards be set at a zero-risk level, but rather at a level that reduces risk sufficiently
so as to protect the public welfare from known or anticipated adverse effects.
In 2020, as in 2015, the Administrator considered the available information regarding the
appropriate O3 exposure metric to employ in assessing adequacy of air quality control in
protecting against RBL. In addition to finding it appropriate to continue to consider the seasonal
W126 index averaged over a 3-year period to estimate median RBL (as was concluded in 2015),
the Administrator in 2020 also judged it appropriate to also consider other metrics including peak
hourly concentrations8 (85 FR 87344, December 2020). With regard to these conclusions, his
6	These limitations included the lack of established E-R functions that would allow prediction of visible foliar injury
severity and incidence under varying air quality and environmental conditions, a lack of consistent quantitative
relationships linking visible foliar injury with other Ch-induced vegetation effects, such as growth or related
ecosystem effects, and a lack of established criteria or objectives relating reports of foliar injury with public
welfare impacts (80 FR 65407, October 26, 2015).
7	With respect to commercial production of commodities, the Administrator noted the difficulty in discerning the
extent to which Ch-related effects on commercially managed vegetation are adverse from a public welfare
perspective, given that the extensive management of such vegetation (which, as the CAS AC noted, may reduce
yield variability) may also to some degree mitigate potential Ch-related effects. Management practices are highly
variable and are designed to achieve optimal yields, taking into consideration various environmental conditions.
Further, changes in yield of commercial crops and commercial commodities, such as timber, may affect producers
and consumers differently, complicating the assessment of overall public welfare effects still further (80 FR
65405, October 26, 2015).
8	Both the 2020 and 2013 IS As reference the longstanding recognition of the risk posed to vegetation of peak hourly
O3 concentrations (e.g., "[h]igher concentrations appear to be more important than lower concentrations in
April 2022
4-7 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
considerations included the extent of conceptual similarities of the 3-year average W126 index to
some aspects of the derivation approach for the established E-R functions, the context of RBL as
a proxy (as recognized above), and limitations associated with a reliance solely on W126 index
as a metric to control exposures that might be termed "unusually damaging"9 (85 FR 877339-40,
December 31, 2020).
With regard to the derivation and application of the established E-R functions, the 2020
review recognized several factors to contribute uncertainty and some resulting imprecision or
inexactitude to RBL estimated from single-year seasonal W126 index values (85 FR 49900-01,
August 14, 2020; 2020 PA sections 4.5.1.2 and 4.5.3).10 Additionally recognized was the
qualitative and conceptual nature of our understanding, in many cases, of relationships of O3
effects on plant growth and productivity with larger-scale impacts, such as those on populations,
communities and ecosystems. From these considerations, it was judged that use of a seasonal
RBL averaged over multiple years, such as a 3-year average, is reasonable, and provides a more
eliciting a response" [ISA, p. 8-180]; "higher hourly concentrations have greater effects on vegetation than lower
concentrations" [2013 ISA, p. 91-4] "studies published since the 2006 O3 AQCD do not change earlier
conclusions, including the importance of peak concentrations, ... in altering plant growth and yield" [2013 ISA,
p. 9-117]). While the evidence does not indicate a particular threshold number of hours at or above 100 ppb (or
another reference point for elevated concentrations), the evidence of greater impacts from higher concentrations
(particularly with increased frequency) and the air quality analyses that document variability in such
concentrations for the same W126 index value led the Administrator to judge such a multipronged approach to be
needed to ensure appropriate consideration of exposures of concern and the associated protection from them
afforded by the secondary standard (85 FR 87340, December 31, 2020).
9	In its discussion regarding the EPA's use of a 3-year average W126 index, the 2019 court decision remanding the
2015 standard back to the EPA referenced advice from the CAS AC in the 2015 review on protection against
"unusually damaging years." Use of this term occurs in the 2014 CASAC letter on the second draft PA (Frey,
2014). Most prominently, the CASAC defined as damage "injury effects that reach sufficient magnitude as to
reduce or impair the intended use or value of the plant to the public, and thus are adverse to public welfare" (Frey,
2014, p. 9). We also note that the context for the CASAC's use of the phrase "unusually damaging years" in the
2015 review is in considering the form and averaging time for a revised secondary standard in terms of a W126
index (Frey, 2014, p. 13), which as discussed below is relatively less controlling of high-concentration years
(whether as a single year index or averaged over three years) than the current secondary standard and its fourth
highest daily maximum 8-hour metric (85 FR 87327, December 31, 2020).
10	The E-R functions were derived mathematically from studies of different exposure durations (varying from
shorter than one to multiple growing seasons) by applying adjustments so that they would yield estimates
normalized to the same period of time (season). Accordingly, the estimates may represent average impact for a
season, and have compatibility with W126 index averaged over multiple growing seasons or years (85 FR 87326,
December 31, 2020; 2020 PA, section 4.5.1.2, Appendix 4A, Attachment 1). The available information also
indicated that the patterns of hourly concentrations (and frequency of peak concentrations, e.g., at/above 100 ppb)
in O3 treatments on which the E-R functions are based differ from the patterns in ambient air meeting the current
standard across the U.S. today (85 FR 87327, December 31, 2020). Additionally noted was the year-to-year
variability of factors other than O3 exposures that affect tree growth in the natural environment (e.g., related to
variability in soil moisture, meteorological, plant-related and other factors), that have the potential to affect O3 E-
R relationships (ISA, Appendix 8, section 3.12; 2013 ISA section 9.4.8.3; PA, sections 4.3 and 4.5). All of these
considerations contributed to the finding of a consistency of the use of W126 index averaged over multiple years
with the approach used in deriving the E-R function, and with other factors that may affect growth in the natural
environment (85 FR 87340, December 31, 2020).
April 2022
4-8 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
stable and well-founded RBL estimate for its use as a proxy to represent the array of vegetation-
related effects identified above. More specifically, the Administrator concluded that the use of an
average seasonal W126 index derived from multiple years (with their representation of
variability in environmental factors) provides an appropriate representation of the evidence and
attention to the identified considerations. In so doing, he found that a sole reliance on single year
W126 estimates for reaching judgments with regard to magnitude of O3 related RBL and
associated judgments of public welfare protection would ascribe a greater specificity and
certainty to such estimates than supported by the evidence. Rather, consistent with the judgment
of the prior Administrator, the Administrator in 2020 found it appropriate, for purposes of
considering public welfare protection from effects for which RBL is used as a proxy, to primarily
consider W126 index in terms of a 3-year average metric (85 FR 87339-87340, December 31,
2020).
With regard to the EPA's use of a 3-year average W126 index to assess protection from
RBL, the 2020 decision additionally took into account the 2019 court remand on this issue,
including the remand's reference to protection against "unusually damaging years." (85 FR
87325-87328, December 31, 2020). Accordingly, the EPA considered air quality analyses of
peak hourly concentrations in the context of considering protection against "unusually damaging
years." With regard to this caution, and in the context of controlling exposure circumstances of
concern (e.g., for growth effects, among others), the EPA considered air quality analyses that
investigated the annual occurrence of elevated hourly O3 concentrations which may contribute to
vegetation exposures of concern (2020 PA, Appendix 2A, section 2A.2; Wells, 2020). These air
quality analyses illustrate limitations of the W126 index (whether in terms of a 3-year average or
a single year) for the purpose of controlling peak concentrations,11 and also the strengths of the
current standard in this regard. The air quality analyses show that the form and averaging time of
the existing standard, in addition to controlling cumulative exposures in terms of W126 (as found
in the 2015 review), is much more effective than the W126 index in limiting peak concentrations
(e.g., hourly O3 concentrations at or above 100 ppb)12 and in limiting number of days with any
such hours (Wells, 2020, e.g., Figures 4, 5, 8, 9 compared to Figures 6, 7, 10 and 11).13 Thus, the
W126 index, by its very definition, and as illustrated by the air quality data analyses, does not
11	The W126 index cannot, by virtue of its definition, always differentiate between air quality patterns with high
peak concentrations and those without such concentrations.
12	As described in section 4.3.3 below, the occurrence of high concentrations (including those at or above 100 ppb
[e.g., Smith, 2012; Smith et al., 2012]), as well as cumulative exposures influence the effects of O3 on plants.
13	With regard to the existing standard, historical air quality data extending back to 2000 additionally show the
appreciable reductions in peak concentrations that have been achieved in the U.S. as air quality has improved
under O3 standards of the existing form and averaging time (Wells, 2020, Figures 12 and 13).
April 2022
4-9 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
provide specificity with regard to year-to-year variability in elevated hourly O3 concentrations
with the potential to contribute to the "unusually damaging years" that the CASAC had identified
for increased concern in the 2015 review. As a result, the 2020 decision found that a standard
based on a W126 index (either a 3-year or a single-year index) would not be expected to provide
effective control of the peak concentrations that may contribute to "unusually damaging years"
for vegetation.14 Based on all of the above, the 2020 decision concluded that control of such
years is a characteristic of the existing standard (the effectiveness of which is demonstrated by
the air quality analyses), and that that use of a seasonal W126 averaged over a 3-year period,
which is the design value period for the current standard, to estimate median RBL using the
established E-R functions, in combination with a broader consideration of air quality patterns,
such as peak hourly concentrations, is appropriate for considering the public welfare protection
provided by the standard (85 FR 87340-87341, December 31, 2020).
With regard to O3 effects on crop yield for which there is long-standing evidence,
qualitative and quantitative, of the reducing effect of O3 on the yield of many crops and a
potential for public welfare significance, the 2020 decision concluded that the existing standard
provides adequate protection of public welfare related to crop yield loss (85 FR 87342,
December 31, 2020). Key considerations in this conclusion included the established E-R
functions for 10 crops and the estimates of RYL derived from them (2020 ISA, 2020 PA,
Appendix 4A, section 4A.1, Table 4A-4), as well as the existence of a number of complexities
related to the heavy management of many crops to obtain a particular output for commercial
purposes, and related to other factors (85 FR 87341-87342, December 31, 2020). For example,
the Administrator considered the extensive management of agricultural crops that occurs to elicit
optimum yields (e.g., through irrigation and usage of soil amendments, such as fertilizer) to be
relevant in evaluating the extent of RYL estimated from experimental O3 exposures that should
be judged adverse to the public welfare. With regard to the E-R functions for RYL for 10 crops,
the Administrator considered the air quality data with regard to the W126 index levels and
corresponding estimated RYL for the median species. He also took into consideration the
extensive management of agricultural crops, and the complexities associated with identifying
adverse public welfare effects for market-traded goods (where producers and consumers may be
impacted differently). Further, he noted that the secondary standard is not intended to protect
against all known or anticipated 03-related effects, but rather those that are judged to be adverse
to the public welfare. The air quality data indicated that the current standard generally maintains
14 From these analyses, the Administrator concluded that the form and averaging time of the current standard is
effective in controlling peak hourly concentrations and that a W126 index based standard would be much less
effective in providing the needed protection against years with such elevated and potentially damaging hourly
concentrations.
April 2022
4-10 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
air quality at a W126 index below 17 ppm-hrs, with few exceptions, and would accordingly limit
the associated estimates of median RYL below 5.1% (based on experimental O3 exposures), a
level which the Administrator judged would not constitute an adverse effect on public welfare.
Therefore, he concluded that the current standard provides adequate protection of public welfare
related to crop yield loss and did not need to be revised to provide additional protection against
this effect (85 FR 87342, December 31, 2020).
With regard to visible foliar injury, the Administrator considered the question of a level
of air quality that would provide protection against visible foliar injury related effects known or
anticipated to cause adverse effects to the public welfare. Based on the evidence and associated
quantitative analyses in this review, summarized in the 2020 PA, the Administrator's judgment
reflected his recognition of less confidence and greater uncertainty in the existence of adverse
public welfare effects with lower O3 exposures (85 FR 87342-87344, December 31, 2020).
While recognizing there to be a paucity of established approaches for interpreting specific levels
of severity and extent of foliar injury in natural areas with regard to impacts on the public
welfare (e.g., related to recreational services), the Administrator recognized that injury to whole
stands of trees of a severity apparent to the casual observer (e.g., when viewed as a whole from a
distance) would reasonably be expected to affect recreational values and thus pose a risk of
adverse effects to the public welfare. He further noted that the available information did not
provide for specific characterization of the incidence and severity that would not be expected to
be apparent to the casual observer, nor for clear identification of the pattern of O3 concentrations
that would provide for such a situation. In recognizing that quantitative analyses and evidence
are lacking that might support a more precise identification of a severity of visible foliar injury
and extent of occurrence that might be judged adverse to the public welfare, the Administrator
considered the USFS system for interpreting visible foliar injury impacts in surveys conducted at
biomonitoring sites (biosites) across the U.S. from 1994 through 2011. At these sites, the USFS
followed a national protocol that includes a scoring system with descriptors for biosite index
(BI)15 scores of differing magnitude for his purposes in this regard. More specifically, he
concluded that findings of BI scores categorized as "moderate to severe" injury by the USFS
scheme would be an indication of visible foliar injury occurrence that, depending on extent and
severity, may raise public welfare concerns. In this framework, the Administrator considered the
2020 PA evaluations of the available information and what that information indicated with
regard to patterns of air quality of concern for such an occurrence, and the extent to which they
are expected to occur in areas that meet the current standard. For example, the incidence of
nonzero BI scores, and particularly of relatively higher scores such as those above 15, classified
15 The BI is a measure of the severity of Ch-induced visible foliar injury observed at each biosite (Smith, 2012).
April 2022
4-11 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
as indicative of "moderate to severe "injury in the USFS scheme appear to markedly increase
only with W126 index values above 25 ppm-hrs. He further took note of the multiple published
studies analyzing the USFS data across multiple years and multiple U.S. regions with regard to
metrics intended to quantify influential aspects of O3 air quality, which indicated a potential role
for an additional metric related to the occurrence of days with relatively high hourly
concentrations (e.g., number of days with a 1-hour concentration at or above 100 ppb [2020 PA,
section 4.5.1.2]). In light of this evidence and the 2020 PA analyses of these data, the
Administrator judged that W126 index values at or below 25 ppm-hrs, when in combination with
infrequent occurrences of hourly concentrations at or above 100 ppb, would not be anticipated to
pose risk of visible foliar injury of an extent and severity so as to be adverse to the public welfare
(85 FR 87343, December 31, 2020).
With these conclusions in mind, the Administrator considered the available air quality
analyses (85 FR 87316-18, December 31, 2020; 2020 PA, Appendix 4C, section 4C.3; Appendix
4D; Wells, 2020). Together these analyses indicated that a W126 index above 25 ppm-hrs (either
as a 3-year average or in a single year) is not seen to occur at monitoring locations where the
current standard is met (including in or near Class I areas), and that, in fact, values above 17 or
19 ppm-hrs are rare and that days with any hourly concentrations at or above 100 ppb at
monitoring sites that meet the current standard are uncommon. Based on these findings, the
Administrator concluded that the current standard provides control of air quality conditions that
contribute to increased BI scores and to scores of a magnitude indicative of "moderate to severe"
foliar injury. Further, he noted the 2020 PA finding that the information from the USFS biosite
monitoring program, particularly in locations meeting the current standard or with W126 index
estimates likely to occur under the current standard, does not indicate a significant extent and
degree of injury (e.g., based on analyses of BI scores in the PA, Appendix 4C) or specific
impacts on recreational or related services for areas, such as wilderness areas or national parks,
thus giving credence to the associated 2020 PA conclusion that the evidence indicates that areas
that meet the current standard are unlikely to have BI scores reasonably considered to be impacts
of public welfare significance (85 FR 87344, December 31, 2020).
Before reaching a final decision on the standard, the Administrator, in returning to his
primary focus on RBL in its role as proxy for the broader array of vegetation-related effects of
O3, further considered the available analyses of both the air quality data newly available in the
2020 review and of historical air quality at sites across the U.S., particularly including those sites
in or near Class I areas, for which the findings were consistent with the air quality analyses
April 2022
4-12 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
available in the 2015 review.16 That is, in virtually all design value periods between 2000 and
2018 and all locations at which the current standard was met across the 19 years and 17 design
value periods (in more than 99.9% of such observations), the 3-year average W126 metric was at
or below 17 ppm-hrs. Further, in all such design value periods and locations the 3-year average
W126 index was at or below 19 ppm-hrs (85 FR 87344, December 31, 2020).
The Administrator additionally considered the protection provided by the current
standard from the occurrence of O3 exposures within a single year with potentially damaging
consequences, including a significantly increased incidence of areas with visible foliar injury that
might be judged moderate to severe. He gave particular focus to BI scores above 15, termed
"moderate to severe injury" by the USFS categorization scheme (85 FR 87344, December 31,
2020; 2020 PA, sections 4.3.3.2, 4.5.1.2 and Appendix 4C). As discussed above, the incidence of
USFS sites with BI scores above 15 markedly increases with W126 index estimates above 25
ppm-hrs, a magnitude of W126 index indicated by the air quality analysis to be scarce at sites
that meet the current standard, with just a single occurrence across all U.S. sites with design
values meeting the current standard in the 19-year historical dataset dating back to 2000 (2020
PA, section 4.4, and Appendix 4D). Further, in light of the evidence indicating that peak short-
term concentrations (e.g., of durations as short as one hour) may also play a role in the
occurrence of visible foliar injury, the Administrator additionally took note of the air quality
analyses of hourly concentrations (2020 PA, Appendix 2A; Wells 2020). These analyses of data
from the past 20 years show a declining trend in 1-hour daily maximum concentrations mirroring
the declining trend in design values, supporting the 2020 PA conclusion that the form and
averaging time of the current standard provides appreciable control of peak 1-hour
concentrations. Furthermore, these analyses for the period from 2000 to 2018 indicate that sites
meeting the current standard had only a few days (up to just seven) with hourly concentrations at
or above 100 ppb (Wells, 2020). In light of these findings from the air quality analyses and
considerations in the 2020 PA, both with regard to 3-year average W126 index values at sites
meeting the current standard and the rarity of such values at or above 19 ppm-hrs, and with
regard to single-year W126 index values at sites meeting the current standard, and the rarity of
such values above 25 ppm-hrs, as well as with regard to the appreciable control of 1-hour daily
maximum concentrations, the Administrator judged that the current standard provides adequate
protection from air quality conditions with the potential to be adverse to the public welfare (85
FR 87344, December 31, 2020).
16 These data are distributed across all nine NOAA climate regions and 50 states, although some geographic areas
within specific regions and states may be more densely covered and represented by monitors than others (2020
PA, Appendix 4D).
April 2022
4-13 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
In reaching his conclusion on the current secondary O3 standard, the Administrator
recognized, as is the case in NAAQS reviews in general, his decision depended on a variety of
factors, including science policy judgments and public welfare policy judgments, as well as the
available information. In the 2020 decision, the Administrator gave primary attention to the
principal effects of O3 as recognized in the current ISA, the 2013 ISA and past AQCDs, and for
which the evidence is strongest (e.g., growth, reproduction, and related larger-scale effects, as
well as visible foliar injury). With regard to growth and the categories of effects identified above
for which RBL has been identified for use as a proxy, based on all of the identified
considerations, including the discussion of air quality immediately above, the Administrator
judged the current standard to provide adequate protection for air quality conditions with the
potential to be adverse to the public welfare. Further, with regard to visible foliar injury, the
Administrator concluded that the available information on visible foliar injury and with regard to
air quality analyses that may be informative to identification of air quality conditions associated
with appreciably increased incidence and severity of BI scores at USFS biomonitoring sites, and
with particular attention to Class I and other areas afforded special protection, indicated the
current standard to provide adequate protection from visible foliar injury of an extent or severity
that might be anticipated to be adverse to the public welfare.
In summary, the 2020 decision was based on consideration of the public welfare
protection afforded by the secondary O3 standard from identified 03-related welfare effects, and
from their potential to present adverse effects to the public welfare, and also on judgments
regarding what the available evidence, quantitative information, and associated uncertainties and
limitations (such as those identified above) indicate with regard to the protection provided from
the array of O3 welfare effects. As a whole, the decision found that this information did not
indicate the current standard to allow air quality conditions with implications of concern for the
public welfare. Based on all of the identified considerations, as well as consideration of advice
from the CASAC17 and public comment, and including consideration of the available evidence
and quantitative exposure/risk information, the Administrator concluded the current secondary
standard to be requisite to protect the public welfare from known or anticipated adverse effects
of O3 and related photochemical oxidants in ambient air, and thus that the standard should be
retained without revision (85 FR 87345, December 31, 2020).
17 Among other things, in the 2020 letter communicating the CASAC's comments on the 2019 draft PA, the CASAC
advised EPA that it "finds, in agreement with the EPA, that the available evidence does not reasonably call into
question the adequacy of the current secondary ozone standard and concurs that it should be retained" (Cox, 2020, p.
1). It further stated that the approach described in the draft PA to considering the evidence for welfare effects ' 'is
laid out very clearly, thoroughly discussed and documented, and provided a solid scientific underpinning for the
EPA conclusion leaving the current secondary standard in place'' (85 FR 87318-87319, December 31, 2020).
April 2022
4-14 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
4.2 GENERAL APPROACH AND KEY ISSUES
As in the case for secondary standard reviews, this reconsideration of the 2020 decision
on the secondary standard is fundamentally based on using the Agency's assessment of the
scientific evidence and associated quantitative analyses to inform the Administrator's judgments
regarding a secondary standard that is requisite to protect the public welfare from known or
anticipated adverse effects. This approach builds on the substantial assessments and evaluations
performed over the course of O3 NAAQS reviews to inform our understanding of the key-policy
relevant issues in this reconsideration of the 2020 decision. As noted above, we are also
considering the court's 2019 decision on the O3 secondary standard, particularly with regard to
issues raised by the court in its remand of the standard (recognized in section 4.1.2 above) as was
also done as part of the 2020 decision on the standard.
The evaluations in the PA, of the scientific assessments in the ISA (building on prior such
assessments) augmented by quantitative air quality and exposure analyses, are intended to inform
the Administrator's public welfare policy judgments and conclusions, including his decisions
regarding the O3 standards. The PA considers the potential implications of various aspects of the
scientific evidence, the air quality, exposure or risk-based information, and the associated
uncertainties and limitations. Thus, the approach for this PA involves evaluating the available
scientific and technical information to address a series of key policy-relevant questions using
both evidence- and exposure/risk-based considerations. Together, consideration of the full set of
evidence and information available will inform the answer to the following initial overarching
question:
Do the available scientific evidence and exposure-/risk-based information support or
call into question the adequacy of the public welfare protection afforded by the current
secondary O3 standard?
In reflecting on this question in the remaining sections of this chapter, we consider the
body of scientific evidence assessed in the ISA, and considered as a basis for developing or
interpreting air quality and exposure analyses, including whether it supports or calls into question
the scientific conclusions reached in the 2020 review regarding welfare effects related to
exposure to O3 in ambient air. Information that may be informative to public policy judgments
on the significance or adversity of key effects on the public welfare is also considered.
Additionally, the available exposure and air quality information is considered, including with
regard to the extent to which it may continue to support judgments made in previous reviews.
Further, in considering this question with regard to the secondary O3 standard, we give particular
attention to exposures and risks for effects with the greatest potential for public welfare
significance. Evaluation of the available scientific evidence and exposure/risk information with
regard to consideration of the current standard and the overarching question above focuses on
April 2022
4-15 External Review Draft - Do Not Quote or Cite

-------
1	key policy-relevant issues by addressing a series of questions on specific topics. Figure 3-1
2	summarizes, in general terms, the approach to considering the available information in the
3	context of policy-relevant questions pertaining to the secondary standard.
4
April 2022
4-16 External Review Draft - Do Not Quote or Cite

-------
Adequacy of Current Standard
Evidence- Based ConsideraSons
Does available evidence and related unoertaWes
strengthen orcalirfc quesaon prior condusions?
•	Evidence of welfare effecs not previously IdenSed?
ฆ	Evidence of elfecs at lower tevefc or for different
exposure circususances?
•	Evidence for vegetaion elfecs from cumulative
exposures allowed by the current standard?
ฆ	Uncertain&es klenifed previously are reduced or
new uncertanfes have emerged ?
1	
Exposure and Risk-Based Considerations
> Nature, magniude, and importance ofessma'ed
exposures and risks associated wiii meeSng the
current ssandard?
VUncertaWes in che exposure and risk esamates?
<

Elernents of Poteraal Aternaave Standards
^ Indicator, Averaging Time, Form, Level
.

Potential Alternative Standards for Consideration
1
2	Figure 4-1. Overview of general approach for the secondary O3 standard.
April 2022
4-17 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
The Agency's approach with regard to the O3 secondary standard is consistent with the
requirements of the provisions of the CAA related to the review of NAAQS and with how the
EPA and the courts have historically interpreted these provisions. As discussed in section 1.2
above, these provisions require the Administrator to establish secondary standards that, in the
Administrator's judgment, are requisite (i.e., neither more nor less stringent than necessary) to
protect the public welfare from known or anticipated adverse effects associated with the presence
of the pollutant in the ambient air. Consistent with the Agency's approach across NAAQS
reviews, the approach of this PA to informing the Administrator's judgments is based on a
recognition that the available evidence generally reflects continuums that include ambient air
exposures for which scientists generally agree that effects are likely to occur through lower
levels at which the likelihood and magnitude of response become increasingly uncertain. The
CAA does not require that standards be set at a zero-risk level, but rather at a level that reduces
risk sufficiently so as to protect the public welfare from known or anticipated adverse effects.
The Agency's decisions on the adequacy of the current secondary standard and, as
appropriate, on any potential alternative standards considered in a review, are largely public
welfare policy judgments made by the Administrator. The four basic elements of the NAAQS
(i.e., indicator, averaging time, form, and level) are considered collectively in evaluating the
protection afforded by the current standard, or by any alternatives considered. Thus, the
Administrator's final decisions in such reviews draw upon the scientific information and
analyses about welfare effects, environmental exposures and risks, and associated public welfare
significance, as well as judgments about how to consider the range and magnitude of
uncertainties that are inherent in the scientific evidence and analyses.
4.3 WELFARE EFFECTS EVIDENCE
The welfare effects evidence on which this PA for the reconsideration of the 2020
decision on the O3 secondary standard will focus is the evidence described in the 2020 ISA and
prior ISAs or AQCDs. As described in section 1.5 above, the EPA has provisionally considered
more recently available studies that were raised in public comments in the 2020 review or were
identified in a literature search that the EPA conducted for this reconsideration of more recently
available controlled human exposure studies (Luben et al., 2020; Duffney et al. 2022). The
provisional consideration of these studies concluded that, taken in context, the associated
information and findings did not materially change any of the broad scientific conclusions of the
ISA regarding the health and welfare effects of O3 in ambient air or warrant reopening the air
quality criteria for this review. Thus, the discussion below focuses on the welfare effects
evidence assessment, with associated conclusions, as described in the 2020 ISA.
April 2022
4-18 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
4.3.1 Nature of Effects
The welfare effects evidence base includes more than fifty years of extensive research on
the phytotoxic effects of O3, conducted both in and outside of the U.S., that documents the
impacts of O3 on plants and their associated ecosystems (1978 AQCD, 1986 AQCD, 1996
AQCD, 2006 AQCD, 2013 ISA, 2020 ISA). As has been long established, O3 can interfere with
carbon gain (photosynthesis) and allocation of carbon within the plant, making fewer
carbohydrates available for plant growth, reproduction, and/or yield (1996 AQCD, pp. 5-28 and
5-29). For seed-bearing plants, reproductive effects can include reduced seed or fruit production
or yield. The strongest evidence for effects from O3 exposure on vegetation was recognized at
the time of the 2015 review to be from controlled exposure studies, which "have clearly shown
that exposure to O3 is causally linked to visible foliar injury, decreased photosynthesis, changes
in reproduction, and decreased growth" in many species of vegetation (2013 ISA, p. 1-15). Such
effects at the plant scale can also be linked to an array of effects at larger spatial scales (and
higher levels of biological organization), with the evidence available in the 2015 review
indicating that "O3 exposures can affect ecosystem productivity, crop yield, water cycling, and
ecosystem community composition" (2013 ISA, p. 1-15, Chapter 9, section 9.4). Beyond its
effects on plants, the evidence in the 2015 review also recognized O3 in the troposphere as a
major greenhouse gas (ranking behind carbon dioxide and methane in importance), with
associated radiative forcing and effects on climate, with accompanying "large uncertainties in the
magnitude of the radiative forcing estimate ... making the impact of tropospheric O3 on climate
more uncertain than the effect of the longer-lived greenhouse gases (2013 ISA, sections 10.3.4
and 10.5.1 [p. 10-30]).
• Does the available evidence alter prior conclusions regarding the nature of welfare
effects attributable to O3 in ambient air? Is there new evidence on welfare effects
beyond those identified in the 2015 review?
The available evidence supports, sharpens, and expands somewhat on the conclusions
reached in the 2015 review (ISA, Appendices 8 and 9). Consistent with the previously available
evidence, the available evidence describes an array of O3 effects on vegetation and related
ecosystem effects, as well as the role of tropospheric O3 in radiative forcing and subsequent
climate-related effects. The ISA concludes there to be causal relationships between O3 and
visible foliar injury, reduced vegetation growth and reduced plant reproduction,18 as well as
reduced yield and quality of agricultural crops, reduced productivity in terrestrial ecosystems,
18 The 2013 ISA did not include a separate causality determination for reduced plant reproduction. Rather, it was
included with the conclusion of a causal relationship of O3 with reduced vegetation growth (ISA, Table IS-12).
April 2022
4-19 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
alteration of terrestrial community composition19, and alteration of belowground biogeochemical
cycles (ISA, section IS.5). The ISA also concludes there likely to be a causal relationship
between O3 and alteration of ecosystem water cycling, reduced carbon sequestration in terrestrial
ecosystems, and with increased tree mortality (ISA, section IS.5). Additionally, newly available
evidence in the 2020 ISA augments more limited previously available evidence related to insect
interactions with vegetation, contributing to the ISA conclusion that the evidence is sufficient to
infer that there are likely to be causal relationships between O3 exposure and alteration of plant-
insect signaling (ISA, Appendix 8, section 8.7) and of insect herbivore growth and reproduction
(ISA, Appendix 8, section 8.6). Thus, prior conclusions continue to be supported and conclusions
are also reached in the 2020 ISA for a few new areas based on the now expanded evidence.
As in the 2015 review, the strongest evidence and the associated findings of causal or
likely causal relationships with O3 in ambient air, and the quantitative characterizations of
relationships between O3 exposure and occurrence and magnitude of effects are for vegetation
effects. The scales of these effects range from the individual plant scale to the ecosystem scale,
with potential for impacts on the public welfare (as discussed in section 4.3.2 below). The
following summary addresses the identified vegetation-related effects of O3 across these scales.
Visible foliar injury has long been used as a bioindicator of O3 exposures, although it is
not always a reliable indicator of other negative effects on vegetation (ISA, sections IS.5.1.2 and
8.2, and Appendix 8, section 8.2; 2013 ISA, section 9.4.2; 2006 AQCD, 1996 AQCD, 1986
AQCD, 1978 AQCD). More specifically, ozone-induced visible foliar injury symptoms on
certain tree and herbaceous species, such as black cherry, yellow-poplar and common milkweed,
have long been considered diagnostic of exposure to elevated O3 based on the consistent
association established with experimental evidence (ISA, Appendix 8, section 8.2; 2013 ISA, p.
1-10).20 The available evidence, consistent with that in past reviews, indicates that "visible foliar
injury usually occurs when sensitive plants are exposed to elevated ozone concentrations in a
predisposing environment," with a major factor for such an environment being the amount of soil
moisture available to the plant (ISA, Appendix 8, p. 8-23; 2013 ISA, section 9.4.2). The
significance of O3 injury at the leaf and whole plant levels also depends on an array of factors
that include the amount of total leaf area affected, age of plant, size, developmental stage, and
degree of functional redundancy among the existing leaf area (ISA, Appendix 8, section 8.2;
19	The 2013 ISA concluded alteration of terrestrial community composition to be likely causally related to 03 based
on the then available information (ISA, Table IS-12).
20	As described in the ISA, "[t]ypical types of visible injury to broadleaf plants include stippling, flecking, surface
bleaching, bifacial necrosis, pigmentation (e.g., bronzing), and chlorosis or premature senescence and [t]ypical
visible injury symptoms for conifers include chlorotic banding, tip burn, flecking, chlorotic mottling, and
premature senescence of needles" (ISA, Appendix 8, p. 8-13).
April 2022
4-20 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
2013 ISA, section 9.4.2). Such modifying factors contribute to the difficulty in quantitatively
relating visible foliar injury to other vegetation effects (e.g., individual tree growth, or effects at
population or ecosystem levels), such that visible foliar injury "is not always a reliable indicator
of other negative effects on vegetation" (ISA, Appendix 8, section 8.2; 2013 ISA, p. 9-39).21
Effects of O3 on physiology of individual plants at the cellular level, such as through
photosynthesis and carbon allocation, can impact plant growth and reproduction (ISA, section
IS.5.1.2, Appendix 8, sections 8.3 and 8.4; 2013 ISA, p. 9-42). The available studies come from
a variety of different study types that cover an array of different species, effects endpoints, and
exposure methods and durations. In addition to studies on scores of plant species that have found
O3 to reduce plant growth, the evidence accumulated over the past several decades documents O3
alteration of allocation of biomass within the plant and plant reproduction (ISA, Appendix 8,
sections 8.3 and 8.4; 2013 ISA, p. 1-10). The biological mechanisms underlying the effect of O3
on plant reproduction include "both direct negative effects on reproductive tissues and indirect
negative effects that result from decreased photosynthesis and other whole plant physiological
changes" (ISA, section IS.5.1.2). A newly available meta-analysis of more than 100 studies
published between 1968 and 2010 summarizes effects of O3 on multiple measures of
reproduction (ISA, Appendix 8, section 8.4.1).
Studies involving experimental field sites have also reported effects on measures of plant
reproduction, such as effects on seeds (reduced weight, germination, and starch levels) that could
lead to a negative impact on species regeneration in subsequent years, and bud size that might
relate to a delay in spring leaf development (ISA, Appendix 8, section 8.4; 2013 ISA, section
9.4.3; Darbah et al., 2007, Darbah et al., 2008). A more recent laboratory study reported 6-hour
daily O3 exposures of flowering mustard plants to 100 ppb during different developmental stages
to have mixed effects on reproductive metrics. While flowers exposed early versus later in
development produced shorter fruits, the number of mature seeds per fruit was not significantly
affected by flower developmental stage of exposure (ISA, Appendix 8, section 8.4.1; Black et al.,
2012). Another study assessed seed viability for a flowering plant in laboratory and field
21 Similar to the 2013 ISA, the 2020 ISA states the following (ISA, pp. 8-23 to 8-24).
Although visible injury is a valuable indicator of the presence of phytotoxic concentrations of
ozone in ambient air, it is not always a reliable indicator of other negative effects on vegetation
[e.g., growth, reproduction; U.S. EPA (2013)]. The significance of ozone injury at the leaf and
whole-plant levels depends on how much of the total leaf area of the plant has been affected, as
well as the plant's age, size, developmental stage, and degree of functional redundancy among the
existing leaf area (U.S. EPA, 2013). Previous ozone AQCDs have noted the difficulty in relating
visible foliar injury symptoms to other vegetation effects, such as individual plant growth, stand
growth, or ecosystem characteristics (U.S. EPA, 2006, 1996). Thus, it is not presently possible to
determine, with consistency across species and environments, what degree of injury at the leaf
level has significance to the vigor of the whole plant.
April 2022
4-21 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
conditions, finding effects on seed viability of O3 exposures (90 and 120 ppb) under laboratory
conditions but less clear effects under more field-like conditions (ISA, Appendix 8, section 8.4.1;
Landesmann et al., 2013).
With regard to agricultural crops, the current evidence base, as in the 2015 review, is
sufficient to infer a causal relationship between O3 exposure and reduced yield and quality (ISA,
section IS.5.1.2). The evidence in the current ISA is augmented by new research in a number of
areas, including studies on soybean, wheat, and other non-soy legumes. The new information
assessed in the ISA remains consistent with the conclusions reached in the 2013 ISA (ISA,
section IS.5.1.2).
The evidence base for trees includes a number of studies conducted at the Aspen free-air
carbon-dioxide and ozone enrichment (FACE) experiment site in Wisconsin (that operated from
1998 through 2011) and also available in the 2015 review (ISA, IS.5.1 and Appendix 8, section
8.1.2.1; 2013 ISA, section 9.2.4). These studies, which occurred in a field setting (more similar
to natural forest stands than open-top-chamber studies), reported reduced tree growth when
grown in single or three species stands within 30-m diameter rings and exposed over one or more
years to elevated O3 concentrations (hourly concentrations 1.5 times concentrations in ambient
air at the site) compared to unadjusted ambient air concentrations (2013 ISA, section 9.4.3;
Kubiske et al., 2006, Kubiske et al., 2007).22
With regard to tree mortality, the 2013 ISA did not include a determination of causality
(ISA, Appendix 8, section 8.4). While the then-available evidence included studies identifying
ozone as a contributor to tree mortality, which contributed to the 2013 conclusion regarding O3
and alteration of community composition (2013 ISA, section 9.4.7.4), a separate causality
determination regarding O3 and tree mortality was not assessed (ISA, Appendix 8, section 8.4;
2013 ISA, Table 9-19). The evidence assessed in the 2013 ISA (and 2006 AQCD) was largely
observational, including studies that reported declines in conifer forests for which elevated O3
was identified as contributor but in which a variety of environmental factors may have also
played a role (2013 ISA, section 9.4.7.1; 2006 AQCD, sections AX9.6.2.1, AX9.6.2.2,
AX9.6.2.6, AX9.6.4.1 and AX9.6.4.2). Since the 2015 review, three additional studies are now
available (ISA, Appendix 8, Table 8-9). Two of these are analyses of field observations, one of
which is set in the Spanish Pyrenees.23 A second study is a large-scale empirical statistical
22	Seasonal (92-day) W126 index values for unadjusted 03 concentrations over six years of the Aspen FACE
experiments ranged from 2 to 3 ppm-hrs, while the elevated exposure concentrations (reflecting addition of O3 to
ambient air concentrations) ranged from somewhat above 20 to somewhat above 35 ppm-hrs (ISA, Appendix 8,
Figure 8-17).
23	The concentration gradient with altitude in the Spanish study, includes - at the highest site - annual average April-
to-September O3 concentrations for the 2004 to 2007 period that range up to 74 ppb (Diaz-de-Quijano et al.,
2016), indicating O3 concentrations likely to exceed the current U.S. secondary standard.
April 2022
4-22 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
analysis of factors potentially contributing to tree mortality in eastern and central U.S. forests
during the 1971-2005 period, which reported O3 (county-level 11-year [1996-2006] average 8
hour metric)24 to be ninth among the 13 potential factors assessed25 and to have a significant
positive correlation with tree mortality (ISA, section IS.5.2, Appendix 8, section 8.4.3; Dietze
and Moorcroft, 2011). A newly available experimental study also reported increased mortality in
two of five aspen genotypes grown in mixed stands under elevated O3 concentrations (ISA,
section IS.5.1.2; Moran and Kubiske, 2013). Coupled with the plant-level evidence of
phytotoxicity discussed above, as well as consideration of community composition effects, this
evidence was concluded to indicate the potential for elevated O3 concentrations to contribute to
tree mortality (ISA, section IS.5.1.2 and Appendix 8, sections 8.4.3 and 8.4.4). Based on the
available evidence, the ISA concludes there is likely to be a causal relationship between O3 and
increased tree mortality (ISA, Table IS-2, Appendix 8, section 8.4.4).
A variety of factors in natural environments can either mitigate or exacerbate predicted
03-plant interactions and are recognized sources of uncertainty and variability. Such factors at
the plant level include multiple genetically influenced determinants of O3 sensitivity, changing
sensitivity to O3 across vegetative growth stages, co-occurring stressors and/or modifying
environmental factors (ISA, Appendix 8, section 8.12).
Ozone-induced effects at the scale of the whole plant have the potential to translate to
effects at the ecosystem scale, such as reduced productivity and carbon storage, and altered
terrestrial community composition, as well as impacts on ecosystem functions, such as
belowground biogeochemical cycles and ecosystem water cycling. For example, under the
relevant exposure conditions, 03-related reduced tree growth and reproduction, as well as
increased mortality, could lead to reduced ecosystem productivity. Recent studies from the
Aspen FACE experiment and modeling simulations indicate that 03-related negative effects on
ecosystem productivity may be temporary or may be limited in some systems (ISA, Appendix 8,
section 8.8.1). Previously available studies had reported impacts on productivity in some forest
types and locations, such as ponderosa pine in southern California and other forest types in the
mid-Atlantic region (2013 ISA, section 9.4.3.4). Through reductions in sensitive species growth,
24	As indicated in Figures 2-11 and 2-12, annual fourth highest daily maximum 8-hour O3 concentrations in these
regions were above 80 ppb in the early 2000s and the median design values at national trend sites was nearly 85
ppb.
25	This statistical analysis, which utilized datasets from within the 1971-2005 period, included an examination of the
sensitivity of predicted mortality rate to 13 different covariates. On average across the predictions for 10 groups
of trees (based on functional type and major representative species), the order of mortality rate sensitivity to the
covariates, from highest to lowest, was: sulfate deposition, tree diameter, nitrate deposition, summer temperature,
tree age, elevation, winter temperature, precipitation, O3 concentration, tree basal area, topographic moisture
index, slope and topographic radiation index (Dietze and Moorcroft, 2011).
April 2022
4-23 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
and related ecosystem productivity, O3 could lead to reduced ecosystem carbon storage (ISA,
IS.5.1.4; 2013 ISA, section 9.4.3). With regard to forest community composition, available
studies have reported changes in tree communities composed of species with relatively greater
and relatively lesser sensitivity to O3, such as birch and aspen, respectively (ISA, section
IS.5.1.8.1, Appendix 8, section 8.10; 2013 ISA, section 9.4.3; Kubiske et al., 2007). As the ISA
concludes, "[t]he extent to which ozone affects terrestrial productivity will depend on more than
just community composition, but other factors, which both directly influence [net primary
productivity] (i.e., availability of N and water) and modify the effect of ozone on plant growth"
(ISA, Appendix 8, section 8.8.1). Thus, the magnitude of O3 impact on ecosystem productivity,
as on forest composition, can vary among plant communities based on several factors, including
the type of stand or community in which the sensitive species occurs (e.g., single species versus
mixed canopy), the role or position of the species in the stand (e.g., dominant, sub-dominant,
canopy, understory), and the sensitivity of co-occurring species and environmental factors (e.g.,
drought and other factors).
The effects of O3 on plants and plant populations also have implications for other
ecosystem functions. Two such functions, effects with which O3 is concluded to be likely
causally or causally related, are ecosystem water cycling and belowground biogeochemical
cycles, respectively (ISA, Appendix 8, sections 8.11 and 8.9). With regard to the former, the
effects of O3 on plants (e.g., via stomatal control, as well as leaf and root growth and changes in
wood anatomy associated with water transport) can affect ecosystem water cycling through
impacts on root uptake of soil moisture and groundwater as well as transpiration through leaf
stomata to the atmosphere (ISA, Appendix 8, section 8.11.1). These "impacts may in turn affect
the amount of water moving through the soil, running over land or through groundwater and
flowing through streams" (ISA, Appendix 8, section 8.11, p. 8-161). Evidence newly available
for the 2020 ISA is supportive of previously available evidence in this regard (ISA, Appendix 8,
section 8.11.6). This evidence, including that newly available, indicates the extent to which the
effects of O3 on plant leaves and roots (e.g., through effects on chemical composition and
biomass) can impact belowground biogeochemical cycles involving root growth, soil food web
structure, soil decomposer activities, soil microbial respiration, soil carbon turnover, soil water
cycling and soil nutrient cycling (ISA, Appendix 8, section 8.9).
Additional vegetation- and insect-related effects with implications beyond individual
plants include the effects of O3 on insect herbivore growth and reproduction and plant-insect
signaling (ISA, Table IS-12, Appendix 8, sections 8.6 and 8.7). With regard to insect herbivore
growth and reproduction, the evidence includes multiple effects in an array of insect species,
although without a consistent pattern of response for most endpoints (ISA, Appendix 8, Table 8-
April 2022
4-24 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
11). As was also the case with the studies available at the time of the 2015 review,26 in the newly
available studies the individual-level responses are highly context- and species-specific and not
all species tested showed a response (ISA, p. IS-64, Table IS-12, section IS.5.1.3 and Appendix
8, section 8.6). Evidence on plant-insect signaling comes from laboratory, greenhouse, open top
chambers (OTC) and FACE experiments (ISA, section IS.5.1.3 and Appendix 8, section 8.7).
The available evidence indicates a role for elevated O3 in altering and degrading emissions of
chemical signals from plants and reducing detection of volatile plant signaling compounds
(VPSCs) by insects, including pollinators. Elevated O3 concentrations degrade some VPSCs
released by plants, potentially affecting ecological processes including pollination and plant
defenses against herbivory. Further, the available studies report elevated O3 conditions to be
associated with plant VPSC emissions that may make a plant either more attractive or more
repellant to herbivorous insects, and to predators and parasitoids that target phytophagous (plant-
eating) insects (ISA, section IS.5.1.3 and Appendix 8, section 8.7).
Ozone welfare effects also extend beyond effects on vegetation and associated biota due
to it being a major greenhouse gas and radiative forcing agent.27 As in the 2015 review, the
available evidence, augmented since the 2013 ISA, continues to support a causal relationship
between the global abundance of O3 in the troposphere and radiative forcing, and a likely causal
relationship between the global abundance of O3 in the troposphere and effects on temperature,
precipitation, and related climate variables28 (ISA, section IS.5.2 and Appendix 9; Myhre et al.,
2013). As was also true at the time of the 2015 review, tropospheric O3 has been ranked third in
importance for global radiative forcing, after carbon dioxide and methane, with the radiative
forcing of O3 since pre-industrial times estimated to be about 25 to 40% of the total warming
effects of anthropogenic carbon dioxide and about 75% of the effects of anthropogenic methane
(ISA, Appendix 9, section 9.1.3.3). Uncertainty in the magnitude of radiative forcing estimated
to be attributed to tropospheric O3 is a contributor to the relatively greater uncertainty associated
with climate effects of tropospheric O3 compared to such effects of the well mixed greenhouse
gases, such as carbon dioxide and methane (ISA, section IS.6.2.2).
26	During the 2015 review, the 2013 ISA stated with regard to O3 effects on insects and other wildlife that "there is
no consensus on how these organisms respond to elevated O3 (2013 ISA, section 9.4.9.4, p. 9-98).
27	Radiative forcing is a metric used to quantify the change in balance between radiation coming into and going out
of the atmosphere caused by the presence of a particular substance. The ISA describes it more specifically as "a
perturbation in net radiative flux at the tropopause (or top of the atmosphere) caused by a change in radiatively
active forcing agent(s) after stratospheric temperatures have readjusted to radiative equilibrium (stratospherically
adjusted RF)" (ISA, Appendix 9, section 9.1.3.3).
28	Effects on temperature, precipitation, and related climate variables were referred to as "climate change" or
"effects on climate" in the 2013 ISA (ISA, p. IS-82; 2013 ISA, pp. 1-14, 10-31).
April 2022
4-25 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
Lastly, the evidence regarding tropospheric O3 and UV-B shielding was evaluated in the
2013 ISA and determined to be inadequate to draw a causal conclusion (2013 ISA, section
10.5.2). The current ISA concludes there to be no new evidence since the 2013 ISA relevant to
the question of UV-B shielding by tropospheric O3 (ISA, IS. 1.2.1 and Appendix 9, section
9.1.3.4).
4.3.2 Public Welfare Implications
The public welfare implications of the evidence regarding O3 welfare effects are
dependent on the type and severity of the effects, as well as the extent of the effect at a particular
biological or ecological level of organization. We discuss such factors here in light of judgments
and conclusions made in NAAQS reviews regarding effects on the public welfare.
As provided in section 109(b)(2) of the CAA, the secondary standard is to "specify a
level of air quality the attainment and maintenance of which in the judgment of the
Administrator ... is requisite to protect the public welfare from any known or anticipated adverse
effects associated with the presence of such air pollutant in the ambient air." The secondary
standard is not meant to protect against all known or anticipated 03-related welfare effects, but
rather those that are judged to be adverse to the public welfare, and a bright-line determination of
adversity is not required in judging what is requisite (78 FR 3212, January 15, 2013; 80 FR
65376, October 26, 2015; see also 73 FR 16496, March 27, 2008). Thus, the level of protection
from known or anticipated adverse effects to public welfare that is requisite for the secondary
standard is a public welfare policy judgment made by the Administrator. The Administrator's
judgment regarding the available information and adequacy of protection provided by an existing
standard is generally informed by considerations in prior reviews and associated conclusions.
• Is there newly available information relevant to consideration of the public welfare
implications of 03-related welfare effects?
The categories of effects identified in the CAA to be included among welfare effects are
quite diverse,29 and among these categories, any single category includes many different types of
effects that are of broadly varying specificity and level of resolution. For example, effects on
vegetation is a category identified in CAA section 302(h), and the ISA recognizes numerous
vegetation-related effects of O3 at the organism, population, community, and ecosystem level, as
summarized in section 4.3.1 above (ISA, Appendix 8). The significance of each type of
vegetation-related effect with regard to potential effects on the public welfare depends on the
29 Section 302(h) of the CAA states that language referring to "effects on welfare" in the CAA "includes, but is not
limited to, effects on soils, water, crops, vegetation, manmade materials, animals, wildlife, weather, visibility, and
climate, damage to and deterioration of property, and hazards to transportation, as well as effects on economic
values and on personal comfort and well-being" (CAA section 302(h)).
April 2022
4-26 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
type and severity of effects, as well as the extent of such effects on the affected environmental
entity, and on the societal use of the affected entity and the entity's significance to the public
welfare. Such factors have been considered in the context of judgments and conclusions made in
some prior reviews regarding public welfare effects. For example, judgments regarding public
welfare significance in two prior O3 NAAQS decisions gave particular attention to O3 effects in
areas with special federal protections (such as Class I areas), and lands set aside by states, tribes
and public interest groups to provide similar benefits to the public welfare (73 FR 16496, March
27, 2008; 80 FR 65292, October 26, 2015).30 In the 2015 review, the EPA recognized the "clear
public interest in and value of maintaining these areas in a condition that does not impair their
intended use and the fact that many of these lands contain Cb-sensitive species" (73 FR 16496,
March 27, 2008).
Judgments regarding effects on the public welfare can depend on the intended use for, or
service (and value) of, the affected vegetation, ecological receptors, ecosystems and resources
and the significance of that use to the public welfare (73 FR 16496, March 27, 2008; 80 FR
65377, October 26, 2015). Uses or services provided by areas that have been afforded special
protection can flow in part or entirely from the vegetation that grows there. Uses or services
provided by areas that have been afforded special protection can flow in part or entirely from the
vegetation that grows there. Ecosystem services range from those directly related to the natural
functioning of the ecosystem to ecosystem uses for human recreation or profit, such as through
the production of lumber or fuel (Costanza et al., 2017; ISA, section IS.5.1). Services of aesthetic
value and outdoor recreation depend, at least in part, on the perceived scenic beauty of the
environment. Additionally, public surveys have indicated that Americans rank as very important
the existence of resources, the option or availability of the resource and the ability to bequest or
pass it on to future generations (Cordell et al., 2008). The spatial, temporal, and social
dimensions of public welfare impacts are also influenced by the type of service affected. For
example, a national park can provide direct recreational services to the thousands of visitors that
come each year, but also provide an indirect value to the millions who may not visit but receive
30 For example, the fundamental purpose of parks in the National Park System "is to conserve the scenery, natural
and historic objects, and wild life in the System units and to provide for the enjoyment of the scenery, natural and
historic objects, and wild life in such manner and by such means as will leave them unimpaired for the enjoyment
of future generations" (54 U.S.C. 100101). Additionally, the Wilderness Act of 1964 defines designated
"wilderness areas" in part as areas "protected and managed so as to preserve [their] natural conditions" and
requires that these areas "shall be administered for the use and enjoyment of the American people in such manner
as will leave them unimpaired for future use and enjoyment as wilderness, and so as to provide for the protection
of these areas, [and] the preservation of their wilderness character ..." (16 U.S.C. 1131 (a) and (c)). Other lands
that benefit the public welfare include national forests which are managed for multiple uses including sustained
yield management in accordance with land management plans (see 16 U.S.C. 1600(l)-(3); 16 U.S.C. 1601(d)(1)).
April 2022
4-27 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
satisfaction from knowing it exists and is preserved for the future (80 FR 65377, October 26,
2015).
The different types of effects on vegetation discussed in section 4.3.1 above differ with
regard to aspects important to judging their public welfare significance. For example, in the case
of crop yield loss, such judgments may consider aspects such as the heavy management of
agriculture in the U.S., while judgments for other categories of effects may generally relate to
considerations regarding natural areas, including specifically those areas that are not managed
for harvest. For example, effects on tree growth and reproduction, and also visible foliar injury,
have the potential to be significant to the public welfare through impacts in Class I and other
areas given special protection in their natural/existing state, although they differ in how they
might be significant.
In this context, it may be important to consider that O3 effects on tree growth and
reproduction could, depending on severity, extent, and other factors, lead to effects on a larger
scale including reduced productivity, altered forest and forest community (plant, insect and
microbe) composition, reduced carbon storage and altered ecosystem water cycling (ISA, section
IS.5.1.8.1; 2013 ISA, Figure 9-1, sections 9.4.1.1 and 9.4.1.2). For example, the composition of
plants and other members of terrestrial communities can be affected through O3 effects on
growth and reproductive success of sensitive plant species in the community, with the extent of
compositional changes dependent on factors such as competitive interactions (ISA, section
IS.5.1.8.1; 2013 ISA, sections 9.4.3 and 9.4.3.1). Impacts on some of these characteristics (e.g.,
forest or forest community composition) may be considered of greater public welfare
significance when occurring in Class I or other protected areas, due to value for particular
services that the public places on such areas.
Agriculture and silviculture provide ecosystem services with clear public welfare
benefits. With regard to agriculture-related effects, however, there are complexities in this
consideration related to areas and plant species that are heavily managed to obtain a particular
output (such as commodity crops or commercial timber production). In light of this, the degree to
which O3 impacts on vegetation that could occur in such areas and on such species would impair
the intended use at a level that might be judged adverse to the public welfare has been less clear
(80 FR 65379, October 26, 2015; 73 FR 16497, March 27, 2008). While having sufficient crop
yields is of high public welfare value, important commodity crops are typically heavily managed
to produce optimum yields. Moreover, based on the economic theory of supply and demand,
increases in crop yields would be expected to result in lower prices for affected crops and their
associated goods, which would primarily benefit consumers. Analyses in past reviews have
described how these competing impacts on producers and consumers complicate consideration of
April 2022
4-28 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
these effects in terms of potential adversity to the public welfare (2014 WREA, sections 5.3.2
and 5.7).
Other ecosystem services valued by people that can be affected by reduced tree growth,
productivity and associated forest effects include aesthetic value, food, fiber, timber, other forest
products, habitat, recreational opportunities, climate and water regulation, erosion control, air
pollution removal, and desired fire regimes, as summarized in Figure 4-2 (ISA, section IS.5.1;
2013 ISA, sections 9.4.1.1 and 9.4.1.2). In considering such services in past reviews, the Agency
the Agency has given particular attention to effects in natural ecosystems, indicating that a
protective standard, based on consideration of effects in natural ecosystems in areas afforded
special protection, would also "provide a level of protection for other vegetation that is used by
the public and potentially affected by O3 including timber, produce grown for consumption and
horticultural plants used for landscaping" (80 FR 65403, October 26, 2015). For example,
locations potentially vulnerable to 03-related impacts might include forested lands, both public
and private, where trees are grown for timber production. Forests in urbanized areas also provide
a number of services that are important to the public in those areas, such as air pollution removal,
cooling, and beautification. There are also many other tree species, such as various ornamental
and agricultural species (e.g., Christmas trees, fruit and nut trees), that provide ecosystem
services that may be judged important to the public welfare.
With its effect on the physical appearance of plants, visible foliar injury has the potential
to be significant to the public welfare, depending on its severity and spatial extent, by impacting
aesthetic or scenic values and outdoor recreation in Class I and other similarly protected areas
valued by the public.31 To assess evidence of injury to plants in forested areas on national and
regional scales, the U.S. Forest Service (USFS) conducted surveys of the occurrence and severity
of visible foliar injury on sensitive (bioindicator) species at biomonitoring sites across most of
the U.S., beginning in 1994 (in the eastern U.S.) and extending through 2011 (Smith, 2012;
Coulston et al., 2003). At these sites (biosites), a national protocol, including verification and
quality assurance procedures and a scoring system, was implemented. The resultant biosite index
(BI) scores may be described with regard to one of several categories ranging from little or no
foliar injury to severe injury. For example, BI scores of zero to five are described as "little or no
31 For example, although analyses specific to visible foliar injury are of limited availability, there have been analyses
developing estimates of recreation value damages of severe impacts related to other types of forest effects, such
as tree mortality due to bark beetle outbreaks (e.g., Rosenberger et al., 2013). Such analyses estimate reductions
in recreational use when the damage is severe (e.g., reductions in the density of live, robust trees). Such damage
would reasonably be expected to also reflect damage indicative of injury with which a relationship with other
plant effects (e.g., growth and reproduction) would be also expected. Similarly, a couple of studies from the
1970s and 1980s indicated potential for differences in recreational use for areas with stands of pine in which
moderate to severe injury was apparent from 30 or 40 feet (1996 AQCD, section 5.8.3).
April 2022
4-29 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
foliar injury," scores above five to 15 as "low" or "light to moderate" foliar injury, scores from
15 to 25 as "moderate foliar injury" and scores above 25 as "severe injury" (Campbell et al.,
2007; Smith et al., 2007; Smith, 2012).32 However, available information does not yet address or
describe the relationships expected to exist between some level of injury severity (e.g., little,
low/light, moderate or severe) and/or spatial extent affected and scenic or aesthetic values. This
gap impedes consideration of the public welfare implications of different injury severities, and
accordingly judgments on the potential for public welfare significance. That notwithstanding,
while minor spotting on a few leaves of a plant may easily be concluded to be of little public
welfare significance, some level of severity and widespread occurrence of visible foliar injury,
particularly if occurring in specially protected areas, where the public can be expected to place
value (e.g., for recreational uses), might reasonably be concluded to impact the public welfare.
The tropospheric Cb-related effects of radiative forcing and subsequent effects on
temperature, precipitation and related climate variables also have important public welfare
implications, although their quantitative evaluation in response to O3 concentrations in the U.S.
is complicated by "[cjurrent limitations in climate modeling tools, variation across models, and
the need for more comprehensive observational data on these effects" (ISA, section IS.6.2.2). An
ecosystem service provided by forested lands is carbon sequestration or storage (ISA, section
IS.5.1.4 and Appendix 8, section 8.8.3; 2013 ISA, section 2.6.2.1 and p. 9-37),33 which has an
extremely valuable role in counteracting the impact of greenhouse gases on radiative forcing and
related climate effects on the public welfare. Accordingly, the service of carbon storage can be of
paramount importance to the public welfare no matter in what location the trees are growing or
what their intended current or future use (e.g., 2013 ISA, section 9.4.1.2). The benefit exists as
long as the trees are growing, regardless of what additional functions and services it provides.
Categories of effects newly identified as likely causally related to O3 in ambient air, such
as alteration of plant-insect signaling and insect herbivore growth and reproduction, also have
potential public welfare implications, e.g., given the role of the plant-insect signaling process in
pollination and seed dispersal (ISA, section IS.5.1.3). Uncertainties and limitations in the
evidence (e.g., as summarized in sections 4.3.3.3 and 4.3.4 below) preclude an assessment of the
extent and magnitude of O3 effects on these endpoints, which thus also precludes an evaluation
of the potential for associated public welfare implications.
32	Authors of studies presenting USFS biomonitoring program data have suggested what might be considered
"assumptions of risk" (e.g., for the forest resource) related to scores in these categories, e.g., none, low, moderate
and high for BI scores of zero to five, five to 15, 15 to 25 and above 25, respectively (e.g., Smith et al., 2003;
Smith et al., 2012). For example, maps of localized moderate to high risk areas may be used to identify areas
where more detailed evaluations are warranted (Smith et al., 2012).
33	While carbon sequestration or storage also occurs for vegetated ecosystems other than forests, it is relatively
larger in forests given the relatively greater biomass for trees compared to some other plants.
April 2022
4-30 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
In summary, several considerations are recognized as important to judgments on the
public welfare significance of the array of welfare effects of different O3 exposure conditions.
These include uncertainties and limitations associated with the consideration of the magnitude of
key vegetation effects that might be concluded to be adverse to ecosystems and associated
services. Additionally, the presence of Cb-sensitive tree species may contribute to a vulnerability
of numerous locations to public welfare impacts from O3 related to tree growth, productivity and
carbon storage and their associated ecosystems and services. Other important considerations
include the exposure circumstances that may elicit effects and the potential for the significance
of the effects to vary in specific situations due to differences in sensitivity of the exposed
species, the severity and associated significance of the observed or predicted Cb-induced effect,
the role that the species plays in the ecosystem, the intended use of the affected species and its
associated ecosystem and services, the presence of other co-occurring predisposing or mitigating
factors, and associated uncertainties and limitations.
April 2022	4-31 External Review Draft - Do Not Quote or Cite

