Responses to Additional Significant Comments on the 2020 Proposed Action on the Ozone National Ambient Air Quality Standards August 14, 2020;


Responses to Additional Significant Comments on the

2020 Proposed Action on the

Ozone National Ambient Air Quality Standards
(August 14, 2020; 85 FR 49830)

Docket Number EPA-HQ-OAR-2020-0279

U.S. Environmental Protection Agency
December 2020

-------
[This page intentionally left blank.]

11

-------
Table of Contents

Frequently Cited Documents	ii

I.	Introduction	1

II.	Comments on the Health and Welfare Effects Evidence	

III.	Comments on Quantitative Air Quality, Exposure and Risk Analyses	

IV.	Legal, Administrative, and Procedural Issues and Misplaced Comments	

V.	References	

Appendix A. List of Abbreviations and Acronyms

111

-------
Frequently Cited Documents

The following documents are frequently cited throughout the EPA's response to
comments, often by means of the short names listed below:

Integrated Science Assessment (ISA)

U.S. EPA (2020a). Integrated Science Assessment of Ozone and Related Photochemical

Oxidants (Final). U.S. Environmental Protection Agency, Washington, DC. EPA/600/R-
20/012. https://cfpub. epa.gov/ncea/isa/recordisplay. cfm ?deid=348522

Policy Assessment (PA)

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, EPA-452/R-20-001.

https://www.epa.gov/sites/production/files/2020-05/documents/o3-final_pa-05-29-
2 Ocompressed.pdf

Proposed Action (Proposal)

Review of the Ozone National Ambient Air Quality Standards: Proposed Action. 85 FR 49830,
August 14, 2020.

Final Action (NFA)

Review of the Ozone National Ambient Air Quality Standards: Final Action. To be published in
the Federal Register.

iv

-------
Responses to Significant Comments on the 2020 Proposed Decision on the
Review of the Ozone National Ambient Air Quality Standards

I. Introduction

This document, together with the Federal Register notice of final decision on the review
of the ozone (O3) national ambient air quality standards (NAAQS), presents the responses of the
Environmental Protection Agency (EPA) to some of the public comments received on the 2020
O3 NAAQS proposal notice (85 FR 49830, August 14, 2020). All significant issues raised in
timely public comments have been addressed. Where comments were submitted after the close of
the public comment period, the EPA has responded to the extent practicable. The responses
presented in this document are intended to address comments not discussed in the final decision
notice. Although portions of the final decision may be paraphrased in this RTC document, to the
extent such paraphrasing introduces any confusion or apparent inconsistency, the preamble itself
remains the definitive statement of the rationale for the decisions in the final action. This
document, together with the preamble to the O3 NAAQS final decision notice and the
information contained in the Integrated Science Assessment (ISA, U.S. EPA, 2020a), the Policy
Assessment (PA, U.S. EPA, 2020b), and related technical support documents, should be
considered collectively as the EPA's response to all of the significant comments submitted on the
EPA's 2020 O3 NAAQS proposal.1

Sections II and III address public comments related to the air quality criteria (and ISA)
and the quantitative analyses (of air quality, exposure and risk), respectively. Section IV includes
responses to legal, administrative, procedural, or misplaced comments.

1 The docket for this review is EPA-HQ-OAR-2018-0279.

1

-------
II. Comments on the Health and Welfare Effects Evidence

This section addresses public comments related to the EPA's Integrated Science
Assessment for Ozone and Related Photochemical Oxidants that are not addressed in the final
decision notice.

A. Health Effects Evidence

This section addresses significant comments on the EPA's interpretation of the health
effects evidence.

(1) Comment: Some commenters disagree with the EPA's causality determinations for O3
exposure and cardiovascular effects and with total mortality, stating that the conclusions
should be reevaluated. Some of these commenters express the view that the evidence is
inadequate for these endpoints.2 The commenters that express the view that the evidence is
inadequate variously state that the term "suggestive" implies that a causal association is more
likely than not, when they claim "this is clearly not the case."

Response: The EPA disagrees with these commenters that the causality determinations for
short-term O3 exposure and cardiovascular effects or total mortality should be "inadequate".
The EPA has integrated the available evidence from animal toxicology, epidemiologic and
controlled human exposure studies for each of these endpoints, and after applying the
causality framework in a manner consistent with how it has applied it to other endpoints, has
determined that the body of evidence is "suggestive of, but not sufficient to infer, a causal
relationship" for each of these endpoints.

With regard to cardiovascular effects, the ISA notes that the evidence from both animal
toxicological and epidemiologic studies is either "limited" or "consistent" depending on the
endpoint evaluated and that the evidence from controlled human exposure studies is either
"limited", "inconsistent" or provides evidence of no effect (ISA, Section 4.1.17, Table 4-1).
Taken together, this evidence is consistent with the way that EPA applied the causality
framework to all other endpoints, which states that "Evidence is suggestive of a causal
relationship with relevant pollutant exposures but is limited, and chance, confounding and
other biases cannot be ruled out" (U.S. EPA, 2015, Table II).