-------
ฆ 

s ซ
Ij
ซ	w
e	c
•— o	m
O \3	2
UJ n  g
T3 O
GS
LU
Li-
ft 5
งซ
1	s
2
s
j Emissions of precursors intoambient air j
4-
03in ambient air
i
Vegetation


Changes in




leaf




chemistry/



1
r ~ nuttents ,
t *
r 1
r
Reductions in plant reproduction and growth
increased tree mortality, and related effects on.
-	Crop yield and quality
-	Terrestrial ecosystem productivity
-	Carbon sequestration in terrestrial ecosystems
-	Terrestrial community composition
• Bel ov/g round biogeochemical cycles
-	Ecosystem water cycling
I
Troposphere O5
Alteration of
herbivore
(insect)
growth and
reproduction
Alteration of
plant-insect
signaling
Radiative
forcing
and effects on
temperature and
related climate
variables
t:
Effects on federally protected Class I and other areas, such as wilderness areas and lands provided similar protections by state,
tribal and other groups
•	Impairment of scenic value, natural vegetation and wildlife
•	Reduced aesthetic and outdoor recreation value
•	Altered ecosystem community composition
Changes to timber, fruit, vegetable and fiber (for fabrics) production, including reductions in somespeciftcproduds/mEteriaSs
Reduced effectiveness of some species of trees/plants for cooling and beautification in residential areas
Erosion in populated areas
Reduced game fish and recreational fisheries and other effects related to reduced late-season stream flow
Climate-related impacts, such as net changes in agricultural productivity, property damage, and changes in heating/cooling costs
Nae: Wefere efecs causaly related x> O5 exposure are in boซJ sext, araj ftety causal effecs are tatazed Dashed weffere efecs boxes conan only kkety causal efecs. Dosed
Ines connect those efecs to other efecs a swhehthey ooukJ poseraaty corOTbuse.
Figure 4-2. Potential effects of O3 on the public welfare.
April 2022
4-32
External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
4.3.3 Exposures Associated with Effects
The types of effects identified in section 4.3.1 above vary widely with regard to the
extent and level of detail of the available information that describes the O3 exposure
circumstances that may elicit them. The information on exposure metric and E-R relationships
for effects related to vegetation growth is long-standing, having been first described in the 1997
review, while such information is much less established for visible foliar injury. The evidence
base for other categories of effects is also lacking in information that might support
characterization of potential impacts of changes in O3 concentrations. The discussion in this
section is organized in recognition of this variation. We focus first on growth and yield effects,
the category of effects for which the information on exposure metric and E-R relationships is
most advanced (section 4.3.3.1). Section 4.3.3.2 discusses the information regarding exposure
metrics and relationships between exposure and the occurrence and severity of visible foliar
injury. The availability of such information for other categories of effects is addressed in section
4.3.3.3.
4.3.3.1 Growth-related Effects
4.3.3.1.1 Exposure Metrics
The longstanding body of vegetation effects evidence includes a wealth of information on
aspects of O3 exposure that influence effects on plant growth and yield, and that has been
described in the scientific assessments across the last several decades (1996 AQCD; 2006
AQCD; 2013 ISA; 2020 ISA). A variety of factors have been investigated, including
"concentration, time of day, respite time, frequency of peak occurrence, plant phenology,
predisposition, etc." (2013 ISA, section 9.5.2). The importance of the duration of the exposure
and the relatively greater importance of higher concentrations over lower concentrations have
been consistently well documented (2013 ISA, section 9.5.3). For example, key conclusions of
the 1996 AQCD, that have been confirmed in the 2006 AQCD, 2013 ISA and 2020 ISA include
that "Ozone effects in plants are cumulative" and "Higher O3 concentrations appear to be more
important than lower concentrations in eliciting a response" (2006 AQCD, p. E-27; 2013 ISA, p.
2-44; 2020 ISA, p. 8-180) These AQCDs and IS As described several mathematical approaches
for a single metric or index that would, to some extent, reflect both conclusions.
The consideration of these different exposure metrics has primarily focused on their
ability to summarize ambient air concentrations of O3 in a way that best correlates with effects
on vegetation, particularly growth-related effects. Metrics based on mean concentrations over
several hours (e.g., a seasonal average 12-hour concentration), have generally been considered to
be less robust as a metric relating exposure to growth effects (2020 ISA, p. 8-181). The
approaches that cumulate exposures over some specified period while weighting higher
April 2022
4-33 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
concentrations more than lower had been evaluated for their predictiveness of growth responses
in a set of crop and tree species assessed in experimental O3 exposure studies for which hourly
O3 concentrations were available for analysis (2013 ISA, sections 9.5.2 and 9.5.3; ISA,
Appendix 8, section 8.2.2.2).
Along with the non-threshold concentration weighted W126 index, two other cumulative
indices that have received greatest attention across the past several O3 NAAQS reviews have
been the threshold weighted indices, AOT6O34 and SUM06 (ISA, section IS.3.2).35 Accordingly,
some studies of O3 vegetation effects have reported exposures in terms of these metrics. Based
on extensive review of the published literature on different types of such E-R metrics, and
comparisons between metrics, and in the context of a single metric, the EPA has generally
focused on cumulative, concentration-weighted indices of exposure that reflect some
consideration of both concern for cumulative effects of O3 exposure and for the greater
importance of higher concentrations than lower concentrations in vegetation effects (1996
AQCD; 2006 AQCD; 2013 ISA).36 Quantifying exposure using such indices has been found to
improve the explanatory power of E-R models with regard to O3 effects in studies of growth and
yield over that of indices based only on mean and peak exposure values (2013 ISA, section
2.6.6.1, p. 2-44).37
The most well-studied datasets in this in this regard are two datasets established two
decades ago (referenced above and described further in section 4.3.3.1.2 below), one for growth
effects on seedlings of a set of 11 tree species and the second for quality and yield effects for a
set of 10 crops (e.g., Lee and Hogsett, 1996, Hogsett et al., 1997). These datasets, which include
growth and yield response information across a range of multiple seasonal cumulative exposures,
were used to develop quantitative E-R functions for reduced growth (termed relative biomass
34	The AOT60 index is the seasonal sum of the difference between an hourly concentration above 60 ppb, minus 60
ppb (2006 AQCD, p. AX9-161). More recently, some studies have also reported O3 exposures in terms of
AOT40, which is conceptually similar but with 40 substituted for 60 in its derivation (ISA, Appendix 8, section
8.13.1).
35	The SUM06 index is the seasonal sum of hourly concentrations at or above 0.06 ppm during a specified daily time
window (2006 AQCD, p. AX9-161; 2013 ISA, section 9.5.2). This may sometimes be referred to as SUM60, e.g.,
when concentrations are in terms of ppb. There are also variations on this metric that utilize alternative reference
points above which hourly concentrations are summed. For example, SUM08 is the seasonal sum of hourly
concentrations at or above 0.08 ppm and SUMO is the seasonal sum of all hourly concentrations.
36	The Agency has focused its analyses in the last several reviews on metrics that characterize cumulative exposures
over a season or seasons: SUM06 in the 1997 review (61 FR 65716, December 13, 1996; 62 FR 38856, July 18,
1997) and W126 in both the 2008 and 2015 reviews (72 FR 37818, July 11, 2007; 73 FR 16436, March 27, 2008;
80 FR 65373-65374, October 26, 2015). This approach to characterizing O3 exposure concentrations with regard
to potential vegetation effects, particularly growth, has been supported by CASAC in the past reviews
(Henderson, 2006; Samet, 2010; Frey, 2014; Cox, 2020).
37	As described in section 4.3.3.2 below, the W126 index and other similar cumulative exposure indices do not
completely describe the relationship of O3 to visible foliar injury in national surveys.
April 2022
4-34 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
loss or RBL) in seedlings of the tree species and E-R functions for RYL for a set of common
crops (ISA, Appendix 8, section 8.13.2; 2013 ISA, section 9.6.2).
The EPA's conclusions regarding cumulative exposure levels of O3 associated with
vegetation-related effects at the time of the 2015 review were based primarily on these
established E-R functions and the W126 index, which is a cumulative, seasonal38 concentration-
weighted index (80 FR 65404, October 26, 2015; ISA, section IS.3.2, Appendix 8, section 8.13).
This metric is a non-threshold approach described as the sigmoidally weighted sum of all hourly
O3 concentrations observed during a specified daily and seasonal time window, where each
hourly O3 concentration is given a weight that increases from zero to one with increasing
concentration (2013 ISA, p. 9-101).
Alternative methods for characterizing O3 exposure to predict various plant responses
have, in recent years, included flux models (models that are based on the amount of O3 that
enters the leaf). However, as was the case in the 2015 review, there remain a variety of
complications, limitations and uncertainties associated with this approach. For example, "[w]hile
some efforts have been made in the U.S. to calculate ozone flux into leaves and canopies, little
information has been published relating these fluxes to effects on vegetation" (ISA, section
IS.3.2). Further, as flux of O3 into the plant under different conditions of O3 in ambient air is
affected by several factors including temperature, vapor pressure deficit, light, soil moisture, and
plant growth stage, use of this approach to quantify the vegetation impact of O3 would require
information on these various types of factors (ISA, section IS.3.2). In addition to these data
requirements, each species has different amounts of internal detoxification potential that may
protect species to differing degrees. The lack of detailed species- and site-specific data required
for flux modeling in the U.S. and the lack of understanding of detoxification processes continues
to make this technique less viable for use in risk assessments in the U.S. (ISA, section IS.3.2).
Among the studies newly available since the 2015 review, no new exposure indices for
assessing effects on vegetation growth or other physiological process parameters have been
identified. In the literature available since the 2013 ISA, the SUM06, AOTx (e.g., AOT60) and
W126 exposure metrics remain the metrics that are most commonly discussed (ISA, Appendix 8,
section 8.13.1). The ISA notes that "[cumulative indices of exposure that differentially weight
hourly concentrations [which would include the W126 index] have been found to be best suited
to characterize vegetation exposure to ozone with regard to reductions in vegetation growth and
yield" (ISA, section ES.3). Accordingly, in this reconsideration of the 2020 decision, as in the
2015 and 2020 reviews, we use the seasonal W126-based cumulative, concentration-weighted
38 In describing the form as "seasonal," the EPA is referring generally to an index focused on a time period of a
duration that may relate to that of a growing season for 03-sensitive vegetation, not to the seasons of the year
(spring, summer, fall, winter).
April 2022
4-35 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
metric in interpreting quantitative exposure analyses, particularly related to growth effects of
cumulative O3 exposures (as summarized in sections 4.3.3.2 and 4.4 below).
The first step in calculating the seasonal W126 index for a specific year is to sum the
weighted hourly O3 concentrations in ambient air during daylight hours (defined as 8:00 a.m. to
8:00 p.m. local standard time) within each calendar month, resulting in monthly index values.
The monthly W126 index values are calculated from hourly O3 concentrations as follows.39
Monthly W126 = Zd-iZft-s	—	
J	t-iri—o i+4403*exp (-126*Cdh)
where,
N is the number of days in the month
d is the day of the month (d = 1, 2, ..., N)
h is the hour of the day (h = 0, 1, ..., 23)
Cdh is the hourly O3 concentration observed on day hour h, in parts per million
The W126 index value for a specific year is the maximum sum of the monthly index values for
three consecutive months within a calendar year (i.e., January to March, February to April, ...
October to December). Three-year average W126 index values are calculated by taking the
average of seasonal W126 index values for three consecutive years (e.g., as described in
Appendix 4D, section 4D.2.2).
4.3.3.1.2 Relationships Between Cumulative Concentration-weighted Exposure
Levels and Effects
Across the array of 03-related welfare effects, consistent and systematically evaluated
information on E-R relationships across multiple exposure levels is limited. Most prominent is
the information on E-R relationships for growth effects on tree seedlings and crops,40 which has
been available for the past several reviews. The information on which these functions are based
comes primarily from the U.S. EPA's National Crop Loss Assessment Network (NCLAN)41
project for crops and the NHEERL-WED project for tree seedlings, projects implemented
primarily to define E-R relationships for major agricultural crops and tree species, thus
advancing understanding of responses to O3 exposures (ISA, Appendix 8, section 8.13.2). These
projects and related studies included a series of experiments that used OTCs to investigate tree
seedling growth response and crop yield over a growing season under a variety of O3 exposures
39	In situations where data are missing, an adjustment is factored into the monthly index (as described in Appendix
4D, section 4D.2.2).
40	The E-R functions estimate 03-related reduction in a year's tree seedling growth or crop yield as a percentage of
that expected in the absence of O3 (Appendix 4A; ISA, Appendix 8, section 8.13.2).
41	The NCLAN program, which was undertaken in the early to mid-1980s, assessed multiple U.S. crops, locations,
and O3 exposure levels, using consistent methods, to provide the largest, most uniform database on the effects of
O3 on agricultural crop yields (1996 AQCD, 2006 AQCD, 2013 ISA, sections 9.2, 9.4, and 9.6; ISA, Appendix 8,
section 8.13.2).
April 2022
4-36 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
and growing conditions (2013 ISA, section 9.6.2; Lee and Hogsett, 1996). These experiments
assessed O3 effects on tree seedling growth and crop yield for a variety of O3 treatments and
growing conditions. The higher exposure levels in these datasets generally included numerous
hours at or above 100 ppb (Lefohn et al., 1997; Appendix 4A, Table 4A-6). Importantly, the
information on exposure includes hourly concentrations across the season (or longer) exposure
period which allowed for derivation of various seasonal metrics that were analyzed for
association with reduced growth. In the initial analyses of these data, exposure was characterized
in terms of several metrics, including seasonal SUM06 and W126 indices (Lee and Hogsett,
1996; 1997 Staff Paper, sections IV.D.2 and IV.D.3; 2007 Staff Paper, section 7.6), while use of
these functions in the 2015 review focused on their implementation in terms of seasonal W126
index (2013 ISA, section 9.6; 80 FR 65391-92, October 26, 2015).42
The 11 species for which robust and well-established E-R functions for RBL were
derived black cherry, Douglas fir, loblolly pine, ponderosa pine, quaking aspen, red alder, red
maple, sugar maple, tulip poplar, Virginia pine, and white pine (Figure 4-3; Appendix 4A; 2020
ISA, Appendix 8, section 8.13.2; 2013 ISA, section 9.6).43 While these 11 species represent only
a small fraction of the total number of native tree species in the contiguous U.S., this small
subset includes eastern and western species, deciduous and coniferous species, and species that
grow in a variety of ecosystems and represent a range of tolerance to O3 (Appendix 4B; 2020
ISA, Appendix 8, section 8.13.2; 2013 ISA, section 9.6.2). The established E-R functions for
most of the 11 species were derived using data from multiple studies or experiments, many of
which employed open top chambers, an established experimental approach, involving a wide
range of exposure and/or growing conditions. For example, many of the experimental treatments
for exposures to elevated O3 on which the established E-R functions for the 11 tree seedling
species are based, involved W126 index levels well above 20 ppm-hrs and had many (tens to
42	This underlying database for the exposure is a key characteristic that sets this set of studies (and their associated
E-R analyses) apart from other available studies.
43	A quantitative analysis of E-R information for an additional species was considered in the 2014 WREA. But the
underlying study, rather than being an OTC controlled exposure study, involves exposure to ambient air along an
existing gradient of O3 concentrations in the New York City metropolitan area, such that O3 and climate
conditions were not controlled (2013 ISA, section 9.6.3.3). Based on comments from the CASAC on the WREA
cautioning against placing too much emphasis on these data (e.g., saying that the eastern cottonwood response
data from a single study "receive too much emphasis," explaining that these "results are from a gradient study
that did not control for ozone and climatic conditions and show extreme sensitivity to ozone compared to other
studies" and that "[although they are important results, they are not as strong as those from other experiments
that developed E-R functions based on controlled ozone exposure") (Frey, 2014, p. 10), the EPA did not include
the E-R function for eastern cottonwood among the set of tree seedling E-R functions given focus in the WREA,
or relied on in decision-making for the 2015 review (80 FR 65292, October 26, 2015.)
April 2022
4-37 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
more than a hundred) of hours of O3 concentrations above 100 ppb (Appendix 4 A, Table 4A-6;
Lefohn et al 1997). 44 45
From the available data, separate E-R functions were developed for each combination of
species and experiment46 (2013 ISA, section 9.6.1; Lee and Hogsett, 1996). For the 11 species,
there are 51 separate "experiment-specific" E-R functions (Appendix 4A, section 4A.1.1; ISA,
section 8.1.2.1.2). For six of the 11 species, the species-specific function is based on just one or
two experimental datasets (e.g., red maple), while for other species there were as many as 11
datasets supporting 11 experiment-specific E-R functions (e.g., ponderosa pine). The exposure
durations varied from periods of 82 to 140 days in a single year to periods of 180 to 555 days
occurring across two years (Lee and Hogsett, 1996; Appendix 4A, Table 4A-5). The
experimental datasets for more than half the 11 species include exposures occurring across two
years. To account for potential for a delayed response, some datasets are for growth
measurements taken in the spring of the year after a prior year growing season exposure and
others are for growth measurements taken immediately after the exposure. From the separate
species-experiment-specific E-R functions, species-specific composite E-R functions were
developed (Appendix 4A). In order to be utilized in deriving a single species-specific function
and to produce species-specific E-R functions of consistent duration, the separate species-
experiment-specific E-R functions were derived first based on the exposure duration of the
experiment and then normalized to 3-month (seasonal) periods47 (see Lee and Hogsett, 1996,
section 1.3; Appendix 4A).
The 11 species-specific composite median functions are presented in Appendix 4A (see
section 4A.1.1). Biomass growth loss predictions using the function for aspen was evaluated in
the 2013 and 2020 ISAs based on a recent study for aspen (2013 ISA, section 9.6.2; ISA,
44	Among the experiments on which the E-R functions are based, N100 values for exposure levels most common at
U.S. sites that meet the current standard (e.g., W126 index less than 25 ppm-hrs for a single season), extend up
above 10, to more than 40. Additionally, in a study that has reported the distributions of hourly concentrations,
the 90th percentile in replicates for one of the elevated O3 treatments ranged from 142 to 156 ppb, and the
maximum ranged from 210 to 260 ppb (Appendix 4A, Table 4A-6; Lefohn et al 1992).
45	Similarly, the experimental exposures in studies supporting some of the established E-R functions for 10 crop
species also include many hours with hourly O3 concentrations at or above 100 ppb (Lefohn and Foley, 1992).
46	Use of the term, experiment, refers to each separate seedling response dataset (from each separate harvest),
including, for example, a 2nd harvest in the spring that received the same growing season exposure as the response
documented for seedlings in the 1st harvest immediately following the growing season. As an initial step in
deriving species-specific E-R functions each of those response datasets were used to derive separate E-R
functions (Appendix 4A, Attachment 1).
47	Underlying the adjustment is a simplifying assumption of uniform W126 distribution across the exposure periods
and of a linear relationship between duration of cumulative exposure in terms of the W126 index and plant
growth response. Some functions for experiments that extended over two seasons were derived by distributing
responses observed at the end of two seasons of varying exposures equally across the two seasons (e.g.,
essentially applying the average to both seasons).
April 2022
4-38 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
Appendix 8, section 8,13.2). The species-specific composite E-R functions developed from the
experiment-specific functions indicate a wide variation in growth sensitivity of the studied tree
species at the seedling stage (Appendix 4A, section 4A.1.1). A stochastic analysis performed for
the 2014 WREA provides a sense of the variability and uncertainty associated with the estimated
E-R relationships among and within species48 (Appendix 4A, section 4A.1.1, Figure 4A-13).
Further, based on the species-specific E-R functions, the studied tree species appear to vary
widely in sensitivity to reduced growth at the seedling stage (Figure 4-3).
With regard to crops, established E-R functions are available for 10 crops: barley, field
corn, cotton, kidney bean, lettuce, peanut, potato, grain sorghum, soybean, and winter wheat
(Figure 4-4; Appendix 4A; ISA, Appendix 8, section 8,13.2). Since the 2015 review, new
evidence is available for seven soybean cultivars that confirms the reliability of the soybean E-R
functions developed from NCLAN data and indicates that they extend in applicability to recent
cultivars (ISA, Appendix 8, section 8.13.2).
o
oo
o
CD
o
m
cc
d
CM
o
o
o
0	10	20	30	40	50
W126 (ppm-hrs)
Figure 4-3. Established RBL functions for seedlings of 11 tree species.
48 The multiple functions derived for each species are derived from separate datasets, some of which may have the
same exposure during the growing season but which reflect response derived from seedlings harvested in the
spring subsequent to the growing season exposure (Lee and Hogsett, 1996). Accordingly, this analysis provides a
sense of both uncertainty in experimental design and environmental and seedling response variability.
Red Maple
•	Sugar Maple
•	Red Alder
Tulip Poplar
Ponderosa Pine
•	White Pine
•	Loblolly Pine
Virginia Pine
•	Aspen
•	Black Cherry
•	Douglas Fir
April 2022
4-39 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
o
00
o
•	Barley
•	Field Corn
•	Cotton
Kidney Bean
Lettuce
_l
>-
IY.
o
o
CD
O
O
C\J
O
0
10
20
30
40
50
W126 (ppm-hrs)
Figure 4-4. Established RYL functions for 10 crops.
Since the initial set of tree seedling studies were completed, several additional studies,
focused on aspen, have been published based on the Aspen FACE experiment in a planted forest
in Wisconsin; the findings were consistent with many of the OTC studies (ISA, Appendix 8,
section 8.13.2). Newly available studies that investigated growth effects of O3 exposures are also
consistent with the existing evidence base, and generally involved particular aspects of the effect
rather than expanding the conditions under which plant species, particularly trees, have been
assessed (ISA, section IS.5.1.2). These publications include a compilation of previously available
studies on plant biomass response to O3 (in terms of AOT40); the compilation reports linear
regressions conducted on the associated varying datasets (ISA, Appendix 8, section 8.13.2).
Based on these regressions, this study describes distributions of sensitivity to O3 effects on
biomass across many tree and grassland species, including 17 species native to the U.S. and 65
introduced species (ISA, Appendix 8, section 8.13.2; van Goethem et al., 2013). Additional
information is needed to describe 03 E-R relationships more completely for these species in the
U.S.49 As was noted in the 2013 ISA, "[i]n order to support quantitative modeling of exposure-
49 The studies compiled in this publication included at least 21 days exposure above 40 ppb 03 (expressed as AOT40
[seasonal sum of the difference between an hourly concentration above 40 ppb and 40 ppb]) and had a maximum
hourly concentration that was no higher than 100 ppb (van Goethem et al., 2013). The publication does not report
study-specific exposure durations, details of biomass response measurements or hourly O3 concentrations, making
it less useful for describing E-R relationships that might support estimation of specific impacts associated with air
quality conditions meeting the current standard (e.g., 2013 ISA, p. 9-118).
April 2022
4-40 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
response relationships, data should preferably include more than three levels of exposure, and
some control of potential confounding or interacting factors should be present in order to model
the relationship with sufficient accuracy" (2013 ISA, p. 9-118). The 2013 ISA further discussed
the differences across available studies, recognizing that the majority of studies contrast only two
(or sometimes three with the addition of a carbon filtration) O3 exposure levels. While such
studies can be important for verifying more extensive studies, they "do not provide exposure-
response information that is highly relevant to reviewing air quality standards" (2013 ISA, p. 9-
118).
4.3.3.2 Visible Foliar Injury
The evidence "continues to show a consistent association between visible injury and
ozone exposure," while also recognizing the role of modifying factors such as soil moisture and
time of day (ISA, section IS.5.1.1). The ISA, in concluding that the newly available information
is consistent with conclusions of the 2013 ISA, also summarizes several recently available
studies that continue to document that O3 elicits visible foliar injury in many plant species,
including a synthesis of previously published studies that categorizes studied species (and their
associated taxonomic classifications) as to whether or not Cb-related foliar injury has been
reported. Although this recent publication identifies many species in which visible foliar injury
has been documented to occur in the presence of elevated O3,50 it does not provide quantitative
information regarding specific exposure conditions or analyses of E-R relationships (ISA,
Appendix 8, section 8.3). Additionally, one recent study is identified as reporting visible foliar
injury in a non-native, yet established, and invasive tree species in a location with O3
concentrations corresponding to a seasonal W126 index of 11.6 ppm-hrs (ISA, Appendix 8,
sections 8.2 and 8.2.1). The annual fourth highest 8-hour daily maximum concentration for the
study year and location of this study (monitoring site 42-027-9991) is 76 ppb. The design value
for the 3-year design period encompassing the year and location of this study exceeds 70 ppb
(monitoring site 42-027-9991 for 2011-2013 design period), indicating that the air quality
associated with the exposure would not have met the current secondary standard.51
As in the past, the available evidence, while documenting that elevated O3 conditions in
ambient air generally result in visible foliar injury in sensitive species (when in a predisposing
50	The publication identifies 245 species across 28 plant genera, many native to the U.S., in which Ch-related visible
foliar injury has been reported (ISA, Appendix 8, section 8.3).
51	Ozone design values fortius period are available at: https://www.epa.gov/air-trends/air-quality-design-values. The
year 2011 is the first year for which data are available and adequate for use in deriving a design value at this
monitoring site.
April 2022
4-41 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
environment)52, it does not include a quantitative description of the relationship of incidence or
severity of visible foliar injury in sensitive species in natural areas of the U.S. with specific
metrics of O3 exposure. Several studies of the extensive USFS field-based dataset of visible
foliar injury incidence in forests across the U.S.53 illustrate the limitations of current
understanding of this relationship. For example, a study that was available in the 2015 review
presents a trend analysis of these data for sites located in 24 states of the northeast and north
central U.S. for the 16-year period from 1994 through 2009 that provides some insight into the
influence of changes in air quality and soil moisture on visible foliar injury and the difficulty
inherent in predicting foliar injury response under different air quality and soil moisture
scenarios (Smith, 2012, Smith et al., 2012; ISA, Appendix 8, section 8.2). This study, like prior
analyses of such data, shows the dependence of foliar injury incidence and severity on local site
conditions for soil moisture availability and O3 exposure. For example, while the authors
characterize the ambient air O3 concentrations to be the "driving force" behind incidence of
injury and its severity, they state that "site moisture conditions are also a very strong influence
on the biomonitoring data" (Smith et al., 2003). In general, the USFS data analyses have found
foliar injury prevalence and severity to be higher during seasons and sites that have experienced
the highest O3 than during other periods (e.g., Campbell et al., 2007; Smith, 2012).
4.3.3.2.1 Exposure Metrics
Although studies of the incidence of visible foliar injury in national forests, wildlife
refuges, and similar areas have often used cumulative indices (e.g., SUM06) to investigate
variations in incidence of foliar injury, studies also suggest an additional role for metrics focused
on peak concentrations (ISA; 2013 ISA; 2006 AQCD; Hildebrand et al., 1996; Smith, 2012).
Other studies have indicated this uncertainty regarding a most influential metric(s), by
recognizing a research need. For example, a study of six years of USFS biosite data for three
western states found that the biosites with the highest cumulative O3 exposure (SUM06 at or
above 25 ppm-hrs) had the highest percentage of biosites with injury and the highest mean
biosite index, with little discernable difference among the lower exposure categories; this study
also identified "better linkage between air levels and visible injury" as an O3 research need
52	As noted in the 2013 and 2020 ISAs, visible foliar injury usually occurs when sensitive plants are exposed to
elevated ozone concentrations in a predisposing environment, with a major modifying factor being the amount of
soil moisture available to a plant. Accordingly, dry periods are concluded to decrease the incidence and severity
of ozone-induced visible foliar injury, such that the incidence of visible foliar injury is not always higher in years
and areas with higher ozone, especially with co-occurring drought (ISA, Appendix 8, p. 8-23; Smith, 2012; Smith
et al., 2003).
53	These data were collected as part of the U.S. Forest Service Forest Health Monitoring/Forest Inventory and
Analysis (USFS FHM/FIA) biomonitoring network program (2013 ISA, section 9.4.2.1; Campbell et al., 2007,
Smith etal., 2012).
April 2022
4-42 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
(Campbell et al., 2007). More recent studies of the complete 16 years of data in 24 northeast and
north central states have suggested that a cumulative exposure index alone may not completely
describe the Cb-related risk of this effect (Smith et al., 2012; Smith, 2012). For example, Smith
(2012) observed there to be a declining trend in the 16-year dataset, "especially after 2002 when
peak ozone concentrations declined across the entire region" thus suggesting a role for peak
concentrations.
Some studies of visible foliar injury incidence data have investigated the role of peak
concentrations quantified by an O3 exposure index that is a count of hourly concentrations (e.g.,
in a year or growing season) above a threshold 1-hour concentration of 100 ppb, N100 (e.g.,
Smith, 2012; Smith et al., 2012). For example, analyses of injury patterns over 16 years at USFS
biosites in 24 states in the Northeast and North Central regions, in the context of the SUM06
index and N100 metrics (although not in statistical combination), suggested that there may be a
threshold exposure needed for injury to occur,54 and that the number of hours of elevated O3
concentrations during the growing season (such as what is captured by a metric like N100) may
be more important than cumulative exposure in determining the occurrence of foliar injury
(Smith, 2012).55 This finding is consistent with statistical analyses of seven years of visible foliar
injury data from a wildlife refuge in the mid-Atlantic (Davis and Orendovici, 2006). The latter
study investigated the fit of multiple models that included various metrics of cumulative O3 (e.g.,
SUM06, SUMO, SUM08), alone and in combination with some other variables (Davis and
Orendovici, 2006). Among the statistical models investigated, the model with the best fit to the
visible foliar injury incidence data was found to be one that included N100 and W126 indices, as
well as drought index (Davis and Orendovici, 2006).56
The established significant role of higher or peak O3 concentrations, as well as pattern of
their occurrence, in plant responses has also been noted in prior ISAs or AQCDs. The evidence
has included studies that use indices to summarize the incidence of injury on bioindicator species
present at specific monitored sites, as well as experimental studies that assess the occurrence of
foliar injury in response to varying O3 concentrations. In identifying support with regard to foliar
54	Authors of the study observed that "injury is minimized when seasonal ozone concentrations, especially peak
(N100) O3 concentrations, drop below a certain threshold as in 2004 through 2009" (Smith et al., 2012).
55	Although the ISA and past assessments have not described extensive evaluations of specific peak concentration
metrics such as the N100 (that might assist in identifying one best suited for such purposes), in summarizing this
study in the last review, the ISA observed that "[o]verall, there was a declining trend in the incidence of foliar
injury as peak O3 concentrations declined" (2013 ISA, p. 9-40).
56	The models evaluated included several with cumulative exposure indices alone. These included SUM60 (i.e.,
SUM06 in ppb), SUMO, and SUM80 (SUM08 in ppb), but not W126. They did include a model with W126 that
did not also include N100. Across all of these models evaluated, the model with the best fit to the data was found
to be the one that included N100 and W126, along with the drought index (Davis and Orendovici, 2006).
April 2022
4-43 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
injury as the response, the 2013 ISA and 2006 AQCD both cite studies that support the
"important role that peak concentrations, as well as the pattern of occurrence, plays in plant
response to O3" (2013 ISA, p. 9-105; 2006 AQCD, p. AX9-169). For example, a study of
European white birch saplings reported that peak concentrations and the duration of the exposure
event were important determinants of foliar injury (2013 ISA, section 9.5.3.1; Oksanen and
Holopainen, 2001). This study also evaluated tree growth, which was found to be more related to
cumulative exposure (2013 ISA, p. 9-105).57 A second study that was cited by both assessments
that focused on aspen, reported that "the variable peak exposures were important in causing
injury, and that the different exposure treatments, although having the same SUM06, resulted in
very different patterns of foliar injury (2013 ISA, p. 9-105; 2006 AQCD, p. AX9-169; Yun and
Laurence, 1999). As noted in the 2006 AQCD, the cumulative exposure indices (e.g., SUM06,
W126) were "originally developed and tested using only growth/yield data, not foliar injury" and
"[t]his distinction is critical in comparing the efficacy of one index to another" (2006 AQCD, p.
AX9-173). It is also recognized that where cumulative indices are highly correlated with the
frequency or occurrence of higher hourly average concentrations, they could be good predictors
of such effects (2006 AQCD, section AX9.4.4.3).
Dose modeling or flux models, discussed in section 4.3.3.1.1 above, have also been
considered for quantifying O3 dose that may be related to plant injury. Among the newly
available evidence is a study examining relationships between short-term flux and leaf injury on
cotton plants that described a sensitivity parameter that might characterize the influence on the
flux-injury relationship of diel and seasonal variability in plant defenses (among other factors)
and suggested additional research might provide for such a sensitivity parameter to "function
well in combination with a sigmoidal weighting of flux, analogous to the W126 weighting of
concentration", and perhaps an additional parameter (Grantz et al., 2013, p. 1710; ISA, Appendix
8, section 8.13.1). However, the ISA recognizes there is "much unknown" with regard to the
relationship between O3 uptake and leaf injury, and relationships with detoxification processes
(ISA, Appendix 8, section 8.13.1 and p. 8-184). These uncertainties have made this technique
less viable for assessments in the U.S., precluding use of a flux-based approach at this time (ISA,
Appendix 8, section 8.13.1 and p. 8-184).
A study (by Wang et al. [2012], newly described in the 2020 ISA) involved a statistical
modeling analysis on a subset of the years of data that were described in Smith (2012). This
analysis, which involved 5,940 data records from 1997 through 2007 from the 24 northeast and
north central states, tested a number of models for their ability to predict the presence of visible
57 The study authors concluded that "high peak concentrations were important for visible injuries and stomatal
conductance, but less important for determining growth responses" (Oksanen and Holopainen, 2001).
April 2022
4-44 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
foliar injury (a nonzero biosite score), regardless of severity, and generally found that the type of
O3 exposure metric (e.g., SUM06 versus N 100) made only a small difference, although the
models that included both a cumulative index (SUM06) and N100 had a just slightly better fit
(Wang et al., 2012). Based on their investigation of 15 different models, using differing
combinations of several types of potential predictors, the study authors concluded that they were
not able to identify environmental conditions under which they "could reliably expect plants to
be damaged" (Wang et al., 2012). This is indicative of the current state of knowledge, in which
there remains a lack of established quantitative functions describing E-R relationships that would
allow prediction of visible foliar injury severity and incidence under varying air quality and
environmental conditions.
4.3.3.2.2 Exposure Levels Associated with Effects
The available information related to O3 exposures associated with visible foliar injury of
varying severity also includes the dataset developed by the EPA in the 2015 review from USFS
BI scores, collected during the years 2006 through 2010 at locations in 37 states (Appendix 4C).
In developing this dataset, the BI scores were combined with estimates of soil moisture58 and
estimates of seasonal cumulative O3 exposure in terms of W126 index59 (Smith and Murphy,
2015; Appendix 4C). This dataset includes more than 5,000 records of which more than 80
percent have a BI score of zero (indicating a lack of visible foliar injury).60 While the estimated
W126 index assigned to records in this dataset (described in Appendix 4C) ranges from zero to
somewhat above 50 ppm-hrs, more than a third of all the records (and also of records with BI
scores above zero or five)61 are at sites with W126 index estimates below 7 ppm-hrs and only 8%
of the records have W126 index values above 15 ppm-hrs. In an extension of analyses developed
58	Soil moisture categories (dry, wet or normal) were assigned to each biosite record based on the NOAA Palmer Z
drought index values obtained from the NCDC website for the April-through-August periods, averaged for the
relevant year; details are provided in Appendix 4C, section 4C.2. There are inherent uncertainties in this
assignment, including the substantial spatial variation in soil moisture and large size of NOAA climate divisions
(hundreds of miles). Uncertainties and limitations in the dataset are summarized in Appendix 4C, section 4C.5).
59	The W126 index values assigned to the biosite locations are estimates developed for 12 kilometer (km) by 12 km
cells in a national-scale spatial grid for each year. The grid cell estimates were derived from applying a spatial
interpolation technique to annual W126 values derived from O3 measurements at ambient air monitoring locations
for the years corresponding to the biosite surveys (details in Appendix 4C, sections 4.C.2 and 4C.5).
60	In the scheme used by the USFS to categorize severity of biosite scores the lowest category encompasses BI
scores from zero to just below 5; scores of this magnitude are described as "little or no foliar injury" (Smith et al.,
2012). The next highest category encompasses scores from five to just below 15 and is described as "light to
moderate foliar injury," BI scores of 15 up to 25 are described as "moderate" and above 25 is described as
"severe" (Smith, 2012; Smith et al., 2012)..
61	One third (33%) of scores above 15 are at sites with W126 below 7 ppm-hrs (Appendix 4C, Table 4C-3).
April 2022
4-45 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
in the 2015 review, the presentation in the Appendix 4C62 describes the BI scores for the records
in the dataset in relation to the W126 index estimate for each record, using "bins" of increasing
W126 index values. The presentation utilizes the BI score breakpoints in the scheme used by the
USFS to categorize severity. This presentation indicates that, across the W126 bins, there is
variation in both the incidence of particular magnitude BI scores and in the average score per
bin. In general, however, the greatest incidence of records with BI scores above zero, five, or
higher - and the highest average BI score (as noted below) - occurs with the highest W126 bin
(i.e., the bin for W126 index estimates greater than 25 ppm-hrs), as seen in Figure 4-5 for records
in the normal soil moisture category63 (see also Appendix 4C, Table 4C-6).
The average BI score per W126 index bin is also variable, although for records
categorized as normal soil moisture, the average BI score in the highest W126 bin is noticeably
greater than for lower W126 bin scores (Figure 4-5). For example, the average BI score for the
normal soil moisture category is 7.9 among records with W126 index estimates greater than 25
ppm-hrs, compared to 1.6 among records for W126 index estimates between 19 and 25 ppm-hrs.
For records categorized as wet soil moisture, the sample size for the W126 bins above 13 ppm-
hrs is quite small (including only 18 of the 1,189 records in that soil moisture category),
precluding meaningful interpretation.64
While for BI scores above zero, the data may indicate a suggestion of increased incidence
among records in the W126 bins just below the highest (e.g., for the dry or normal soil moisture
categories), for BI scores above 5, there is little or no difference across the W126 bins except for
the highest bin, which is for W126 above 25 ppm-hrs (Appendix 4C, Table 4C-6). For example,
among records in the normal soil category, the proportion of records with BI above five
fluctuates between 5% and 13% across all but the highest W126 bin (>25 ppm-hrs) for which the
proportion is 41% (Appendix 4C, Table 4C-6). The same pattern is observed for BI scores above
15 at sites with normal and dry soil moisture conditions, albeit with lower incidences. For
example, the incidence of normal soil moisture records with BI score above 15 in the bin for
W126 index values above 25 ppm-hrs was 20% but fluctuates between 1% and 4% in the bin
with W126 index values at or below 25 ppm-hrs (Appendix 4C, Table 4C-6).
62	Beyond the presentation of a statistical analysis developed in the last review (Appendix 4C, section 4C.4.1), the
PA presentations are primarily descriptive (as compared to statistical) in recognition of the limitations and
uncertainties of the dataset (Appendix 4C, section 4C.5).
63	The number of records per W126 bin in Figure 4-5 ranges from a low of 15 in the ">19-25" bin to 158 in the "<7"
bin (Appendix 4C, Table 4C-4).
64	In the full database for the wet soil moisture category, there are only 18 records at sites with a W126 index value
above 13 ppm-hrs, with 9 or fewer (less than 1%) in each of them (Appendix 4C, Table 4C-3). Across the W126
bins in which at least 1% of the wet soil moisture records are represented, differences of incidence or average
score of lower bins from the highest bin is less than a factor of two (Appendix 4C, section 4C.4.2).
April 2022
4-46 External Review Draft - Do Not Quote or Cite