With regard to total mortality, the evidence base includes primarily epidemiologic studies,
which, consistent with the 2006 AQCD and the 2013 ISA, continue to provide consistent,
positive associations between short-term O3 exposure and total mortality (U.S. EPA, 2006;
U.S. EPA, 2013; ISA, Section 6.1.8). However, the evidence for biological plausibility for O3
to result in mortality, especially cardiovascular mortality, which comprises a large percentage
of total mortality, remains limited. In addition, while recent studies have further examined
potential confounding by copollutants and other variables, confounding cannot be ruled out.
Thus, the evidence base for short-term O3 exposure and total mortality is limited, and chance,
confounding, and other biases cannot be ruled out, making the "suggestive of, but not

2 Comments expressing the view that the evidence supports a conclusion of likely causality are addressed
in the NFA.

-------
sufficient to infer, a causal relationship" determination appropriate under the causality
framework.

Finally, the EPA disagrees with the commenters that the term "suggestive" implies that a
causal association is more likely than not. The term "suggestive" is clearly defined in the
presentation of the causality framework in the Preamble to the ISA as limited evidence for
which chance, confounding, and other biases cannot be ruled out (U.S. EPA, 2015).

(2) Comment: One comment states that, in discussing the health research in the current review,
the ISA dwells on exposure uncertainties (e.g., measurement error, activity patterns) and
ignores "more certain" underestimation of health effects, especially with regard to respiratory
health effects. The commenter seems to be claiming that an underestimation of respiratory
effects is "more certain" than either an under- or over-estimation. The basis provided by the
commenter is that epidemiologic studies focused on severe outcomes such as mortality,
hospitalizations, and emergency department visits reflect just a fraction of relevant effects.
The commenter states that ignoring the other effects may bias estimates of O3 risk toward the
null.

Response: The EPA disagrees with this comment that too much weight has been placed on
the uncertainties related to exposure estimation in its assessment of epidemiologic studies.
The implications of exposure measurement error on the results of epidemiologic studies and
a characterization of the uncertainties due to such exposure measurement error are detailed in
Section 2.6, as well as Tables 2.6 and 2.7 of the ISA. As stated in the ISA, "[t]he importance
of exposure measurement error depends on the spatial and temporal aspects of the study
design" (ISA, Section 2.6), which tend to vary based on whether short- or long-term
exposures are being examined. When evaluating both short- and long-term exposure studies,
the evidence demonstrates that "[f]or epidemiologic studies of short-term exposure to ozone,
the effect estimates potentially have decreased precision and negative bias" (ISA, p. 2-55)
and that "[f]or epidemiologic studies of long-term exposure to ozone, when concentrations
measured at fixed-site monitors are used as exposure surrogates, effect estimates have the
potential to be biased in either direction" (ISA, p. 2-56). We additionally note that there is
greater uncertainty for the results of long-term O3 exposure studies, for which effect
estimates may be biased in either direction, than for the results of short-term O3 exposure
studies, for which effect estimates are bias toward the null. Recognition of this greater
uncertainty is reflected in the different causality determinations for short- and long-term
respiratory effects, which are causal and likely to be causal, respectively (ISA, Appendix 3).
The potential for underestimating the risk of respiratory health effects is greater for
epidemiologic studies of short-term O3 exposure compared to long-term O3 exposure and is
not dependent on the severity of the respiratory effect, but instead on the study design, as
reflected in Tables 2.6 and 2.7 of the ISA.

Further, we disagree that the ISA ignores the evidence for effects other than the more severe
effects, such as those emphasized by the commenter. The EPA considers the weight of
evidence across all respiratory endpoints, from sub-clinical effects measured in animal
toxicological, controlled human exposure and epidemiologic studies, to more severe
endpoints, such as ED visits, hospital admissions or mortality evaluated in epidemiologic
studies, and considers the strengths and limitations of the evidence from different scientific

3

-------
disciplines and study designs. Bias due to exposure measurement error is a single limitation
that is considered when evaluating epidemiologic evidence.

(3)	Comment: A comment stated that ISA presentation of recent studies should, and does not,
facilitate evaluation with regard to informing level and averaging time of standard.
Commenters suggest that the final ISA should include additional tables that facilitate
evaluation of studies with respect to informing the level and averaging time that would
protect public health.

Response: The EPA disagrees with the premise of the comment that the ISA does not present
information useful to considering the adequacy of the standard in all its elements, while
noting that the PA with its policy-relevant evaluations builds on the evidence in the ISA with
quantitative exposure and risk analyses and identification of policy-relevant considerations.
As described in the PA and summarized in the Notice of Final Action (NFA), the exposure
and risk analyses builds on information from controlled human exposure studies assessed in
the ISA to characterize exposure and risk under air quality conditions that just meet the
current standard, in all its elements (and two alternative standards). By their very designs,
these analyses consider the effectiveness of the current form and averaging time in providing
protection for effects documented in the controlled human exposure studies. The evidence
from these studies is considered the "strongest evidence" that O3 causes respiratory effects
(ISA, p. IS-1).