-------
1
2
3
i
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
100
o
90
80 tป
70
0
O
s
o
o
Q
o
o
o
o
o
0
<7 >7-9 >9-11 >11-13 >13-15 >15-17 >17-19 >19-25 >25
W12I Index Bin
Key: The boxes denote the 25th, 50th and 75th percentiles, the x's the arithmetic mean, and the whiskers
denote the value equal to the 75th percentile plus 1.5 times the interquartile range (75th minus 25th percentile).
Circles show scores higher than that.
Figure 4-5. Distribution of nonzero BI scores at USFS biosites (normal soil moisture)
grouped by assigned W126 index estimates.
Overall, the dataset described in Appendix 4C generally indicates the risk of injury, and
particularly injury considered at least light, moderate or greater injury, to be higher at the highest
W126 index values, with appreciable variability in the data for the lower bins. This appears to be
consistent with the conclusions of the detailed quantitative analysis studies, summarized above,
that the pattern is stronger at higher O3 concentrations. A number of factors may contribute to the
observed variability in BI scores and lack of a clear pattern with W126 index bin; among others,
these may include uncertainties in assignment of W126 estimates and soil moisture categories to
biosite locations, variability in biological response among the sensitive species monitored, and
the potential role of other aspects of O3 air quality not captured by the W126 index. Thus, the
dataset has limitations affecting associated conclusions, and uncertainty remains regarding the
tools for and the appropriate metric (or metrics) for quantifying influence of O3 exposures, as
well as perhaps for quantifying soil moisture conditions, with regard to their influence on extent
and/or severity of injury in sensitive species in natural areas, as quantified via BI scores (Davis
and Orendovici, 2006; Smith et al., 2012; Wang et al., 2012). Accordingly, the limitations
recognized in the past remain in our ability to quantitatively estimate incidence and severity of
visible foliar injury likely to occur in areas across the U.S. under different air quality conditions
over a year, or over a multi-year period (Appendix 4C, section 4C.5).
April 2022
4-47 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
4.3.3.3 Other Effects
With regard to radiative forcing and subsequent climate effects associated with the global
tropospheric abundance of O3, the available evidence does not provide more detailed quantitative
information regarding O3 concentrations at the national scale than was available in the 2015
review (ISA, Appendix 9). Rather, it is noted that "the heterogeneous distribution of ozone in the
troposphere complicates the direct attribution of spatial patterns of temperature change to ozone
induced [radiative forcing]" and there are "ozone climate feedbacks that further alter the
relationship between ozone [radiative forcing] and temperature (and other climate variables) in
complex ways" (ISA, Appendix 9, section 9.3.1, p. 9-19). Further, "precisely quantifying the
change in surface temperature (and other climate variables) due to tropospheric ozone changes
requires complex climate simulations that include all relevant feedbacks and interactions" (ISA,
section 9.3.3, p. 9-22). Yet, there are limitations in current climate modeling capabilities for O3;
an important one is representation of important urban- or regional-scale physical and chemical
processes, such as O3 enhancement in high-temperature urban situations or O3 chemistry in city
centers where NOx is abundant. Such limitations impede our ability to quantify the impact of
incremental changes in ground-level O3 concentrations in the U.S. on radiative forcing and
subsequent climate effects.
With regard to tree mortality, the evidence available in the last several reviews included
field studies of pollution gradients that concluded O3 damage to be an important contributor to
tree mortality although "several confounding factors such as drought, insect outbreak and forest
management" were identified as potential contributors (2013 ISA, p. 9-81, section 9.4.7.1).
Among the newly available studies, there is only limited experimental evidence that isolates the
effect of O3 on tree mortality65 and might be informative regarding O3 concentrations of interest
in the review, and evidence is lacking regarding exposure conditions closer to those occurring
under the current standard and any contribution to tree mortality.
With regard to alteration of herbivore growth and reproduction, although "there are
multiple studies demonstrating ozone effects on fecundity and growth in insects that feed on
ozone-exposed vegetation", "no consistent directionality of response is observed across studies
and uncertainties remain in regard to different plant consumption methods across species and the
exposure conditions associated with particular severities of effects" (ISA, pp. ES-18, IS-64, IS91
and Appendix 8, section 8.6.3). Such limitations and uncertainties in the evidence base for this
category of effects preclude broader characterization, as well as quantitative analysis related to
65 Of the three new studies on tree mortality described in the ISA is another field study of a pollution gradient that,
like such studies in prior reviews, recognizes O3 exposures as one of several contributing environmental and
anthropogenic stressors (ISA, p. 8-55).
April 2022
4-48 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
air quality conditions meeting the O3 standard. As characterized in the ISA, uncertainties remain
in the evidence; these relate to the different plant consumption methods across species and the
exposure conditions associated with particular responses, as well as variation in study designs
and endpoints used to assess O3 response (ISA, IS.6.2.1 and Appendix 8, section 8.6). Thus,
while the evidence describes changes in nutrient content and leaf chemistry following O3
exposure (ISA, p. IS-73), the effect of these changes on herbivores consuming the leaves is not
well characterized or clear.
The evidence for a second newly identified category of effects, alteration of plant-insect
signaling, draws on new research yielding clear evidence of O3 modification of VPSCs and
behavioral responses of insects to these modified chemical signals (ISA, section IS.6.2.1). While
the evidence documents effects on plant production of signaling chemicals and on the
atmospheric persistence of signaling chemicals, as well as on the behaviors of signal-responsive
insects, it is limited with regard to characterization of mechanisms and the consequences of any
modification of VPSCs by O3 (ISA, section IS.6.2.1). Further, the evidence includes a relatively
small number of plant species and plant-insect associations66 and is limited to short, controlled
exposures, posing limitations for our purposes of considering the potential for associated impacts
to be elicited by air quality conditions that meet the current standard (ISA, section IS.6.2.1 and
Appendix 8, section 8.7).
For categories of vegetation-related effects that were recognized in past reviews, other
than growth and visible foliar injury (e.g., reduced plant reproduction, reduced productivity in
terrestrial ecosystems, alteration of terrestrial community composition and alteration of below-
ground biogeochemical cycles), the newly available evidence includes a variety of studies that
quantify exposures of varying duration in various countries using a variety of metrics (ISA,
Appendix 8, sections 8.4, 8.8 and 8.10). The ISA additionally describes publications that
summarize previously published studies in several ways. For example, a meta-analysis of
reproduction studies categorized the reported O3 exposures into bins of differing magnitude,
grouping differing concentration metrics and exposure durations together, and performed
statistical analyses to reach conclusions regarding the presence of an 03-related effect (ISA,
Appendix 8, section 8.4.1). While such studies continue to support conclusions of the ecological
hazards of O3, they do not improve capabilities for characterizing the likelihood of such effects
under patterns of environmental O3 concentrations occurring with air quality conditions that meet
the current standard (e.g., factors such as variation in exposure assessments and limitations in
response information preclude detailed analysis for such conditions).
66 The available studies vary with regard to the experimental exposure circumstances in which the different types of
effects have been reported; most of the studies have been carried out in laboratory conditions rather than in
natural environments (ISA, section IS.6.2.1).
April 2022
4-49 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
As at the time of the 2015 review, growth impacts, most specifically as evaluated by RBL
for tree seedlings and RYL for crops, remain the type of vegetation-related effects for which we
have the best understanding of exposure conditions likely to elicit them. Accordingly, as was the
case in the 2015 review, the quantitative analyses of exposures occurring under air quality that
meets the current standard (summarized in section 4.4 below) is focused primarily on the W126
index, given its established relationship with growth effects.
4.3.4 Key Uncertainties
The type of uncertainties for each category of effects generally tends to vary in relation to
the maturity of the associated evidence base from those associated with overarching
characterizations of the effects to those associated with quantification of the cause-and-effect
relationships. For example, given the longstanding nature of the evidence for many of the
vegetation effects identified in the ISA as causally or likely causally related to O3 in ambient air,
the key uncertainties and limitations in our understanding of these effects relate largely to the
implications or specific aspects of the evidence, as well as to current understanding of the
quantitative relationships between O3 concentrations in the environment and the occurrence and
severity (or relative magnitude) of such effects or understanding of key influences on these
relationships. For more newly identified categories of effects, the evidence may be less
extensive, thus precluding consideration of such details.
• What are important uncertainties in the evidence? To what extent have important
uncertainties in the evidence identified in the past been reduced and/or have new
uncertainties been recognized?
Among the categories of effects identified in past reviews, key uncertainties remain in the
evidence. The category of O3 welfare effects for which current understanding of quantitative
relationships is strongest is reduced plant growth. As a result, this category was the focus of
decision-making on the standard in the 2015 review, with RBL in tree seedlings playing the role
of surrogate (or proxy) for the broader array of vegetation-related effects that range from the
individual plant level to ecosystem services. Limitations in the evidence base and associated
uncertainties recognized then remain and include a number of uncertainties that affect
characterization of the magnitude of cumulative exposure conditions that might be expected to
elicit growth reductions in U.S. forests. These limitations and uncertainties relate both to aspects
related to the extent and precision of the E-R evidence for the O3 concentration patterns and
associated cumulative seasonal exposures common in areas of the U.S. that meet the current
standard, and with regard broader interpretation of RBL estimates with regard to longer term and
population and ecosystem scale impacts.
April 2022
4-50 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
Uncertainties in RBL estimates for today's O3 air quality stem from limitations and
imprecision in our tools, and aspects of the underlying data. While the tree seedling E-R
relationships for the 11 species are long-established, there is large variation among the species
regarding the number of experimental datasets supporting each, and among the species and
experiments in the duration of the controlled exposures assessed. For example, the E-R function
for aspen (representing a mixture of wild type and four specific clones) is based on functions for
13 experimental datasets (for six different exposure studies), while the E-R functions for the red
maple and Virginia pine were each derived from a single experimental study (of 55 days for red
maple and 159 days for Virginia pine) (Appendix 4A, section 4A.2, Table 4A-6; 1996 AQCD,
Table 5-28; Lee and Hogsett, 1996).
Across these varied datasets, the controlled exposure periods vary in duration both within
and across years (e.g., from exposure periods of 82 to 140 days in a single year to periods of 180
to 555 days distributed across two years) and in whether measurements were made immediately
following exposure period or in the subsequent spring. The final set of E-R functions were
derived first for the exposure duration of the experiment and, then adjusted or normalized to 3-
month periods based on assumptions regarding relationships between duration, cumulative
exposure in terms of W126 index and plant growth response (Lee and Hogsett, 1996, section 1.3;
Appendix 4A, Attachment 1). For example, while the functions are defined as describing a
seasonal response, some were derived by distributing responses observed at the end of two
seasons of varying exposures equally across the two seasons (essentially applying the average to
both seasons). Uncertainty associated with this variation in durations and assumptions inherent in
the adjustment step is contributed to RBL estimates derived through application of the resultant
functions.
Further, there is uncertainty associated with estimates of effects across multiple years
related to the limited availability of studies of seasonal growth effects on trees across multiple
years (particularly more than two) that have also reported detailed O3 concentration data
throughout the exposure. This contributes uncertainty, and accordingly a lack of precision, to an
understanding of the quantitative impacts of seasonal O3 exposure, including its year-to-year
variability, on tree growth and annual biomass accumulation. This uncertainty limits our
understanding of the extent to which tree biomass would be expected to appreciably differ at the
end of multi-year exposures for which the overall average exposure is the same, yet for which
the individual year exposures vary in different ways (e.g., as analyzed in Appendix 4D).67
67 Variation in annual W126 index values is described in Appendix 4D, indicating for the period, 2016-2018, that the
amount by which annual W126 index values at a site differ from the 3-year average varies, but generally falls
below 10 ppm-hrs across all sites and generally below 5 ppm-hrs at sites with design values at or below 70 ppb
(Appendix 4D, Figure 4D-7).
April 2022
4-51 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
One available study of multi-year growth effects for aspen, was summarized and assessed
in the 2020 and 2013 IS As with regard the extent to which it confirmed Cb-related biomass
impacts estimated using the established E-R functions for aspen (King et al., 2005; 2013 ISA,
section 9.6.3.2; ISA, Appendix 8, section 8.13.2). The 2013 ISA applied the E-R functions to O3
exposure (quantified as cumulative average seasonal W126 index) at each of six consecutive
years and compared the estimated aboveground biomass to estimates based on data reported for
each year by the study (2013 ISA, section 9.6.3.2). The conclusions reached were that the
experimental observations are "very close" to estimates based on the established E-R function
for aspen, and that "the function based on one year of growth was shown to be applicable to
subsequent years" (2013 ISA, p. 9-135; ISA, Appendix 8, p. 8-186). A similar assessment in the
2020 ISA that applied the E-R functions to O3 exposure, quantified individually after a 92-day
season in each of six consecutive years similarly also concluded that predictions based on the E-
R functions generally agreed with the observations, given generally similar pattern and
magnitude of cumulative response (with some variation). In addition to indicating general
support for the E-R functions based on the cumulative W126 index, these assessments also
indicated uncertainty associated with the relative influences of individual seasonal exposures and
longer-term exposures, as represented by a cumulative average, given that either multiyear
average or single year W126 estimates provided general agreement with experimental
observations (Appendix 4A, section 4A.3.1; 2013 ISA, Figure 9-20; 2020 ISA, Appendix 8,
Figure 8-17).
Another area of important uncertainties relates to the extent to which the E-R functions
for reduced growth in tree seedlings are also descriptive of such relationships during later
lifestages, for which there is a paucity of established E-R relationships. Although such
information is limited with regard to mature trees, the analyses in the 2013 and 2020 IS As
(summarized above) indicated that reported growth response of young aspen over six years was
similar to the reported growth response of seedlings (ISA, Appendix 8, section 8.13.2; 2013 ISA,
section 9.6.3.2). Evidence is lacking, however, on the shape of such relationships for older,
mature trees, or the extent to which these relationships in seedlings might also reflect responses
in older, mature trees
Additionally, there are uncertainties with regard to the extent to which various factors in
natural environments can either mitigate or exacerbate predicted 03-plant interactions and
contribute variability in vegetation-related effects, including reduced growth. Such factors
include multiple genetically influenced determinants of O3 sensitivity, changing sensitivity to O3
across vegetative growth stages, co-occurring stressors and/or modifying environmental factors.
Such factors contribute uncertainties to interpretations of potential impacts in a season as well as
over multi-year periods. With regard to the latter, there is variability in ambient air O3
April 2022
4-52 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
concentrations from year to year, as well as year-to-year variability in environmental factors,
including rainfall and other meteorological factors that affect plant growth and reproduction,
such as through changes in soil moisture. These variabilities contribute uncertainties to estimates
of the occurrence and magnitude of Cb-related effects in any year, and to such estimates over
multi-year periods, as well as related effects in associated communities and ecosystems. All the
factors identified here contribute uncertainty and an associated imprecision or inexactitude to
estimates for trees in natural areas derived from the E-R functions and W126 index values in a
single year/season.
The uncertainties identified here are important for our interpretation of potential impacts
under air quality conditions that meet the current standard, which as described in section 4.4.2
below are generally associated with cumulative seasonal exposures lower than 20 or 25 ppm-hrs,
in terms of W126 index, and with quite lowNlOO values in a year. Such conditions are not
extensively represented in the datasets on which the tree seedling E-R functions are based. While
the functions have been concluded to provide a good fit to the underlying experimental datasets,
the datasets vary with regard to their representation of relatively lower O3 treatment levels,68 in
terms of W126 index (e.g., below 20 ppm-hrs). Additionally, the experimental datasets include
patterns of hourly concentrations that differ markedly from those common in areas meeting the
current standard (e.g., with greater prevalence of peak hourly concentrations). With regard to
W126 index level, the W126 index levels across the experiments range as high as 109.5 ppm-hrs
across a 121-day exposure (which, assuming a constant daily cumulative exposure would
correspond to 83 ppm-hrs across a 92-day season). Three of the eleven species include just one
of their treatment levels below a W126 index value of 20 ppm-hrs , with the other levels ranging
from 25.6 ppm-hrs (over 112 days) to 109.5 ppm-hrs (over 121 days), corresponding to 21 to 83
ppm-hrs for a 92-day season, based on assuming uniform cumulative exposure distribution
across the period (Appendix 4A, Table 4A-5).69 With regard to peak concentrations, for the
experimental treatments with W126 index levels of a magnitude common at U.S. sites that meet
the current standard (e.g., less than 20 ppm-hrs for a single season), the values for N100 extend
up above 10, to more than 40 in one instance (Appendix 4A, Table 4A-6, black cherry and
aspen). Across the full set of treatments, values for N100 extend into the hundreds up to 515 in a
single treatment over 121 days. As discussed in section 4.4.1 below, such occurrences of
concentrations at or above 100 ppb are not common for U.S. sites that meet the current standard,
68	As noted in Appendix 4A, section 4A.2, the baseline, untreated, ambient air was treated with O3 to develop
exposure levels for comparison to charcoal-filtered air and the baseline ambient air.
69	For three of the five species in Table 4A-5 in Appendix 4A for which only one treatment exposure is for a W126
index below 20 ppm-hrs, there are three other treatments that range from a W126 index of 25 ppm-hrs up to one
of 109.5 ppm-hrs (Appendix 4A, Table 4A-6).
April 2022
4-53 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
at which N100 is virtually always less than 10 (and generally less than 5 [see Figure 4-7
below])70 Collectively, all of the factors identified above contribute uncertainty and an
associated imprecision or inexactitude to estimates based on the E-R functions for W126 index
levels at sites in the U.S. with air quality meeting the current standard.
We also note, as recognized in the 2015 review, uncertainties in the extent to which the
11 tree species for which there are established E-R functions encompass the range of O3 sensitive
species in the U.S., and also the extent to which they represent U.S. vegetation as a whole. These
11 species include both deciduous and coniferous trees with a wide range of sensitivities and
species native to every NOAA climate region across the U.S. and in most cases are resident
across multiple states and regions. While recognizing this uncertainty, the available information
does not lead us to assume any difference in the range of sensitivity indicated by the species with
E-R functions.71
There are also uncertainties associated with our consideration of the magnitude of tree
growth effects, quantified as RBL, that might cause or contribute to adverse effects for trees,
forests, forested ecosystems, or the public welfare, these are related to various uncertainties or
limitations in the evidence base, including those associated with relating magnitude of tree
seedling growth reduction to larger-scale forest ecosystem impacts. Additionally, several factors
can also influence the degree to which Cb-induced growth effects in a sensitive species affect
forest and forest community composition and other ecosystem service flows (e.g., productivity,
belowground biogeochemical cycles, and terrestrial ecosystem water cycling) from forested
ecosystems. These include (1) the type of stand or community in which the sensitive species is
found (i.e., single species versus mixed canopy); (2) the role or position the species has in the
stand (i.e., dominant, sub-dominant, canopy, understory); (3) the O3 sensitivity of the other co-
occurring species (O3 sensitive or tolerant); and (4) environmental factors, such as soil moisture
and others. The lack of such established relationships with O3 complicates consideration of the
extent to which different estimates of impacts on tree seedling growth would indicate
significance to the public welfare. Further, efforts to estimate O3 effects on carbon sequestration
are handicapped by the large uncertainties involved in attempting to quantify the additional
70	Among published studies of the datasets for the eleven E-R functions, the findings for at least one study (black
cherry) reported statistical significance only for biomass effects observed for the highest O3 exposure, which had
a seasonal W126 index of 23 ppm-hrs and 77 hours with an O3 concentration at or above 100 ppb (e.g., Appendix
4A, Table 4A-6, black cherry).
71	The CASAC in the 2015 review recognized this uncertainty, expressing the view that it should be anticipated that
there are highly sensitive vegetation species for which we do not have E-R functions and others that are
insensitive (Frey, 2014, p. 15), and concluding it to be more appropriate to assume that the sensitivity of species
without E-R functions might be similar to the range of sensitivity for those species with E-R functions (Frey,
2014, p. 11).
April 2022
4-54 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
carbon uptake by plants as a result of avoided Cb-related growth reductions. Such analyses
require complex modeling of biological and ecological processes with their associated sources of
uncertainty.
With regard to crop yield effects, as at the time of the 2015 review, we recognize the
potential for greater uncertainty in estimating the impacts of O3 exposure on agricultural crop
production than that associated with O3 impacts on vegetation in natural forests. This relates to
uncertainty in the extent to which agricultural management methods influence potential for O3-
related effects and accordingly, the applicability of the established E-R functions for RYL in
current agricultural areas. Additionally, as changes in yield of commercial crops and commercial
commodities may affect producers and consumers differently, consideration of these effects in
terms of potential adversity to the public welfare impacts is limited.
With regard to visible foliar injury, for which longstanding evidence documents a causal
role for O3, important uncertainties and limitations fall into two categories. The first category
relates to our understanding of the key aspects of O3 concentrations - and other key variables
(e.g., soil moisture) - that have a direct bearing on the severity and incidence of vegetation
injury, while the second concerns the impacts on aesthetic and recreational values of various
severities and incidences of injury. With regard to the former, there is a lack of detailed
understanding of specific patterns of O3 concentrations over a growing season and the key
aspects of those patterns (e.g., incidence of concentrations of particular magnitude) that
contribute to an increased incidence and severity of injury occurrence in the U.S. For example,
"the incidence of visible foliar injury is not always higher in years and areas with higher ozone,
especially with co-occurring drought" (ISA, Appendix 8, p. 8-24). Accordingly, there are no
established, quantitative E-R functions that document visible foliar injury severity and incidence
under varying air quality and environmental conditions (e.g., soil moisture). As discussed in
section 4.3.3.2 above, the available studies that have investigated the role of different variables,
including different metrics for characterizing O3 concentrations over a growing season, do not
provide a basis for a single metric that would characterize the potential for different patterns of
O3 concentrations to contribute to different incidences and severity of foliar injury in U.S.
forests. Further, while several studies of the USFS biosite dataset indicate a role for two metrics -
one reflecting cumulative, concentration-weighted exposures and a second that reflects peak
concentrations, statistical analyses of a number of models containing various metrics and
combinations of metrics have not been able to identify environmental conditions under which
visible foliar injury could be reliably expected (Smith, 2012; Wang et al., 2012). The second
category of uncertainties and limitations concerns the information that would support associated
judgments on the public welfare significance of different patterns of and severity of foliar injury,
such as the extent to which such effects in areas valued by the public for different uses may be
April 2022
4-55 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
considered adverse to public welfare. In considering this issue, we note that some level of
severity of injury to a tree stand would be obvious to the casual observer (e.g., when viewing a
stand covering a hillside from a distance), and some level of severity of injury (e.g., leaf and
crown damage that appreciably affects overall plant physiology) would also be expected to affect
plant growth and reproduction. The extent to which recreational values are affected by lesser
levels of injury severity and incidence is not clear from the available information. Thus,
limitations and uncertainties in the available information, such as those described above,
complicate our ability to comprehensively estimate the potential for visible foliar injury, its
severity or extent of occurrence for specific air quality conditions, and associated public welfare
implications, thus affecting a precise identification of air quality conditions that might be
expected to provide a specific level of protection for this effect.
During the 2015 review, the 2013 ISA did not assess the evidence of O3 exposure and
tree mortality with regard to its support for inference of a causal relationship. Evidence available
in the last several reviews included field studies of pollution gradients that concluded O3 damage
to be an important contributor to tree mortality although several confounding factors such as
drought, insect outbreak and forest management were identified as potential contributors (2013
ISA, section 9.4.7.1). Since the 2015 review, three additional studies have been identified, as
summarized in section 4.3.1 above, contributing to the ISA conclusion of sufficient evidence to
infer a likely causal relationship for O3 with tree mortality (ISA, Appendix 8, section 8.4). As
noted in the ISA, there is only limited evidence from experimental studies that isolate the effect
of O3 on tree mortality, with the recently available Aspen FACE study of aspen survival
involving cumulative seasonal exposures above 30 ppm-hrs during the first half of the 11-year
study period (ISA, Appendix 8, Tables 8-8 and 8-9). Evidence is lacking regarding exposure
conditions closer to those occurring under the current standard and any contribution to tree
mortality.
In the case of the two newly identified categories of effects, the key uncertainties relate to
comprehensive characterization of the effects. For example, with regard to alteration of herbivore
growth and reproduction, although "there are multiple studies demonstrating ozone effects on
fecundity and growth in insects that feed on ozone-exposed vegetation", "no consistent
directionality of response is observed across studies and uncertainties remain in regard to
different plant consumption methods across species and the exposure conditions associated with
particular severities of effects" (ISA, pp. ES-18, IS-64, IS91 and Appendix 8, section 8.6.3).
Such limitations and uncertainties in the evidence base for this category of effects preclude
broader characterization, as well as quantitative analysis related to air quality conditions meeting
the O3 standard. As characterized in the ISA, uncertainties remain in the evidence; these relate to
the different plant consumption methods across species and the exposure conditions associated
April 2022
4-56 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
with particular responses, as well as variation in study designs and endpoints used to assess O3
response (ISA, IS.6.2.1 and Appendix 8, section 8.6). Thus, while the evidence describes
changes in nutrient content and leaf chemistry following O3 exposure, the effect of these changes
on herbivores consuming the leaves is not well characterized or clear (ISA, p. IS-73).
The evidence for a second newly identified category of effects, alteration of plant-insect
signaling, draws on new research that has provided clear evidence of O3 modification of VPSCs
and behavioral responses of insects to these modified chemical signals. Most of these studies,
however, have been carried out in laboratory conditions rather than in natural environments, and
involve a relatively small number of plant species and plant-insect associations. While the
evidence documents effects on plant production of signaling chemicals and on the atmospheric
persistence of signaling chemicals, as well as on the behaviors of signal-responsive insects, it is
limited with regard to characterization of mechanisms and the consequences of any modification
of VPSCs by O3 (ISA, section IS.6.2.1). Further, the available studies vary with regard to the
experimental exposure circumstances in which the different types of effects have been reported
(most of the studies have been carried out in laboratory conditions rather than in natural
environments), and many of the studies involve quite short controlled exposures (hours to days)
to elevated concentrations, posing limitations for our purposes of considering the potential for
impacts associated with the studied effects to be elicited by air quality conditions that meet the
current standard (ISA, section IS.6.2.1 and Appendix 8, section 8.7).
With regard to radiative forcing and climate effects, "uncertainty in the magnitude of
radiative forcing estimated to be attributed to tropospheric ozone is a contributor to the relatively
greater uncertainty associated with climate effects of tropospheric ozone compared to such
effects of the well mixed greenhouse gases (e.g., carbon dioxide and methane)" (ISA, section
IS.6.2.2). With regard to O3 effects on temperature, "the heterogeneous distribution of ozone in
the troposphere complicates the direct attribution of spatial patterns of temperature change to
ozone induced RF" and the existence of O3 climate feedbacks "further alter the relationship
between ozone RF and temperature (and other climate variables) in complex ways" (ISA,
Appendix 9, section 9.3.1). Thus, various uncertainties "render the precise magnitude of the
overall effect of tropospheric ozone on climate more uncertain than that of the well-mixed
GHGs" (ISA, Appendix 9, section 9.3.3). Further, "[cjurrent limitations in climate modeling
tools, variation across models, and the need for more comprehensive observational data on these
effects represent sources of uncertainty in quantifying the precise magnitude of climate responses
to ozone changes, particularly at regional scales" (ISA, Appendix 9, section 9.3.3).
April 2022
4-57 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
4.4 EXPOSURE AND AIR QUALITY INFORMATION
In general, decision-making in the 2015 review placed greatest weight on estimates of
cumulative exposures to vegetation based on ambient air monitoring data and consideration of
those estimates in light of E-R functions for Cb-related reduction in tree seedling growth
(summarized in section 4.3.3 above). These analyses supported the consideration of the potential
for O3 effects on tree growth and productivity, as well as its associated impacts on a range of
ecosystem services, including forest ecosystem productivity and community composition (80 FR
65292, October 26, 2015). These analyses were recognized as involving relatively reduced
uncertainty (compared to the national or regional-scale modeling performed in the 2015 review)
for the purposes of informing a characterization of cumulative O3 exposure (in terms of the
W126 index) associated with air quality just meeting the existing standard (IRP, section 5.2.2).
The lesser uncertainty of these air quality monitoring-based analyses contributed to their being
more informative in the 2015 review and to their being updated in the 2020 PA. A second set of
air quality analyses was also considered in the 2020 decision; these analyses investigated the
occurrence of peak concentrations at sites for which the O3 concentrations meet different design
values or contribute to different cumulative exposure levels in terms of the W126 index (Wells,
2020). Both sets of analyses have been updated for this reconsideration of the 2020 decision
using the more recently available air quality data now available (Appendices 4D and 4F).
The first set of analyses are air quality and exposure analyses. They are an update of the
analyses considered in the 2015 decision establishing the current standard, and in the 2020
decision to retain that standard. This set of analyses, in 2015 and 2020, as well as the current
updated analyses presented here, evaluate W126-based cumulative exposure estimates at all U.S.
monitoring locations, nationwide, and at the subset of sites in or near Class I areas, during 3-year
periods that met the then-current standard and potential alternatives (80 FR 65485-86, Table 3,
October 26, 2015; Wells, 2015; 2020 PA, section 4.4). For the 2015 and 2020 decisions, W126
index values72 occurring in locations with air quality meeting the then-current standard (or
potential alternatives) were considered in the context of the magnitude of W126 exposure index
associated with an estimate of 6% RBL in tree seedlings for the median tree species among the
11 species for which there are established E-R relationships (80 FR 65391-92, Table 4, October
26, 2015; 2020 PA, section 4.4). That magnitude of W126 index is 19 ppm-hrs (80 FR 65391-
65392). This set of analyses also includes an evaluation of relationships between W126 index
72 Based on judgments in the last review, the W126 metric analyzed and considered in the 2015 decision was the 3-
year average of consecutive year seasonal W126 index values (derived as described in section 4.3.3.1 above).
April 2022
4-58 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
values and design values73 based on the form and averaging time of the then-current secondary
standard (Wells, 2015; 2020 PA, section 4.4).
The second set of analyses (initially performed for consideration in the 2020 decision and
updated here) focus on the occurrence of peak concentrations, investigating the occurrence of
peak concentrations at sites for which the O3 concentrations meet different design values or
contribute to different cumulative exposure levels in terms of the W126 index. The metrics used
for these analyses are the number of hours in a year for which the O3 concentration was at or
above 100 ppb (N100), and the number of days in a year in which there was at least one hour
with an O3 concentration at or above 100 ppb (D100). The value of 100 ppb is used here as it has
been in some studies focused on O3 effects on vegetation (and discussed in section 4.3.3 2
above), simply as an indicator of elevated or peak hourly O3 concentrations (e.g., Lefohn et al.,
1997, Smith, 2012; Davis and Orendovici, 2006; Kohut, 2007). Other values that have also been
considered in this way in other studies are 95 ppb and 110 ppb (2013 ISA, section 9.5.3.1). These
analyses provided additional information for the 2020 review beyond that provided by the first
set of analyses that focused only on W126 index.
Both sets of analyses described here have been performed with the expanded set of air
monitoring data now available,74 which includes 1,578 monitoring sites with sufficient data for
derivation of design values (Appendix 4D, section 4D.2.2; Appendix 4F). Both sets of analyses
include a component based on data for the most recent periods, and a second component
considering data across the full historical period back to 2000, which is now expanded from that
previously available.75 The most recent data analyzed are those for the design value period from
2018 to 2020. The first set of analyses include a focus on all sites in the U.S., as well as on the
subset of sites in or near Class I areas is described in detail in Appendix 4D. The second set of
analyses, which investigate the occurrence of peak concentrations at sites varying by design
value and W126 index, are described in detail in Appendix 4F.
For all monitoring sites with valid design values for the recent period of 2018 through
2020, Figure 4-6 presents the 3-year average seasonal W126 index and also denotes whether
each site meets the current standard. Similarly, Figure 4-7 and Figure 4-8 present N100 and
D100 values, respectively, for these sites. Consideration of all three figures indicates that the
73	As described in earlier chapters, a design value is a statistic that describes the air quality status of a given area
relative to the level of the standard, taking the averaging time and form into account. For example, a design value
of 75 would have indicated O3 concentrations that just met the prior standard in a specific 3-yr period.
74	In addition to being expanded with regard to data for more recent time periods than previously available, the
current dataset also includes a small amount of newly available older data for some monitoring sites that are now
available in the AQS.
75	In the 2015 review, the dataset analyzed included data from 2000 through 2013 (Wells, 2015).
April 2022
4-59 External Review Draft - Do Not Quote or Cite

-------
1	monitoring sites with design values above the level of the current standard have the higher W126
2	index values and also the higher values of N100 and D100 (compared to monitoring sites not
3	meeting the current standard, denoted with triangles). It can also be seen that there are some sites
4	that have relatively lower W126 index values, e.g., in the Northwest, Northeast and Midwest,
5	while recording N100 or D100 values of more than 5 (including some N100 values above 1010
6	and 5 respectively. The sections below summarize more completely the findings of all the air
7	quality analyses involving these three metrics.
8
• 0 - 7 ppm-hrs (720 sites) O 14-15 ppm-hrs (29 sites) • 18-58 ppm-hrs (77 sites)
9	O 8-13 ppm-hrs (241 sites) ฉ 16 -17 ppm-hrs (21 sites) A 4th Max Value > 70 ppb
10 Figure 4-6. W126 index at monitoring sites with valid design values (2018-2020 average).
April 2022
4-60 External Review Draft - Do Not Quote or Cite

-------
• 0 (814 sites)	01.1- 5.0 (94 sites) • > 10.0 (22 sites)
1	Q 0.1-1.0 (154 sites) ฉ 5.1 -10.0 (21 sites) A 4th Max Metric > 70 ppb
2	Figure 4-7. N100 values at monitoring sites with valid design values (2018-2020 average).
• 0(814 sites)	G 1.1-2.0
4	ฉ 0.1-1.0 (202 sites) ฉ2.1-5.0
5	Figure 4-8. D100 values at monitoring sites
(44 sites) • > 5.0 (21 sites)
(24 sites) A 4th Max Metric > 70 ppb
with valid design values (2018-2020 average).
April 2022	4-61 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
4.4.1 Influence of Form and Averaging Time of Current Standard on W126 Index and
Peak Concentration Metrics
In revising the standard in 2015 to the now-current standard, the Administrator concluded
that, with revision of the standard level, the existing form and averaging time provided the
control needed to achieve the cumulative seasonal exposure circumstances identified for the
secondary standard (80 FR 65408, October 26, 2015). The focus on cumulative seasonal
exposure primarily reflected the evidence on E-R relationships for plant growth. The 2015
conclusion was based on the air quality data analyzed at that time (80 FR 65408, October 26,
2015). Analyses of the now expanded set of air monitoring data, which includes 1,578
monitoring sites with sufficient data for derivation of design values (Appendix 4D, section
4D.2.2), document similar findings as from the analysis of data from 2000-2013 described in the
2015 review, and the 2020 analysis of 2000-2018 data. The current (updated) analyses, which
now span 21 years and 19 3-year periods, are described in detail in Appendix 4D.
These analyses document the positive nonlinear relationship that is observed between
cumulative seasonal exposure, quantified using the W126 index, and design values, based on the
form and averaging time of the current standard. This is shown for both the average W126 index
across the 3-year design value period (Figure 4-9, left) and for annual index values within the
period (Figure 4-9, right). For both annual and 3-year average index values, it is clear that
cumulative seasonal exposures, assessed in terms of W126 index, are lower at monitoring sites
with lower design values. This is seen both for design values above the level of the current
standard (70 ppb), where the slope is steeper (due to the sigmoidal weighting of higher
concentrations by the W126 index function), as well as for lower design values that meet the
current standard (Figure 4-9; Appendix 4D). These presentations also indicate some regional
differences. For example, as shown in Figure 4-6 and Figure 4-9 for the 2018-2020 period, sites
meeting the current standard in the regions outside of the West and Southwest regions, all 3-year
average W126 index values (and virtually all annual values) are at or below 13 ppm-hrs. Ozone
concentrations, and W126 index values, are generally higher in the West and Southwest regions
(Figure 4-6). However, the positive relationship between the W126 index and the design value is
evident in all regions (Figure 4-9).
April 2022
4-62 External Review Draft - Do Not Quote or Cite

-------
ฎ Central
o EastNorthCentral






ฉ NorthEast
Q NorthWest
Q South
O SouthEast
O Southwest




•
ซ
•
•	West
•	WestNorthCentral


•

•





•
- •
•
o . ซ•
* • |
• • •
•
• ••




0
1 J
•
•




I

•
MM



-






40
50
60
70	80	90
4th Max Metric Value (ppb)
100
110
E ^
ฎ Central
o EastNorthCentral
* 2018
ฆ 2019



A
t
ฆ
ฉ NorthEast
o NorthWest
o south
o SouthEast
o Southwest
• 2020



ฆ
•
•
•	West
•	WestNorthCentral


•
A
A
ฐ
•
•
1
& A




A #
• A4
Ai m
• ! j
8"
A



A
A t
* ฃ

A ^
ฆ
ฆ
ฆ



A j

q, ฆ A
ฆ
A


* ซsl
i$S5S


ฃ



40
50
60
70	80	90
4th Max Metric Value (ppb)
100
110
Figure 4-9. Relationship between the W126 index and design values for the current standard (2018-2020). The W126 index is
analyzed in terms of averages across the 3-year design value period (left) and annual values (right).
April 2022
4-63
External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
An additional analysis, which was also performed in the 2015 review with the then-
available data, assesses the relationship between long-term changes in design value and long-
term changes in the W126 index (presented in detail in Appendix 4D, section 4D.3.2.3). Ozone
monitoring data have well documented reductions in O3 design values in response to national
programs to control O3 precursors (see section 2.4.2 above). The current analysis explores the
extent to which the W126 index has responded to these declines by focusing on the relationship
between changes (at each monitoring site) in the 3-year design value (termed "4th max" in
Appendix 4D, Figure 4-10 and Figure 4-10) across the 19 design value periods from 2000-2002
to 2018-2020 and changes in the W126 index over the same period.76 This analysis, performed
using either the 3-year average W126 index or annual values, shows there to be a positive, linear
relationship between the changes in the W126 index and the changes in the design value at
monitoring sites across the U.S. (Figure 4-10). This means that a change in the design value at a
monitoring site was generally accompanied by a similar change in the W126 index (e.g., a
reduction in design value accompanied by a reduction in W126 index). Nationally, the W126
index (in terms of 3-year average) decreased by approximately 0.59 ppm-hrs per ppb decrease in
design value over the full period from 2000 to 2020. This relationship varies across the NOAA
climate regions, with the greatest change in the W126 index per unit change in design value
observed in the Southwest and West regions. Thus, the regions which had the highest W126
index values at sites meeting the current standard (Figure 4-10) also showed the greatest
improvement in the W126 index per unit decrease in their design values over the past 21 years
(Appendix 4D, Table 4D-12 and Figure 4D-12). This indicates that going forward as design
values are reduced in areas that are presently not meeting the current standard, the W126 index
in those areas would also be expected to decline (Appendix 4D, section 4D.3.2.3 and 4D.4).
Thus, the air quality analyses indicate control by the form and averaging time of the
current standard of W126 index exposures, both in terms of 3-year average and single-year
values. The overall trend showing reductions in the W126 index concurrent with reductions in
the design value metric for the current standard is positive whether the W126 index is expressed
in terms of the average across the 3-year design value period or the annual value (Appendix 4D,
section 4D.3.2.3). This similarity is consistent with the relationship between the W126 index and
the design value metric for the current standard summarized above, which shows a strong
positive relationship between those metrics (Figure 4-9, Appendix 4D, section 4D.3.1.2).
76 At each site, the trend in values of a metric (W126 or 4th max), in terms of a per-year change in metric value, is
calculated using the Theil-Sen estimator, a type of linear regression method that chooses the median slope among
all lines through pairs of sample points. For example, if applying this method to a dataset with metric values for
four consecutive years (e.g., W 126i, W1262, W1263, W1264), the trend would be the median of the different per-
year changes observed in the six possible pairs of values ([W1264- W1263]/l, [W1263- W1262]/l, [W1262-
W126i]/1, [W1264- W1262]/2, [W1263- W126i]/2, [W1264- W126i]/3).
April 2022
4-64 External Review Draft - Do Not Quote or Cite

-------
• Central
o EastNorthCentral
ฉ NorthEast
o Northwest
ฉ	South
o SouthEast
o Southwest
•	West
•	WestNorthCentral
-2	-1
Trend in 4th Max Metric Value (ppb/yr)
•	Central	o SouthEast
o	EastNorthCentral	ฉ Southwest
ฉ	NorthEast	• West
o	Northwest	• WestNorthCentral
o	South
-2	-1
Trend in 4th Max Metric Value (ppb/yr)
Figure 4-10. Relationship between trends in the W126 index and trends in design values across a 21-year period (2000-2020)
at U.S. monitoring sites. W126 is analyzed in terms of averages across 3-year design value periods (left) and
annual values (right).
April 2022
4-65
External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
In considering the control of the current form and averaging time on vegetation exposures
of potential concern, we additionally take note of the evidence discussed in section 4.3.3.2 above
regarding the potential for days with particularly high O3 concentrations to play a contributing
role in vegetation effects. While the occurrence and severity of visible foliar injury indicates
some relationship with cumulative concentration-weighted indices such as SUM06 and W126,
the evidence also indicates a contributing role for occurrences of peak concentrations. We note
that the current standard's form and averaging time, by their very definition, limit such
occurrences. For example, the peak 8-hour average concentrations are lower at sites with lower
design values, as illustrated by the declining trends in annual fourth highest MDA8
concentrations that accompany the declining trend in design values described in chapter 2 (e.g.,
Figure 2-11). Additionally, peak hourly concentrations are also lower with lower design values.
As shown in Figure 4-11, the 99th through 25th percentile daily maximum 1-hour concentrations
(MDA1) are lower with lower design values. This is true both for the most recent three design
value periods and the three periods in 2000 through 2004. Additionally Figure 4-11 shows that
for sites with design values below the level of the standard (i.e., at or below 70 ppb) the 99th
percentile of daily maximum 1-hour ozone concentrations is less than 80 ppb. Further analyses
summarized in Appendix 2A document many fewer hourly concentrations at or above 100 ppb at
sites that meet the current standard compared to sites that do not. For example, the average
number of hours at or above 100 ppb per site in a 3-year period was well below one for sites
meeting the current standard compared to approximately 10 occurrences per site for sites not
meeting the current standard (Appendix 2A, Table 2A-2). This pattern also holds for hourly
concentrations at or above 120 or 160 ppb and is true for the recent air quality as well as past air
quality (Appendix 2A, Tables 2A-2 through 2A-4).
April 2022
4-66 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
a
n
o
E 80
<60
61-70
71-84
> 84
8-hour 03 Desian Value (cpb)
Figure 4-11. Distributions of MDA1 concentrations for the three design value periods in
2000-2004 (red) and 2016-2020 (blue), binned by the design value at each
monitoring site. Boxes represent the 25th, 50th and 75th percentiles; whiskers
represent the 1st and 99th percentiles; and circles are outlier values.
An additional investigation into the extent of control the current standard exerts on peak
concentrations is described in the set of analyses presented in Appendix 4F. This investigation
tallied the number of hours at or above 100 ppb (N100), and the number of days with an hour at
or above 100 ppb (D100), at sites meeting different criteria with regard to seasonal W126 index,
in a single year and as an average across three years, and also at sites with varying design values.
The strong control of these peak concentration metrics exerted by the current standard is
illustrated in Figure 4-12 by the low values common at sites meeting the current standard (design
value of 70 ppb or lower). The parallel presentation for varying values of W126 index suggests
that this metric has generally less potential for control of such peaks (Figure 4-12). For example,
the distributions for N100 and D100 observed for monitoring sites meeting the current standard
are more compressed and have lower maximum values than any of the W126 bins, with the
lowest bins (for W126 index values at or below 7 ppm-hrs) being most similar (Figure 4-12).
April 2022
4-67 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
200-
150 -
100-
50-
60-
50-
40'
30-
20-
10-

<60 61-70 71-84 >84
Design Value (ppb)
<60
61-70
Design Value (ppb)
71-84 >84
200-
150 -
100
50-
60-
50-
40-
TO
>
O
30-
20-
10-
Annual W126
3-year W126
4- -?	f-
<7
8-13	14-19
W126 Index (ppm-hrs)
> 19
Annual W126
3-year W126
<7
8-13	14-19
W126 Index (ppm-hrs)
> 19
Figure 4-12. Distributions of N100 (top panels) and D100 (bottom panels) values at
monitoring sites differing by design values (left panels) and W126 index values
(right panels) based on 2018-2020 monitoring data. The boxes represent the
25th, 50th and 75th percentiles and the whiskers extend to the 1st and 99th.
April 2022	4-68 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
In considering the prevalence of peak concentrations occurring at monitoring sites, it can
be seen that O3 concentrations at or above 100 ppb occur at lower prevalence at sites that meet
the current standard than at sites that meet a range of W126 index values. As shown in Table 4-1,
during the highest year for the different N100 or D100 thresholds, the percentage of sites
exceeding those thresholds is greater for the sites restricted to meet the different annual W126
levels, with the exception of 7 ppm-hrs, than it is for sites meeting the current standard (design
values [3-year 4th Max] at or below 70 ppb) for which the percentages are similar to those for the
sites meeting a W126 of 7 ppm-hrs. This observation can also be made for the average
percentages across the 3-year period. Further, in looking at the three most recent 3-year periods
(extending from 2016 through 2020), a similar finding holds (Table 4-2).
Table 4-1. Percent of monitoring sites during the 2018 to 2020 period with 4th max or
W126 metrics at or below various thresholds that have N100 or D100 values
above various thresholds.

Total
Number of
Sites
Num
N100 > 0
ber of sites v
N100 > 5
/here:
N100 >10
Nut
D100 > 0
Tiber of sites w
D100 > 2
lere:
D100 > 5

A verage percent of sites exceeding N100 or D100 threshold per year*
3-year 4th Max < 70
877
6%
0.4%
<0.1%
6%
0.3%
0%
Annual W126< 25
1134-1144
11%
1,7%
0.5%
11%
1.7%
0.3%
Annual W126< 19
1091-1129
10%
1.3%
0.3%
10%
1.3%
0.2%
Annual W126< 17
1067-1117
9.3%
1.3%
0.2%
9.3%
1.3%
0.2%
Annual W126< 15
1031-1091
9%
1.2%
0.2%
9%
1.2%
0.1%
Annual W126< 7
626-860
5.3%
0,4%
0%
5.3%
0.4%
0%
Annual 4"' Max < 70
802-1000
3.7%
0%
0%
3.7%
0%
0%

Percent of sites exceeding N100 or D100 threshold in maximum year of the three
3-year 4th Max < 70

9%
0.6%
0.1%
9%
0.5%
0%
Annual W126< 25
See above
15%
2%
0.6%
15%
2%
0.4%
Annual W126< 19
13%
2%
0.4%
13%
28%
0.3%
Annual W126< 17
13%
2%
0.3%
13%
2%
0.3%
Annual W126< 15
13%
2%
0.3%
13%
2%
0.3%
Annual W126< 7
8%
1%
0%
8%
1%
0%
Annual 4th Max < 70

4%
0%
0%
4%
0%
0%
* For the annual metrics, the entries for each N100 or D100 column may be for different years in the 3-year period. Thus the
"Total Number of Sites" column presents the range in number of sites that meet the annual 4th Max or W126 thresholds in
each of the three years (as presented in Table 4F-2, Appendix 4F).
April 2022
4-69 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
Table 4-2. Average percent of monitoring sites per year during 2016-2020 with 4th max or
W126 metrics at or below various thresholds that have N100 or D100 values
above various thresholds.

Total
Number
of Sites
Perceri
N100 > 0
t of sites wher
N100 > 5
e:
N100 >10
Perc
D100 > 0
ent of sites wl
D100 > 2
lere:
D100 > 5


A verage percent of sites exceeding N100orD100 threshold per year (2016 - 2020)
3-year 4th Max < 70

5.1%
0.3%
0.01%
5.1%
0.2%
0%
Annual W126<25

11.0%
1.7%
0.5%
11.0%
1.8%
0.4%
Annual W126<19
10.0%
1.4%
0.3%
10.0%
1.4%
0.2%
Annual W126<17
9.5%
1.2%
0.2%
9.5%
1.2%
0.1%
Annual W126<15
9.1%
1.2%
0.2%
9.1%
1.1%
0.1%
Annual W126<7
5.1%
0.4%
0%
5.1%
0.3%
0%
Annual 4th Max < 70

3.3%
0.02%
0%
3.3%
0.3%
0%
Drawn from Appendix 4F, Table 4F-3.
These air quality analyses illustrate limitations of the W126 index for purposes of
controlling peak concentrations, and also the strengths of the current standard in this regard. As
discussed more fully in section 4.5.1.1 below, the W126 index cannot, by virtue of its definition,
always differentiate between air quality patterns with high peak concentrations and those without
such concentrations. This is demonstrated in the air quality analyses referenced above which
indicate that the form and averaging time of the existing standard is much more effective than the
W126 index in limiting peak concentrations (e.g., hourly O3 concentrations at or above 100 ppb)
and in limiting number of days with any such hours (e.g., Appendix 4F, Figures 4F-4, 4F-5, 4F-
8, 4F-9 compared to Figures 4F-6, 4F-7, 4F-10 and 4F-11). A similar finding is evidenced in the
historical data extending back to 2000. These data show the appreciable reductions in peak
concentrations that have been achieved in the U.S. as air quality has improved under O3
standards of the existing form and averaging time (Appendix 4F, Figures 4F-12 and 4F-13).
From the analyses, it can be seen that the form and averaging time of the current standard is
effective in controlling peak hourly concentrations and that a W126 index-based standard would
be much less effective in providing the needed protection against years with such elevated and
potentially damaging hourly concentrations.
In summary, monitoring sites with lower O3 concentrations as measured by the design
value metric (based on the current form and averaging time of the secondary standard) have
lower cumulative seasonal exposures, as quantified by the W126 index, and also lower short-
term peak concentrations, thus indicating a level of control exerted by the current standard on
these other metrics. As the form and averaging time of the secondary standard have not changed
April 2022
4-70 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
since 1997, the decreasing trends in W126 index and in hourly and 8-hour daily maximum
concentrations over time also support the finding that a change in level (i.e., from 80 ppb in 1997
to 75 ppb in 2008 to 70 ppb in 2015) for a standard of the current form and averaging time
contributes to reductions in the level on cumulative seasonal exposures in terms of W126 index
(and on the magnitude of short-term peak concentrations). That is, that reductions in design
value, presumably associated with implementation of the revised standards, have been
accompanied by reductions in cumulative seasonal exposures in terms of W126 index, as well as
reductions in short-term peak concentrations. Further, the analyses focused on N100 and D100
metrics provide additional evidence of the control of the current standard on peak concentrations,
and also indicate a likely lesser effectiveness of the W126 index metric in providing such
control. Altogether, the analyses summarized here demonstrate the form and averaging time of
the current standards to be effective in controlling cumulative, concentration-weighted exposures
as well as peak hourly concentrations (e.g., concentrations at/above 100 ppb), two metrics that
have been found to be important to O3 effects on vegetation (as discussed in section 4.3 above).
4.4.2 Environmental Exposures in Terms of W126 Index
Given the evidence indicating the W126 index to be strongly related to growth effects
and its use in the E-R functions for tree seedling RBL, exposure in the analyses described here is
quantified using the W126 metric (Figure 4-13). These analyses are intended to inform
conclusions regarding the magnitude of cumulative, concentration-weighted exposures, in terms
of W126 index, likely to occur in areas that meet the current standard. In light of the importance
placed on Class I areas in past secondary standard reviews and the greater public welfare
significance of O3 related impacts in such areas, as discussed in section 4.3.2 above, a separate
evaluation is conducted on cumulative O3 exposure at monitoring sites in or near Class I areas77,
in addition to that at all monitoring sites nationwide. The potential for impacts of interest is
assessed through considering the magnitude of estimated exposure in light of current information
and, in comparison to levels given particular focus in the 2015 decision on the current standard
(80 FR 65292; October 26, 2015) 78
77	Included are monitors sited within Class I areas or the closest monitoring site within 15 km of the area boundary.
78	The W126 index values were rounded to the nearest unit ppm-hr for these comparisons to a specific whole-
number W126 level (Appendix 4D, section 4D.2).
April 2022
4-71 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
s ซ
2
<
Figure 4-13. Analytical approach for characterizing vegetation exposure with W126 index.
The updated analyses discussed here and described in greater detail in Appendix 4D
include assessment of all monitoring sites nationally and also a focused evaluation in Class I
areas for which such monitoring data are available. The analyses include air quality monitoring
data for the most recent 3-year period (2018 to 2020) for which data were available when the
analyses were performed, and also all 3-year periods going back as far as the 2000-2002 period.
Design values (3-year average annual fourth-highest 8-hour daily maximum concentration, also
termed "4tb max metric" in this analysis) and W126 index values (in terms of the 3-year average)
were calculated at each site where sufficient data were available.79 Across the nineteen 3-year
periods from 2000-2002 to 2018-2020, the number of monitoring sites with suffici ent data for
calculation of valid design values and W126 index values ranged from a low of 992 in 2000-
2002 to a high of 1,118 in 2015-2017. As specific monitoring sites differed somewhat across the
21 years, there were 1,578 sites with sufficient data for calculation of valid design values and
W126 index values for at least one 3-year period between 2000 and 2020, and 510 sites had such
data for all nineteen 3-year periods. The sections below discuss key aspects of these analyses and
what they indicate with regard to protection from vegetation-related effects of potential public
welfare significance.
The analyses of cumulative seasonal exposures included a focus on the W126 index in
terms of the average seasonal index across the 3-year design value period, with additional
analyses also characterizing the annual W126 index. Among the analyses performed is an
evaluation of the variability of annual W126 index values across the 3-year period (Appendix
79 Data adequacy requirements and methods for these calculations are described in Appendix 4D, section 4D.2.
April 2022
4-72 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
4D, section 4D.3.1.2). This evaluation was performed for all monitoring sites in the most recent
3-year period, 2018 to 2020. This analysis indicates the extent to which single-year values within
the 3-year period deviate from the average for the period. Across the 877 sites (Appendix 4D,
Table 4D-1) meeting the current standard (design value at or below 70 ppb), 99% of single-year
W126 values in this subset differ from the 3-year average by no more than 5 ppm-hrs, and 78%
by no more than 2 ppm-hrs (Appendix 4D, Figure 4D-7).
The following discussion is framed by a key policy-relevant question based on those
identified in the IRP. The question considers all areas nationally, with particular focus on air
quality data for Class I areas.
• What are the nature and magnitude of vegetation exposures associated with
conditions meeting the current standard at sites across the U.S., particularly in
specially protected areas, such as Class I areas, and what do they indicate regarding
the potential for C>3-related vegetation impacts?
To address this question, we considered both recent air quality (2018-2020) and air
quality since 2000. These air quality analyses of cumulative seasonal exposures associated with
conditions meeting the current standard nationally provide conclusions generally similar to those
based on the data available at the time of the 2015 review when the current standard was set,
when the most recent data available for analysis were 2011 to 2013 (Wells, 2015). Cumulative
exposures vary across the U.S, with the highest W126 index values for sites that met the current
standard being located exclusively in the Southwest and West climate regions (Figure 4-6). In all
other NOAA climate regions, average W126 index values (for the 3-year period, 2018-2020) at
sites meeting the current standard are generally at or below 13 ppm-hrs (Figure 4-6). In the
Southwest and West, W126 index values at all sites meeting the current standard are at or below
17 ppm-hrs in the most recent 3-year period (Figure 4-6) and virtually all sites meeting the
current standard are at or below 17 ppm-hr across all of the nineteen 3-year periods in the full
dataset evaluated80 (Table 4-3). Additionally, the historical dataset includes no occurrences of a
3-year average W126 index above 19 ppm-hrs at sites meeting the current standard, and just a
small number of occurrences (limited to eight [less than 0.08% of values], all but one from a
period prior to 2011) of a W126 index above 17 ppm-hrs, with the highest just equaling 19 ppm-
hrs (Table 4-3; Appendix 4D, section 4D.3.2.1).
80 On over 99.9 percent of occasions across all sites with valid design values at or below 70 ppb during the 2000 to
2020 period, the W126 metric (seasonal W126, averaged over three years) was at or below 17 ppm-hrs (Table 4-
1). All but one of the eight occasions when it was above 17 ppm-hrs (the highest was 19 ppm-hrs) occurred in the
Southwest region during a period before 2011. The eighth occasion occurred at a site in the West region when the
3-year average W126 index value was 18 ppm-hrs. On more than 97 percent of occasions in the full dataset with
valid design values at or below 70 ppb, the 3-year average W126 index was at or below 13 ppm-hrs (Appendix
4D, section4D.3.2).
April 2022
4-73 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
Given the recognition of more significant public welfare implications of effects in
protected areas, such as Class I areas (as discussed in section 4.3.2 above), we give particular
attention to Class I areas (Appendix 4D, section 4D.3.2.4). In so doing, we consider the updated
air quality analysis presented in Appendix 4D for 65 Class I areas. The findings for these sites,
which are distributed across all nine NOAA climate regions in the contiguous U.S., as well as
Alaska and Hawaii, mirror all U.S. sites. Among the Class I area sites meeting the current
standard (i.e., having a design value at or below 70 ppb) in the most recent period of 2018 to
2020, there are none with a W126 index (averaged over design value period) above 17 ppm-hrs
(Table 4-3). The historical dataset includes just seven occurrences (all dating from the 2000-2010
period) of a Class I area site meeting the current standard and having a 3-year average W126
index above 17 ppm-hrs, and no such occurrences above 19 ppm-hrs (Table 4-1). Additionally,
across the full 21-year dataset for 56 Class I areas with monitors meeting the current standard
during at least one or as many as nineteen 3-year periods since 2000, there are no more than 15
occurrences of a single-year W126 index above 19 ppm-hrs, the majority occurring during the
earlier years of the period (Appendix 4D, section 4D.3.2.4, Tables 4D-14 and 4D-16). For
example, the highest values were equal to 23 ppm-hrs, all occurring before 2012 (Appendix 4D,
4D-16).
Across the complete dataset (2000-2020), the W126 index, averaged over a 3-year
period, at sites with design values above 70 ppb (i.e., that would not meet the current standard)
ranges up to approximately 60 ppm-hrs (Appendix 4D, Table 4D-17). Focusing on the most
recent period, among all sites across the U.S. that do not meet the current standard in the 2018 to
2020 period, more than a quarter have average W126 index values above 19 ppm-hrs and more
than a third exceed 17 ppm-hrs (Table 4-3).81 A similar situation exists for Class I area sites
(Table 4-3). Thus, as at the time of the 2015 decision, the available quantitative information
continues to indicate appreciable control of seasonal W126 index-based cumulative exposure at
all sites with air quality meeting the current standard.
81 As described above and in detail in Appendix 4D, W126 index values were rounded to the nearest unit ppm-hr for
comparisons to a specific whole-number W126 level.
April 2022
4-74 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
Table 4-3. Distribution of 3-yr average seasonal W126 index for sites in Class I areas and
across U.S. that meet the current standard and for those that do not.
3-year periods
Number of Occurrences or Site-DVs A
In Class I Areas
Across All Monitoring Sites (urban and rural)
Total
W126 (ppm-hrs)
>19 | >17 | <17
Total
W126 (ppm-hrs)
>19 | >17 | <17

At sites that meet the current standard (design value at or below 70 ppb)
2018-2020
47
0
0 47
877
0 0 877
All from 2000 to 2020
589
0
7 582
10,039
0 8 10,031