To the extent that this comment is asserting that the ISA should include additional tables that
facilitate evaluation of epidemiologic studies with respect to considering the level and
averaging time of the standard, the EPA disagrees. While epidemiologic studies evaluate the
relationship between health effects and specific ambient air O3 concentrations during a
defined study period and the generally consistent and coherent associations observed in these
epidemiologic studies contribute to the causality determinations and the conclusions
regarding the causal nature of the effect of O3 exposure on health effects, "they do not
provide information about which averaging times or exposure metrics may be eliciting the
health effects under study" (ISA, section IS.6.1, p. IS-87). In general, the epidemiologic
studies using ambient air quality measurements provide less information on details of the
specific O3 exposure circumstances that may be eliciting health effects, and whether these
occur under air quality conditions that meet the current standard. For example, when
considering 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 (ISA, Appendix 3,
Tables 3-13, 3-14, 3-39, 3-41, 3-42 and Appendix 6, Tables 6-5 and 6-8; PA, Appendix 3B,
Table 3B-1). When considering other lines of health evidence, such as that from controlled
human exposure studies, the EPA has compiled the evidence from relevant studies into
figures which can be used to inform the level and averaging time of the standard that would
protect public health (e.g., see ISA, Figure 3-3).

(4)	Comment: The ISA failed to include a relevant study (Bell et al., 2014) which describes
increased risk of emergency department visits, hospital admissions, and mortality to older
adults.

4

-------
Response: The study by Bell et al. (2014) is a systematic review and meta-analysis, which
does not include any original results. As such, it was purposely excluded from the ISA
consistent with the Population, Exposure, Comparator, Outcome, Study Design (PECOS)
statement, which does not include reviews or meta-analyses as study designs to be considered
(ISA, p. 3-4, p. 6-4).

(5) Comment: The EPA did not and should have identified people with any lung impairment
(other than asthma) as at risk.

Response: The EPA can only make determinations about groups at increased risk when there
is evidence available to evaluate and support such a determination. In the evaluation in the
ISA of the available evidence, the EPA evaluated populations potentially at increased risk of
health effects due to O3 exposure for which evidence was available. With regard to pre-
existing respiratory disease, this included those with asthma or COPD. In fact, the evidence
supported a conclusion that populations with asthma are at increased risk, though there was
inadequate evidence to make any determination for populations with COPD (ISA, Table IS-
10). As no evidence was identified comparing the health effects associated with O3 among
those with other lung impairments (other than asthma or COPD) to those without such
impairments, no determinations could be made for people with any other lung impairment.

(6) Comment: One comment stated that the ISA does not recognize how anthropogenic O3 has
been controlled.

Response: Section 108 of the Clean Air Act requires the EPA to issue air quality criteria,
which 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 such
pollutant in the ambient air. Consistent with Section 108, the ISA describes precursor
emissions, precursor chemistry, and meteorological effects and related scientific information,
but does not describe how this information has been used to date for air quality management
decisions. Rather, the discussion in the ISA of the atmospheric sciences of the subject criteria
air pollutant provides a succinct description of general context for the discussion of human
and ecosystems exposure, human dosimetry, and epidemiology.

The general photochemistry of urban-scale O3 is understood and well-documented in the
previous O3 ISA and preceding assessments. These sources were cited in the Appendix 1 of
the ISA as providing the details to support the overview discussions of precursor emissions
and their trends, atmospheric chemistry and observed ambient air concentrations and their
trends provided as necessary context for the discussion of U.S. background (USB) O3,3 and
in support of the central discussion of the ISA, i.e. the health and welfare impacts of O3 in
ambient air. Beyond that, the ISA atmospheric sciences discussion was focused on new
developments in scientific understanding of the sources and formation of USB O3.

The commenters misrepresented a relatively minor group of sources as total anthropogenic
volatile organic compounds (VOCs). What is stated in the ISA is: "U.S. industrial and related

3 Concentrations of O3 in ambient air that result from natural and non-U.S. anthropogenic sources are
collectively referred to as USB.

-------
VOC emissions have increased by approximately 20% since 2012, while other anthropogenic
emissions have declined over the same period" (ISA, Appendix 1, p. 1-14). This decline is
described graphically in Figure 1-3, where it is clear that the dominant sources of VOC's,
highway and non-highway vehicles, have steadily declined from 2011 to 2014 to 2017, and
that the industrial sources that show increases are relatively minor in comparison.