At sites that exceed the current standard (design value above 70 ppb)
2018-2020
10
7
8 2
213
58 77 136
All from 2000 to 2020
391
174
219 172
11,142
2,424 3,317 7,825
AThe counts presented here are drawn from Appendix D, Tables 4D-2, 4D-4, 4D-5, 4D-6, 4D-9, 4D-10, and 4D-14 through 17.
In summary, as discussed in section 4.3.3 above, the evidence available leads us to
similar conclusions regarding exposure levels associated with effects as in the 2015 review.
Based largely on this evidence in combination with the use of RBL as a surrogate or proxy for all
vegetation-related effects, the value of 17 ppm-hrs, as an average W126 index (over three years)
was generally identified as a target level for protection in the 2015 decision (80 FR 65393;
October 26, 2015). The available information continues to indicate that average cumulative
seasonal exposure levels at virtually all sites and 3-year periods with air quality meeting the
current standard fall at or below this level of 17 ppm-hrs. Additionally, at sites meeting the
current standard, single-year W126 index values are less than or equal to 19 ppm-hrs well over
99% of the time (Appendix 4D, section 4D.3.2.1). In Class I area sites that meet the current
standard for the most recent 3-year period, the average W126 index is below 17 ppm-hrs
(Appendix 4D, Table 4D-16). Further, across the full 21-year dataset, with the exception of
seven values that occurred prior to 2011, Class I area W126 index values (averages for each 3-
year period) were no higher than 17 ppm-hrs during periods that met the current standard. This
contrasts with the occurrence of much higher seasonal W126 index values in sites when the
current standard was not met. For example, out of the 10 Class I area sites with design values
above 70 ppb during the most recent period, seven had a W126 index (based on 3-year average)
above 19 ppm-hrs (ranging up to 47 ppm-hrs) and eight sites had a W126 index above 17 ppm-
hrs (Table 4-3; Appendix 4D, Table 4D-17). This same pattern is exhibited at all sites in the full
dataset, as shown in Table 4-3, including both urban and rural sites.
April 2022
4-75 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
4.4.3 Limitations and Uncertainties
• What are the important uncertainties associated with any exposure estimates and
associated characterization of potential for public welfare effects?
The analyses described above in sections 4.1 and 4.2 are based primarily on the hourly air
monitoring dataset that is available at O3 monitoring sites nationwide. While there are inherent
limitations in any air monitoring network, the monitors for O3 are distributed across the U.S.,
covering all NOAA regions and all states (e.g., Figure 4-6).
That distribution notwithstanding, there is uncertainty about whether areas that are not
monitored would show the same patterns of exposure as areas with monitors. There are
limitations in the distributions of the monitors, such that some geographical areas are more
densely covered than others. For example, only about 40% of all Federal Class I Areas have or
have had O3 monitors within 15 km with valid design values, thus allowing inclusion in the Class
I area analysis. Even so, the dataset includes sites in 27 states distributed across all nine NOAA
climatic regions across the contiguous U.S, as well as Hawaii and Alaska. Some NOAA regions
have far fewer numbers of Class I areas with monitors than others. For instance, the Central,
Northeast, East North Central, and South regions all have three or fewer Class I areas in the
dataset. However, these areas also have appreciably fewer Class I areas in general when
compared to the Southwest, Southeast, West, and West North Central regions, which are more
well represented in the dataset. The West and Southwest regions are identified as having the
largest number of Class I areas, and they have approximately a third of those areas represented
with monitors, which include locations where W126 index values are generally higher, thus
playing a prominent role in the analysis. We also recognize a limitation that accompanies any
analysis, i.e., that it is based on information available at this time. Thus, it may or may not reflect
conditions far out into the future as air quality and patterns of O3 concentrations in ambient air
continue to change in response to changing circumstances, such as changes in precursor
emissions to meet the current standard across the U.S. That said, we note that for the air quality
analyses (e.g., involving W126 index) that were also conducted in the 2015 review, the findings
are largely consistent.
In considering the estimates of exposure represented by the W126 index, we note a
limitation in this index in its ability to distinguish among air quality conditions with differing
prevalence of peak concentrations (e.g., hourly concentrations at or above 100 ppb). As indicated
in the analyses in Appendix 4F, summarized above in section 4.1.1, two different locations or
years may have appreciably different patterns of hourly concentrations but the same W126 index
value. To the extent that these concentrations influence vegetation responses, this may contribute
an uncertainty to applications of the tree seedling E-R functions (as recognized by Lefohn et al.,
1997).
April 2022
4-76 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
Further, we note the discussion in section 4.4.1 above of how changes in O3 patterns in
the past have affected the relationship between W126 index and the averaging time and form of
the current standard, as represented by design values (section 4.4.1, and Appendix 4D, section
4D.3.2.3). This analysis finds a positive, linear relationship between trends in design values and
trends in the W126 index (both in terms of single-year W126 index and averages over 3-year
design value period), as was also the case for similar analyses conducted for the data available at
the time of the 2015 review (Wells, 2015). While this relationship varies across NOAA regions,
the regions showing the greatest potential for exceeding W126 index values of interest (e.g., with
3-year average values above 17 and/or 19 ppm-hrs) also showed the greatest improvement in the
W126 index per unit decrease in design value over the historical period assessed (Appendix 4D,
section 4D.3.2.3). Thus, the available data and this analysis appear to indicate that as design
values are reduced to meet the current standard in areas that presently do not, W126 values in
those areas would also be expected to decline (Appendix 4D, section 4D.4).
4.5 KEY CONSIDERATIONS REGARDING THE CURRENT
SECONDARY STANDARD
In considering what the available evidence and exposure/risk information indicate with
regard to the current secondary O3 standard, the overarching question we address is:
• Does the available scientific evidence and air quality and exposure analyses support
or call into question the adequacy of the protection afforded by the current
secondary O3 standard?
To assist us in interpreting the available scientific evidence and the results of recent
quantitative analyses to address this question, we have focused on a series of more specific
questions. In considering the scientific and technical information, we consider both the
information available at the time of the 2015 review and information newly available since then
which has been critically analyzed and characterized in the current ISA, the 2013 ISA and prior
AQCDs. In this context, an important consideration is whether the newly available information
alters the EPA's overall conclusions from the 2015 review regarding welfare effects associated
with photochemical oxidants, including O3, in ambient air. We also consider the available
quantitative information regarding environmental exposures, characterized by the pertinent
metric, likely to occur in areas of the U.S. where the standard is met. Additionally, we consider
the significance of these exposures with regard to the potential for 03-related vegetation effects,
their potential severity, and any associated public welfare implications.
April 2022
4-77 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
4.5.1 Evidence and Exposure/Risk-based Considerations
In considering first the available evidence with regard to the overarching question posed
above regarding the protection provided by the current standard from welfare effects, we address
a series of more specific questions that focus on policy-relevant aspects of the evidence. These
questions relate to three main areas of consideration: (1) the available evidence on welfare
effects associated with exposure to photochemical oxidants, and particularly O3 (section 4.5.1.1);
(2) the risk management framework or approach for reaching conclusions on the adequacy of
protection provided by the secondary standard (section 4.5.1.2); and (3) findings from the air
quality and exposure analyses pertaining to public welfare protection under the current standard
(section 4.5.1.3).
4.5.1.1 Welfare Effects Evidence
• Is there newly available evidence that indicates the importance of photochemical
oxidants other than O3 with regard to abundance in ambient air, and potential for
welfare effects?
No newly available evidence has been identified regarding the importance of
photochemical oxidants other than O3 with regard to abundance in ambient air, and potential for
welfare effects.82 As summarized in section 2.1 above, O3 is one of a group of photochemical
oxidants formed by atmospheric photochemical reactions of hydrocarbons with nitrogen oxides
in the presence of sunlight, with O3 being the only photochemical oxidant other than nitrogen
dioxide that is routinely monitored in ambient air (ISA, Appendix 1, section 1.1).83 Data for
other photochemical oxidants are generally derived from a few special field studies, such that
national scale data for these other oxidants are scarce (ISA, Appendix 1, section 1.1; 2013 ISA,
sections 3.1 and 3.6). Moreover, few studies of the welfare effects of other photochemical
oxidants beyond O3 have been identified by literature searches conducted for the 2013 ISA and
prior AQCDs (ISA; Appendix 1, section 1.1). As stated in the current ISA, "the primary
literature evaluating the health and ecological effects of photochemical oxidants includes ozone
almost exclusively as an indicator of photochemical oxidants" (ISA, section IS. 1.1). Thus, as was
the case for previous reviews, the evidence base for welfare effects of photochemical oxidants
does not indicate an importance of any other photochemical oxidants. For these reasons,
discussion of photochemical oxidants in this document focuses on O3.
82	Close agreement between past ozone measurements and the photochemical oxidant measurements upon which the
early NAAQS (for photochemical oxidants including O3) was based indicated the very minor contribution of
other oxidant species in comparison to O3 (U.S. DHEW, 1970).
83	Consideration of welfare effects associated with nitrogen oxides in ambient air is addressed in the review of the
secondary NAAQS for ecological effects of oxides of nitrogen, oxides of sulfur and particulate matter (U.S. EPA,
2018).
April 2022
4-78 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
• Does the available evidence alter prior conclusions regarding the nature of welfare
effects attributable to O3 in ambient air?
The current evidence documented in the 2020 ISA, including that newly available,
supports, sharpens, and expands somewhat on the conclusions reached in the 2015 review (ISA,
sections IS.1.3.2 and IS.5 and Appendices 8 and 9). A wealth of scientific evidence, spanning
more than six decades, demonstrates effects on vegetation and ecosystems of O3 in ambient air
(ISA, section IS.6.2.1; 2013 ISA, 2006 AQCD, 1997 AQCD, 1986 AQCD; U.S. DHEW, 1970).
Accordingly, consistent with the evidence in the 2015 review, the available evidence describes
an array of O3 effects on vegetation and related ecosystem effects. The evidence also describes
climate effects of tropospheric O3, through a role in radiative forcing and subsequent effects on
temperature, precipitation, and related climate variables. Evidence newly available in the 2020
ISA strengthens previous conclusions, provides further mechanistic insights and augments
current understanding of varying effects of O3 among species, communities, and ecosystems
(ISA, section IS.6.2.1). The current evidence, including a wealth of longstanding evidence,
supports conclusions reached in the 2015 review of causal relationships between O3 and visible
foliar injury, reduced yield and quality of agricultural crops, reduced vegetation growth and plant
reproduction,84 reduced productivity in terrestrial ecosystems, and alteration of belowground
biogeochemical cycles. The current evidence, including that previously available, also supports
conclusions reached in the 2015 review of likely causal relationships between O3 and reduced
carbon sequestration in terrestrial systems, and alteration of terrestrial ecosystem water cycling
(ISA, section IS.1.3.2). Additionally, as in the 2015 review, the current ISA determines there to
be a causal relationship between tropospheric O3 and radiative forcing and a likely causal
relationship between tropospheric O3 and temperature, precipitation, and related climate
variables (ISA, section IS. 1.3.3). Further, the current evidence has led to an updated conclusion
on the relationship of O3 with alteration of terrestrial community composition to causal (ISA,
sections IS.1.3.2). Lastly, the current ISA concludes the current evidence sufficient to infer likely
causal relationships of O3 with three additional categories of effects (ISA, sections IS.1.3.2).
While previous recognition of O3 as a contributor to tree mortality in a number of field studies
was a factor in the 2013 conclusion regarding composition, it has been separately assessed in the
current ISA, with the conclusion that the evidence is sufficient to infer a likely causal
relationship with O3. Additionally, evidence newly available since the last ISA on two additional
plant-related effects augments more limited previously available evidence related to insect
interactions with vegetation, contributing to additional conclusions that the body of evidence is
84 As noted in section 4.3.1 above, the 2020 ISA includes a causality determination specific to reduced plant
reproduction, while this category of effects was considered in combination with reduced plant growth in the 2015
review (ISA, Table IS. 13).
April 2022
4-79 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
sufficient to infer likely causal relationships between O3 and alterations of plant-insect signaling
and insect herbivore growth and reproduction (ISA, Appendix 8, sections 8.6 and 8.7).85
As in the 2015 review, the strongest evidence and the associated findings of causal or
likely causal relationships with O3 in ambient air, and quantitative characterizations of
relationships between O3 exposure and occurrence and magnitude of effects, are for vegetation-
related effects, and particularly those identified in the 2015 review. The evidence base for the
newly identified category of increased tree mortality includes previously available evidence
largely comprised of field observations from locations and periods of O3 concentrations higher
than are common today and three more recently available publications assessing O3 exposures
not expected under conditions meeting the current standard. Among the three more recent
publications, one assessed survival of aspen clones across an 11-year period under O3 exposures
that included single-year seasonal W126 index values ranging above 30 ppm-hrs during the first
four years, and the other two were analyses based on field observations during periods when O3
concentrations were such that they would not be expected to meet the current standard, as
summarized in section 4.3.1 above (ISA, Appendix 8, section 8.4.3).
The information available regarding the newly identified categories of plant-insect
signaling and insect herbivore growth and reproduction does not provide for a clear
understanding of the specific environmental effects that may occur in the natural environment
under specific exposure conditions (as discussed in sections 4.3.1, 4.3.3.2 and 4.3.4 above). For
example, while the evidence base for effects on herbivore growth and reproduction is expanded
since the 2013 ISA, "there is no clear trend in the directionality of response for most metrics,"
such that some show an increased effect and some show reductions (ISA, p. IS-64; section
IS.5.1.3 and section 8.6). More specifically "no consistent directionality of response is observed
across the literature, and uncertainties remain in regard to different plant consumption methods
across species and the exposure conditions associated with particular severities of effects" (ISA,
p. IS-91). Additionally, while the available evidence documents effects of O3 on some plant
VPSCs (e.g., changing the floral scent composition and reducing dispersion), and indicates
reduced pollinator attraction, decreased plant host detection and altered plant-host preference in
some insect species in the presence of elevated O3 concentrations, characterization of such
effects is still "an emerging area of research with information available on a relatively small
number of insect species and plant-insect associations," and with gaps remaining in the
consequences of modification of signaling compounds by O3 in natural environments (ISA, p.
85 As in the 2015 review, the 2020 ISA again concludes that the evidence is inadequate to determine if a causal
relationship exists between changes in tropospheric ozone concentrations and UV-B effects (ISA, Appendix 9,
section 9.1.3.4; 2013 ISA, section 10.5.2).
April 2022
4-80 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
IS-91 and section IS.6.2.1). Accordingly, we focus on other vegetation effects described above,
rather than these two newly identified categories.
With regard to tropospheric O3 and effects on climate, we recognize the strength of the
ISA conclusion that tropospheric O3 is a greenhouse gas at the global scale, with associated
effects on climate (ISA, section 9.1.3.3). Accordingly, as indicated by the ISA causal
determinations, O3 abundance in the troposphere contributes to radiative forcing and likely also
to subsequent climate effects. There is appreciable uncertainty, however, associated with
understanding quantitative relationships involving regional O3 concentrations near the earth's
surface and climate effects of tropospheric O3 on a global scale As recognized in the ISA (and
summarized in sections 4.3.3.3 and 4.3.4 above), there are limitations in our modeling tools and
associated uncertainties in interpretations related to capabilities for quantitatively estimating
effects of regional-scale lower tropospheric O3 concentrations on climate. Thus, while additional
characterizations of tropospheric O3 and climate have been completed since the 2015 review,
uncertainties and limitations in the evidence that were also recognized in the 2015 review
remain. As summarized in sections 4.3.3.3 and 4.3.4 above, these affect our ability to make a
quantitative characterization of the potential magnitude of climate response to changes in O3
concentrations in ambient air, particularly at regional (vs global) scales, and thus our ability to
assess the impact of changes in ambient air O3 concentrations in regions of the U.S. on global
radiative forcing or temperature, precipitation, and related climate variables. Consequently, the
evidence in this area is not informative to our consideration of the adequacy of public welfare
protection of the current standard.
• To what extent does the available evidence provide E-R information (e.g.,
quantitative E-R relationships) for 03-related effects that can inform judgments on
the likelihood of occurrence of such effects in areas with air quality that meets the
current standard? Does the available evidence provide new or altered such
information since the 2015 review?
In considering what the available information indicates with regard to exposures
associated with welfare effects and particularly in the context of what is indicated for exposures
associated with air quality conditions that meet the current standard, we focus particularly on the
availability of quantitatively characterized E-R relationships for key effects. While the ISA
describes additional studies of welfare effects associated with O3 exposures since the 2015
review, the established E-R functions for tree seedling growth and crop yield that have been
available in the last several reviews continue to be the most robust descriptions of E-R
relationships for welfare effects. These well-established E-R functions for seedling growth
reduction in 11 tree species and yield loss in 10 crop species are based on response information
across multiple levels of cumulative seasonal exposure (estimated from extensive records of
April 2022
4-81 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
hourly O3 concentrations across the exposure periods). Studies of some of the same species,
conducted since the E-R function derivation, provide supporting information for these functions
(ISA, Appendix 8, section 8.13.2; 2013 ISA, sections 9.6.3.1 and 9.6.3.2). The E-R functions
provide for estimation of growth-related effects for a range of cumulative seasonal exposures.
The newly available evidence does not include new studies that assessed reductions in
tree growth or crop yield responses across multiple O3 exposures and for which sufficient data
are available for analyses of the shape of the E-R relationship across the range of cumulative
exposure levels (e.g., in terms of W126 index) relevant to conditions associated with the current
standard. For example, among the newly available studies are several that summarize previously
available studies or draw from them, such as for linear regression analyses.86 However, as
discussed in section 4.3.3.2 above, these do not provide robust E-R functions or cumulative
seasonal exposure levels associated with important vegetation effects that define the associated
exposure circumstances in a consistent manner, limiting their usefulness for our purposes here
with regard to considering the potential for occurrence of welfare effects in air quality conditions
that meet the current standard. Thus, robust E-R functions are not available for growth or yield
effects on any additional tree species or crops.
Based on these established E-R functions for tree seedling growth reductions in 11
species, the tree seedling RBL for the median tree species is 5.3% for a W126 index of 17 ppm-
hrs, rising to 5.7% for 18 ppm-hrs, 6.0% for 19 ppm-hrs and 6.4% for 20 ppm-hrs. Below 17
ppm-hrs, the median estimates include 4.9% for 16 ppm-hrs, 4.5% for 15 ppm-hrs, 4.2% for 14
ppm-hrs and 3.8% for 13 ppm-hrs (Appendix4A, Table 4A-5). These RBL estimates are
unchanged from what was indicated by the evidence in the 2015 review. As summarized in
section 4.1 above, the RBL estimates were used in the 2015 decision as a surrogate or proxy for
the broader array of vegetation-related effects.
With regard to visible foliar injury, as in the 2015 review, we lack established E-R
relationships that would quantitatively describe relationships between visible foliar injury
(occurrence and incidence, as well as injury severity) and O3 exposure, as well as factors
influential in those relationships, such as soil moisture conditions. As discussed in section
4.3.3.2 above, the available evidence continues to include both experimental studies that
86 For example, among the newly available publications cited in the ISA is a publication on tree and grassland
species that compiles EC10 values (estimated concentration at which 10% lower biomass [compared to zero 03] is
predicted) derived using linear regression of previously published data on plant growth response and O3
concentration quantified as AOT40. The data were from studies of various experimental designs, that involved
various durations ranging up from 21 days, and involving various concentrations no higher than 100 ppb as a
daily maximum hourly concentration. More detailed analyses of consistent, comparable E-R information across a
relevant range of seasonal exposure levels, accompanied by detailed records of O3 concentrations, that would
support derivation of robust E-R functions for purposes discussed here are not available (ISA, Appendix 8,
section 8.10.1.2).
April 2022
4-82 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
document foliar injury in specific plants in response to O3 exposures, and quantitative analyses
of the relationship between environmental O3 exposures and occurrence of foliar injury. The
analyses involving environmental conditions, while often using cumulative exposure metrics to
quantify O3 exposures (e.g., the W126 and SUM06 indices), have additionally reported there to
also be a role for a metric that quantifies the frequency or incidence of "high" O3 days, such as
N100 (2013 ISA, p. 9-10; Smith, 2012; Wang et al., 2012). However, such analyses have not
resulted in the establishment of specific air quality metrics and associated quantitative functions
for describing the influence of ambient air O3 on incidence and severity of visible foliar injury.
Multiple studies have involved quantitative analysis of data collected as part of the USFS
biosite biomonitoring program (e.g., Smith, 2012). These analyses continue to indicate the
limitations in capabilities for predicting the exposure circumstances under which visible foliar
injury would be expected to occur, as well as the circumstances contributing to increased injury
severity (Smith, 2012; Wang et al., 2012). As noted in section 4.3.3.2 above, expanded
summaries of the dataset compiled in the 2015 review from several years of USFS biosite
records does not clearly and consistently describe the shape of a relationship between incidence
of foliar injury or severity (based on individual site scores) and W126 index estimates (as a sole
representative of exposure). Overall, however, the dataset indicates that the proportion of records
having different levels of severity score is generally highest in the group of records for sites with
the highest W126 index (e.g., greater than 25 ppm-hrs for the normal and dry soil moisture
categories). Thus, the available evidence indicates increased occurrence and severity at the
highest category of exposures in the dataset (above 25 ppm-hrs in terms of a W126 index), but
does not provide for identification of air quality conditions, in terms of O3 concentrations
associated with the relatively lower environmental exposures most common in the USFS dataset
that would correspond to a specific magnitude of injury incidence or severity scores across
locations.
Thus, based on considering the available information for the array of O3 welfare effects,
we again recognize the E-R relationships available in the 2015 review for purposes of
considering O3 exposure levels associated with growth-related impacts to be the most robust E-R
information available. The available evidence for growth-related effects, including that newly
available, does not indicate the occurrence of growth-related responses attributable to cumulative
O3 exposures lower than was established at the time of the 2015 review. With regard to visible
foliar injury, the available information continues to be limited with regard to estimating
occurrence and severity (e.g., as quantified by BI score) across a range of air quality conditions
quantified by W126 index, such that a clear shape for a relationship between these variables is
not evident with the available data. Thus, the available information provides for only limited and
somewhat qualitative conclusions related to potential occurrence and/or severity under different
April 2022
4-83 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
air quality conditions. The quantitative information for other effects is still more limited, as
recognized in sections 4.3.3 and 4.3.4 above. Thus, the newly available evidence does not
appreciably address key limitations or uncertainties needed to expand capabilities for estimating
welfare impacts that might be expected as a result of differing patterns of O3 concentrations in
the U.S.
• Does the evidence continue to support a cumulative, seasonal exposure index, such as
the W126 function, as a biologically relevant and appropriate metric for assessment
of vegetation-related effects of O3 in ambient air?
As in the 2015 review, the available evidence continues to support a cumulative, seasonal
exposure index as a biologically relevant and appropriate metric for assessment of the evidence
of exposure/risk information for vegetation, most particularly for growth-related effects. The
most commonly used such metrics are the SUM06, AOT40 (or AOT60) and W126 indices (ISA,
section IS.3.2).87 The evidence for growth-related effects continues to support important roles for
cumulative exposure and for weighting higher concentrations over lower concentrations. Thus,
among the various such indices considered in the literature, the cumulative, concentration-
weighted metric, defined by the W126 function, continues to be best supported for purposes of
relating O3 air quality to growth-related effects.
We additionally note that while in its approach to emphasizing higher concentrations, the
W126 index assigns greater weights to higher hourly concentrations, it cannot, given its
definition as an index that sums three months of weighted hourly concentrations into one, single
value, always differentiate between air quality patterns with frequent high peak concentrations
and those without such concentrations.88 While the metric describes the pattern of varying
growth response observed across the broad range of cumulative exposures examined in the tree
seedling E-R studies (see Appendix 4A), given the way it is calculated the W126 index can
conceal peak concentrations that can be of concern. More specifically, one season or location
87	While the evidence includes some studies reporting Ch-reduced soybean yield and perennial plant biomass loss
using AOT40 (as well as W126) as the exposure metric, no newly available analyses are available that compare
AOT40 to W126 in terms of the strength of association with such responses. Nor are studies available that
provide analyses of E-R relationships for AOT with reduced growth or RBL with such extensiveness as the
analyses supporting the established E-R functions for W126 with RBL and RYL.
88	This is illustrated by the following two hypothetical examples. In the first example, two air quality monitors have
a similar pattern of generally lower average hourly concentrations but differ in the occurrence of higher
concentrations (e.g., hourly concentrations at or above 100 ppb). The W126 index describing these two monitors
would differ. In the second example, one monitor has appreciably more hourly concentrations above 100 ppb
compared to a second monitor; but the second monitor has higher average hourly concentrations than the first. In
the second example, the two monitors may have the same W126 index, even though the air quality patterns
observed at those monitors are quite different, particularly with regard to the higher concentrations, which have
been recognized to be important in eliciting responses (as noted above).
April 2022
4-84 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
could have few, or even no, hourly concentrations above 100 ppb89 and the second could have
many such concentrations; yet (due to greater prevalence of more mid-range concentrations, e.g.,
contributing to a generally higher average hourly concentration in the second) each of the two
seasons or locations could have the identical W126 index (e.g., equal to 25 or 15 or 10 ppm-hrs,
or some other value), as discussed in section 4.4.1 above.
Accordingly, in our consideration of the potential for vegetation-related effects to occur
under air quality conditions associated with the current standard, we continue to focus on the
W126 index as the appropriate metric, while also being aware of the importance of considering
the occurrence and frequency of particularly high concentrations. We also recognize that this
metric may not well describe the key circumstances of O3 exposure for occurrences of other
effects, particularly, visible foliar injury. As discussed in section 4.3.3.2 above, the evidence
indicates an important role for peak concentrations (e.g., N100) in influencing the occurrence
and severity of visible foliar injury. Thus, while we continue to recognize the W126 index as an
appropriate and biologically relevant focus for assessing air quality conditions with regard to
potential effects on vegetation growth and related effects, we also recognize the need for
attention to the pattern and magnitude of peak concentrations.
4.5.1.2 General Approach for Considering Public Welfare Protection
The general approach and risk management framework applied in 2015 for making
judgements and reaching conclusions regarding the adequacy of public welfare protection
provided by the newly established secondary standard is summarized in section 4.1 above. In
light of the available evidence and air quality information, we discuss here key considerations in
judging public welfare protection provided by the O3 secondary standard in the context of a
series of questions.
• Does the newly available information continue to support the use of tree seedling
RBL as a proxy for the broad array of vegetation-related effects?
As summarized in section 4.3 above, the available evidence is largely consistent with that
available in the 2015 review and does not call into question conceptual relationships between
plant growth impacts and the broader array of vegetation effects. Rather, the ISA describes (or
relies on) conceptual relationships in considering causality determinations for ecosystem-scale
effects such as altered terrestrial community composition and reduced productivity, as well as
reduced carbon sequestration, in terrestrial ecosystems (ISA, Appendix 8, sections 8.8 and 8.10).
89 As noted in section 4.4 above, the value of 100 ppb is used here as it has been in some studies focused on O3
effects on vegetation, simply as an indicator of elevated or peak hourly O3 concentrations (e.g., Lefohn et al.,
1997, Smith, 2012; Davis and Orendovici, 2006; Kohut, 2007). Values of 95 ppb and 110 ppb have also been
considered in this way (2013 ISA, section 9.5.3.1).
April 2022
4-85 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
Thus, the evidence continues to support the use of tree seedling RBL as a proxy for a broad array
of vegetation-related effects, most particularly those conceptually related to growth.
Beyond these relationships of plant-level effects and ecosystem-level effects,90 RBL can
be appropriately described as a scientifically valid surrogate of a variety of welfare effects based
on consideration of ecosystem services and the potential for adverse impacts on public welfare,
as well as conceptual relationships between vegetation growth-related effects (including carbon
allocation) and ecosystem-scale effects. Beyond tree seedling growth (on which RBL is
specifically based), two other vegetation effect categories with extensive evidence bases are crop
yield and visible foliar injury, both types of effects, their evidence bases and key considerations
with regard to protection afforded by the current standard (which go beyond a RBL target for
tree seedlings) are separately addressed in section 4.5.1.3 below.
• To what extent does the available information alter our understanding of an
appropriate magnitude of RBL, in its role as a surrogate or proxy, reasonably
expected to be of public welfare significance?
The available information does not differ from that available in the 2015 review with
regard to a magnitude of RBL in the median species appropriately considered a reference for
judgments concerning potential vegetation-related impacts to the public welfare. Based on the
available information, a 6% RBL median estimate from the established species-specific E-R
functions continues to be appropriate for such a reference point. We note this in the context of
RBL's role as a surrogate or proxy of a larger array of vegetation effects for which it was judged
that isolated rare instances of cumulative exposures that correspond to 6% (as the median of the
11 E-R functions) were not indicative of adverse effects to the public welfare (80 FR 65409,
October 26, 2015). The available evidence continues to indicate conceptual relationships
between reduced growth and the broader array of vegetation-related effects (as discussed above).
Quantitative representations of such relationships have been used to study potential impacts of
tree growth effects on such larger-scale effects as community composition and productivity with
the results indicating the array of complexities involved (e.g., ISA, Appendix 8, section 8.8.4).
Given their purpose in exploring complex ecological relationships and their responses to
environmental variables, as well as limitations of the information available for such work, these
analyses commonly utilize somewhat general representations. This work indicates how
90 As summarized in the ISA, 03 can mediate changes in plant carbon budgets (affecting carbon allocation to leaves,
stems, roots and other biomass pools) contributing to growth impacts, and altering ecosystem properties such as
productivity, carbon sequestration and biogeochemical cycling. In this way, O3 mediated changes in carbon
allocation can "scale up"to population, community and ecosystem-level effects including changes in soil
biogeochemical cycling, increased tree mortality, shifts in community composition, changes in species
interactions, declines in ecosystem productivity and carbon sequestration and alteration of ecosystem water
cycling (ISA, section 8.1.3).
April 2022
4-86 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
established the existence of such relationships is, while also identifying complexities inherent in
quantitative aspects of such relationships and interpretation of estimated responses. Thus, the
currently available evidence, as characterized in the 2020 ISA, is little changed from the 2015
review with regard to informing identification of an RBL reference point reflecting ecosystem-
scale effects with public welfare impacts elicited through such linkages.
• What does the available information indicate with regard to the roles of seasonal
cumulative and peak exposures on O3 vegetation effects, and accordingly regarding
the uses of cumulative and peak exposure metrics in assessing air quality conditions
that may pose risk of harm to vegetation?
As summarized in section 4.3.3, longstanding conclusions regarding O3 effects on
vegetation recognize both the cumulative effect of O3 on plants and the importance of higher
concentrations in eliciting responses (1996 and 2006 AQCDs; 2013 and 2020 IS As). As a result,
there has been substantial research into identification of an air quality exposure-related metric
that might address both aspects of potentially harmful O3 conditions. As discussed in section
4.3.3.1.1 above, the metrics explored have included, among others, those that sum the portion of
a concentration above a reference point (e.g., AOT06), those that sum only those concentrations
above a reference point (e.g., SUM06), and also, the W126 index, a non-threshold approach
described as the sigmoidally weighted sum of hourly O3 concentrations (2013 ISA, p. 9-101).
These indices (designed to address both cumulative effects and the importance of higher
concentrations) have been analyzed with regard to the extent to which they may describe the
growth response of plants (e.g., crops and tree seedlings) in studies assessing multiple exposure
levels and have been found to improve the explanatory power of E-R models over those based
only on mean (e.g., seasonal mean of 7-hour daily means) or peak exposure values (e.g., seasonal
maximum of maximum daily 7-hour and/or 1-hour averages) (2020 ISA, p. IS-79; 2013 ISA, p.
2-44; 2006 AQCD 1996 AQCD).
The explanatory strength of these cumulative, concentration-weighted approaches with
regard to plant response to O3 indicates the influence of the various dimensions of exposure (e.g.,
concentration, duration, frequency) on plant response. With regard to the role of concentrations,
the 2020 and 2013 IS As and past AQCDs generally recognize higher O3 concentrations to be
associated with relatively greater risk of vegetation damage, in terms growth-related effects
(and/or visible foliar injury, which is discussed more specifically in response to a question
below) and emphasize the risk posed to vegetation from higher hourly average O3
concentrations.91 With regard to duration and cumulative effects, analyses of the controlled
91 For example, as stated in the 2020 and 2013 ISAs, "[h]igher concentrations appear to be more important than
lower concentrations in eliciting a response" [ISA, p. 8-180]; "higher hourly concentrations have greater effects
April 2022
4-87 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
exposure datasets also supported conclusions in the 1996 and 2006 AQCDs (retained in more
recent ISAs) that a model focused only on a peak-concentration based metric (found to be an
improvement over earlier use of a long-term average to summarize exposure), without
consideration of duration was less descriptive of response (e.g., 1996 AQCD, Volume II, section
5.5.1.1). Accordingly, metrics that cumulated concentrations, e.g., through summing, as is the
case for those identified above, were developed, with preference to those that emphasized higher
concentrations (1996 and 2006 AQCDs; 2020 ISA, IS 5.1.9).
As recognized across several past reviews, the strength of the cumulative, concentration-
weighted approaches, including the continuously weighted W126 index function, is in describing
variation in response documented in controlled exposure studies or crops and tree seedlings for
which extensive hourly O3 datasets are available. We note that in these exposures studies, the
higher cumulative exposure levels (e.g., W126 index levels) were generally accompanied by an
appreciable prevalence of high concentrations (e.g., Appendix 4A, Table 4A-6; Lefohn et al
1997; Lefohn and Foley, 1992). While these were part of the patterns of O3 concentrations to
which the plants were exposed, another exposure circumstance may have the same W126 index,
yet with a different pattern of peak concentrations that may contribute to differences in risk of
vegetation effects. In an example highlighted in the 2006 AQCD and 2013 ISA, a study by Yun
and Lawrence (1999) used exposure regimes constructed from 10 U.S. cities to demonstrate that
in regimes with similar values of cumulative, concentration-weighted metrics, differences in the
magnitude and occurrence of peak concentrations were influential with regard to injury in tree
seedlings (2006 AQCD, p. AX9-176; 2013 ISA, section 9.5.3.1; Yun and Lawrence, 1999).92
Given this, we recognize that the seasonal cumulative metrics may not always differentiate
between air quality patterns that include particularly high peak concentrations and those without
or with relatively fewer such concentrations.
For example, while the W126 index preferentially weights higher hourly concentrations,
given its definition as an index that sums three months of weighted hourly concentrations into a
single value, it can estimate the same value for very different incidence of elevated O3
concentrations. As described in section 4.5.1.1, at two sites with the same W126 index value, the
air quality patterns may differ such that one site may have appreciably more hourly
on vegetation than lower concentrations" [2013 ISA, p. 91-4] "studies published since the 2006 03 AQCD do not
change earlier conclusions, including the importance of peak concentrations, ... in altering plant growth and
yield" [2013 ISA, p. 9-117]).
92 The 2013 ISA, in examining trends (1970s through 1990s) in an areas of the San Bernardino Mountains in
California, noted the reductions in ponderosa pine growth impacts occurring with reductions in SUM06,
maximum peak concentration and hourly concentrations over 95 ppb. In observing that there had been little
change in mid-range O3 concentrations over the same period, the 2013 ISA noted the lesser role indicated for the
mid-range concentration ranges compared to the higher values (2013 ISA, p. 9-106).
April 2022
4-88 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
concentrations at or above 100 ppb compared to the other site. This is also supported by the
analyses of available air quality data summarized in section 4.4.1 (e.g., Appendix 4F, Figure 4F-
10). Focusing on the data for the most recent five years (2016 through 2020), the distribution of
N100 or D100 values at monitoring sites meeting different W126 index values also shows this
variability, which contrasts with the much lesser variability in N100 and D100 values for sites
meeting the current standard (see Figure 4-12, W126 index bins at/below 19 ppm-hrs compared
to design value bins for 70 ppb or lower). It can be seen that (1) there is little difference in D100
at sites with W126 index ranging from 8 to 19 ppm-hrs (single-year or 3-year average index);
and (2) the form and averaging time of the existing standard is much more effective than the
W126 index in limiting the number of hours with O3 concentrations at or above 100 ppb (N100)
and in limiting the number of days with any such hours.93
Given the considerations raised here, we recognize that focusing solely on W126 index
for considering the public welfare protection provided by the current standard would not be
considering all the relevant scientific information. Further, we note that such a sole focus, given
the damaging potential for repeated elevated hourly concentrations (e.g., at or above 100 ppb), as
discussed in sections 4.3.3 and 4.5.1.1 above (ISA, p. 8-180; 2013 ISA, section 9.5.3.1)94, may
not give adequate attention to ensuring protection against "unusually damaging years." As a
result, we find that focusing solely on the W126 index may not ensure protection is provided
from potentially damaging air quality, such as that associated with exposure patterns marked by
repeated occurrences of elevated concentrations. Thus, we conclude it is important to consider
both cumulative, concentration-weighted and peak exposure metrics in assessing air quality with
regard to the potential for specific exposure conditions that might be harmful to vegetation."95
93	As one example contained in Table 4-1 above, across all sites that met the current standard during the recent
period (2018-2020), few sites had more than 5 hours at or above 100 ppb in a year (0.6% in the highest year,
Appendix 4F, Table 4F-2). Among the sites with any such hours, all had fewer than five days in any one year
with any such concentrations (Table 4-1, Appendix 4F, Figure 4F-5). In comparison, across all sites with an
annual W126 index below 15 ppm-hrs, 2% of them had more than 5 hours with a concentration at or above 100
ppb, and this included sites with as many as eight days with such a concentration (Table 4-1, Appendix 4F, Figure
4F-1 l).We note that we are not intending to ascribe specific significance to five days with an hour at or above
100 ppb or ten such hours, per se. Rather, these are used simply as reference points to facilitate comparison and
to illustrate the point that such high concentrations, which based on toxicological principles, pose greater risk to
biota than lower concentrations, are not necessarily limited at sites meeting particular W126 index values.
94	The section of the 2013 ISA titled "Role of Concentration," summarizes the experimental evidence base on which
the significant role of peak O3 concentrations was established (2013 ISA, section 9.5.3.1).
95	With regard to air quality occurring under the current standard, we note analyses presented in section 4.4 above
that show the current standard to provide control of both cumulative exposures and of peak concentrations
indicating the potential to address both aspects of potentially harmful O3 conditions noted here.
April 2022
4-89 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
• What does the available information indicate regarding the use of W126 index in a
single year or averaged over three years in considering cumulative seasonal exposure
protection objectives for the secondary standard?
In setting the current standard in 2015, as described in section 4.1 above, the decision
focused on control of seasonal cumulative exposures in terms of a 3-year average W126 index
based on consideration of several factors.96 We again consider here the extent to which the
available evidence supports the 3-year average W126 index as a reasonable metric for assessing
the level of protection provided by the current standard from cumulative seasonal exposures
related to RBL, or whether an alternate approach is more appropriate for use with the E-R
functions.
We first consider the evidence and information underlying the E-R functions and the
extent to which they can be said to better describe or predict growth reductions specific to single
season exposures, as compared to growth reductions generally reflecting an average seasonal
exposure. With regard to the established tree seedling E-R functions themselves, we note there
are aspects of the datasets and methodology on which the E-R functions are based which provide
support for a multiyear (e.g. 3-year) average approach. As summarized in section 4.3.4 above,
the E-R functions were derived from studies of durations that varied from shorter than 92 days to
as many as 140 days in a single year, and up to 555 days distributed across multiple years or
growing seasons, with the results normalized to the duration of a single 92-day seasonal period
(Appendix 4A, pp. 4A-31 to 4A- 32). Inherent in this approach is an assumption that the growth
impacts relate generally to the cumulative O3 exposure across the full time period (which may
include multiple growing seasons), i.e., with little additional influence related to any seasonal or
year to year differences in the exposures. Consequently, given this step in their derivation
approach, the E-R functions cannot provide precise estimates of response from a single year's
seasonal exposure (e.g., vs averages over a period longer than 92 days or one that spans multiple
growing seasons). Thus, the use of a multiyear (e.g. 3-year) average in assessing RBL using the
established tree seedling E-R functions is reasonably described as compatible with the
normalization step taken to derive functions for a seasonal 90-day period from the underlying
data with its varying exposure durations.
96 These factors include consideration of the strengths and limitations of the evidence and of the information on
which to base judgments regarding adversity of effects on the public welfare (80 FR 65390, October 26, 2015).
Also recognized was year-to-year variability, not just in O3 concentrations, but also in environmental factors,
including rainfall and other meteorological factors, that influence the occurrence and magnitude of 03-related
effects in any year (e.g., through changes in soil moisture), contributing uncertainties to projections of the
potential for harm to public welfare based on a single year, particularly at the exposure levels of interest (80 FR
65404, October 26, 2015).
April 2022
4-90 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
We also take note of aspects of the evidence that reflect variability in organism response
under different experimental conditions and the extent to which this variability is represented in
the available data, which might indicate an appropriateness of assessing environmental
conditions using a mean across seasons in recognition of the existence of such year-to-year
variability in conditions and responses. For example, among the species for which there are more
than two or three experimental datasets comprising the support for the species' E-R function (14
experimental datasets for aspen [seven for which the E-R function for wild aspen has been
derived and seven supporting a function for aspen clones] and 11 for ponderosa pine) illustrate
appreciable variability in response across experiments (Appendix 4A, Figure 4A-10).
Contributions to this variability may come from several factors, including variability in seasonal
response related to variability in non-03 related environmental influences on growth, such as
rainfall, temperature and other meteorological variables, as well as biological variability across
individual seedlings. An additional variability could also be due to influential aspects of the O3
air quality on plant growth that are not completely captured by the W126 index, e.g., different
patterns of hourly concentrations that yield the same W126 index (see section 4.4.1 and below).
Such variability in the data underlying these E-R functions may further support a multiyear (e.g.
3-year) average approach.
An additional aspect of considering the evidence and information, is how well the data
underlying the E-R functions represent and reflect conditions that are currently being
experienced in the U.S., and most importantly, conditions that reflect current air quality patterns
when meeting the current standard. On a related note, it is also important to understand the extent
to which E-R predictions are extrapolated beyond the tested exposure conditions. As noted in
section 4.3.4 above, the O3 concentrations and cumulative exposures for the experimental
datasets from which the tree seedling E-R functions were derived include conditions that do not
occur in ambient air at sites the meet the current standard (section 4.4; Appendix 4A, Table 4A-
6; section 4.4). A similar issue was discussed in a previously available publication that observed
appreciable differences between the prevalence of hourly concentrations at or above 100 ppb in
exposures on which the E-R functions are based and those common in ambient air at that time, a
difference which is in many ways only increased with today's air quality (Lefohn et al., 1997).97
97 For example, many of the experimental exposure of elevated O3 on which the established E-R functions for the 11
tree seedling species are based, had hundreds of hours of O3 concentrations above 100 ppb, far more than are
common in (unadjusted) ambient air, including in areas that meet the current standard (Lefohn et al., 1997,
Appendix 2A, section 2A.2, Appendix4 F). To illustrate, in the most recent 2018-2020 design value period, the
mean number of observations per site at or above 100 ppb was well below one. In contrast, across most of the O3
treatments in the experiments comprising the E-R function database, well below half had an N100 value less than
20 hours through the exposure period (Appendix 4A, Table 4A-6). Similarly, the experimental exposures in
studies supporting some of the established E-R functions for 10 crop species also include many hours with O3
concentrations at or above 100 ppb (Lefohn and Foley, 1992)
April 2022
4-91 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
This issue is also discussed in section 4.3.4 above, where it is noted that in the E-R tree seedling
datasets, the O3 treatments for W126 index levels observed in areas that meet the current
standard had N100 counts ranging up above 40. And for many of the treatments, N100 values
range up to several hundred (Appendix 4A, Table 4A-6). We find it reasonable to interpret this
information, and its contribution to uncertainty in the application of the underlying E-R
functions, as arguing for a less precise interpretation, such as an average across multiple seasons.
In further considering the evidence and information and its support for use of a single or
multiple year W126 index, the concept of cumulative multiyear exposures and associated
impacts should be considered. In particular, we ask the question of whether applying the E-R
functions to a W126 index averaged over multiple years would over- or under-estimate
cumulative exposure response, whereas use of a single seasonal exposure metric would not. The
evidence relevant to this question, e.g., that allows for specific evaluation of the predictability of
growth impacts from single-year versus multiple-year average exposure estimates, is limited.
Multi-year studies reporting results for each year of the study are the most informative to the
question of plant annual and cumulative responses to individual years (high and low) over
multiple-year periods. However, as summarized in section 4.3.4 above, the evidence is quite
limited with regard to studies of O3 effects that report seasonal observations across multi-year
periods and that also include detailed hourly O3 concentration records (to allow for derivation of
cumulative exposure index values). One such study, which tracked exposures across six years, is
available for aspen (King et al., 2005; 2013 ISA, section 9.6.3.2; ISA, Appendix 8, section
8.13.2). This study is presented in the 2013 and 2020 ISAs in an evaluation of predicted growth
impacts compared to observations from the multiple years of the study.
For this evaluation, the ISAs considered the 6-year experimental dataset of O3 exposures
and aspen growth effects with regard to correspondence of E-R function predictions with study
observations (2020 ISA, Appendix 8, section 8.13.2 and Figure 8-17; 2013 ISA, section 9.6.3.2,
Table 9-15, Figure 9-20). The analysis in the 2013 ISA compared observed reductions in growth
for each of the six years to those predicted by applying the established E-R function for Aspen to
cumulative multi-year average W126 index values (2013 ISA, section 9.6.3.2).98 99 The
evaluation in the 2020 ISA applied the E-R functions to the single-year W126 index for each
year rather than the cumulative multi-year W126 (2020 ISA, Appendix 8, Figure 8-17), with this
98	Although not emphasized or explained in detail in the 2013 ISA, the W126 index estimates used to generate the
predicted growth response were cumulative averages. For example, the growth impact estimate for year 1 used
the W126 index for year 1; the estimate for year 2 used the average of W126 index in year 1 and W126 index in
year 2; the estimate for year 3 used the average of W126 index in years 1,2 and 3; and so on.
99	One finding of this evaluation was that "the function based on one year of growth was shown to be applicable to
subsequent years" (2013 ISA, p. 9-135).
April 2022
4-92 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
approach indicating a somewhat less tight fit to the experimental observations (2020 ISA,
Appendix 8, p. 8-192),100 Both IS As reach similar conclusions regarding general support for the
E-R functions across a multiyear study of trees in naturalistic settings (ISA, Appendix 8, section
8.13.3 and p. 8-192; 2013 ISA, p. 9-135).101 We additionally note that an illustrative
mathematical exercise that explored estimates of above ground biomass of an aspen stand when a
multi-year O3 exposure was quantified in terms of a single year varying W126 index or as a
repeated yearly exposure equal to the associated 3-year average. These analyses suggest that the
two approaches may yield generally similar total biomass estimates after multiple years'
exposure (Appendix 4A, section 4A.3).
Thus, while the E-R functions are based on strong evidence of cumulative seasonal O3
exposure reducing tree growth, and while they provide for quantitative characterization of the
extent of such effects across cumulative seasonal O3 exposure levels of appreciable magnitude,
there is uncertainty associated with the resulting RBL predictions that might be described as an
imprecision or inexactitude. Further, as summarized above, the evidence does not indicate
single-year seasonal exposure in combination with the established E-R functions to be a better
predictor of RBL than a seasonal exposure based on a multi-year average. Accordingly, it is
reasonable to conclude that the evidence provides support for use of a 3-year average in
assessing the level of protection provided by the current standard from cumulative seasonal
exposures related to RBL of concern based on the established E-R functions.102 The 3-year
average metric also appears to be reasonable for use in the context of the use of RBL as a proxy
to represent an array of vegetation-related effects. Accordingly, upon consideration of all of the
factors raised above, we find the use of a multiyear average, and more specifically a 3-year
average, W126 index in assessing protection for RBL based on the established tree seedling E-R
functions to be reasonable. We also note, as discussed in response to the prior question, the
importance of also considering an additional aspect of O3 air quality, specifically the occurrence
100	Based on information drawn from Figure 8-17 in the 2020 ISA, the correlation metric (r2) for the percent
difference (estimated vs observed biomass) and year of growth can be estimated to be approximately 0.7, while
using values reported in Table 9-15 of the 2013 ISA (which are plotted in Figure 9-20), the r2 for predicted O3
impact versus observed impact is 0.99 and for the percent difference versus year is approximately 0.85.
101	For the 2013 ISA, the conclusions reached were that the agreement between the set of predictions and the Aspen
FACE observations were "very close" (2013 ISA, p. 9-135). The results indicate that when considering O3
impacts across multiple years, a multi-year average index yields predictions close to observed measurements
(2013 ISA, section 9.6.3.2 and Figure 9-20; Appendix 4A, section 4.A.3). For the 2020 ISA, the conclusion
reached was that results from the aspen study were "exceptionally close" to predictions from the E-R model (ISA,
p. 8-192
102	Three years (versus two or four years) was selected based on its compatibility with the multiyear duration often
used in forms for NAAQS to account for year-to-year variability in air quality.
April 2022
4-93 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
of elevated hourly concentrations that influence vegetation exposures of potential concern, in
reaching conclusions about the adequacy of the current standard.
• What does the available information indicate for considering potential public welfare
protection from C>3-related visible foliar injury?
In establishing the current secondary standard in 2015 and its underlying public welfare
protection objectives, as summarized in section 4.1, above, the Administrator focused primarily
on RBL in tree seedlings as a proxy or surrogate for the full array of vegetation related effects of
O3 in ambient air, from sensitive species to broader ecosystem-level effects. At that time, the
Administrator also concluded the information regarding visible foliar injury to also provide
support for strengthening the standard at that time, taking note of the available analyses of USFS
biosite data (80 FR 65407-65408, October 26, 2015). She also concluded, however, that, due to
associated uncertainties and complexities, the evidence was not conducive to use for identifying
a quantitative public welfare protection objective focused specifically on visible foliar injury. In
reaching this conclusion, she recognized significant challenges in judging the specific extent and
severity at which such effects should be considered adverse to public welfare, in light of the
variability in the occurrence of visible foliar injury and the lack of clear quantitative relationships
for prediction of visible foliar injury severity and incidence or extent under varying air quality
and environmental conditions, as well as the lack of established criteria or objectives that might
inform consideration of potential public welfare impacts related to this vegetation effect (80 FR
65407, October 26, 2015).
As an initial matter, we note that, as recognized in the 2015 review, some level of visible
foliar injury can impact public welfare and thus might reasonably be judged adverse to public
welfare.103 As summarized in section 4.3.2 above, depending on its spatial extent and severity,
there are many locations in which visible foliar injury can adversely affect the public welfare.
For example, significant, readily perceivable (or obvious) and widespread injury in national
parks and wilderness areas can adversely impact the perceived scenic beauty of these areas,
impacting the aesthetic experience for both outdoor enthusiasts and the occasional park visitor.104
103	As stated in the Federal Register notice for the 2015 decision: "[depending on the extent and severity, O3-
induced visible foliar injury might be expected to have the potential to impact the public welfare in scenic and/or
recreational areas during the growing season, particularly in areas with special protection, such as Class I areas.
(80 FR 65379, October 26, 2015); "[t]he Administrator also recognizes the potential for this effect to affect the
public welfare in the context of affecting values pertaining to natural forests, particularly those afforded special
government protection (80 FR 65407, October 26, 2015). The CASAC in the 2015 review also stated that visible
foliar injury "can impact public welfare" (Frey, 2014, p. 10).
104	In the discussion of the need for revision of the 1997 secondary standard, the 2008 decision noted that "[i]n
considering what constitutes a vegetation effect that is adverse from a public welfare perspective, ... the
Administrator has taken note of a number of actions taken by Congress to establish public lands that are set aside
April 2022
4-94 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
Thus, as aesthetic value and outdoor recreation depend, at least in part, on the perceived scenic
beauty of the environment, judgments related to the extent of public welfare impacts of visible
foliar injury depend on the severity and extent of the injury, as well as the location where the
effects occur and the associated intended use. Beyond the limitations associated with the
evidence for descriptive quantitative relationships for O3 concentrations and visible foliar injury
(as summarized in sections 4.3.3.2 and 4.3.4 above), there is little information clearly relating
differing severity and prevalence of injury to conditions in natural areas that would reasonably be
concluded to impact public use and enjoyment in a way that might suggest adversity to the public
welfare. The available information does not yet address or describe the relationships expected to
exist between some level of severity and/or extent of location affected and scenic or aesthetic
values (e.g., reflective of visitor enjoyment and likelihood of frequenting such areas). However,
while minor spotting on a few leaves of a plant may easily be concluded to be of little public
welfare significance, it might reasonably be expected that in cases of widespread and relatively
more severe injury during the growing season (particularly when sustained across multiple years
and accompanied by obvious impacts on the plant canopy), Cb-induced visible foliar injury could
adversely impact the public welfare in scenic and/or recreational areas, particularly in parks and
other areas with special protection, such as Class I areas.
In the face of the paucity of established approaches that might be informative to the
Administrator in judging severity and extent of visible foliar injury in a natural area that may be
appropriate to consider of public welfare significance, we take note of the USFS scheme,
summarized in section 4.3.2 above, for categorizing areas based on BI scores (e.g., Smith, 2012).
In this scheme, BI scores may be described with regard to one of several categories ranging from
little or no foliar injury to severe injury (e.g., Smith et al., 2003; Campbell et al., 2007; Smith et
al., 2007; Smith, 2012). However, the available information does not yet address or describe the
relationships expected to exist between some level of severity of foliar injury (e.g., little or
severe) and/or a spatial extent affected and scenic or aesthetic values. This gap impedes
consideration of the public welfare implications of different injury severities, and accordingly,
judgments on the potential for public welfare significance.
With regard to the USFS BI program, we further note that authors of studies presenting
USFS biomonitoring program data have suggested what might be "assumptions of risk" (e.g., for
the forest resource) related to scores in these categories, e.g., as described in section 4.3.2 above.
for specific uses that are intended to provide benefits to the public welfare, including lands that are to be protected
so as to conserve the scenic value and the natural vegetation and wildlife within such areas, and to leave them
unimpaired for the enjoyment of future generations" (73 FR 16496, March 27, 2008). This passage of the Federal
Register notice announcing the 2008 decision clarified that "[s]uch public lands that are protected areas of
national interest include national parks and forests, wildlife refuges, and wilderness areas" (73 FR 16496, March
27, 2008).
April 2022
4-95 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
One suggestion has been that maps of localized moderate to high-risk areas may be used to
identify areas (for scores of 15 or higher) where more detailed evaluations are warranted (Smith
et al., 2012). While these are not explicitly related to consideration of the public values described
above (e.g., with regard to public aesthetic or recreational value), the description of the BI score
categories as well as these corresponding judgments related to risk for the forest resource may
both be informative for the Administrator's purposes. For example, it might be reasonable to
conclude that a small discoloring on a single leaf of a plant that might yield a quite low, nonzero
BI score in the USFS system is not adverse to the public welfare. On the other hand, BI scores
corresponding to a high risk to the resource may reasonably be concluded to indicate the need for
attention and, perhaps a public welfare adversity potential. Thus, while the available evidence
does not include characterization of USFS biosite scores with regard to public perception and
potential impacts on public enjoyment, we find that they may be useful for the Administrator's
purposes in considering the potential public welfare significance of different severities and
extents of visible foliar injury, as scored by BI. That notwithstanding, limitations remain in our
tools for characterizing the air quality conditions at sites that elicit scores of a particular severity
level, thus continuing to challenge our ability to precisely identify conditions that might provide
particular levels of public welfare protection for this effect.
In considering the available information regarding a relationship between W126 index
and the severity of visible foliar injury, we consider the presentation of USFS biosite data in
Appendix 4C, summarized in section 4.3.3.2.2 above. While recognizing limitations in the
dataset105 and considering the records for the normal or dry soil moisture categories, for which
there is somewhat better representation of W126 index levels above 13 ppm-hrs,106 we note the
lack of a clear trend in the percentage of USFS records recording visible foliar injury (of any
severity level) W126 index estimates below 17 ppm-hrs. Focusing on the magnitude of BI score,
we note that among records in the normal soil category, BI scores are noticeably increased in the
highest W126 index bin (above 25 ppm-hrs) compared to the others. The percentages of records
in the greater than 25 ppm-hrs bin that have BI scores above 15 ("moderate" and "severe" injury)
and above 5 ("light," "moderate" and "severe" injury) are more than three times greater than
percentages for these score levels in any of the lower W126 bins. Additionally, the average BI of
7.9 in the greater-than-25-ppm-hrs bin is more than three times the average BI for the next
105	For example, the majority of these data are records with W126 index estimates at or below 9 ppm-hrs, and fewer
than 10% of the records have W126 estimates above 15 ppm-hrs. Additionally, the BI scores are quite variable
across the full dataset, with even the bin for the lowest W126 index estimates (below 7 ppm-hrs) including BI
scores well above 15 (Appendix 4C, section 4C.4.2).
106	In the case of records in the wet soil moisture category, nearly 90% of the records are for W126 estimates at or
below 9 ppm-hrs, limiting interpretations for higher W126 bins (Appendix 4C, Table 4C.4 and section 4C.6).
April 2022
4-96 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
highest W126 index bin. The average BI in the next two lower W126 bins (which vary inversely
with W126 index) are just slightly higher than average Bis for the rest of the bins, and the
average BI for all bins at or below 25 ppm-hrs are well below 5. Among records in the dry soil
moisture category, the two highest W126 bins (which together include the W126 index estimates
above 19 ppm-hrs) exhibit percentages of records with BI above 15 or above 5 that are
appreciably greater than that for the lower W126 bins. With regard to average scores across all
dry soil moisture records, average BI for all W126 index bins is below 5, although the three
highest W126 index bins (above 17 ppm-hrs) are markedly greater than the lower bins (e.g.,
average Bis greater than versus less than 1).
Thus, the strongest conclusions that can be reached from the USFS dataset described in
Appendix 4C are that the incidence of sites with more severe injury (e.g., BI score above 15 or 5)
is also lower at sites with W126 index values below 25 ppm-hrs than at sites with higher W126
index values and that clear trends in such incidence related to increasing W126 index levels are
not evident across the bins for lower W126 index estimates (all of which are below 5%). As
discussed in section 4.3.3.2 above, variability in the data across sites, and uncertainty, with
regard to the role of peak O3 concentrations as an influence on occurrence of visible foliar injury
separate from cumulative W126 index, lead to the conclusion that the available information does
not support precise conclusions as to the severity and extent of such injury associated with the
lower values of W126 index most common at USFS sites during the time of the dataset (2006-
2010). Notwithstanding this, records categorized as normal soil moisture indicate there to be an
appreciable difference in severity of injury between records with W126 index estimates above 25
ppm-hrs and those with estimates at or below 25 ppm-hrs (e.g., Appendix 4C, Figures 4C-5 and
4C-6 and Table 4C-5). The records categorized as dry soil moisture do not indicate such a clear
pattern. The records categorized as wet soil moisture are too limited (and variable) for W126
index estimates above 13 ppm-hrs to support a conclusion (Appendix 4C). Thus, we conclude,
based primarily on the BI scores records categorized as having normal soil moisture, that under
conditions that maintain W126 index values below 25 ppm-hrs a reduced severity (average BI
score below 5) and incidence of visible foliar injury, as quantified by biosite index scores, would
be expected. The observation of a lack of clear relationship between levels of a cumulative
seasonal index and BI scores until reaching a higher value is conceptually similar to findings of
the study by Campbell et al. (2007), identified in the 2013 ISA that focused on visible foliar
injury in west coast states. This study observed that both percentage of USFS biosites with injury
and the average BI were higher for sites with average cumulative O3 concentrations above 25
ppm-hrs in terms of SUM06 as compared to groups of sites with lower average cumulative
exposure levels, with little difference apparent between the two lower exposure groups (80 FR
April 2022
4-97 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
65395, October 26, 2015; Smith and Murphy, 2015; Campbell et al., 2007, Figures 27 and 28
and p. 30).107
Such findings of variability in scores at lower values of a cumulative seasonal index and
a lack of clear relationship with exposure may relate to patterns of peak concentrations at sites
with similar cumulative seasonal index values. As discussed in section 4.3.3.2 above, several
studies of the USFS data have concluded that inclusion of a metric for quantifying peak
concentrations, in combination with one for cumulative seasonal exposures, may yield a more
predictive description of the relationship between 03 air quality and the occurrence of visible
foliar injury. Similarly, a county-scale analysis of USFS biosite data in the 2007 Staff Paper
(from earlier years than those analyzed in the 2015 review) indicated a somewhat smaller
incidence of biosites with nonzero BI scores in counties with air quality meeting a fourth-high
metric of 74 ppb as compared to larger groups that also included sites with air quality meeting a
fourth-high metric up to 84 ppb (U.S. EPA 2007, pp. 7-63 to 7-64; 80 FR 65395, October 26,
2015). Given the control of the averaging time and form of the current standard on peak
concentrations (as discussed in section 4.4.1 above), this observation is consistent with a role for
peak concentrations in eliciting visible foliar injury. Although given that lower design values for
the current standard also yield lower W126 index values, the relative influence of peak
concentrations and cumulative seasonal exposures cannot be distinguished. With regard to the
control of the current standard on peak concentrations, however, we note the conceptual
similarity to the finding of the most recent and extensive USFS data analysis that reductions in
peak 1-hour concentrations have influenced the declining trend in visible foliar injury since 2002
(Smith, 2012).
In consideration of all of the above, we recognize the appreciable limitations of the
available information touched on above with regard to providing a foundation for judgments on
public welfare protection objectives specific to visible foliar injury. In light of such limitations
and in light of the above discussion, we recognize that while the evidence continues to show a
consistent association between the occurrence of visible injury and ozone, "visible foliar injury is
not always a reliable indicator of other negative effects on vegetation" (ISA, Appendix 8, section
8.2), and we do not have a precise understanding of the appropriate metrics for quantifying O3 air
quality conditions for the purposes of informing the Administrator's consideration of this
endpoint. Based on studies and analyses of the USFS biosite data, the conditions associated with
visible foliar injury in locations with sensitive species appear to relate to peak concentration
107 In considering their findings, the authors expressed the view that "[although the number of sites or species with
injury is informative, the average biosite injury index (which takes into account both severity and amount of
injury on multiple species at a site) provides a more meaningful measure of injury" for their assessment at a
statewide scale (Campbell et al., 2007).
April 2022
4-98 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
(e.g., hours above a concentration such as 100 ppb) as well as sustained exposure to higher
concentrations over the growing season, such that cumulative exposure metrics may not well or
completely describe or predict the occurrence and severity of injury. Thus, in making judgments
regarding air quality conditions of concern and those providing protection with regard to impacts
associated with incidence and severity of visible foliar injury, we find it appropriate to consider
both cumulative concentration-weighted seasonal exposures and the occurrence of peak
concentrations. In this context, we note the control of these metrics achieved by the form and
averaging time of the current standard, as discussed in section 4.4 above. Lastly, we take note of
the USFS BI scheme as potentially useful to informing the Administrator's consideration of the
potential public welfare significance of differing magnitudes of BI scores.
• What does the available information indicate for considering potential public welfare
protection from C>3-related climate effects?
In considering the available information for the effects of the global abundance of O3 in
the troposphere on radiative forcing, and temperature, precipitation and related climate variables,
we note as an initial matter that, as summarized in section 4.3.3 above, there are limitations and
uncertainties in the associated evidence bases with regard to assessing potential for occurrence of
climate-related effects as a result of varying ground-level O3 concentrations in ambient air of
locations in the U.S. Specifically, such limitations and uncertainties affect our ability to
characterize the extent of any relationships between O3 concentrations in ambient air in the U.S.
and climate-related effects, thus precluding a quantitative characterization of climate responses
to changes in ground-level O3 concentrations in ambient air at regional (vs global) scales that
might inform considerations related to the current standard. While the evidence supports a causal
relationship between the global abundance of O3 in the troposphere and radiative forcing, and a
likely causal relationship between the global abundance of O3 in the troposphere and effects on
temperature, precipitation, and related climate variables (ISA, section IS.5.2 and Appendix 9;
Myhre et al., 2013), the non-uniform distribution of O3 (spatially and temporally) makes the
development of quantitative relationships between the magnitude of such effects and differing
ground-level O3 concentrations in the U.S. challenging (ISA, Appendix 9). Additionally, "the
heterogeneous distribution of ozone in the troposphere complicates the direct attribution of
spatial patterns of temperature change to ozone induced [radiative forcing]" and there are "ozone
climate feedbacks that further alter the relationship between ozone [radiative forcing] and
temperature (and other climate variables) in complex ways" (ISA, Appendix 9, section 9.3.1, p.
9-19). Thus, various uncertainties "render the precise magnitude of the overall effect of
tropospheric ozone on climate more uncertain than that of the well-mixed GHGs" and "[c]urrent
limitations in climate modeling tools, variation across models, and the need for more
comprehensive observational data on these effects represent sources of uncertainty in quantifying
April 2022
4-99 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
the precise magnitude of climate responses to ozone changes, particularly at regional scales"
(ISA, section IS.6.2.2, Appendix 9, section 9.3.3, p. 9-22).
As one example, current limitations in modeling tools include "uncertainties associated
with simulating trends in upper tropospheric ozone concentrations" (ISA, section 9.3.1, p. 9-19),
and uncertainties such as "the magnitude of [radiative forcing] estimated to be attributed to
tropospheric ozone" (ISA, section 9.3.3, p. 9-22). Further, "precisely quantifying the change in
surface temperature (and other climate variables) due to tropospheric ozone changes requires
complex climate simulations that include all relevant feedbacks and interactions" (ISA, section
9.3.3, p. 9-22). An important specific limitation in current climate modeling capabilities for O3 is
representation of important urban- or regional-scale physical and chemical processes, such as O3
enhancement in high-temperature urban situations or O3 chemistry in city centers where NOx is
abundant. Because of such limitations in the available information, we lack the ability to quantify
or judge the impact of incremental changes in ground-level O3 concentrations in the U.S. on
radiative forcing and subsequent climate effects, thus precluding a consideration of potential
public welfare protection provided by the existing O3 standard from 03-related climate effects.108
4.5.1.3 Public Welfare Implications of Air Quality under the Current Standard
Our consideration of the available scientific evidence in this reconsideration, as at the
time of the 2015 review, is informed by results from a quantitative analysis of air quality and
associated exposure. An overarching consideration is whether this information calls into question
the adequacy of protection provided by the current standard. As in our consideration of the
evidence above, we have organized the discussion regarding the information related to exposures
and potential risk around a key question to assist us in considering the quantitative analyses of air
quality at U.S. locations nationwide, particularly including those in Class I areas. We first
consider analyses particular to cumulative O3 exposures, in terms of the W126 index, given the
established E-R relationships with growth-related effects, and specifically RBL as the identified
proxy or surrogate for the full array of such effects.
To understand the cumulative O3 exposures likely occurring under the current standard
nationally, including in Class I areas, we consider the air quality analyses summarized in section
4.4 above. Nationwide in the most recent 3-year period, seasonal W126 index values are at or
below 17 ppm-hrs, as assessed by the 3-year average, when the current standard is met (Table 4-
108 While these complexities inhibit our ability to analyze and quantitatively climate-related effects of 03, such as
radiative forcing, we note that our consideration of O3 growth-related impacts on trees inherently encompasses
consideration of the potential for O3 to reduce carbon sequestration in terrestrial ecosystems (e.g., through
reduced tree biomass as a result of reduced growth). That is, limiting the extent of 03-related effects on growth
would be expected to also limit reductions in carbon sequestration, a process that can reduce the tropospheric
abundance of CO2, the greenhouse gas ranked highest in importance (section 4.3.3.3 above; ISA, section 9.1.1).
April 2022
4-100 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
3). With very few exceptions, this is also true across the full historical period. Further, such
exposures are generally well below 17 ppm-hrs across most of the U.S. Additionally, the overall
pattern for single-year seasonal W126 index values at monitors meeting the current standard in
the most recent period is generally similar, with few sites (about a dozen of the 877 sites
nationwide) having a single-year W126 index above 19 ppm-hrs (and under two dozen above 17
ppm-hrs).109 The frequency of such higher single-year W126 index values at Class I area
monitors is also low during periods when the current standard is met. During the most recent
three years, the average seasonal W126 index is at or below 17 ppm-hrs at all Class I area
monitors meeting the current standard, just two single-year W126 index values above 17 ppm-
hrs and none above 19 ppm-hrs (Appendix 4D, Table 4D-16).110
Combining this information regarding likely W126-based exposure levels with the
established E-R functions for 11 tree seedling species indicates that based on monitoring data for
locations meeting the current standard during the most recent design period, the median species
RBL for tree seedlings, based on the 3-year average W126, would be at or below 5.3%, with
very few exceptions; the highest estimates are associated with W126 index values occurring in
areas that are not near or within Class I areas. Looking at the data over a longer time period
(2000-2018) confirms this general pattern for the bulk of the data, with some infrequent higher
occurrences, such that virtually all RBL estimates would be below 6%.m Further, given the
variability and uncertainty associated with the data underlying the E-R functions (as discussed in
section 4.5.1.2 above), the few higher single-year occurrences are reasonably considered to be of
less significance than 3-year average values.
With regard to visible foliar injury, as discussed earlier, the evidence is somewhat limited
and unclear with regard to the metric and quantitative approach that well describes a relationship
between incidence or severity of injury in U.S. forests across a broad range of air quality
conditions. However, we note several key findings of the evidence and quantitative analyses.
First, the increased incidence of BI scores associated with injury considered greater than "a
little" by the USFS scheme appears most consistently with higher W126 estimates, with greatest
109	These highest W126 index values occur in the Southwest and West regions in which there are nearly 150 monitor
locations meeting the current standard (Figure 4-6; Appendix 4D, Table 4D-1).
110	Across the full 21-year dataset for Class I area monitors meeting the current standard (57 monitors with at least
one such period), there are 15 design value periods with single-year W126 index values above 19 ppm-hrs, all of
which are prior to the 2013-2015 period (Appendix 4D, section 4D.3.2.4).
111	Although potential for effects on crop yield was not given particular emphasis in the 2015 review (for reasons
similar to those summarized earlier), we additionally note that combining the exposure levels summarized for
areas across the U.S. where the current standard is met with the E-R functions established for 10 crop species