The analysis provided by the commenters showing New York City biogenic VOC
contributions are lower than the national average overlooks some important context and
complexity for regional 03 chemistry. The EPA-funded "Northeast Ozone and Particulate
Study,"4 now more than 15 years old, clearly showed with light detection and ranging
(LIDAR) and balloon measurements that high 03 episodes in the Northeast U.S. were largely
due to downmixing of O3 formed and transported aloft, rather than from local ground level
precursors. This study was completed well before completion of the 2013 ISA. The overview
presented in the current ISA as important reference material to provide context for the recent
major advances in 03 chemistry (i.e. winter O3, halogen chemistry, and compression of the
O3 concentration distributions) provides citations to the 2013 ISA and earlier Air Quality
Criteria documents in which results from these and other studies on regional Northeast O3
studies are individually cited (ISA, section 1.4; 2013 ISA, section 3.2).

We agree with the comment that anthropogenic VOC sources are important to O3
concentrations in urban areas and downwind of VOC sources. Excluding oil and gas, which
is not quantified in the figure provided by the commenter, and making allowance for
terminology differences, the three greatest anthropogenic VOC sources - highway vehicles
(on-road mobile sources), non-highway vehicles (non-road mobile sources), and solvent use
are shown to be the same in the ISA Figure 1-2 (trends plot), ISA Figure 1-3 (pie chart), and
the commenter's figure. Since this pattern is usual for many urban areas, and we did not find
there to be sufficient evidence to support a separate focus on the topic of anthropogenic
VOC's as relevant precursors in urban areas, such as New York, with proportionally low
biogenic contributions, such separate presentations were not included.

(7) Comment: One commenter stated that the ISA does not address all of the questions for the
ISA in the Integrated Review Plan (IRP) including "what data are available to characterize
precursor emissions of non-background O3" and "how does recent evidence contribute to
what is known about photochemical production of non-background O3"

Response: The statement that the ISA does not address the question of what data are
available to characterize precursor emissions of non-background O3 is not correct. This is
exactly what Section 1.3.1.1 "Ozone precursor emissions: anthropogenic sources and trends
in the U.S." is intended to address (ISA, section 1.3.1.1). The U.S. National Emissions
Inventory, the primary source of information used in this section to describe the sources and
trends of ozone precursor emissions within the U.S., represents the synthesis of the best

4 The results of this study appear in 20 peer-reviewed publications; citations for these publications can be
found here:

https://cfpub.epa.gOv/ncer_abstracts/index.cfm/fuseaction/display.publications/abstract/909#22732

6

-------
available data on emissions, emissions factors, and source activity factors. The statement that
the ISA does not address recent evidence on what is known about photochemical production
of non-background O3 is incorrect. Section 1.4 of the ISA ("Ozone Photochemistry")
addresses this topic. The section includes two extensive subsections on the major new
research areas of winter O3 formation and halogen chemistry (ISA, Appendix 1, section 1.4).
Additionally, the commenter is correct in noting that the concentrations of O3 coming near to
or exceeding the NAAQS at urban scales is largely formed from precursors emitted at local
scales. This point is emphasized in the ISA discussion of the importance of the USB O3
contributions to total ground level O3 in urban settings (ISA, Appendix 1, section 1.8.2).

(8) Comment: One commenter expresses the view that the finding of a 20% increase in industrial
and related VOCs since 2012 should be studied. In addition, the commenter stated that the
effects of VOC releases associated with natural gas extraction activities on ambient air O3 in
regions other than the Uinta Basin (i.e., Wyoming's Upper Green River Basin, the Colorado
Front Range, and Pennsylvania) are also in need of examination.

Response: The ISA presents a graphical description of the national-scale trends in VOC
emissions from industrial, petroleum, and related industries (ISA, Appendix 1, Figure 1-3C).
Sector-based emissions at regional scales are readily available from the NEI as cited in the
ISA. Studies of oil and gas-derived O3 production in the Upper Green River Basin are
discussed in the same section as those addressing the Uinta Basin (ISA, Appendix 1, section
1.4.1). Figure 1-10 of Appendix 1 of the ISA shows the average O3 concentration trends for
each region, including a decreasing trend for the Northeast. Figure 1-11 of Appendix 1 of the
ISA, derived from the Air Quality Monitoring System database, provides detailed
information concerning trends in the 4th highest daily maximum 8-hour average
concentration O3 trends at monitoring sites throughout the Northeast Corridor between 2008-
2010 and 2015-2017. Given the objective of the air quality section of the ISA (see the
response to the previous comment), the presentation of annual emissions, their trends, and
examples of the influence of emissions from industrial, petroleum, and related industries on
ambient air O3 concentrations is considered sufficient.

B. Welfare Effects Evidence

This section addresses significant comments on the EPA's interpretation of the welfare

effects evidence.

(1) Comment: In support of their views opposing the EPA's use of a 3-year average W126 index
of 17 ppm-hrs as a target for assessing protection provided by the current standard, some
commenters cite a meta-analysis by Wittig et al. (2009), also cited in comments addressed
with the 2015 decision, stating that RBL estimates based on the established E-R functions for
seedlings of 11 tree species "do not stand on their own" and are supported by studies such as
Wittig et al., (2009). In making this statement, they quote a statement from the ISA ("A
meta-analysis by Wittig et al. (2009) found that average O3 exposure of 40 ppb significantly
decreased annual total biomass by 7% across 263 studies").