indicates a median RYL across crops to be at or below 5.1%, on average, with very few exceptions. Further,
estimates based on W126 index at the great majority of the areas are below 5%.
April 2022
4-101 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
incidence for the highest exposure level (W126 index above 25 ppm-hrs), a magnitude not seen
to occur in Class I area monitoring sites, or in virtually any sites nationwide, that meet the
current standard (Appendix 4C, section 4C.3). Further, we note a decline in frequency of peak
hourly concentrations, including those at/above 100 ppb, at U.S. monitoring sites over the past
15 years. The analyses of hourly concentrations summarized in section 4.4.1 above, also
demonstrate substantial control of peak 1-hour concentrations by the current standard. Thus, we
lack an established metric or combination of metrics that well describes the relationship between
occurrence and severity of visible foliar injury across a broad range of O3 concentration patterns
from those more common in the past to those in areas recently meeting today's standard, the
current information indicates air quality conditions of concern for this endpoint to generally
include cumulative seasonal exposures, in terms of seasonal single-year W126 index, at/above 25
ppm-hrs, in addition to appreciable occurrence of peak hourly concentrations at/above 100 ppb.
Based on this information, the available air quality information indicates that the exposure
conditions occurring at sites with air quality meeting the current standard are not those that might
reasonably be concluded to elicit the occurrence of significant foliar injury (with regard to
severity and extent).
• Are such exposures (in terms of W126 index) that occur in areas that meet the
current standard indicative of welfare effects reasonably judged important from a
public welfare perspective? What are important associated uncertainties?
Given the findings summarized in section 4.4 above regarding W126 index values in
areas where the current standard is met, we reflect on the potential public welfare significance of
vegetation-related effects that may be associated with such exposures. This consideration is
important to judgments regarding the secondary standard, which is not meant to protect against
all known or anticipated Cb-related welfare effects, but rather those that are judged to be adverse
to the public welfare (as noted in section 4.3.2 above). Accordingly, for the purposes of
informing that judgment, we consider here the exposures indicated to occur under conditions that
meet the current standard, the associated potential for effects and the potential public welfare
implications.
As an initial matter, we recognize the increased significance to the public welfare of
effects in areas that have been accorded special protection, such as Class I areas. In this context,
we note some general similarities of the exposure estimates in Class I areas for periods when the
current standard was met to such estimates at monitoring sites in other areas, as documented in
the larger air quality data analysis. Across both datasets, and extending back 21 years, the
cumulative exposure estimates, averaged over the design value period, for these air quality
conditions were virtually all at or below 17 ppm-hrs, with most of the W126 index values below
13 ppm-hrs (Appendix 4D, Table 4D-10), corresponding to median RBL estimates of 3.8% or
April 2022
4-102 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
less (based on the established tree seedling E-R relationships detailed in Appendix 4A). We
additionally note that single-year W126 index values in Class I areas over the 21-year dataset
evaluated were generally at or below 19 ppm-hrs, particularly in the more recent years
(Appendix 4D, section 4D.3.2.4). Regarding the potential for effects associated with commonly
occurring exposures, we consider first the categories of effects for which the quantitative
information related to exposure and associated effects is most well developed. In this
reconsideration, as in the 2015 review, these are effects on plant growth. Based on the median of
RBL estimates derived from the established E-R functions for 11 tree species seedlings, W126
index values at or below 17 ppm-hrs correspond to median species tree seedling RBL estimates
at or below 5.3% (Appendix 4A, Table 4A-5). Judgments in the 2015 review (in the context of
the framework considered in section 4.5.1.2 above) concluded isolated rare occurrences of
exposures for which median RBL estimates might be at or just above 6% to not be indicative of
conditions adverse to the public welfare, particularly considering the variability in the array of
environmental factors that can influence O3 effects in different systems, and the uncertainties
associated with estimates of effects in the natural environment.
In the 2015 review, the Administrator focused on cumulative exposure estimates derived
as the average W126 index over the 3-year design value period, concluding variations of single-
year W126 index from the average to be of little significance. This focus generally reflected the
judgment that estimates based on the average adequately, and appropriately, reflected the
precision of the current understanding of Cb-related growth reductions, given the various
limitations and uncertainties in such predictions. Additional analyses have been explored since
the 2015 to further examine this issue, as summarized in section 4.5.1.2 above. The current air
quality data indicate single-year W126 index values generally to vary by less than 5 ppm-hrs
from the 3-year average when the 3-year average is below 20 ppm-hrs (which is the case for
locations meeting the current standard). With such variation, year-to-year differences in tree
growth responding to each year's seasonal exposure from estimated response based on the 3-year
average of those seasonal exposures would, given the offsetting impacts of seasonal exposures
above and below the average, reasonably be expected to generally be small over tree lifetimes.
Additionally, we have also further considered the experimental data underlying the E-R
functions for estimating RBL, particularly those pertaining to cumulative exposures on the order
of 17 ppm-hrs and informing estimates of multiyear impacts. We note limitations in the evidence
base in these regards, as discussed further in section 4.5.1.2 above, that contribute to imprecision
or inexactitude to estimates of growth impacts associated with multi-year exposures in this range.
Further, the information available since 2015 does not appreciably address these limitations and
uncertainties to improve the certainty or precision in RBL estimates for such exposures.
April 2022
4-103 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
With regard to visible foliar injury, as discussed in sections 4.3.3.2 and 4.5.1.2 above, a
quantitative description of the relationship between O3 concentrations and visible foliar injury
extent or incidence, as well as severity, that would support estimation of injury under varying air
quality and environmental conditions (e.g., moisture), most particularly for locations that meet
the current standard is not yet established. In light of the potential role of peak O3 concentrations
(e.g., hourly concentrations at or above 100 ppb) as an influence on visible foliar injury
occurrence and severity (that may not be fully captured by a focus on cumulative seasonal O3
indices), we take note of analyses of peak concentrations summarized in section 4.4.1. These
indicate that the magnitude of daily maximum 1-hour concentrations has declined appreciably
since 2000. For example, the median annual 2nd highest MDA1 concentration across U.S. trend
monitoring sites declined by 27% from 2002 to 2013 (Figure 2-17 above), and the 99th percentile
MDA1 for all sites meeting the current standard in 2020 is below 80 ppb (Figure 4-11)). The
analysis in Appendix 2A of three recent design value periods (covering 2016 through 2020) and
three periods more than ten years prior (covering 2000 through 2004) show that the mean
number of observations per site at or above 100 ppb was well below one (0.22) for sites meeting
the current standards compared to well above one (10.04) for sites not meeting the current
standard. Further, the number of days with an hour at or above 100 ppb is below five at sites
meeting the current standard, and 99% are well below five (Figure 4-11, Appendix 2A, section
2A.2). These data and analyses indicate that the current standard provides appreciable control of
peak 1-hour concentrations, and thus, to the extent that such peak concentrations play a role in
the occurrence and severity of visible foliar injury, the current standard also provides appreciable
control.
In considering protection for visible foliar injury impacts provided by the standard, we
note, as discussed in section 4.3.2 above, that the public welfare implications associated with
visible foliar injury (when considered as an effect separate from effects on plant physiology)
relate largely to effects on scenic and aesthetic values. The available information does not yet
address or describe the relationships expected to exist for some level of visible foliar injury
severity (below that at which broader physiological effects on plant growth and survival might
also be expected) and/or extent of location or site injury (e.g., BI) scores with values held by the
public and associated impacts on public uses of the locations.112 As discussed in section 4.3.2
above, this gap limits our ability to identify air quality conditions that might be expected to
provide a specific level of protection from public welfare effects of this endpoint (e.g., separate
from effects that might relate to plant growth and reproduction under conditions where foliar
112 Information with some broadly conceptual similarity to this has been used forjudging public welfare implications
of visibility effects of PM in setting the PM secondary standard (78 FR 3086, January 15, 2012).
April 2022
4-104 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
injury may also be severe).113 Thus, key considerations of this endpoint in past reviews have
related to qualitative consideration of potential impacts related to the plant's aesthetic value in
protected forested areas and the somewhat general, nonspecific judgment that a more restrictive
standard is likely to provide increased protection. Nevertheless, while minor spotting on a few
leaves of a plant may easily be concluded to be of little public welfare significance, it is
reasonable to conclude that cases of widespread and relatively severe injury during the growing
season (particularly when sustained across multiple years and accompanied by obvious impacts
on the plant canopy) would likely impact the public welfare in scenic and/or recreational areas,
particularly in areas with special protection, such as Class I areas. In this context, we note the
potential usefulness of the USFS scheme for the purposes of informing the Administrator's
judgments with regard to public welfare significance of such effects.
In light of the discussions here and in sections 4.3.3.2 and 4.5.1.2 (with consideration of
presentations in Appendix 4C and air quality analyses in Appendices 2A, 4D and 4F) we find
that the available information does not indicate that a situation of widespread and relatively
severe visible foliar injury is likely associated with air quality that meets the current standard.
More specifically, the air quality data for areas meeting the standard do not indicate conditions
associated with BI scores reasonably considered of concern in the context described above
(concerning potential for public welfare significance). For example, we note that the air quality
analyses indicate that virtually all seasonal W126 index values at locations meeting the current
standard are below 25 ppm-hr. Further, the average number of observations of 1-hour
concentrations at or above 100 ppb per site and design value period are well below one during
periods when the current standard is met. Thus, while the current evidence is limited for the
purposes of identifying public welfare protection objectives related to visible foliar injury in
terms of specific air quality metrics, the current information indicates that the occurrence of
injury categorized as more severe than "little" by the USFS categorization (i.e., a BI score above
5 or above 15) would be expected to be infrequent in areas that meet the current standard. Based
on the USFS dataset presentations as well as the air quality analyses of W126 index values and
frequency of 1-hour observations at or above 100 ppb, the prevalence of injury scores
categorized as severe, which, depending on spatial extent, might contribute to impacts of public
welfare significance do not appear likely to occur under air quality conditions that meet the
current standard.
With regard to other vegetation-related effects, including those at the ecosystem scale,
such as alteration in community composition or reduced productivity in terrestrial ecosystems, as
113 Further, no criteria have been established regarding a level or prevalence of visible foliar injury considered to be
adverse to the affected vegetation as the current evidence does not provide for determination of a degree of leaf
injury that would have significance to the vigor of the whole plant (ISA, Appendix 8, p. 8-24).
April 2022
4-105 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
recognized in section 4.5.1.1, the available evidence is not clear with regard to the risk of such
impacts (and their magnitude or severity) associated with the environmental O3 exposures
estimated to occur under air quality conditions meeting the current standard (e.g., W126 index
generally at or below 17 ppm-hrs). In considering effects on crop yield, the air quality analyses at
monitoring locations that meet the current standard indicate estimates of RYL for such
conditions to be at and below 5.1%, based on the median estimate derived from the established
E-R functions for 10 crops (Appendix 4A, Table 4A-5). We additionally recognize there to be
complexities involved in interpreting the significance of such small estimates in light of the
factors identified in section 4.3.2 above. These include the extensive management of crops in
agricultural areas that may to some degree mitigate potential Cb-related effects, as well as the use
of variable management practices to achieve optimal yields, while taking into consideration
various environmental conditions. We also recognize that changes in yield of commercial crops
and commercial commodities may affect producers and consumers differently, further
complicating consideration of these effects in terms of potential adversity to the public welfare
impacts. In light of these factors complicating conclusions regarding crop yield impacts, in
combination with the relatively low RYL estimates associated with W126 index values occurring
in areas meeting the current standard, as well as the relative scarcity of peak hourly
concentrations at or above 100 ppb, a situation which differs from the extensive occurrences
associated with the exposure treatments on which the established E-R functions for the 10 crop
species are based (e.g., Lefohn and Foley, 1992), the current information does not indicate
exposures occurring in areas meeting the current standard to be of public welfare significance
with regard to crop yield.
4.5.2 Preliminary Conclusions
This section describes preliminary conclusions for the Administrator's consideration with
regard to the current secondary O3 standard. These preliminary conclusions are based on
consideration of the assessment and integrative synthesis of the evidence (as summarized in the
ISA, and the 2013 ISA and AQCDs from prior reviews), and the information on quantitative
exposure and air quality analyses summarized above. Taking into consideration the discussions
above in this chapter, this section addresses the following overarching policy question.
• Do the scientific evidence and air quality and exposure analyses support or call into
question the adequacy of the protection afforded by the current secondary O3
standard?
In considering this question, we first recognize what the CAA specifies with regard to
protection to be provided by the secondary standard. Under section 109(b)(2) of the CAA, a
secondary standard must "specify a level of air quality the attainment and maintenance of which,
April 2022
4-106 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
in the judgment of the Administrator, based on such criteria, is requisite to protect the public
welfare from any known or anticipated adverse effects associated with the presence of [the]
pollutant in the ambient air." Accordingly, as noted in section 4.3.2 above, the secondary
standard is meant to protect against Cb-related welfare effects that are judged to be adverse to the
public welfare (78 FR 8312, January 15, 2013; see also 73 FR 16496, March 27, 2008). Thus,
our consideration of the available information regarding welfare effects of O3 is in this context,
while recognizing that the level of protection from known or anticipated adverse effects to public
welfare that is requisite for the secondary standard is a public welfare policy judgment made by
the Administrator.
As is the case in NAAQS reviews in general, the extent to which the protection provided
by the current secondary O3 standard is judged to be adequate will depend on a variety of factors,
including science policy judgments and public welfare policy judgments. These factors include
public welfare policy judgments concerning the appropriate benchmarks on which to place
weight, as well as judgments on the public welfare significance of the effects that have been
observed at the exposures evaluated in the welfare effects evidence. The factors relevant to
judging the adequacy of the standard also include the interpretation of, and decisions as to the
weight to place on, different aspects of the quantitative analyses of air quality and cumulative O3
exposure and any associated uncertainties. Thus, we recognize that the Administrator's
conclusions regarding the adequacy of the current standard will depend in part on public welfare
policy judgments, science policy judgments regarding aspects of the evidence and exposure/risk
estimates, as well as judgments about the level of public welfare protection that is requisite under
the Clean Air Act.
As an initial matter, we recognize the continued support in the current evidence for O3 as
the indicator for photochemical oxidants (as summarized in section 4.5.1.1 above). We note that
no newly available evidence has been identified since the 2015 decision regarding the
importance of photochemical oxidants other than O3 with regard to abundance in ambient air,
and potential for welfare effects, and that, as stated in the current ISA, "the primary literature
evaluating the health and ecological effects of photochemical oxidants includes ozone almost
exclusively as an indicator of photochemical oxidants" (ISA, section IS. 1.1). Thus, we recognize
that, as was the case for the 2015 and prior reviews, the evidence base for welfare effects of
photochemical oxidants does not indicate an importance of any other photochemical oxidants.
Thus, we conclude that the evidence continues to support O3 as the indicator for the secondary
NAAQS for photochemical oxidants.
Our response to the overarching question above takes into consideration the discussions
that address the specific policy-relevant questions in prior sections of this document and the
approach described in section 4.2. We consider the evidence and the extent to which it alters key
April 2022
4-107 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
conclusions supporting the current standard. We also consider the quantitative analyses,
including associated limitations and uncertainties, and what they may indicate regarding level of
protection provided by the current standard from adverse effects. We additionally consider the
key aspects of the evidence and air quality/exposure information emphasized in establishing the
now-current standard, and the associated public welfare policy judgments and judgments about
inherent uncertainties that are integral to decisions on the adequacy of the current secondary O3
standard. Together these considerations contribute to our preliminary conclusion as to whether
the available scientific evidence and air quality and exposure analyses support or call into
question the adequacy of the protection afforded by the current secondary O3 standard.
In considering the available evidence, we recognize the longstanding evidence base of the
vegetation-related effects of O3, augmented in some aspects since the 2015 review. Consistent
with the evidence in the 2015 review, the existing evidence describes an array of effects on
vegetation and related ecosystem effects causally or likely causally related to O3 in ambient air,
as well as the causal relationship of tropospheric O3 with radiative forcing and subsequent likely
causally related effects on temperature, precipitation, and related climate variables. As was the
case in the 2015 review, a category of effects for which the evidence supports quantitative
description of relationships between air quality conditions and response is plant growth or yield.
The evidence base continues to indicate growth-related effects as sensitive welfare effects, with
the potential for ecosystem-scale ramifications. For this category of effects, there are established
E-R functions that relate cumulative seasonal exposure of varying magnitudes to various
incremental reductions in expected tree seedling growth (in terms of RBL) and in expected crop
yield (in terms of RYL). Many decades of research also recognize visible foliar injury as an
effect of O3, although uncertainties continue to hamper efforts to quantitatively characterize the
relationship of its occurrence and relative severity with O3 exposures. The evidence for these
categories of vegetation-related O3 effects is discussed further below. But before focusing further
on these key vegetation-related effects, we address two endpoints newly identified in the 2020
ISA, as well as tropospheric O3 effects related to climate.
With regard to categories of effects newly identified in the 2020 ISA as likely causally
related to O3 in ambient air, such as alteration of plant-insect signaling and insect herbivore
growth and reproduction, we recognize that uncertainties limit our consideration of the
protection that might be provided by the current standard against these effects. Depending on a
number of factors, such effects may have a potential for adverse effects to the public welfare,
e.g., given the role of plant-insect signaling in such important ecological processes as pollination
and seed dispersal, as well as natural plant defenses against predation and parasitism (as
discussed in section 4.3.2 above Uncertainties in the evidence, however, preclude a sufficient
understanding to support a focus on such effects in considering protection provided by the
April 2022
4-108 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
current standard. Areas of uncertainty and limitations in the evidence include key aspects of such
effects, the air quality conditions that might elicit them (and the magnitude or severity), the
potential for impacts in a natural ecosystem and, consequently, the potential for such impacts
under air quality conditions associated with meeting the current standard, as discussed in section
4.5.1.1 above. Thus, we do not find the evidence to provide sufficient information to support
judgments related to how particular patterns of O3 concentrations in ambient air may relate to the
occurrence of such effects in natural systems or, accordingly, to any related impacts to the public
welfare.
We next recognize the strong evidence documenting tropospheric O3 as a greenhouse gas
causally related to radiative forcing, and likely causally related to subsequent effects on variables
such as temperature and precipitation. In so doing, however, we take note of the limitations and
uncertainties in the evidence base that affect our ability to characterize the extent of any
relationships between O3 concentrations in ambient air in the U.S. and climate-related effects,
thus precluding a quantitative characterization of climate responses to changes in O3
concentrations in ambient air at regional (vs global) scales (as summarized in sections 4.3.3.3
and 4.3.4 above).114 As a result, we recognize the lack of important quantitative tools with which
to consider such effects in the context of protection provided by the current secondary O3
standard, such that it is not feasible to relate different patterns of O3 concentrations at the
regional (or national) scale in the U.S. with specific risks of alterations in temperature,
precipitation and other climate-related variables. We find these significant limitations and
uncertainties together to contribute to an insufficiency in the available information for the
purposes of supporting the Administrator's judgments particular to a secondary O3 NAAQS and
protection of the public welfare from adverse effects linked to O3 influence on radiative forcing,
and related climate effects.115 Thus, as is the case for the two newly identified categories of
insect-related effects discussed above, we conclude that the available evidence does not support a
focus on radiative forcing and related climate effects in considering the extent to which the
available evidence supports or calls into question the adequacy of protection afforded by the
current secondary standard.
Turning next to consideration of visible foliar injury, the available information has been
examined and analyzed as to what it indicates and supports with regard to adequacy of protection
114	With regard to radiative forcing and effects on temperature, precipitation, and related climate variables, while
additional characterizations have been completed since the 2015 review, uncertainties and limitations in the
evidence that were also recognized at that time remain.
115	Notwithstanding consideration of these effects, we note that a focus on the protection offered by the standard
against vegetation-related effects is expected to also have positive implications for climate change protection
through the protection of terrestrial ecosystem carbon storage.
April 2022
4-109 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
provided by the current standard (e.g., as discussed in section 4.5.1 above). Visible foliar injury
is an effect for which an association with O3 in ambient air is well documented. The public
welfare significance of visible foliar injury of vegetation in areas not closely managed for
harvest, particularly specially protected natural areas, has generally been considered in the
context of potential effects on aesthetic and recreational values, such as the aesthetic value of
scenic vistas in protected natural areas such as national parks and wilderness areas (e.g., 73 FR
16496, March 27, 2008). Accordingly, depending on its severity and spatial extent, as well as the
location(s) and the associated intended use, its effects on the physical appearance of the plant
have the potential to be significant to the public welfare. For example, while limited occurrences
(e.g., of severity of prevalence) may easily be concluded to be of little public welfare
significance, cases of widespread and relatively severe injury during the growing season
(particularly when sustained across multiple years and accompanied by obvious impacts on the
plant canopy) might reasonably be expected to have the potential to adversely impact the public
welfare in scenic and/or recreational areas, particularly in areas with special protection, such as
Class I areas.
In considering existing approaches for categorizing the severity of injury in natural areas,
we take note of the system developed by the USFS for its monitoring program116 to categorize BI
scores of visible foliar injury at biosites (sites with Cb-sensitive vegetation assessed for visible
foliar injury) in natural vegetated areas by severity levels (described in section 4.3.2 above). We
recognize, however, that quantitative analyses and evidence are lacking that might support a
precise conclusion - and associated judgment - as to a magnitude of BI score coupled with an
extent of occurrence that might be specifically identified as adverse to the public welfare. That
notwithstanding, we additionally note that the scale of the USFS biosite monitoring program's
objectives, which focus on natural settings in the U.S. and forests as opposed to individual
plants, may be informative to the Administrator with regard to his judgments concerning the
public welfare protection afforded by the current standard for such effects.
In considering the availability of established approaches that might be employed for
considering degrees of public welfare impacts related to the occurrence of visible foliar injury of
differing severity and extent (e.g., as summarized in sections 4.3.3.2 and 4.5.1.1 above), we note
the paucity of established approaches for interpreting specific levels of severity and extent of
foliar injury in protected forests with regard to impacts on public welfare effects (e.g., related to
116 During the period from 1994 (beginning in eastern U.S.) through 2011, the USFS conducted surveys of the
occurrence and severity of visible foliar injury on sensitive species at sites across most of the U.S. following a
national protocol (Smith, 2012).
April 2022
4-110 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
recreational services).117 In this context, we recognize a potential usefulness of the USFS system,
including its descriptors for BI scores of differing magnitudes intended for that Agency's
consideration in identifying areas of potential impact to forest resources. As described in section
4.3.2 above, very low BI scores (at or below 5) are described by the USFS scheme as "little or no
foliar injury" (Smith et al., 2007; Smith et al., 2012),118 and BI scores above 15 are categorized
as moderate to severe (and scores above 25 as severe). The lower categories of BI scores are
described by the USFS descriptions as indicative of injury of generally lesser risk to the natural
area, which we would suggest may also indicate lesser risk to public enjoyment. Accordingly, to
the extent that the USFS ranking system is of value to the Administrator's judgments in this
context, it may be reasonable to conclude that occurrence of BI scores categorized as "moderate
to severe" injury by the USFS scheme would be an indication of visible foliar injury occurrence
that, depending on extent and severity, may be indicative of conditions of public welfare
significance. Thus, this framework may be informative to the Administrator's consideration of
the evidence and analyses summarized in the sections above and what they indicate with regard
to patterns of air quality of concern for such an occurrence, and the extent to which they are
expected to occur in areas that meet the current standard.
We additionally consider the USFS biosite monitoring program studies of the occurrence,
extent, and severity of visible foliar injury in indicator species in defined plots or biosites in
natural areas across the U.S. Some of these studies, particularly those examining such data across
multiple years and multiple regions of the U.S., have reported that variation in cumulative O3
exposure, in terms of metrics such as SUM06 or W126 index, does not completely explain the
patterns of occurrence and severity of injury observed. Although the availability of detailed
analyses that have explored multiple exposure metrics and other influential variables is limited,
multiple studies have indicated a potential role for an additional metric, one related to the
occurrence of days with relatively high concentrations (e.g., number of days with a 1-hour
concentration at or above 100 ppb), as summarized in section 4.5.1.2 above. Also noteworthy are
the publications related to the USFS biosite monitoring program that provide extensive evidence
of trends across the past nearly 20 years that indicate reductions in severity of visible foliar
injury that parallel reductions in peak concentrations that have been suggested to be influential in
the severity of visible foliar injury. For example, observations of such reductions in the incidence
of the higher BI scores over the 16-year period of the program (1994 through 2010), especially
117	This contrasts with another welfare effect, visibility, for which there is evidence relating to levels of visibility
found to be acceptable by the public that was considered in judging the public welfare protection provided by the
particulate matter secondary standard (78 FR 3226-3228, January 15, 2013).
118	Studies that consider such data for purposes of identifying areas of potential impact to the forest resource suggest
this category corresponds to "none" with regard to "assumption of risk" (Smith et al., 2007; Smith et al., 2012).
April 2022
4-111 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
after 2002, have led to researcher conclusions of a "declining risk of probable impact" on the
monitored forests over this period (e.g., Smith, 2012). These reductions parallel the O3
concentration trend information nationwide that show clear reductions in cumulative seasonal
exposures, as well as in peak O3 concentrations, both in terms of 8-hour and hourly
concentrations (e.g., Figures 2-11 and 2-17, and as summarized in section 4.4.1 above). . That is,
the extensive air quality evidence of trends across the past nearly 20 years indicate reductions in
peak concentrations that some studies have suggested to be influential in the severity of visible
foliar injury, as discussed in section 4.5.1 above.
In considering the available information that might inform the Administrator's judgments
regarding visible foliar injury, we note a paucity of established approaches to inform the
Administrator's judgment of a magnitude, severity or extent of visible foliar injury related effects
appropriately concluded to be known or anticipated to cause adverse effects to the public
welfare. However, some general conclusions or observations may be supported. For example,
based on the available evidence and associated quantitative analyses, we have less confidence
and greater uncertainty in the existence of adverse public welfare effects with lower O3
exposures. More specifically, as discussed in the prior sections, the available information
suggests that O3 air quality associated with W126 index values below 25 ppm-hrs (in a single
year), particularly when in combination with infrequent occurrences of hourly concentrations at
or above 100 ppb, is not likely to pose a risk of visible foliar injury in natural areas of an extent
and severity that might reasonably be considered to be of public welfare significance.
Support for this conclusion is seen in the air quality analyses that inform our
understanding of the occurrence and magnitude of cumulative seasonal exposures, in terms of
W126 index, and peak concentrations, in terms of the N100 and D100 metrics, in areas that meet
the current standard. These analyses indicate that virtually all W126 index values in a single year
are below 25 ppm-hrs at all monitoring locations (including in or near Class I areas) where the
current standard is met, and that, in fact, such values above 19 ppm-hrs are rare, as summarized
in section 4.4.2 above (Appendix 4D, sections 4D.3.1.24 and 4D.3.2.4). Thus, the analyses of air
quality since 2000 for areas that meet the current standard do not indicate the occurrence of
cumulative seasonal exposure, in terms of W126 index, of a magnitude that might be expected,
based on the available information (e.g., based on analyses of BI scores considered in sections
4.5.1.2 and 4.5.1.3 above), to contribute to a significant extent and degree of injury or specific
impacts on recreational or related services for areas, such as wilderness areas or national parks.
Further, we take note of the uncommonness of days with any hours at or above 100 ppb at
monitoring sites that meet the current standard, as well as the minimal number of hours on any
such days (as summarized in section 4.4.1). Based on these considerations, it would appear that
the current standard provides control of air quality conditions that contribute to increased BI
April 2022
4-112 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
scores and to scores of a magnitude indicative of "moderate to severe" foliar injury. Thus, we
conclude that the evidence indicates that areas that meet the current standard are unlikely to have
BI scores reasonably considered to pose a risk of impacts of public welfare significance.
Accordingly, based on all of the considerations raised here, and in the sections above, we find it
reasonable to conclude that the available evidence and quantitative exposure information for
visible foliar injury do not call into question the adequacy of protection provided by the current
standard.
We turn now to consideration of the other vegetation-related effects, the evidence for
which as a whole is extensive, spans several decades, and supports the Agency's conclusions of
causal or likely to be causal relationship for O3 in ambient air with an array of effect categories
(as noted above). As an initial matter, we note the new ISA determination that the current
evidence is sufficient to infer likely causal relationships of O3 with increased tree mortality,
while also noting that the evidence does not indicate a potential for O3 concentrations that occur
in locations that meet the current standard to cause increased tree mortality, as summarized in
section 4.3.1 above.
As we turn our focus now to the more sensitive effect of vegetation growth and
conceptually related effects with a focus on RBL (described in section 4.5.1.2 above), we
recognize that public welfare policy judgments play an important role in decisions regarding a
secondary standard, just as public health policy judgments have important roles in primary
standard decisions. One type of public welfare policy judgment focuses on how to consider the
nature and magnitude of the array of uncertainties that are inherent in the scientific evidence and
analyses. These judgments are traditionally made with a recognition that current understanding
of the relationships between the presence of a pollutant in ambient air and associated welfare
effects is based on a broad body of information encompassing not only more established aspects
of the evidence but also aspects in which there may be substantial uncertainty. This may be true
even of the most robust aspect of the evidence base. In the case of the available evidence base, as
an example, we recognize increased uncertainty, and associated imprecision, at lower cumulative
exposures in application of the established and well-founded E-R functions, and in the current
understanding of aspects of relationships of such estimated effects with larger-scale impacts,
such as those on populations, communities, and ecosystems, as summarized in sections 4.5.1.3
and 4.3.4 above. Further, we recognize uncertainties in the details and quantitative aspects of
relationships between plant-level effects such as growth and reproduction, and ecosystem
impacts, the occurrence of which are influenced by many other ecosystem characteristics and
processes. These examples illustrate the role of public welfare policy judgments, both with
regard to the Administrator's consideration of the extent of protection that is requisite and
April 2022
4-113 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
concerning the weighing of uncertainties and limitations of the underlying evidence base and
associated quantitative analyses.
As summarized in section 4.1 above, the decisions that established the current standard in
2015, and retained it in 2020, involved a series of judgments contributing to the standard's
foundation with regard to growth-related effects. The first of these judgments relates to
consideration of the O3 effect of reduced growth (quantified using the metric, RBL) as a proxy
for an array of other vegetation-related effects to the public welfare. The category of effects for
which the evidence is most certain with regard to quantitative functions describing relationships
between O3 in ambient air and response continues to be reduced plant growth or yield. The
evidence base includes established E-R functions for seedlings of 11 tree species that relate
cumulative seasonal exposure of varying magnitudes to various incremental reductions in
expected tree seedling growth (in terms of RBL) and in expected crop yield. These functions are
well established and have been recognized across multiple O3 NAAQS reviews. Uncertainties
related to use of the RBL estimates include the limited information regarding the extent to which
they reflect growth impacts in mature trees, and the fact that the 11 species represent a very small
portion of the tree species across the U.S.
While recognizing these and other uncertainties, RBL estimates based on the median of
the 11 species were used in the 2015 and 2020 decisions as a surrogate for comparable
information on other species and lifestages, as well as a proxy or surrogate for other vegetation-
related effects, including larger-scale effects. Use of this approach continues to appear to be a
reasonable judgment in this reconsideration of the 2020 decision. More specifically, the currently
available information continues to support (and does not call into question) the consideration of
RBL as a useful and evidence-based approach for consideration of the extent of protection from
the broad array of vegetation-related effects associated with O3 in ambient air. As discussed in
section 4.5.1.2 above, these categories of effects include reduced vegetation growth,
reproduction, productivity, and carbon sequestration in terrestrial systems, and also alteration of
terrestrial community composition, belowground biogeochemical cycles, and ecosystem water
cycling. The current evidence base and available information (qualitative and quantitative), as in
the 2015 review, continue to support consideration of the potential for 03-related vegetation
impacts in terms of the RBL estimates from established E-R functions as a quantitative tool
within a larger framework of considerations pertaining to the public welfare significance of O3
effects. Such consideration would include effects that are associated with effects on vegetation,
and particularly those that conceptually relate to growth, and that are causally or likely causally
related to O3 in ambient air, yet for which there are greater uncertainties affecting estimates of
impacts on public welfare. This approach to weighing the available information in reaching
judgments regarding the secondary standard additionally takes into account uncertainties
April 2022
4-114 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
regarding the magnitude of growth impact that might be expected in mature trees, and of related,
broader, ecosystem-level effects for which the available tools for quantitative estimates are more
uncertain and those for which the policy foundation for consideration of public welfare impacts
is less well established. (80 FR 65389, October 26, 2015). The currently available evidence,
while somewhat expanded since the 2015 review, does not indicate an alternative metric for such
a use; nor is an alternative approach evident.
In considering tree growth effects, we take note of the other public welfare policy
judgments inherent in the Administrators' decisions in establishing the current standard in 2015,
and in retaining it in 2020. In addition to adoption of the median tree seedling RBL estimate for
the studied species as a surrogate for the broad array of vegetation related effects that extend to
the ecosystem scale, the decisions in 2015 and 2020 both incorporated the judgment that
cumulative seasonal exposures (in terms of the average W126 index across the 3-year design
period for the standard) associated with a median RBL somewhat below 6% is an appropriate
focus for considering target levels of protection for the secondary standard.
Decisions on the adequacy of secondary NAAQS require judgments on the extent to
which particular welfare effects (e.g., with regard to type, magnitude/severity, or extent) are
important from a public welfare perspective. In the case of O3, such a judgment includes
consideration of the public welfare significance of small magnitude estimates of RBL and
associated unquantified potential for larger-scale related effects. In establishing the current
standard in 2015 with a focus on RBL as a proxy or surrogate for the broad array of vegetation
effects, the Administrator took note of the 2014 CASAC characterization of 6% RBL (in
seedlings of median tree species). As described in section 4.1 above, the rationale provided by
the CASAC with this characterization was primarily conceptual and qualitative, rather than
quantitative. The conceptual characterization recognized linkages between effects at the plant
scale and broader ecosystem impacts, with the CASAC recommending that the Administrator
consider RBL as a surrogate or proxy for the broader impacts that could be elicited by O3. In the
2015 decision, the Administrator took note of this CASAC advice regarding use of RBL as a
proxy and set the standard with an underlying objective of limiting cumulative exposures (in
terms of W126 index, averaged over three years) "in nearly all instances to those for which the
median RBL estimate would be somewhat lower than 6%" (80 FR 65407, October 26, 2015).119
The information available in this reconsideration of the 2020 decision does not appear to call into
question such judgments, indicating them to continue to appear reasonable.
119 The 2015 decision additionally noted that "the Administrator does not judge RBL estimates associated with
marginal higher exposures [at or above 19 ppm-hrs] in isolated, rare instances to be indicative of adverse effects
to the public welfare" (80 FR 65407, October 26, 2015).
April 2022
4-115 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
In considering what the available information indicates regarding the level of protection
for growth-related effects provided by the current standard, we recognize the importance of
considering the extent of both cumulative seasonal O3 exposures and of elevated hourly
concentrations, as discussed in section 4.5.1.2 above. These aspects of O3 air quality can
contribute to damaging conditions for vegetation. Thus, in considering the extent of protection
provided by the current standard, in addition to considering seasonal W126 index to estimate
median RBL using the established E-R functions, we also consider metrics that convey
information regarding peak hourly concentrations. While we recognize that the evidence does
not indicate a particular threshold number of hours at or above 100 ppb (or another reference
point for elevated concentrations), we take particular note of the evidence of greater impacts
from higher concentrations (particularly with increased frequency) and of the air quality analyses
that document variability in such concentrations for the same W126 index value. In light of these
factors, a multipronged approach is reasonably concluded to be appropriate for considering
exposures of concern and the protection from them that may be afforded by the secondary
standard.
The air quality analyses summarized in section 4.4 above describe the air quality
conditions that occur under the current standard and also the conditions in areas where the
standard is not met. We consider what is indicated regarding protection overall and protection
against "unusually damaging years" (an issue raised in the court remand of the 2015 decision on
the secondary standard). With regard to this issue, we take note of the air quality analyses
summarized in section 4.4.1, as also considered in section 4.5.1.2 above, that investigate the
annual occurrence of elevated hourly O3 concentrations which may contribute to vegetation
exposures of concern (Appendix 2A, section 2A.2; Appendix 4F).120 These air quality analyses
illustrate limitations of the W126 index for purposes of controlling peak concentrations, and also
the strengths of the current standard in this regard, showing that the form and averaging time of
the existing standard is much more effective than the W126 index in limiting peak concentrations
(e.g., hourly O3 concentrations at or above 100 ppb) and in limiting number of days with any
such hours. As noted in prior sections, the W126 index, by virtue of its definition, does not
provide specificity with regard to year-to-year variability in elevated hourly O3 concentrations
with the potential to contribute to the increased risk of vegetation effects, and the air quality
analyses illustrate this limitation. These analyses additionally document the control exerted by
120 The ISA references the longstanding recognition of the risk posed to vegetation of peak hourly O3 concentrations
(e.g., "[h]igher concentrations appear to be more important than lower concentrations in eliciting a response"
[ISA, p. 8-180]; "higher hourly concentrations have greater effects on vegetation than lower concentrations"
[2013 ISA, p. 91-4] "studies published since the 2006 O3 AQCD do not change earlier conclusions, including the
importance of peak concentrations, ... in altering plant growth and yield" [2013 ISA, p. 9-117]).
April 2022
4-116 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
the current standard, through all of its elements, on both cumulative seasonal O3 exposures and
peak hourly concentrations.
In considering cumulative seasonal 03 exposures occurring in areas that meet the current
standard with regard to growth-related effects represented by RBL (as discussed more fully
earlier, including in section 4.5.1.2), we focus, as was done in the 2015 decision, on a seasonal
W126 index, averaged across three years. In do so based on consideration of the extent of
conceptual similarities of the 3-year average W126 index with some aspects of the derivation
approach for the established E-R functions, the context of RBL as a proxy (as recognized above)
and other factors. With regard to the established E-R functions used to describe the relationship
of RBL with O3 in terms of a seasonal W126 index, we recognize that the functions were derived
mathematically from studies of different exposure durations (varying from shorter than one to
multiple growing seasons) by applying adjustments so that they would yield estimates
normalized to the same period of time (season), such that the estimates may conceptually
represent average impact for a season. We note the compatibility of W126 index averaged over
multiple growing seasons or years with these adjustments. We also note that the exposure levels
represented in the data underlying the E-R functions are somewhat limited with regard to the
relatively lower cumulative exposure levels most commonly associated with the current standard
(e.g., at or below 17 ppm-hrs), with generally greater representation for higher exposures (e.g.,
ranging up to W126 index levels above 100 ppm-hrs), indicating additional uncertainty for
applications of the E-R functions to the lower cumulative exposure levels. We additionally note
the differing patterns of hourly concentrations of the elevated exposure levels (particularly with
regard to peak hourly concentrations, such as those at/above 100 ppb) in the datasets from which
the E-R functions from the patterns in ambient air meeting the current standard across the U.S.
today, as summarized in section 4.5.1.2 above. With these considerations regarding the E-R
functions and their underlying datasets in mind, we also take note of year-to-year variability of
factors other than O3 exposures that affect tree growth in the natural environment (e.g., related to
variability in soil moisture, meteorological, plant-related and other factors), that have the
potential to affect O3 E-R relationships, as noted in sections 4.3 and 4.5 above (ISA, Appendix 8,
section 3.12; 2013 ISA section 9.4.8.3). Thus, the use of the W126 index averaged over multiple
years has a compatibility with the approach used in deriving the E-R functions, and reflects
consideration of other aspects of the E-R function datasets and other factors that may affect
growth in the natural environment.
We additionally recognize the qualitative and conceptual nature of our understanding, in
many cases, of relationships of O3 effects on plant growth and productivity with larger-scale
impacts, such as those on populations, communities and ecosystems. Based on these
considerations, use of a seasonal RBL averaged over multiple years, such as a 3-year average,
April 2022
4-117 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
appears to be a reasonable approach, and provides a stable and well-founded RBL estimate for its
purposes as a proxy to represent the array of vegetation-related effects identified above. In light
of these considerations, we conclude there is support in the available information for use of an
average seasonal W126 index derived from multiple years (with their representation of
variability in environmental factors), and that the use of such averaging may provide an
appropriate representation of the evidence and attention to considerations summarized above.
Thus, we conclude that application of the multipronged approach referenced above would assess
anticipated exposures and protection afforded by the current secondary standard using a seasonal
W126 averaged over a 3-year period, which is the design value period for the current standard, to
estimate median RBL via the established E-R functions, in combination with a broader
consideration of air quality patterns, such as peak hourly concentrations.
In considering the quantitative analyses available in this review with regard to the control
of air quality conditions that might pose risks to the public welfare by the current standard, we
note the findings from the analysis of recent air quality at sites across the U.S., including in or
near 65 Class I areas, and also analyses of historical air quality. Findings from the analysis of the
air quality data from the most recent period and from the larger analysis of historical air quality
data extending back to 2000 are consistent with the air quality analysis findings that were part of
the basis for the current standard. That is, in virtually all design value periods and all locations at
which the current standard was met (more than 99.9% of the observations), the 3-year average
W126 metric was at or below 17 ppm-hrs, the target identified by the Administrator in
establishing the current standard and, in all such design value periods and locations, the W126
metric was at or below 19 ppm-hrs, as was also the case for the earlier and smaller dataset (80
FR 65404-65410, October 26, 2015). Additionally, across the full 21-year dataset for 56 Class I
areas with monitors meeting the current standard during at least one or as many as nineteen 3-
year periods since 2000, there are no more than 15 occurrences of a single-year W126 index
above 19 ppm-hrs, the majority occurring during the earlier years of the period (Appendix 4D,
section 4D.3.2.4, Tables 4D-14 and 4D-16). For example, the highest values were equal to 23
ppm-hrs, all occurring before 2012. Additionally, as emphasized in earlier sections, the current
standard better controls for peak concentrations (at or above 100 ppm-hrs), which may pose risks
of vegetation effects, than would be expected by either a single-year or three-year average
W126.121 Based on the evidence and air quality analyses described in sections 4.3 and 4.4 above,
as well as considerations summarized in section 4.5.1 above, the occurrences of 3-year average
W126 index values allowed by the current standard in Class I areas, including such infrequent
121 The historical dataset also shows the appreciable reductions in peak concentrations (via either the N100 or D100
metric) that have been achieved in the U.S. as air quality has improved under O3 standards of the existing form
and averaging time (Appendix 4F, Figures 4F-13 and 4F-14).
April 2022
4-118 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
single-year deviations of the magnitude recognized here, above the average, can reasonably be
concluded not to raise concerns of adverse effects on the public welfare.
With regard to O3 effects on crop yield, we take note of the long-standing evidence,
qualitative and quantitative, of the reducing effect of O3 on the yield of many crops, as
summarized in the ISA and characterized in detail in past reviews (e.g., 2013 ISA, 2006 AQCD,
1997 AQCD, 2014 WREA). We also note the established E-R functions for 10 crops and the
estimates of RYL derived from them (Appendix 4A, section 4A.1, Table 4A-4), and the potential
public welfare significance of reductions in crop yield, as summarized in section 4.3.2 above. We
additionally recognize, however, that not every effect on crop yield will be judged adverse to
public welfare. In the case of crops in particular there are a number of complexities related to the
heavy management of many crops to obtain a particular output for commercial purposes, and
related to other factors, that are relevant to consider in evaluating potential Cb-related public
welfare impacts, as summarized in sections 4.3.2 and 4.5.1.3). For example, the extensive
management of agricultural crops that occurs to elicit optimum yields (e.g., through irrigation
and usage of soil amendments, such as fertilizer) is relevant to judgments concerning evaluation
of the extent of RYL estimated from experimental O3 exposures reasonably considered to be
adverse to the public welfare. Such considerations include opportunities in crop management for
market objectives, as well as complications in judging relative adversity that relate to market
responses and their effects on producers and consumers in evaluating the potential impact on
public welfare of estimated crop yield losses.
In light of such complexities, uncertainties, and limitations, we have considered how
RYL estimates relate to RBL estimates identified above for evaluating protection provided by
the current standard. In this context, we note that W126 index values (3-year average) were at or
below 17 ppm-hrs in virtually all monitoring sites with air quality meeting the current standard.
Based on the established E-R functions, the median RYL estimate corresponding to 17 ppm-hrs
is 5.1%. In considering single-year index values, as discussed in section 4.4.2 above, the vast
majority are similarly low (with more than 99% less than or equal to 17 ppm-hrs), and the higher
values predominantly occur in urban areas. We additionally take note of the role of elevated
hourly concentrations in effects on vegetation growth and yield. In this context we also note the
extensive management of agricultural crops, and the complexities associated with identifying
adverse public welfare effects for market-traded goods (where producers and consumers may be
impacted differently). We also recognize that the current standard generally maintains air quality
at a W126 index below 17 ppm-hrs, with few exceptions, and accordingly would limit the
estimated RYL (based on experimental O3 exposures) to this degree. In light of all of these
factors, we do not find the available information to call into question the adequacy of protection
afforded by the current standard for crop yield-related effects.
April 2022
4-119 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
Thus, the available information leads us to conclude that the combined consideration of
the body of evidence and the quantitative air quality and exposure analyses, including associated
uncertainties, does not call into question the adequacy of the protection provided by the current
secondary standard. Rather, this information provides support for the current standard, and thus
supports consideration of retaining the current standard, without revision. In reaching these
conclusions, we recognize that the Administrator's decisions in secondary standard reviews, in
general, are largely public welfare judgments, as described above. We further note that different
public welfare policy judgments (e.g., from those in both 2020 and 2015) could lead to different
conclusions regarding the extent to which the current standard provides the requisite protection
of the public welfare. Such public welfare judgments include those related to the appropriate
level of protection that should be afforded to protect against vegetation-related effects of public
welfare significance, as well as with regard to the appropriate weight to be given to differing
aspects of the evidence and air quality information, and how to consider their associated
uncertainties and limitations. For example, different judgments might give greater weight to
more uncertain aspects of the evidence or reflect a differing view with regard to public welfare
significance. Such judgments are left to the discretion of the Administrator. We note, however,
that the scientific evidence and quantitative air quality, exposure and risk information in the
record on which this reconsideration is based are largely unchanged. Staff conclusions regarding
the adequacy of the current standards thus remain unchanged from those reached in the 2020 PA.
In summary, the evidence characterized in the 2020 ISA is consistent with that available
in the 2015 review for the principal effects for which the evidence is strongest (e.g., plant
growth, reproduction, and related larger-scale effects, as well as visible foliar injury) and for key
aspects of the current standard. The evidence regarding RBL and air quality in areas meeting the
current standard does not appear to call into question the adequacy of public welfare protection
afforded by the standard. With regard to visible foliar injury, the currently available evidence for
forested locations across the U.S., such as studies of USFS biosites, does not indicate an
incidence of significant visible foliar injury that might reasonably be concluded to be adverse to
the public welfare under air quality conditions meeting the current standard. For the insect-
related effects that the ISA newly concludes likely to be causally related to O3, the new
information does not support an understanding of the potential for the occurrence of such effects
in areas that meet the current standard to an extent that they might reasonably be judged
significant to public welfare. Thus, we do not find the current information for these newly
identified categories to call into question the adequacy of the current standard. Similarly, key
uncertainties recognized in the 2015 review remain in the evidence for O3 contribution to
radiative forcing or effects on temperature, precipitation and related climate variables, including
specifically uncertainties that limit quantitative evaluations that might inform consideration of
April 2022
4-120 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
these effects (as discussed above). Based on all of the above considerations, we conclude that the
currently available evidence and quantitative exposure/risk information does not call into
question the protection afforded by the current secondary standard, such that it is appropriate to
consider retaining the current standard without revision. In light of this conclusion, we have not
identified any potential alternative standards for consideration.
4.6 KEY UNCERTAINTIES AND AREAS FOR FUTURE RESEARCH
In this section, we highlight key uncertainties associated with reviewing and establishing
the secondary O3 standard and additionally recognize that research in these areas may
additionally be informative to the development of more efficient and effective control strategies.
The list in this section includes key uncertainties and data gaps thus far highlighted in this review
of the secondary standard. Additional information in several areas would reduce uncertainty in
our interpretation of the available information and, accordingly, reduce uncertainty in our
characterization of Cb-related welfare effects. For example, the items listed below generally
include uncertainties associated with the extrapolation to plant species and environments outside
of specific experimental or field study conditions and the assessment of ecosystem-scale impacts,
such as structure and function. Additional E-R studies in different species or for responses other
than reduced growth over multiple exposure conditions over growing seasons, that include
details on exposure circumstances (e.g., hourly concentrations throughout the exposure), and
exposure history, etc. would improve on and potentially expand characterizations of the potential
for and magnitude of the identified vegetation effects under different seasonal exposures.
Accordingly, in this section, we highlight areas for future welfare effects research, model
development, and data collection activities to address these uncertainties and limitations in the
current scientific evidence. These areas are similar to those highlighted in past reviews.
• While national visible foliar injury surveys have provided an extensive dataset on the
incidence of such effects at sites across the country that experienced differing cumulative
seasonal O3 exposures and soil moisture conditions, there remain uncertainties in the
current understanding of the relationship between seasonal O3 exposures (and other
influential factors, such as relative soil moisture) and the incidence and relative severity
of visible foliar injury. Further research investigating the role of peak concentrations, in
addition to cumulative seasonal exposures (particularly for W126 index values below 25
ppm ) is also needed to improve consideration of the occurrence and variability of higher
hourly O3 concentrations associated with vegetation effects. Research to better
characterize the relationship between O3, soil moisture and foliar injury and specifically a
quantifiable relationship between these (and any other influential) factors. Additionally,
research would assist in interpreting connections between 03-related foliar injury and
other physiological effects and ecosystem services. For example, research is needed on
the extent and severity of visible foliar injury that might impact ecosystem services (e.g.,
tourism), and the extent of impact it might have.
April 2022
4-121 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
•	Additional controlled exposure studies of effects, such as biomass impacts, that include
multiple exposure levels within the lower range of exposures associated with ambient air
quality conditions common today, extend over multiple years, and include the collection
of detailed O3 concentration data over the exposure would reduce uncertainty in estimates
of effects across multiple-year periods and at the O3 exposures common today. Also
needed is evaluation of such datasets with regard to the role of peak concentrations in
combination with that of cumulative seasonal exposures (e.g., as quantified by metrics
such as the W126 and SUM06 indices).
•	Evidence newly available since the 2015 review includes studies on insect-plant
interactions that have established some statistically significant effects, but the evidence is
still limited with regard to discerning a pattern of responses in growth, reproduction, or
mortality, and a directionality of responses for most effects. More research is needed to
investigate the degree of response and directionalities of these relationships, and to
investigate potential effects on pollination. The evidence is also limited with regard to the
species represented (i.e., currently confined to three insect orders).
•	Some evidence provides for linkages of effects on tree seedlings with larger trees and
similarities in results between exposure techniques. Uncertainties remain in this area as
well as uncertainties in extrapolating from O3 effects on young trees (e.g., seedlings
through a few years of age) to mature trees and from trees grown in the open versus
those within the forest canopy.
•	Uncertainties that remain in extrapolating individual plant response spatially or to higher
levels of biological organization, including ecosystems, could be informed by research
that explores and better quantifies the nature of the relationship between O3, plant
response and multiple biotic and abiotic stressors, including those associated with the
affected ecosystem services (e.g., hydrology, productivity, carbon sequestration).
•	Other uncertainties are associated with estimates of the effects of O3 on the ecosystem
processes of water, carbon, and nutrient cycling, particularly at the stand and community
levels. These below- and above-ground processes include interactions of roots with the
soil or microorganisms, effects of O3 on structural or functional components of soil food
webs and potential impacts on plant species diversity, changes in the water use of
sensitive trees, and if the sensitive tree species is dominant, potential changes to the
hydrologic cycle at the watershed and landscape level. Research on competitive
interactions under different O3 exposures and any associated impacts on biodiversity or
genetic diversity would improve current understanding.
•	Uncertainties related to characterizing the potential public welfare significance of O3-
induced effects and impacts to associated ecosystem services could also be informed by
research. Research relating effects such as those on plant reproduction and propagation to
effects on production of non-timber forest products, and research to characterize public
preferences including valuation related to non-use and recreation for foliar injury, could
also help inform consideration of the public welfare significance of these effects.
April 2022
4-122 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
REFERENCES
Black, VJ, Stewart, CA, Roberts, JA and Black, CR (2012). Timing of exposure to ozone affects
reproductive sensitivity and compensatory ability in Brassica campestris. Environ Exp
Bot 75: 225-234.
Campbell, SJ, Wanek, R and Coulston, JW (2007). Ozone injury in west coast forests: 6 years of
monitoring - Introduction. U.S. Department of Agriculture. Portland, OR.
Cordell, H, Betz, FM, Mou, S and Green, G (2008). How do Americans View Wilderness. A
WILDERNESS Research Report in the Internet Research Information Series. Natinoal
Survey on Recreation and the Environment. This research is a collaborative effort
between the U.S. Department of Agriculture Forest Service's Southern Research Station
and its Forestry Sciences Laboratory in Athens, Georgia; the University of Georgia in
Athens; and the University in Tennessee in Knoxville, Tennessee. .
Costanza, R, De Groot, R, Braat, L, Kubiszewski, I, Fioramonti, L, Sutton, P, Farber, S and
Grasso, M (2017). Twenty years of ecosystem services: How far have we come and how
far do we still need to go? Ecosyst Serv 28: 1-16.
Cox, LA. (2020). Letter from Louis Anthony Cox, Jr., Chair, Clean Air Scientific Advisory
Committee, to Administrator Andrew R. Wheeler. Re:CASAC Review of the EPA's
Policy Assessment for the Review of the Ozone National Ambient Air Quality Standards
(ExternalReview Draft - October 2019). February 19, 2020. EPA-CASAC-20-003.
Available at:
https://yosemite. epa.gov/sab/sabproduct.nsf/264cb122 7d55e02c85257402007446a4/4 713
D217BC07103485258515006359BA/$File/EPA-CASAC-20-003.pdf.
Darbah, INT, Kubiske, ME, Neilson, N, Oksanen, E, Vaapavuori, E and Karnosky, DF (2007).
Impacts of elevated atmospheric C02 and 03 on paper birch (Betula papyrifera):
Reproductive fitness. ScientificWorldJournal 7: 240-246.
Darbah, INT, Kubiske, ME, Nelson, N, Oksanen, E, Vapaavuori, E and Kamosky, DF (2008).
Effects of decadal exposure to interacting elevated C02 and/or 0-3 on paper birch
(Betula papyrifera) reproduction. Environ Pollut 155(3): 446-452.
Davis, DD and Orendovici, T (2006). Incidence of ozone symptoms on vegetation within a
National Wildlife Refuge in New Jersey, USA. Environ Pollut 143(3): 555-564.
Diaz-de-Quijano, M, Kefauver, S, Ogaya, R, Vollenweider, P, Ribas, A and Penuelas, J (2016).
Visible ozone-like injury, defoliation, and mortality in two Pinus uncinata stands in the
Catalan Pyrenees (NE Spain). Eur J Forest Res 135(4): 687-696.
Dietze, MC and Moorcroft, PR (2011). Supplemental information 1: Tree mortality in the eastern
and central United States: Patterns and drivers. Global Change Biol 17.
Duffney, PF, Brown, JS, and Stone, SL (2022). Memorandum to the Review of the Ozone
National Ambient Air Quality Standards (NAAQS) Docket (EPA-HQ-ORD-2018-
April 2022
4-123 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
0279). Re: Provisional Evaluation of Newly Identified Controlled Human Exposure
Studies in the context of the 2020 Integrated Science Assessment for Ozone and Related
Photochemical Oxidants. April 15, 2020.
Frey, HC. (2014). Letter from Dr. H. Christopher Frey, Chair, Clean Air Scientific Advisory
Committee to Honorable Gina McCarthy, Administrator, US EPA. Re: CASAC Review
of the EPA's Second Draft Policy Assessment for the Review of the Ozone National
Ambient Air Quality Standards. June 26, 2014. EPA-CASAC-14-004. Available at:
https://nepis. epa.gov/Exe/ZyPURL. cgi?Dockey=P 100JR6F. Ixl.
Haefele, M, Kramer, RA and Holmes, TP (1991). Estimating the Total Value of a Forest Quality
in High-Elevation Spruce-Fir Forests. The Economic Value of Wilderness: Proceedings
of the Conference, Southeastern For Exper. Station. Asheville, NC, USD A Forest
Service.
Heck, WW and Cowling, EB (1997). The need for a long term cumulative secondary ozone
standard - An ecological perspective. Environ Manager January: 23-33.
Henderson, R. (2006). Letter from Dr. Rogene Henderson, Chair, Clean Air Scientific Advisory
Committee to Honorable Stephen L. Johnson, Administrator, US EPA. Re: CASAC Peer
Review of the Agency's 2nd Draft Ozone Staff Paper October 24, 2006. EPA-CASAC-
07-001. Available at: https://nepis.epa.gov/Exe/ZyPURL.cgi?Dockey=P1000W07.txt.
Hildebrand, E, Skelly, JM and Fredericksen, TS (1996). Foliar response of ozone-sensitive
hardwood tree species from 1991 to 1993 in the Shenandoah National Park, Virginia. Can
J For Res 26(4): 658-669.
Hogsett, WE, Weber, JE, Tingey, D, Herstrom, A, Lee, EH and Laurence, JA (1997).
Environmental auditing: An approach for characterizing tropospheric ozone risk to
forests. J Environ Manage 21(1): 105-120.
King, JS, Kubiske, ME, Pregitzer, KS, Hendrey, GR, McDonald, EP, Giardina, CP, Quinn, VS
and Karnosky, DF (2005). Tropospheric 03 compromises net primary production in
young stands of trembling aspen, paper birch and sugar maple in response to elevated
atmospheric C02. NewPhytol 168(3): 623-635.
Kohut, R (2007). Handbook for Assessment of Foliar Ozone Injury on Vegetation in the National
Parks: Revised Second Edition. Kohut, R.
Kubiske, ME, Quinn, VS, Heilman, WE, McDonald, EP, Marquardt, PE, Teclaw, RM, Friend,
AL and Karnoskey, DF (2006). Interannual climatic variation mediates elevated C02 and
03 effects on forest growth. Global Change Biol 12(6): 1054-1068.
Kubiske, ME, Quinn, VS, Marquardt, PE and Karnosky, DF (2007). Effects of elevated
atmospheric C02 and/or 03 on intra- and interspecific competitive ability of aspen. Plant
Biol 9(2): 342-355.
April 2022
4-124 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
Landesmann, JB, Gundel, PE, Martinez-Ghersa, MA and Ghersa, CM (2013). Ozone exposure of
a weed community produces adaptive changes in seed populations of Spergula arvensis.
PLoS ONE 8(9): e75820.
Lee, EH and Hogsett, WE (1996). Methodology for calculating inputs for ozone secondary
standard benefits analysis part II. Office of Air Quality Planning and Standards. Research
Triangle Park, NC.
Lefohn, A, Shadwick, D, Somerville, M, Chappelka, A, Lockaby, B and Meldahl, R (1992). The
characterization and comparison of ozone exposure indices used in assessing the response
of loblolly pine to ozone. Atmospheric Environment, Part A: General Topics 26(2): 287-
298.
Lefohn, AS, Jackson, W, Shadwick, DS and Knudsen, HP (1997). Effect of surface ozone
exposures on vegetation grown in the southern Appalachian Mountains: Identification of
possible areas of concern. Atmos Environ 31(11): 1695-1708.
Luben, T, Lassiter, M and Herrick, J (2020). Memorandum to Ozone NAAQS Review Docket
(EPA-HQ-ORD-2018-0279). RE: List of Studies Identified by Public Commenters That
Have Been Provisionally Considered in the Context of the Conclusions of the 2020
Integrated Science Assessment for Ozone and Related Photochemical Oxidants.
December 2020. Docket Document ID: EPA-HQ-OAR-2018-0279-0560.
Moran, EV and Kubiske, ME (2013). Can elevated C02 and ozone shift the genetic composition
of aspen (Populus tremuloides) stands? New Phytol 198(2): 466-475.
Myhre, G, Shindell, D, Breon, FM, Collins, W, Fuglestvedt, J, Huang, J, Koch, D, Lamarque, JF,
Lee, D, Mendoza, B, Nakajima, T, Robock, A, Stephens, G, Takemura, T and Zhang, H,
Eds. (2013). Anthropogenic and natural radiative forcing. Cambridge University Press
Cambridge, UK.
Oksanen, E and Holopainen, T (2001). Responses of two birch (Betula pendula Roth) clones to
different ozone profiles with similar AOT40 exposure. Atmos Environ 35(31): 5245-
5254.
Rosenberger, RS, Bell, LA, Champ, PA and White, EM (2013). Estimating the economic value
of recreation losses in Rocky Mountain National Park due to a mountain pine beetle
outbreak. Western Economics Forum 12(1): 31-39.
Samet, JM. (2010). Letter from Jonathan Samet, Chair, Clean Air Scientific Advisory
Committee, to Administrator Lisa Jackson. Re: CAS AC Review of EPA's Proposed
Ozone National Ambient Air Quality Standard (Federal Register, Vol. 75, Nov. 11,
January 19, 2010). . February 19, 2010. EPA-CASAC-10-007. Available at:
https:/fnepis.epa.gov/Exe/ZyPURL.cgi?Dockey=P10072Tl.txt.
Smith, G (2012). Ambient ozone injury to forest plants in Northeast and North Central USA: 16
years of biomonitoring. Environ Monit Assess(184): 4049-4065.
April 2022
4-125 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
Smith, G, Coulston, J, Jepsen, E and Prichard, T (2003). A national ozone biomonitoring
program: Results from field surveys of ozone sensitive plants in northeastern forests
(1994-2000). Environ Monit Assess 87(3): 271-291.
Smith, GC, Morin, RS and McCaskill, GL (2012). Ozone injury to forests across the Northeast
and North Central United States, 1994-2010. General Technical Report NRS-103. United
States Department of Agriculture, US Forest Service, Northern Research Station.
Smith, GC, Smith, WD and Coulston, JW (2007). Ozone bioindicator sampling and estimation.
General Technical Report NRS-20. United States Department of Agriculture, US Forest
Service, Northern Research Station.
Smith, JT and Murphy, D. (2015). Memorandum to Ozone NAAQS Review Docket (EPA-HQ-
OAR-2008-0699). Additional Observations from WREA Datasets for Visible Foliar
Injury. September 24, 2015. Docket Document ID EPA-HQ-OAR-2008-0699-4250.
Available at: https://www.regulations.gov/document/EPA-HQ-OAR-2008-0699-4250.
U.S. DHEW (1970). Air Quality Criteria for Photochemical Oxidants. National Air Pollution
Control Administration, . Washington, DC. U.S. DHEW. publication no. AP-63. NTIS,
Springfield, VA; PB-190262/BA.
U.S. EPA (1996). Air Quality Criteria for Ozone and Related Photochemical Oxidants. Volume I
- III. Office of Research and Development Research Triangle Park, NC. U.S. EPA. EPA-
600/P-93-004aF, EPA-600/P-93-004bF, EPA-600/P-93-004cF. July 1996. Available at:
https://nepis. epa.gov/Exe/ZyPURL. cgi?Dockey=300026GN. txt
https://nepis. epa.gov/Exe/ZyPURL. cgi?Dockey=300026SH. Ixl
https://nepis.epa.gov/Exe/ZyPURL.cgi?Dockey=10004RHL. txt.
U.S. EPA (2007). Review of the National Ambient Air Quality Standards for Ozone: Policy
Assessment of Scientific and Technical Information: OAQPS Staff Paper. Office of Air
Quality Planning and Standards. Research Triangle Park, NC. U.S. EPA. EPA-452/R-07-
003. January 2007. Available at:
https://nepis. epa.gov/Exe/ZyPURL.cgi?Dockey=P10083 VX. txt.
U.S. EPA (2018) Review of the Secondary Standards for Ecological Effects of Oxides of
Nitrogen, Oxides of Sulfur, and Particulate Matter: Risk and Exposure Assessment
Planning Document. Office of Air Quality Planning and Standards, Health and
Environmental Impacts Division. Research Triangle Park, N.C. EPA-452/D-18-001.
Available at: https://www. epa.gov/naaqs/nitrogen-dioxide-no2-and-sulfur-dioxide-so2-
secondary-standards-planning-documents-current.
U.S. Forest Service, NPS, and U.S. Fish and Wildlife Service (2010). Federal land managers' air
quality related values work group (FLAG): phase I report—revised (2010). National Park
Service, Denver, CO.
USFS (2013). Forest Inventory and Analysis: Fiscal Year 2012 Business Report. United States
Department of Agriculture, http://www.fia.fs.fed.us/library/bus-org-
documents/docs/FIA Annual Report 2013.pdf.
April 2022
4-126 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
USFS (2017). Forest Inventory and Analysis: Fiscal Year 2016 Business Report. United State
Department of Agriculutre.
https://www.fs.fed.us/sites/default/files/fs media/fs document/publication-15817-usda-
forest-service-fia-annual-report-508.pdf.
van Goethem, TM, Azevedo, LB, van Zelm, R, Hayes, F, Ashmore, MR and Huijbregts, MA
(2013). Plant species sensitivity distributions for ozone exposure. Environ Pollut 178: 1-
6.
Wang, P, Baines, A, Lavine, M and Smith, G (2012). Modelling ozone injury to U.S. forests.
Environ Ecol Stat 19(4): 461-472.
Wells, B. (2015). Memorandum to Ozone NAAQS Review Docket (EPA-HQ-OAR-2008-0699).
Expanded Comparison of Ozone Metrics Considered in the Current NAAQS Review.
September 28, 2015. Docket Document Identifier EPA-HQ-OAR-2008-0699-0163.
Available at: https://www.regulations.gov/contentStrearner?documentId=EPA-HQ-OAR-
2008-0699-4325&contentType=pdf.
Wells, B (2020). Memorandum to Ozone NAAQS Review Docket (EPA-HQ-OAR-2018-0279).
Additional Analyses of Ozone Metrics Related to Consideration of the Ozone Secondary
Standard. December 2020. Docket Document Identifier EPA-HQ-OAR-2018-0279-0557.
Yun, S-C and Laurence, JA (1999). The response of clones of Populus tremuloides differing in
sensitivity to ozone in the field. New Phytol 141(3): 411-421.
April 2022
4-127 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
APPENDIX 2A
ADDITIONAL DETAILS ON DATA ANALYSIS
PRESENTED IN SECTION 2.4
TABLE OF CONTENTS
2A. 1 Analyses of 8-Hour Concentrations	2A-2
2A.2 Analyses of 1-Hour Concentrations	2A-3
TABLE OF FIGURES
Figure 2A-1. Boxplots comparing the distribution of MDA1 concentrations for 2000-2004
(red) to the distribution of MDA1 concentrations for 2016-2020 (blue), binned
by the 8-hour design value at each monitoring site. The boxes represent the 25th,
50th and 75th percentiles and the whiskers represent the 1st and 99th percentiles.
Outlier values are represented by circles	2A-5
Figure 2A-2. Map showing the average number of days with MDA1 >100 ppb,
2000-2004	2A-7
Figure 2A-3. Map showing the average number of days with MDA1 >100 ppb,
2016-2020	2A-7
Figure 2A-4. Number of days in 2018-2020 at each monitoring site with a MDA1
concentration greater than or equal to 100 ppb and an 8-hour design value less
than 98 ppb. Sites with higher design values had more days, up to a maximum
of 164 (at a site in southern CA)	2A-8
TABLE OF TABLES
Table 2A-1. Summary of criteria describing the sites for which 8-hour metrics are presented
in section 2.4 of main document	2A-3
Table 2A-2. Summary statistics for MDA1 concentrations at sites with differing design
values for 2018-2020	2A-4
Table 2A-3. Summary statistics for MDA1 concentrations at differing design values for
2000-2004	2A-6
Table 2A-4. Summary statistics for MDA1 concentrations at differing design values for
2016-2020	2A-6
April 2022
2A-1 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
2A.1 ANALYSES OF 8-HOUR CONCENTRATIONS
The analyses presented in section 2.4 of the main document are based on hourly O3
concentration data from the EPA's Air Quality System (AQS) database (retrieved on August 12,
2021) for the years 2000 to 2020 for the sites meeting data completeness criteria as summarized
in Table 2A-1 below. The daily maximum 8-hour (hr) average (MDA8) values, annual fourth
highest MDA8 values, and design values (DVs) for the current standards were calculated
according to Appendix U to 40 CFR Part 50. Those steps are generally as follows.
8-hr average concentrations are derived as the average of concentrations during eight
consecutive hours for the:
o 8-hr periods which have at least six hourly concentrations;1 and
o 8-hr periods which have fewer than six hourly concentrations and the sum of
concentrations divided by eight, after truncation of the digits after the third
decimal place, is greater than 0.070 parts per million (ppm)
The digits for the resultant 8-hr average concentration are truncated after the third
decimal place.
MDA8 concentrations are derived as the highest of the consecutive 8-hr averages
beginning with the 8-hr period from 7am to 3pm and ending with the period from
11pm to 7am the following day for those days with:
o 8-hr concentrations for at least 13 of the 17 8-hr periods that begin with the
7am-to-3pm period and end with the 1 lpm-to-7am (next day) period, or
o 8-hr concentrations for fewer than 13 of the 17 8-hr periods if the maximum
8-hr concentration, after truncation of the digits after the third decimal place,
is greater than 0.070 ppm.
Design Values in ppm are derived as average of the annual 4th highest MDA8
concentrations in three consecutive years, with digits after the third decimal place
truncated.
o Design values greater than 0.070 ppm are always considered valid,
o Design values less than or equal to 0.070 ppm must have MDA8 values for at
least 90% of the days in the ozone monitoring season3, on average over the 3-
year period, with a minimum of 75% of those days in any individual year.
1	When there are at least six hours with a concentration reported, the 8-hr average is the average calculated using the
number of hours with concentrations in the denominator.
2	When there are fewer than six hours with a concentration reported, the 8-hr average is the average calculated using
eight in the denominator and substituting zero for the missing hourly concentrations.
3	Ozone monitoring seasons are defined for each State in Table D-2 of Appendix D to 40 CFR Part 58.
April 2022
2A-2 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
Table 2A-1. Summary of criteria describing the sites for which 8-hour metrics are
presented in section 2.4 of main document.
Presentation of 8-hour
metrics in section 2.4
Time
Period
Data included
Figure 2-8, DVs
2018-2020
Design values are presented for all sites with valid design values,
which are sites having at least 75% data completeness in each of the
three years and at least 90% completeness on average across the
three years (per Appendix U)
Figure 2-9, DVs
2000-2020
Figure 2-10, Trends
1980-2020
Annual fourth highest MDA8 values are based on all sites with at least
75% annual data completeness for at least 31 of the 41 years, with no
more than two consecutive years having less than 75% complete data
(n = 188 sites)
Figure 2-11, Trends
2000-2020
Annual fourth highest MDA8 values are based on all sites with at least
75% annual data completeness for at least 16 of the 21 years, with no
more than two consecutive years having less than 75% complete data
(n = 822 sites)
Design values are presented for sites with valid DVs for at least 15 of
the 19 3-year periods, with no more than two consecutive periods
having invalid DVs (n = 658 sites)
Figure 2-12, Trends
2000-2020
Figure 2-13, Diurnal
Patterns
2015-2017
All hourly concentrations are presented for 2015-2017 for these four
monitoring sites
Figure 2-14, Seasonal
Pattern
2015-2017
All valid MDA8 values are presented for 2015-2017 for these four
monitoring sites
2A.2 ANALYSES OF 1-HOUR CONCENTRATIONS
Figure 2-15 of Chapter 2 presents hourly concentrations available in AQS (at the time of
the data query on August 12, 2021) from any site with such data during the 2018-2020 period.
The daily maximum 1-hr (MDA1) values presented in section 2.4.5 and (summary statistics
shown in Table 2A-2 below) were calculated according to Appendix H to 40 CFR Part 50 for all
sites with valid 2018-2020 design values for the current 8-hour standards. Generally, MDA1
values are derived (as the maximum 1-hr concentration during a day) for days for which at least
18 hourly concentrations are available in AQS or for which at least one hourly concentration
greater than 0.12 ppm has been reported in AQS. For this most recent design value period, the
mean number of observations per site at or above 100 parts per billion (ppb) was well below one
(0.22) for sites meeting the current standards compared to well above one (10.53) for sites not
meeting the current standards.
April 2022	2A-3 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
Table 2A-2. Summary statistics for MDA1 concentrations at sites with differing design
values for 2018-2020.