Response: We disagree with implication of the commenters' statement that the numbers in
the ISA sentence cited by the commenters relate directly to the established E-R functions or

-------
their supporting data. Rather, the analysis by Wittig et al. (2009) provides support to the ISA
determination of a causal relationship between O3 exposure and reduced plant growth (ISA,
section 8.3.4). But it does not provide information that can be directly compared to RBL
estimates based on the established E-R functions or to their supporting data. This is because
it is a very different and not directly applicable type of study. The study by Wittig et al.
(2009) is not itself a carefully controlled experimental exposure study such as those on which
the E-R functions are based. Rather, this study is a meta-analysis that groups together a broad
array of controlled exposure studies of widely varying exposure circumstances and tested
species.

The reference to a 7% decrease in annual total biomass refers to the central tendency (based
on analysis of 263 publications5 reporting on O3 exposures of widely varying duration and
magnitude) of the total biomass difference estimated between experiments involving tree
exposure to charcoal-filtered air and exposure to untreated ambient air. The 40 ppb "average
ozone exposure" refers to the overall average of the concentrations during all the hours
across all study "ambient air" exposures. The exposures in these studies varied from four to
24 hours per day across seven to 365 days. The study authors characterize the value of 40
ppb average exposure (in comparison to negligible O3 in charcoal filtered treatments) as "a
measure of how the elevation of [O3] that has occurred since the Industrial Revolution has
reduced tree productivity" (Wittig et al., 2009). Not only is the 40 ppb estimate not directly
comparable to values for commonly cited O3 metrics, including metrics commonly used to
represent cumulative exposures for assessing growth effects, the value is derived from
studies of multiple species that involved differing numbers of hours exposure per day (e.g.,
potentially from four to 24 hours) and differing exposure durations (e.g., potentially from
seven days to more than a year). Thus, we disagree with the commenters that the Wittig et al.
(2009) study provides any quantitative support to the studies on which the E-R functions are
based.

(2) Comment: In expressing their disagreement with EPA's conclusions regarding the potential
impact of considering the W126 index as a 3-year average rather than each single year and
with a description in the PA (Appendix 4A, pp. 4A-22, 4A-23) of the evaluation of the
predicted growth impacts compared to observations from the multiyear study of O3 impact on
aspen by King et al. (2005), as presented in the 2013 and 2020 IS As, one commenter objects
to the observation made in the 2020 ISA (and repeated in the PA) that aspen growth
observations (from a FACE multiyear O3 exposure study) are "exceptionally close" to
predictions based on single year W126 index (ISA, p. 8-192). The commenter states that,
based on their analysis of 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 is approximately 0.7. This commenter additionally states that over the years of the
study, the predictions (in the 2020 ISA) increasingly underestimate the observed biomass,
describing the percent difference for the last year as approximately 40%, implying that that
analysis is lacking in its ability to predict cumulative impacts of the multi-year exposure.

5 Given the 99 degrees of freedom for the central tendency, less than or equal to 99 publications may be
of greatest influence (Wittig et al., 2009).

-------
This commenter suggests that the predictions are ignoring the cumulative growth effect of
03.

Response: As an initial matter, we note that the aspect of the commenter's statements
specific to the PA Appendix 4A, section 4A.3 (e.g., regarding cumulative growth effects
and size of the tree) are addressed in section III.B(l) below.6 Further, we note that the
intention of the reference in the PA Appendix 4A to the ISA was in part to indicate that
similar conclusions with regard to use of a function derived to describe growth response
for a single season across multiple years have been reached by both ISAs based on
slightly different analyses (2013 ISA, pp. 9-133 to 9-135; 2020 ISA, Appendix 8, pp. 8-
192 to 8-193). The commenter describes the phrase, "exceptionally close," used in the
2020 ISA (ISA, p. 8-192) to describe its analysis, as inconsistent with specific
quantitative estimates they provide based on their interpretation of the ISA's Figure 8-17
(in which the analysis utilized single-year W126 index). We note that the section of the
PA Appendix 4A for which the commenter raises concerns also summarized the findings
of the 2013 ISA analysis (on p. 4A-22 of the PA). In consideration of the calculations
offered by the commenter with regard to Figure 8-17 in the ISA, we have used the values
reported in Table 9-15 of the 2013 ISA (which are plotted in Figure 9-20), to derive
correlation coefficients related to that analysis. The r2 for predicted O3 impact versus
observed impact is 0.99 and for the percent difference versus year is approximately 0.85.
We note that this indicates a much better correlation from the 2013 ISA analysis
(compared to 2020 analysis, based on commenter's estimates) of the same observations
with predictions based on a cumulative multiyear W126 index, suggesting a better fit for
the exposure metric reflecting cumulative multiyear exposure.

III. Comments on Quantitative Air Quality, Exposure and Risk Analyses

A. Population Exposure and Health Risk Analyses

This section addresses significant comments on the EPA's quantitative exposure and risk
analyses (presented in detail in the PA, Appendices 3C and 3D) that are not addressed in the final
decision notice.

(1) Comment: One comment resubmitted a comment on the draft PA that expressed the view that
the PA was constrained from conducting a comprehensive review of studies due to an
accelerated timeline. In support of their claim the commenter cites a statement on p. 3D-146
of the Draft PA that "[g]iven the limited time schedule for review, we evaluated the
contribution to risk from low [ozonejexposures using only three of the eight study areas,
selected at random (i.e., Atlanta, Dallas, and St. Louis) and for a single year (2016)" and
imply that for the main exposure and risk analysis, the EPA randomly selected three study
cities that the commenter states were not representative of the various regions of the U.S. nor
representative of hourly O3 distributions.

6 As noted in section III.B(l) below, the simple example presented in Appendix 4A of PA was not
recognized as a consideration in the Administrator's decision in this review.

-------
Response: As an initial matter, we note that the section of the draft PA cited by the
commenter was describing a sensitivity analysis rather than the full risk analysis. As is
common for sensitivity analyses, this analysis focuses on a specific aspect of the full analysis
and explores the influence of certain parameters in a more limited context than the full
analysis. That is what is described by the section of the draft PA quoted by the commenters.
This analysis and the reasons for focusing on three study areas are also described in the final
PA (PA, Appendix 3D, p. 3D-165). Thus, we disagree with the commenters' suggestion that
the main risk and exposure analysis was incomplete or based on only three non-
representative study areas. The main risk and exposure analysis provided comprehensive
details on the data, tools, and methods employed, specific results based on those data, tools,
and methods, and the sensitivity, uncertainty and variability analyses conducted. In making
this comment the commenter referred to a sensitivity analysis that was not part of the results
for the primary analysis. For the main risk exposure analysis, we analyzed eight study areas
(PA, appendix 3D) which reflect the full range of air quality and exposure variation expected
across major urban areas in the U.S and seven different NOAA climate regions (PA, section
3.4.1).

The sensitivity analysis cited by the commenter informed the characterization of uncertainty
associated with the full analysis (PA, section 3.4, Appendix 3D, section 3D.3.4). This
analysis was focused on evaluating the contribution to estimates of lung function risk (by
both of the approaches used for such estimates) resulting from the lower exposure
concentrations in light of the lesser data (between 50 ppb and 60 ppb) or lack of data (less
than 40 ppb) for these low levels (PA, Appendix 3D, section 3D.3.4.2.3). To perform the
analyses for the E-R function approach that were the basis for the main lung function risk,
results reported in section 3D.3.3 of the PA were in a format useful for calculating the risk
contribution from each 7-hr average exposure bin (0 to 160 ppb, in 10 ppb increments). Thus
no new APEX simulations were needed for this evaluation. However, for the evaluations of
the MSS model approach, new simulations were required. For purposes of efficiency and
given the objectives for this evaluation, we focused on three of the eight study areas for this
evaluation. These areas were selected at random (i.e., Atlanta, Dallas, and St. Louis), and
simulations were performed for three air quality scenarios using a single year (2016) of data;
the analysis focused on the risk contribution to lung function decrements occurring at least
one and two days per year (PA, Appendix 3D, section 3D.3.4.2.3). These analyses were
informative to the PA consideration of the lung function risk estimates by the two approaches
employed. As discussed in the PA, these analyses indicated estimates via both approaches to
have contributions from low exposures concentrations for which there are fewer data and the
MSS model estimates to have appreciable such concentrations. These findings and other
factors, as discussed in the PA and proposal contributed to the lesser weight placed on these
estimates relative to estimates from the comparison-to-benchmarks analysis (PA, section
3.4.4; 85 FR 49857-49859, 49871-72, August 14, 2020).

(2) Comment: Some commenters express the view that the EPA failed to demonstrate that the
APEX model provides "meaningful" estimates of population-level exposures or individual-
level exposures. In so doing, they claim that APEX uses age, gender, race, and work status to
establish user profiles and that the EPA does not provide analyses to establish that these
variables alone or in combination reasonably and accurately explain the amount of time an
individual spends outdoors at moderate exertion (e.g., "data [APEX] uses... have not been

10

-------
shown to reflect" variables correlating with people's outdoor physical activity) or meaningful
individual exposure estimates).