Design Value (ppb)
Statistic
31-60
61-70
71-84
85-114
Number of observations (obs)
261,302
554,712
164,988
27,958
Number of sites
287
590
170
26
25th percentile concentration (ppb)
34
36
40
44
Median concentration (ppb)
40
44
48
57
Mean concentration (ppb)
40.7
44.5
49.7
61.2
75th percentile concentration (ppb)
48
52
59
75
95th percentile concentration (ppb)
58
65
76
101
99th percentile concentration (ppb)
67
76
90
121
# of obs (# of sites) > 240 ppb
0(0)
0(0)
0(0)
0(0)
# of obs (# of sites) > 200 ppb
0(0)
0(0)
0(0)
0(0)
# of obs (# of sites) > 160 ppb
0(0)
0(0)
4(4)
14(6)
# of obs (# of sites) > 120 ppb
2(2)
22 (17)
46 (29)
328(21)
# of obs (# of sites) > 100 ppb
15(12)
180 (112)
526 (127)
1,538(26)
Mean # of obs > 100 ppb per siteA
0.05
0.31
3.09
59.15
A This is the number of obs at or above 100 ppb divided by the number of sites in this bin (column). For the two lowest bins
combined (i.e., all sites with a design value < 70 ppb), the mean is 0.22 obs > 100 ppb per site, and for the two highest bins
combined (i.e., all sites with a design value > 70 ppb), the mean is 10.53 obs > 100 ppb per site.
The figures and tables presented below contain additional analyses based on the MDA1
concentrations for years 2000-2004 and 2016-2020. Figure 2A-1 compares the distribution of
MDA1 concentrations for each 8-hour design value bin between the earlier (2000-2004; red
boxes) and latter (2016-2020; blue boxes) periods. The comparison shows a slight upward shift
in the mid-range concentrations for the highest (> 85 ppb) and lowest (< 60 ppb) DV bins, while
the two middle bins show little change. The range between the 1st and 99th percentiles as
represented by the whiskers shrinks slightly between the earlier and latter periods in all four bins.
Finally, the very highest concentrations (shown as dots above the top whisker) are reduced in the
two highest DV bins. This is also reflected in Table 2A-3 and Table 2A-4, which show summary
statistics similar to Table 2A-2 for the 2000-2004 and 2016-2020 periods, respectively. These
tables show, as might be expected, that sites with higher design values have a larger number of
days with MDA1 values at or above 100 ppb than sites with lower design values. This statistic is
over 35 times higher for sites not meeting the current standard compared to sites meeting the
current standard in 2000-2004, and over 45 times higher in 2016-2020. Across the three design
value periods in 2016 to 2020, sites not meeting the current standards have on average over 9
observations at or above 100 ppb per 3-year period, while the average for sites meeting the
current standards is about 0.2.
Figure 2A-2 and Figure 2A-3 show maps of the average number of days where the
MDA1 concentrations were greater than or equal to 100 ppb (also known as the D100 metric, see
April 2022
2A-4 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
Appendix 4F) for the 2000-2004 and 2016-2020 periods, respectively. These maps show that
nearly all sites in the U.S. have seen a large reduction in the number of days with high MDA1
concentrations since the beginning of the century. This is also reflected in the final rows of Table
2A-3 and Table 2A-4, which indicate a decrease of 83% in the total number MDA1 values
greater than or equal to 100 ppb between 2000-2004 and 2016-2020.
_Q
Q.
Q-
C
o
ป
CD
-*—ป
ฃ=
 84
8-hour 03 Design Value (ppb)
Figure 2A-1. Boxplots comparing the distribution of M DA 1 concentrations for 2000-2004
(red) to the distribution of M DA 1 concentrations for 2016-2020 (blue),
binned by the 8-hour design value at each monitoring site. The boxes
represent the 25th, 50th and 75th percentiles and the whiskers represent the 1st and
99th percentiles. Outlier values are represented by circles.
April 2022
2A-5 External Review Draft - Do Not Quote or Cite