Response: As an initial matter, contrary to the implication or assertion of the commenters the
EPA is not relying on age, gender, race, and work status as sole basis for estimation of
population exposure in the study areas. As described in the PA, population exposures were
estimated using the APEX, version 5 model, which probabilistically generates a large sample
of hypothetical individuals from a population database 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 (PA, section 3.4.1 and Appendix
3D). The variables age, gender, race, and work status are used to identify a sample activity
pattern to draw from the CHAD for simulation in APEX. In this way, the APEX creates,
through stochastic sampling, a population with age, gender, race, and work status reflective
of the specific study area. 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 ventilation rate (PA, Appendix 3D, section 3D.2). For
example, each modeled person is assigned anthropometric and physiological attributes by
APEX, as described in detail in Appendix 3D of the PA, and associated attachments and
cited references. All of these variables are treated probabilistically, accounting for
interdependences where possible, and reflecting variability in the population (PA, Appendix
3D, section 3D.2.2.3). Such variables include those that influence ventilation rate.

As explained in the sentences that the commenter only partially quotes conveys, "while a
single profile does not, in isolation, provide information about the study population, a
distribution of profiles represents a random sample drawn from the study area population. As
such, the statistical properties of the distribution of simulated profiles are meant to reflect
statistical properties of the population in the study area." (PA, Appendix 3D, p. 3D-20).

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 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 (PA, Appendix 3D, section 3D.2). The analyses described in the PA, include a
variety of updates from prior such assessments, including new statistical distributions for
estimating body weight, equations for estimating resting metabolic rate, and equations for
estimating activity-specific ventilation rate (PA, Appendix 3D).

Thus, the EPA disagrees with the commenter's implication that the APEX model does not
provide "meaningful" estimates of population-level exposures or individual-level exposures
with respect to time spent outdoors and physical activity data used in estimating exposures
for the exposure and risk analyses. The APEX model is not focused on exposures for specific
individuals. Rather it is intended to estimate distributions of population-level exposures.
APEX probabilistically generates a large sample of hypothetical individuals from a
population database and simulates each individual's movements through time and space to

11

-------
estimate their time-series of O3 exposures occurring within indoor, outdoor, and in-vehicle
microenvironments (PA, Appendix 3D, section 3D.2). 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. By incorporating individual activity patterns from the CHAD,7 the
model estimates physical exertion associated with each exposure event. After the basic
demographic variables are identified by APEX for a simulated individual in the study area,
values for the other variables are selected as well as the development of the activity patterns
that account for the places the simulated individual visits (such as indoors or outdoors) and
the exertion level of activities they perform (PA, appendix 3D). For example, to account for
the variability in activity patterns dependent on age, such as time spent outdoors and
associated activities performed, the APEX model assigns age-specific diaries from CHAD to
simulate age-specific individuals' locations visited and physical activities performed and thus
accounts for when time expenditure varies due to influential individual attributes such as age.
APEX is built upon accepted first principles and has high quality input data and therefore the
results are meaningful. The "meaningfulness" of the results of the risk and exposure analysis
is further supported by our consideration of the uncertainties associated with the quantitative
estimates of exposure and risk, including those recognized by the characterization of
uncertainty in Appendix 3D of the PA (PA, Appendix 3D, section 3D.3.4). This is further
supported by an analysis of the uncertainty of APEX when used to model 03 exposures
(Langstaff, 2007) and an evaluation of APEX in microenvironments (Johnson et al., 2018).

(3) Comment: One commenter states that the APEX model (used in the exposure analysis)
underestimates weather-dependent behavior changes (e.g., playing outside when the
temperature is warm), which they argue results in the exposure analyses scenarios including
too few children breathing at elevated respiratory rates.

Response: The EPA disagrees with the comment. In support of their view, the commenter
provides descriptive statements about children's habits for playing outdoors in various
locations every day, weather permitting, and does not also provide any evidence on which the
statements are based. The exposure assessment, however, is based on extensive data and
established methods. The APEX model used in the assessment 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
involving different levels of exertion on breathing rate for the exposed individuals of
different sex, age, and body mass in the study area (PA, Appendix 3D, section 3D.2). To

7 To represent personal time-location-activity patterns of simulated individuals, the APEX model draws
from the consolidated human activity database (CHAD) developed and maintained by the EPA
(McCurdy, 2000; U.S. EPA, 2019a). The CHAD provides data on human activities through a database
system of human diaries or daily time series or daily time location activity logs collected in surveys at
city, state, and national levels. Included are personal attributes of survey participants (e.g., age, sex),
along with the locations they visited, activities performed throughout a day, time-of-day the activities
occurred and activity duration (PA, Appendix 3D, section 3D.2.5.1).