-------
1	Table 2A-3. Summary statistics for MDA1 concentrations at differing design values for
2	2000-2004.

Design Value (ppb)
Statistic
35-60
61-70
71-84
85-131
Number of observations (obs)
117,848
288,396
1,312,716
912,178
Number of design values (DVs)A
130
313
1,518
1,151
25th percentile concentration (ppb)
29
35
37
39
Median concentration (ppb)
36
44
48
52
Mean concentration (ppb)
36.5
44.3
49.3
54.8
75th percentile concentration (ppb)
44
53
60
68
95th percentile concentration (ppb)
56
68
79
95
99th percentile concentration (ppb)
68
79
94
116
# of obs (# of DVsA)>240 ppb
0(0)
0(0)
0(0)
0(0)
# of obs (# of DVsA)>200 ppb
0(0)
0(0)
0(0)
4(4)
# of obs (# of DVsA)> 160 ppb
0(0)
0(0)
15(12)
252 (100)
# of obs (# of DVsA)> 120 ppb
0(0)
8(6)
623 (339)
7,203 (940)
# of obs (# of DVsA)> 100 ppb
26 (16)
161 (87)
7,078 (1,277)
32,133(1,151)
Mean # of obs > 100 ppb per DVB
0.20
0.51
4.66
27.92
A Since this table covers three design value periods, individual sites may be counted up to three times.
B This is the number of obs at or above 100 ppb divided by the number of site-DVs in this bin (column). For the two lowest bins
combined (i.e., sites with a design value <
combined (i.e., sites with a design value >
70 ppb), the mean is 0.40 obs > 100 ppb per site, and for the two highest bins
70 ppb), the mean is 14.69 obs > 100 ppb per site.
3
4	Table 2A-4. Summary statistics for MDA1 concentrations at differing design values for
5	2016-2020.
Statistic
Design Value (ppb)
29-60
61-70
71-84
85-114
Number of observations (obs)
582,220
1,824,438
558,927
99,742
Number of design values (DVs)A
637
1,969
579
93
25th percentile concentration (ppb)
33
37
39
45
Median concentration (ppb)
40
44
48
57
Mean concentration (ppb)
40.6
44.8
49.2
60.6
75th percentile concentration (ppb)
48
53
59
74
95th percentile concentration (ppb)
58
65
76
99
99th percentile concentration (ppb)
66
75
89
118
# of obs (# of DVsA)>240 ppb
0(0)
0(0)
0(0)
0(0)
# of obs (# of DVsA)>200 ppb
0(0)
0(0)
0(0)
0(0)
# of obs (# of DVsA)> 160 ppb
0(0)
1(1)
4(4)
15(7)
# of obs (# of DVsA)> 120 ppb
8(6)
51 (42)
101 (77)
904 (69)
# of obs (# of DVsA)> 100 ppb
41 (32)
486 (335)
1,591 (423)
4,761 (93)
Mean # of obs > 100 ppb per DVB
0.06
0.25
2.75
51.19
A Since this table covers three design value periods, individual sites may be counted up to three times.
B This is the number of obs at or above 100 ppb divided by the number of site-DVs in this bin (column). For the two lowest bins
combined (i.e., sites with a design value < 70 ppb), the mean is 0.20 obs > 100 ppb per site, and for the two highest bins
combined (i.e., sites with a design value > 70 ppb), the mean is 9.45 obs > 100 ppb per site.
6
April 2022
2A-6 External Review Draft - Do Not Quote or Cite

-------
Average Number of Days with MDA1 > 100 ppb, 2000 - 2004
2	*0	o 0.1-1.0 o 1.1-3.0 ฉ 3.1 - 10.0 •> 10.0
2 Figure 2A-2. Map showing the average number of days with M DA 1 > 100 ppb, 2000-2004.
Average Number of Days with MDA1 s 100 ppb, 2016 - 2020
3	• 0	ฎ 0.1-1.0 ฉ 1.1-3.0 ฉ 3.1-10.0 • >10.0
4	Figure 2A-3. Map showing the average number of days with M DA 1 > 100 ppb, 2016-2020.
April 2022	2A-7 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
Figure 2A-4 below shows the number of days in 2018-2020 with an MDA1 concentration
at or above 100 ppb and 8-hour design values (similar to Figure 2-16), for all sites with a 2018-
2020 design value less than 102 ppb. All sites meeting the current standard had seven or fewer
(i.e., two or fewer per year) MDA1 values at or above 100 ppb, and all but eight sites meeting
the current standard had three or fewer (i.e., one or fewer per year) VIDAI values at or above
100 ppb.
80-
70
& 60-
O
O
Al

t_
13
O
sz
X
TO
50-
40-
5 30'

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
APPENDIX 2B
ADDITIONAL DETAILS ON BACKGROUND OZONE
MODELING AND ANALYSIS
TABLE OF CONTENTS
2B. 1 Photochemical Modeling Methodology	2B-2
2B. 1.1 Modeling Platform Overview	2B-4
2B.1.2 Emissions Overview	2B-5
2B1.2.1 Natural Emission Inventory	2B-3
2B1.2.2 Anthropogenic Emission Inventory	2B-4
2B.2 Evaluation	2B-8
2B.3 International Contributions	2B-40
References 	2B-44
TABLE OF FIGURES
Figure 2B-1. NOAA U.S. climate regions	2B-9
Figure 2B-2. (a) Normalized Mean Bias (%) and (b) Mean Bias (ppb) of maximum daily
average 8-hr ozone (MDA8) by NOAA climate region (y-axis) and by season
(x-axis) at AQS monitoring sites	2B-17
Figure 2B-3. NMB (a) and MB (b) of MDA8 O3 greater than or equal to 60 ppb from the
12km resolution CONUS simulation by NOAA climate region (y-axis) and
by season (x-axis) at AQS monitoring sites	2B-17
Figure 2B-4. Density scatter plots of observed versus predicted MDA8 O3 from the 12km
resolution CONUS simulation for the Northeast region by season	2B-18
Figure 2B-5. Density scatter plots of observed versus predicted MDA8 O3 from the 12km
resolution CONUS simulation for the Central region by season	2B-19
Figure 2B-6. Density scatter plots of observed versus predicted MDA8 O3 from the 12km
resolution CONUS simulation for the EastNorthCentral region by season.... 2B-20
Figure 2B-7. Density scatter plots of observed versus predicted MDA8 O3 from the 12km
resolution CONUS simulation for the Southeast region by season	2B-21
Figure 2B-8. Density scatter plots of observed versus predicted MDA8 O3 from the 12km
resolution CONUS simulation for the South region by season	2B-22
Figure 2B-9. Density scatter plots of observed versus predicted MDA8 O3 from the 12km
resolution CONUS simulation for the Southwest region by season	2B-23
April 2022
2B-1 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
Figure 2B-10. Density scatter plots of observed versus predicted MDA8 O3 from the 12km
resolution CONUS simulation for the WestNorthCentral region by season. . 2B-24
Figure 2B-11. Density scatter plots of observed versus predicted MDA8 O3 from the 12km
resolution CONUS simulation for the Northwest region by season	2B-25
Figure 2B-12. Density scatter plots of observed versus predicted MDA8 O3 from the 12km
resolution CONUS simulation for the West region by season	2B-26
Figure 2B-13. Mean Bias (ppb) from the 12km resolution CONUS simulation of MDA8 O3
greater than or equal to 60 ppb over the period May through September 2016
at AQS and CASTNET monitoring sites in the continental U.S. modeling
domain	2B-27
Figure 2B-14. Mean Error (ppb) from the 12km resolution CONUS simulation of MDA8 O3
greater than or equal to 60 ppb over the period May through September 2016 at
AQS and CASTNET monitoring sites in the continental U.S. modeling domain.
	2B-27
Figure 2B-15. NMB (%) from the 12km resolution CONUS simulation of MDA8 O3 greater
than or equal to 60 ppb over the period May through September 2016 at AQS
and CASTNET monitoring sites in the continental U.S. modeling domain. .. 2B-28
Figure 2B-16. NME (%) from the 12km resolution CONUS simulation of MDA8 O3 greater
than or equal to 60 ppb over the period May through September 2016 at AQS
and CASTNET monitoring sites in the continental U.S. modeling domain. .. 2B-28
Figure 2B-17. WOUDC sonde locations and sampling frequency used in evaluation of
hemispheric model simulation	2B-29
Figure 2B-18. WOUDC sonde releases averaged by release location over 2016; observations
(left), predictions from the hemispheric CMAQ simulation (middle), ratio
(right). Observations are ordered with increasing latitude (South to North).. 2B-30
Figure 2B-19. WOUDC sonde releases averaged by day with a 20-point moving average;
observations (left), predictions from the hemispheric CMAQ simulation
(middle), ratio (right)	2B-31
Figure 2B-20. WOUDC sonde releases averaged by release location over March, April, May
in 2016; observations (left), predictions from the hemispheric CMAQ simulation
(middle), ratio (right)	2B-32
Figure 2B-21. WOUDC sonde releases averaged by release location over June, July, August in
2016; observations (left), predictions from the hemispheric CMAQ simulation
(middle), ratio (right)	2B-33
Figure 2B-22. OMI O3 (OMPROFOZ v003, left) compared to simulated (hemispheric CMAQ
simulation, center), and ratios (right) of vertical column densities for January
(top) and April (bottom)	2B-34
Figure 2B-23. OMI O3 (OMPROFOZ v003, left) compared to simulated (hemispheric CMAQ
simulation, center), and ratios (right) of vertical column densities for July (top),
and October (bottom)	2B-35
April 2022
2B-2 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
Figure 2B-24
Figure 2B-25
Figure 2B-26
Figure 2B-27
Figure 2B-28
Figure 2B-29
Figure 2B-30
Table 2B-1.
OMI Nitrogen Dioxide (0MN02D HR v003, left) compared to simulated
(hemispheric CMAQ simulation, center), and ratios (right) of vertical column
densities for January (top) and April (bottom)	2B-36
OMI Nitrogen Dioxide (0MN02D HR v003, left) compared to simulated
(hemispheric CMAQ simulation, center), and ratios (right) of vertical column
densities for July (top) and and October (bottom)	2B-37
OMI Formaldehyde (OMHCHO v003, left) compared to simulated (hemispheric
CMAQ simulation, center), and ratios (right) of vertical column densities for
January (top) and April (bottom)	2B-38
OMI Formaldehyde (OMHCHO v003, left) compared to simulated (hemispheric
CMAQ simulation, center), and ratios (right) of vertical column densities for
July (top), and October (bottom)	2B-39
Total predicted MDA8 O3 and contributions (see legend) over time in the West
(top), and all East (bottom) averaged over all grid cells and days in the U.S. 2B-41
International contribution (black line) to predicted MDA8 O3 and components
(see legend) over time in the West (top), and all East (bottom) averaged over
all grid cells and days in the U.S	2B-42
International contribution (black line) to predicted MDA8 O3 and components
(see legend) over time averaged over all grid cells in the West at high elevation
(top), near-border sites (middle), and Low/Interior sites (bottom)	2B-43
TABLE OF TABLES
Summary of 12km resolution CONUS CMAQ 2016 model performance
statistics for MDA8 O3 by NOAA climate region, by season and monitoring
Network	2B-15
April 2022
2B-3 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
This appendix for the background ozone (O3) modeling and analysis, which presents the
analysis that was also presented in Appendix 2B of the 2020 PA (and is virtually identical to that
appendix), includes a description of the methodology for photochemical modeling, an evaluation
of the modeling, and a more detailed analysis of the predicted contributions from international
anthropogenic emissions. The methodology section includes a description of the modeling
platform and emissions. The evaluation section includes comparisons against surface, sondes and
satellite measurements. The international component analysis separately estimates O3 impacts
from China, India, Canada/Mexico, and global shipping at the hemispheric scale.
2B.1 PHOTOCHEMICAL MODELING METHODOLOGY
2B.1.1 Modeling Platform Overview
A multiscale modeling system is applied at both hemispheric and regional scales with
consistent methodologies for emissions inputs, meteorological inputs, model chemistry, and
photochemical models. Consistency across spatial scales reduces the number of assumptions that
have to be made in integrating predictions from the global and the regional modeling. However,
methodological consistency does not address sources of uncertainty associated with individual
inputs used by the modeling system.
The modeling system uses one emission model, one meteorological model, and one
chemical transport model. The meteorological model is the Weather Research and Forecasting
model (WRF v3.8). The emissions model is the Sparse Matrix Operating Kernel for Emissions
(SMOKE v4.5). The chemical transport model is the Community Multiscale Air Quality model
(CMAQ) version 5.2.1 with the Carbon Bond mechanism (CB6r3) and the non-volatile aerosol
option (AE6). Each of these models is applied at hemispheric and regional scales. The regional
meteorology components of the modeling system are described in more detail in section 3C.4.1.4
of Appendix 3C, while emissions inputs are summarized here.
The models identified above are configured differently for the hemispheric and regional
scales as appropriate for the intended purpose. The hemispheric scale model uses a polar
stereographic projection at 108 kilometer (km) resolution to completely and continuously cover
the Northern Hemisphere. At the regional scale, the model employs a Lambert conic conformal
projection at 36 km resolution to cover North America and at 12 km resolution to cover the
lower 48 contiguous states. The hemispheric scale allows for long-range free tropospheric
transport with 44 layers between the surface and 50 hPa (-20 km asl). The 36 km and 12 km
regional modeling has 35 vertical layers between the surface and 50 hPa. The hemispheric
modeling system was initiated on May 1, 2015 and run continuously through December 31,
2016. The regional model was initialized using the hemispheric result on December 21, 2015 and
run continuously through December 31, 2016.
April 2022
2B-4 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
2B.1.2Emissions Overview
The emissions inventories are summarized here and more information is available in the
Emissions Technical Support Documents (U.S. EPA, 2019a, U.S. EPA, 2019b) and in Appendix
3C. The emissions model inputs are discussed separately for natural and anthropogenic
emissions. The stratospheric fluxes (section 2.5.1.1 of main document) are not discussed here
because, although they are a source of ozone, they are not emissions. The regional inventories
over North America are based on the Inventory Collaborative 2016 emissions modeling platform
(ihttp://views.cira.colostate.edu/wiki/wiki/9169), which was developed through the summer of
2019. Three versions of the 2016 inventory developed: "alpha" (also known as the 2016v7.1
platform) - which consisted of data closely related to the 2014 National Emissions Inventory
(NEI) version 2 and 2016-specific data for some sectors; "beta" (also known as the 2016v7.2
platform) - which incorporated data from state and local agencies and adjustments to better
represent the year 2016; and "version 1" (also known as the 2016v7.3 platform) - which has the
completed representation of 2016 and some elements from the 2017 NEI. For any regional
inventories, this analysis used the 2016 "alpha release" (specifically the modeling case
abbreviated 2016fe) that is publicly available from https://www.epa.gov/air-emissions-
modeling/2016-alpha-platform. Any changes in the 2016 "beta" or "version 1" platforms are not
included in this modeling and therefore are not captured in the subsequent analysis.
2B.1.2.1 Natural Emission Inventory
The natural emission inventory databases cover all the sources discussed in section 2.5.1
except the International Anthropogenics. The databases that are available depend upon the scale.
At the global scale, lightning NOx emissions are based on monthly climatological data; biogenic
VOC emissions have hourly and day-specific (MEGAN v2.1, Guenther et al., 2012) temporal
scales; soil NOx also has hourly and day-specific temporal scales (Berkeley Dalhousie Soil NOx
Parameterization, as implemented by Hudman et al., 2012); and fire emissions are based on day-
specific data (FINN vl.5, Wiedinmyer et al., 2011). Over our regional domain, regional
inventories supersede the biogenic VOCs, soil NOx, and fire emissions using estimates
consistent with the 2016 collaborative emissions modeling platform {https://www.epa.gov/air-
emissions-modeling/2016-alpha-platform). The regional biogenic VOCs and soil NOx are
derived from the Biogenic Emission Inventory System (BEIS v3.61). Of the natural inventories,
only fires are expected to change significantly in future versions of the 2016 emissions platform.
The biogenic VOC and NOx changes will be minor due to small changes to the land use data
input to BEIS3.
April 2022
2B-5 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
Emissions of NOx are of particular importance to this study and the natural inventory is
summarized here. The total natural NOx emissions4 in this platform is 56 megatons NOx
(reported as equivalent NO2 mass) which is approximately 15.5 TgN. The contributors in order
of magnitude are lightning (55%), soil (33%), and wildfires (12%). Lightning is treated as a
climatological monthly mean contribution, while soils and wildfires are day-specific. It is
important to note that outside North America, prescribed fires are not identified distinctly from
wildfires. Therefore, all wildland fires outside North America are treated as natural. Though not
directly comparable, the lightning and soil magnitudes are consistent with the ranges reported by
(Lamarque et al., 2012). Consistent with previous regional modeling platforms, the lightning
emissions are not included in the emissions inputs to the regional modeling platform. At the
regional scale, the representation of lightning as a monthly mean rate would add lightning on
days where it may not have occurred. At the hemispheric scale, omitting lightning would remove
an important contribution to the well-mixed background O3.
2B.1.2.2 Anthropogenic Emission Inventory
Anthropogenic emissions inputs include both domestic and international sources. The
domestic inventory includes a high-level of detail that is consistent with previous EPA emissions
platforms such as those used to model the year 2011 (,https://www.epa.gov/air-emissions-
modeling/201 l-version-6-air-emissions-modeling-platforms). For the hemispheric emissions
modeling platform, there are over thirty anthropogenic sector of emission files. The traditional
regional platform covers North America including the U.S. sectors, Canadian sectors, and
Mexican sectors. In addition to the typical regional platform sectors, there are nine sectors based
on the Hemispheric Transport of Air Pollution Version 2 (EDGAR-HTAPv2) inventory and 15
sectors that represent emissions in China which together comprise the anthropogenic emissions
outside of North America. The international emission inventories are synthesized from the
EDGAR-HTAP v2 harmonized emission inventory and country specific databases where updates
were likely to be influential. Previous assessments like HTAP (2010, Phase 1) and HTAP (Phase
2) have shown that the anthropogenic portion of USB is most sensitive to emissions in Mexico,
Canada, and China. For Mexico and Canada, the hemispheric platform relies on the same
country-specific databases as the regional platform. For China, as mentioned above, the
hemispheric platform uses a new country specific database. The sources are detailed further
below.
The EDGAR-HTAP v2 inventories were projected to represent the year 2014. Projection
factors were calculated from the Community Emissions Data System (CEDS) inventory at a
4 We refer to wildfires and soil NOx as natural for the purposes of this section even though both may be impacted to
various degrees by human activity.
April 2022
2B-6 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
country-sector level. This allowed our inventory to evolve without the risks associated with
transitioning to a new inventory system. Especially because EDGAR-HTAP v2 is superseded for
critical counties, this was the optimal approach. Details of scaling factor development are
described in Section 2.1.5 of the 2016v7.1 Hemispheric Modeling Platform Technical Support
Document (U.S. EPA, 2019a).
Emissions estimates over Mexico are a combination of emissions supplied by the
Mexican government and emissions developed by the EPA. For the 2016 platform, emissions for
point, nonpoint, and nonroad sources were developed based on projections of Secretariat of
Environment and Natural Resources (SEMARNAT)-supplied data for the year 2008. For the
onroad mobile sources, the EPA developed year-specific inventories for 2014 and 2017 by
applying the MOVES-Mexico model and interpolating to the year 2016. More details are
available in the 2016v7.1 emissions platform TSD (U.S. EPA, 2019b).
Emissions for Canada were supplied by Canadian agencies and reprocessed by the EPA
for the domains and model years used in this analysis. Environment and Climate Change Canada
(ECCC) supplied data for four broad inventory sectors (point, on-road mobile, fugitive dust, and
area and non-road mobile sources, the latter including commercial marine vessels). The ECCC
emissions were interpolated to 2016 based on inventories from the years 2013 and 2025.
The China emission inventory was developed at Tsinghua University (THU) and
documented in Zhao et al., 2018 (see supplement). This inventory was extensively compared to
the EDGAR-HTAP v2 and EDGAR v4.3 inventories before use. The largest differences for NOx
in 2016 occurred in individual emissions sectors rather than inventory totals. The SO2 emissions
were more different than NOx emissions between the two inventories because the THU
inventory applies controls to the metal industry that have been adopted by China. The difference
between emissions, primarily NOx emissions, causes small decrease in the spring time surface
O3 over the U.S. compared to using EDGAR-HTAP v2. Comparisons of this update are
summarized by Henderson et al.(2019).
Emissions for the United States representing the year 2016 were developed using the
2014 National Emissions Inventory version 2 (2014NEIv2) as the starting point, although
emissions for some data categories were updated to better represent the year 2016. The point
source emission inventories for the platform are partially updated to represent 2016. Because
2016 is not a year for which a full NEI is compiled, states are only required to submit emissions
for their larger point sources. For units without 2016-specific emissions, the emissions were
carried forward from the 2014 NEIv2. For electric generating units, 2016-specific Continuous
Emissions Monitoring System (CEMS) data are used where the data can be matched to units in
the NEI. Point and nonpoint oil and gas emissions were projected from 2014 to 2016 using
factors based on historic production levels.
April 2022
2B-7 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
Other sectors are briefly summarized here and the reader is directed to the TSD for more
details (U.S. EPA, 2019a). Agricultural and wildland (including prescribed) fire emissions were
developed for the year 2016 using methods similar to those used to develop the 2014 NEI, except
that the input data relied on nationally-available data sets and did not benefit from state-
submitted data as are used for NEI year emissions. The assignment of wildland fires to wild or
prescribed is a complex process that is documented in the regional platform emissions TSD (U.S.
EPA, 2019b). Most area source sectors for this platform use unadjusted 2014 NEIv2 emissions
estimates except for commercial marine vehicles (CMV), fertilizer emissions, oil and gas
emissions, and onroad and nonroad mobile source emissions. For CMV, SO2 emissions were
updated to reflect new rules for the North American Emission Control Area (regulation 13.6.1
and appendix VII of MARPOL Annex VI) on sulfur emissions that took effect in the year 2015.
For fertilizer ammonia emissions, a 2016-specific emissions inventory is used in this platform,
while animal ammonia emissions were the same as those in 2014 NEIv2. Onroad and nonroad
emissions were developed based on MOVES2014a outputs for the year 2016, and the activity
data used to compute the onroad emissions were projected from 2014 to 2016 based on distinct
state-specific factors for urban and rural roads. Emissions from 2014 NEIv2 were used directly
for residential wood combustion, fugitive dust, and other nonpoint sources, although
meteorological-based adjustments for dust sources and temporal allocation for residential wood
and agricultural ammonia sources were based on 2016 meteorology. Additional details on the
development of the U.S., Canada, and Mexico emissions are provided in the 2016v7.1 (U. S.
EPA, 2019b).
2B.2 EVALUATION
An operational model performance evaluation for O3 was conducted for the 2016fe
simulation (as referred to in Section 2.5.2.2) using monitoring data, ozone sonde data, and
satellite data in order to estimate the ability of the CMAQv5.2.1 modeling system to replicate the
2016 base year O3 concentrations for the 12 km continental U.S. domain and the 108 km
Northern Hemispheric domain. The purpose of this evaluation is to examine the ability of the
2016 air quality modeling platform to represent the magnitude and spatial and temporal
variability of measured (i.e., observed) O3 concentrations within the modeling domain. The
model evaluation for O3 focuses on comparisons of model-predicted 8-hour daily maximum
concentrations (MDA8) to the corresponding concentrations from monitoring data (for 2016)
collected at monitoring sites in the AQS. The evaluation divided these data into two datasets, one
limited to only CASTNET sites (described in section 2.3.1), and the second comprised of all
other sites. We refer to this second dataset as "AQS."
April 2022
2B-8 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
Included in the evaluation are statistical measures of model performance based upon
model-predicted versus observed MDA8 O3 concentrations that were paired in space and time.
Statistics were generated for each of the nine National Oceanic and Atmospheric Administration
(NOAA) climate regions of the 12-km U.S. modeling domain (Figure 2B-1). The regions include
the Northeast, Central, EastNorthCentral, Southeast, South, Southwest, WestNorthCentral,
Northwest and West as were originally identified in Karl and Koss (1984). Note that most
monitoring sites in the West region are located in California, therefore statistics for the West will
be mostly representative of California O3 model performance.
5 Each monitoring site had to have 75% of MDA8 values within any seasonal subset to be included in that subset.
Thus individual monitors may be included in one evaluation of season, but not another.
NorthWest	ฆ WestNorthCentral ~ EastNorthCentral ฆ Central	0 NorthEast
West	~ Southwest	~ South	~ SouthEast
Source: http://www.ncdc.noaa.gOv/monitoring-references/maps/us-climate-regions.php#references
Figure 2B-1. NOAA U.S. climate regions.
For MDA8 O3, model performance statistics were calculated for each climate region by
season and for the May through September O3 season of 2016. Seasons were defined as: winter
(December-January-February), spring (March-April-May), summer (June-July-August), and fall
(September-October-November). Observational data were excluded from the analysis and model
evaluations for sites that did not meet a 75% completeness criterion.5 In addition to the
performance statistics, several graphical presentations of model performance were prepared for
MDA8 O3 concentrations. These graphical presentations include:
April 2022	2B-9 External Review Draft - Do Not Quote or Cite

-------
1	(1) density scatter plots of observations obtained from the AQS system excluding CASTNET
2	(hereafter AQS) and predicted MDA8 O3 concentrations for May through September;
3	(2) regional maps that show the mean bias and error as well as normalized mean bias and
4	error calculated for MDA8 > 60 ppb for May through September at individual AQS and
5	CASTNET monitoring sites;
6	(3) tile plots that show normalized mean bias (%) and mean bias (ppb) of MDA8 and MDA8
7	>60 ppb by NOAA climate region (y-axis) and by season (x-axis) at AQS monitoring
8	sites;
9	(4) O3 sonde evaluations comparing vertically resolved ozone model predictions to ozone
10	sondes measurements from the World Ozone and Ultraviolet Data Centre (woudc.org).
11	(5) satellite evaluation comparing simulated tropospheric vertical column densities of O3,
12	nitrogen dioxide, and formaldehyde to OMI retrievals.
13	The Atmospheric Model Evaluation Tool (AMET) was used to calculate the model
14	performance statistics used in this evaluation (Gilliam et al., 2005). For this evaluation of the O3
15	predictions in the 2016fe CMAQ modeling platform, we have selected the mean bias, mean
16	error, normalized mean bias, and normalized mean error to characterize model performance,
17	statistics which are consistent with the recommendations in Simon et al. (2012) and the
18	photochemical modeling guidance (U.S. EPA, 2018).
19	Mean bias (MB) is used as average of the difference (predicted - observed) divided by
20	the total number of replicates (ri). Mean bias is defined as:
21	MB = ~IIi(f — O) , where P = predicted and O = observed concentrations for every site
22	and day included in the evaluation.
23	Mean error (ME) calculates the absolute value of the difference (predicted - observed)
24	divided by the total number of replicates (ri). Mean error is defined as:
25	ME = ฑฃI|P-0|
26	Normalized mean bias (NMB) is used as a normalization to facilitate a range of
27	concentration magnitudes. This statistic averages the difference (predicted - observed) over the
28	sum of observed values. NMB is a useful model performance indicator because it avoids
29	overinflating the observed range of values, especially at low concentrations. Normalized mean
30	bias is defined as:
April 2022
2B-10 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
IsiP-O)
NMB = —	*100, where P = predicted concentrations and O = observed
i(o)
i
Normalized mean error (NME) is also similar to NMB, where the performance statistic is
used as a normalization of the mean error. NME calculates the absolute value of the difference
(model - observed) over the sum of observed values. Normalized mean error is defined as
i\p-dl
NME = 	*100
n
t(o)
1
As described in more detail below, the model performance statistics indicate that the
MDA8 O3 concentrations predicted by the 2016 CMAQ modeling platform closely reflect the
corresponding monitoring data-based MDA8 O3 concentrations in space and time in each region
of the U.S. modeling domain. The acceptability of model performance was judged for the 2016
CMAQ O3 performance results considering the range of performance found in recent regional O3
model applications (NRC, 2002; Phillips et al., 2008; Simon et al., 2012; U.S. EPA, 2009; U.S.
EPA, 2018). These other modeling studies represent a wide range of modeling analyses that
cover various models, model configurations, domains, years and/or episodes, chemical
mechanisms, and aerosol modules. Overall, the 2016 CMAQ O3 model performance results are
within the range found in other recent peer-reviewed and regulatory applications. The model
performance results, as described in this document, demonstrate the predictions from the 2016
modeling platform closely replicate the corresponding observed concentrations in terms of the
magnitude, temporal fluctuations, and spatial differences for 8-hour daily maximum O3.
The model performance bias and error statistics for MDA8 O3 predictions in each of the
nine NOAA climate regions and each season are provided in Table 2B-1. As noted above, seasons
were defined as: winter (December-January-February), spring (March-Apri 1 -May), summer
(June-July-August), and fall (September-October-November). As indicated by the statistics in
Table 2-7, mean bias and error for 8-hour daily maximum O3 are relatively low in each
subregion, not only in the summer when concentrations are highest, but also during other times of
the year. Generally, MB for MDA8 O3 > 60 ppb is less than + 10 ppb. Generally, MDA8 O3 at the
AQS sites in the summer and fall is over predicted except in the Southwest, with the greatest over-
prediction in the EastNorthCentral and WestNorthCentral. Likewise, MDA8 O3 at the
CASTNET sites in the summer and fall is typically over predicted except in the West, Southwest
and WestNorthCentral where the bias shows an under-prediction. In the winter and spring.
April 2022
2B-11 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
MDA8 0} is under predicted at AQS and CASTNET sites in all the climate regions (with NMBs
less than approximately + 25 percent in each sub region).
Figure 2B-2 and Figure 2B-3 are tile plots that summarize to provide an overview of
model performance by region and by season. Figure 2B-2 shows NMB (%) and MB (ppb) of
MDA8 by NOAA climate region (y-axis) and by season (x-axis) at AQS monitoring sites.
Likewise, Figure 2B-3 shows the NMB (%) and MB (ppb) of MDA8 > 60 ppb by NOAA climate
region (y-axis) and by season (x-axis) at AQS monitoring sites. Figure 2B-2 shows that for the
majority of the nine climate regions throughout each year the NMB is within ฑ10 percent. There
is greater over-prediction (<20%) during the fall in the South, EastNorthCentral (aka Upper
Midwest), and Central (aka Ohio Valley) regions and during the summer in the South, Southeast
and Central (aka Ohio Valley) regions. However, there is greater under-prediction (up to 30
percent) during the winter in the Northwest, Southwest, WestNorthCentral (aka
NRockiesPlains), EastNorthCentral (aka Upper Midwest), Central (aka Ohio Valley), and
Northeast regions as well during the spring in the Northwest.
The density scatterplots in Figure 2B-4 to Figure 2B-12 provide a qualitative comparison
of model-predicted and observed MDA8 O3 concentrations for each climate region by season. In
these plots the intensity of the colors indicates the density of individual observed/predicted
paired values. The greatest number of individual paired values is denoted by locations in the plot
denoted in warmer colors. The plots indicate that the predictions correspond closely to the
observations in that a large number of observed/predicted paired values lie along or close to the
1:1 line shown on each plot. The model is more likely to over-predict the observed values at low
and mid-range concentrations generally < 60 ppb in each of the regions. There are some
relatively infrequent very large over predictions at high concentrations. Preliminary review of
these biases finds that some are related to fire impacts.
Spatial plots of the MB, ME, NMB and NME for individual monitors are shown in Figure
2B-13 through Figure 2B-16, respectively. The statistics shown in these two figures were
calculated over the May through September period, using data pairs on days with observed 8-hr
O3 of greater than or equal to 60 ppb. Model bias at individual sites during the O3 season is
similar to that seen on a sub-regional basis for the summer. Figure 2B-13 shows the mean bias
for 8-hr daily maximum O3 greater than 60 ppb is under predicted overall, but generally within
ฑ10 ppb across the AQS and CASTNET sites. The greatest exceptions are most evident at certain
near-coastal sites where, on average, the model over predicts MDA8 observed O3 > 60 ppb.
Likewise, the information in Figure 2B-15 indicates that the normalized mean bias for days with
observed 8-hr daily maximum O3 greater than 60 ppb is within ฑ 10% at the vast majority of
monitoring sites across the U.S. domain. Model error, as seen from Figure 2B-14 and Figure 2B-
16, is generally 2 to 10 ppb and 20 percent or less at most of the sites across the U.S. modeling
April 2022
2B-12 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
domain. Somewhat greater error is evident at sites in several areas most notably in the West,
WestNorthCentral, Northeast, EastNorthCentral, Southeast, and along portions of the Gulf Coast
and Great Lakes coastlines.
Sonde evaluations are shown for the 108 km Northern Hemisphere domain in Figure 2B-
18 through Figure 2B-21. The sondes used in this analysis and their release frequencies are
shown in Figure 2B-17. Figure 2B-18 shows that the annual mean prediction is generally within
20% of the measured sonde data, except for near the tropopause. Figure 2B-19 shows that the
performance of all sites is generally not as good in the spring (March, April, May ) than in the
summer (June, July, August). The seasonal performance of each monitor is shown in Figure 2B-
20 for spring and Figure 2B-21 for summer. By comparison. Figure 2B-20 shows that low biases
extend deeper into the troposphere in spring than in summer. The structure of the bias seems to
suggest a stratospheric causal mechanism because the bias is near the tropopause.
Satellite evaluations in this analysis include tropospheric vertical columns of O3, nitrogen
dioxide (an ozone precursor as described in chapter 2), and formaldehyde (a VOC reaction
product which is an indicator of VOCs and total reactivity of the atmosphere). At this time, only
formaldehyde comparison includes the application of the scattering weights and air mass factor
to the model, which are often used to create an averaging kernel. Similar processing for O3 and
NO2 was not available at the time this appendix was completed. Satellite evaluations focus
exclusively on the 108 km results over the Northern Hemisphere.
Simulated O3 tropospheric vertical column densities are compared to the O3 product
described and evaluated by Huang et al. (2017). Figure 2B-22 and Figure 2B-23 compares the
model to the retrieved column data without application of the averaging kernel. Omitting the
averaging kernel introduces some error into the comparison (Huang et al., 2017; see Figure 9 for
details). Even so, the comparison shows reasonable performance within the mid-latitudes. There
is a notable low bias in January mid-latitudes and near the north pole in April. In addition, high
biases are consistently seen near the corners of the domain in January and April. This cause of
this high-bias pattern will require further analysis. Within the mid latitudes, the model is
performing well with notable low biases in January and scattered high biases in Asia in July.
Given the limitations of the comparison, the performance is quite good.
Simulated nitrogen dioxide (NO2) vertical columns are compared is the OMN02d
(Krotkov et al., 2017, as processed by Lok Lamsal called OMN02D HR). Similar to O3, the
averaging kernel is not being applied for NO2. Figure 2B-24 and Figure 2B-25 show larger
relative biases for NO2 than O3, particularly in low NO2 regions like over the oceans. Best
performance was over land during July. Model comparisons to NO2 have commonly shown
biases and research in the broader community continues to resolve this issue.
April 2022
2B-13 External Review Draft - Do Not Quote or Cite

-------
1	Formaldehyde retrieval comparisons are shown in Figure 2B-26 and Figure 2B-27 using
2	the OMHCHO files, but using the recommended product described by Gonzalez Abad et al.
3	(2015). The formaldehyde retrievals show a seasonal cycle in the evaluation with a low bias for
4	the northern-most retrievals in January and October. During April there are high biases that seem
5	to migrate northward by July. Though we note this bias feature, the main result is reasonable
6	spatial consistency between the satellite product and the model results. Future work should
7	explore this evaluation further.
8
9
April 2022	2B-14 External Review Draft - Do Not Quote or Cite

-------
1	Table 2B-1. Summary of 12km resolution CONUS CMAQ 2016 model performance
2	statistics for MDA8 O3 by NOAA climate region, by season and monitoring
3	Network.
Climate region
Monitor
Network
Season
No. of
Obs
MB (ppb)
ME (ppb)
NMB (%)
NME (%)
Northeast
AQS
Winter
11,462
-5.9
6.9
-18.1
21.2
Spring
15,701
-4.3
6.7
-9.8
15.2
Summer
16,686
4.6
7.7
10.0
17.0
Fall
13,780
3.3
5.8
9.5
16.9
CASTNET
Winter
1,195
-6.7
7.3
-19.6
21.3
Spring
1,246
-5.0
6.9
-11.0
15.2
Summer
1,224
2.9
6.5
6.7
15.1
Fall
1,215
3.4
5.6
9.9
16.5
Central
AQS
Winter
4,178
-3.8
5.7
-12.5
18.8
Spring
15,498
-1.1
5.5
-2.5
12.1
Summer
20,501
5.5
8.1
12.1
17.9
Fall
14,041
4.9
6.1
12.6
15.7
CASTNET
Winter
1,574
-3.1
5.4
-9.6
16.3
Spring
1,600
-2.2
5.5
-4.8
12.0
Summer
1,551
3.9
7.1
9.0
16.2
Fall
1,528
2.7
5.1
6.9
12.8
EastNorthCentral
AQS
Winter
1,719
-8.5
9.2
-27.3
29.5
Spring
6,892
-3.8
6.8
-8.4
15.2
Summer
9,742
3.2
6.9
7.7
16.3
Fall
6,050
5.6
3.4
17.6
20.2
CASTNET
Winter
435
-9.6
10.1
-28.6
30.1
Spring
434
-6.5
7.8
-14.4
17.4
Summer
412
0.2
5.5
0.5
13.4
Fall
426
2.9
5.1
9.2
16.0
Southeast
AQS
Winter
7,196
-1.4
5.0
-3.9
14.0
Spring
14,569
-1.5
5.3
-3.2
11.3
Summer
15,855
5.1
7.1
12.9
17.9
Fall
12,589
3.4
5.4
8.4
13.3
CASTNET
Winter
887
-3.5
5.3
-9.3
14.3
Spring
947
-3.6
5.6
-7.5
11.7
Summer
926
3.9
6.2
9.9
16.0
Fall
928
1.7
5.0
4.0
11.9
South
AQS
Winter
11,342
-1.0
5.0
-3.1
15.0
Spring
13,093
1.3
6.1
2.8
13.9
Summer
12,819
6.0
7.8
15.7
20.4
Fall
12,443
4.8
6.3
12.1
16.0
April 2022
2B-15 External Review Draft - Do Not Quote or Cite

-------
Climate region
Monitor
Network
Season
No. of
Obs
MB (ppb)
ME (ppb)
NMB (%)
NME (%)

CASTNET
Winter
516
-1.7
5.0
-4.8
13.7
Spring
532
-1.2
5.6
-2.6
12.3
Summer
508
2.6
6.1
6.7
15.8
Fall
520
3.5
5.0
9.0
12.9
Southwest
AQS
Winter
9,695
-4.2
6.2
-11.0
16.1
Spring
10,608
-4.8
6.5
-9.4
12.7
Summer
10,549
-1.2
6.0
-2.3
11.2
Fall
10,298
2.5
4.9
6.0
12.0
CASTNET
Winter
757
-8.1
8.5
-18.0
18.9
Spring
810
-6.9
7.6
-13.1
14.5
Summer
812
-2.8
5.5
-5.3
10.3
Fall
791
-0.1
3.6
-0.3
8.3
WestNorthCentral
AQS
Winter
4,740
-9.3
9.6
-24.9
25.9
Spring
5,066
-3.1
5.9
-7.2
13.5
Summer
5,134
0.7
4.9
1.4
10.6
Fall
4,940
3.3
5.2
9.8
15.3
CASTNET
Winter
568
-9.1
9.8
-23.1
25.0
Spring
607
-5.8
7.3
-12.4
15.6
Summer
600
-1.8
4.6
-3.7
9.4
Fall
505
1.7
4.8
4.4
12.8
Northwest
AQS
Winter
677
-5.7
7.5
-17.5
23.1
Spring
1,288
-4.3
7.3
-10.5
18.2
Summer
2,444
1.2
6.6
3.3
17.5
Fall
1,236
2.8
5.9
9.0
18.7
CASTNET
Winter
-
-
-
-
-
Spring
-
-
-
-
-
Summer
-
-
-
-
-
Fall
-
-
-
-
-
West
AQS
Winter
14,550
-2.1
5.3
-6.1
15.3
Spring
17,190
-4.0
6.1
-8.8
13.3
Summer
18,046
0.6
8.1
1.2
15.2
Fall
16,163
0.4
5.5
0.9
12.8
CASTNET
Winter
506
-3.4
5.6
-8.7
14.1
Spring
519
-5.7
6.6
-11.8
13.7
Summer
526
-5.3
8.1
-8.7
13.3
Fall
530
-2.2
4.7
-4.6
10.0
April 2022
2B-16 External Review Draft - Do Not Quote or Cite

-------
1
2
3
4
5
6
7
8
9
10
11
(a) NMB
Ohio Valley-
Upper Midwest-_
South-
NRockiesPlains-*
Westi















{_



ฆft*



ฆ











(b) MB
Fall Wtr Spr Sum
-50 to -40
-40 to -30
-