12

-------
represent personal time-location-activity patterns of simulated individuals, the APEX model
draws from CHAD, which provides time series data on human activities through a database
system of collected human diaries, or daily time location activity logs. This information
indicates that children spend more time outdoors than all other age groups while at elevated
exertion, and consistently do so when considering the most important influential factors such
as day-of-week and outdoor temperature, all of which APEX takes into account. The APEX
model stochastically generates a user-specified number of simulated people to represent the
population in the study area. The number of simulated individuals can vary and is dependent
on the size of the population to be represented. For the current analysis, the number of
simulated individuals was set at 60,000 for each of the children and adult study groups
(which includes people with asthma for both of these study groups) to represent population
residing within each study area (i.e., between 2 and 10 million). While precisely 60,000
children and 60,000 adults were simulated as part of each APEX model run, the number of
individuals estimated to be exposed are appropriately weighted to reflect the actual
population residing within the census tracts that comprise each respective study area. 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, children spend about 2 hours of afternoon time outdoors per day and 80% of which is at
elevated exertion levels (PA, section 3.3.2 and Appendix 3D, section 3D.2.5.3 and Figure
3D9, p. 3D-56).

(4) Comment: Some commenters expressed the view that the exposure assessment "irrationally
and arbitrarily" presumes that people with asthma will experience identical lung function
decrements as healthy individuals at a given exposures of concern.

Response: As an initial matter, we note that the exposure assessment does not make any
presumptions with regard to lung function decrements; it simply estimates exposure.
Accordingly, we assume that the commenter is referring to the lung function risk assessment.
As described in the PA, that assessment uses the APEX model estimates of population
exposures for simulated individuals with 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% (PA, section 3.4).
These estimates are derived by two different approaches, one using E-R functions and the
other based on a more mechanistic model (the MSS model). In both cases, the underlying
data are from controlled human exposure studies comprised of largely healthy subjects. As
recognized in the PA, the evidence is quite limited with regard to such studies conducted
with subjects with asthma. Further, as described in the ISA, the evidence that does exist
indicates generally similar responses for the two populations (ISA, Appendix 3, section
3.1.5.4.4).

(5) Comment: Some commenters claim that the exposure and risk analysis use of the National
Health Interview Survey (NIHS) responses biased the size of vulnerable populations
downward by asserting that the EPA only considered adults who responded that they "still
have asthma" versus including respondents who had "ever" had an asthma diagnosis.

-------
Response: We disagree that our approach significantly biased the size of sensitive
populations with asthma in the exposure and risk analysis. We note that focusing on adults
and children who "still have asthma," is consistent with the characterization of asthma
prevalence in the ISA (ISA, Table IS-11), and as such, this approach provides us with the
most appropriate estimate of the population of individuals that have asthma (PA, appendix
3D, p. 3D-150). As discussed in the uncertainty characterization documented in the PA,
based on analyses also documented there, we conclude that using the response for the "Still"
question may underestimate asthma prevalence for those not having a physician determined
diagnosis. Those analyses indicate such an underestimation to be at most about 0.6
percentage points (i.e., with an "ever had" variable, the overall 'current' asthma prevalence
for children would be about 9.0% rather than the 8.4% used in the simulations), as discussed
in detail in the PA (PA, appendix 3D, Attachment 1). Thus, while it is likely that using the
response for the "Still" question underestimates asthma prevalence for those not having a
physician determined diagnosis, the magnitude of underestimation is likely quite small. We
note that with regard to asthma prevalence, the data are used to identify if a simulated
individual residing within a modeled census geographic area has asthma. The data are not
used for selection of any other personal attribute nor in the selection of activity pattern data
(PA, Appendix 3D, section 3D.2.2.1).

(6) Comment: One commenter claimed that the EPA failed to consider several issues in exposure
assessment and as such, they imply that reliance on this analysis is unlawful. In particular,
the commenter asserts that the EPA did not adequately consider averting behavior, the
correlation between asthma and race, and that the EPA failed to perform the quantitative
uncertainty analysis recommended by CASAC on the draft PA. As a basis for their claim on
averting behavior, they state that in contrast to the 2014 HREA, the EPA arbitrarily failed to
discuss how averting behavior influences the activity diaries in the APEX modeling in the
current review and as such the exposure estimates may be inaccurate. With regard to the
issue on asthma and race, the commenter holds that the EPA failed to address the correlations
between asthma and race in attributing asthma prevalence to simulated study populations.
With respect to the comment on CASAC and the uncertainty analysis, they express the view
that the EPA failed to provide uncertainty bounds on its exposure and risk estimates, noting
that the CASAC observed that the ranges of estimates "represent variability between cities,
not uncertainty."

Response: The EPA disagrees that the exposure assessment does not adequately consider the
issues raised by the commenter and consequently disagree with the view that reliance on
results of the assessment is unlawful. As an initial matter, we note that the NAAQS must "be
established at a level necessary to protect the health of persons," not the health of persons
refraining from normal activity or resorting to medical interventions to ward off adverse
effects of poor air quality (S. Rep. No. 91-1196, 91st Cong. 2d Sess. at 10). We additionally
note that the issue of averting behavior and potential modifying effect on O3 exposure
estimates was investigated in the 2014 HREA through a very limited sensitivity analysis.8

8 This analysis was limited to a single urban study area, a 2-day period, and a single air quality adjustment
scenario (2014 HREA, section 5.4.3.3 and Appendix 5G). The analyses include a number of limitations to
and uncertainties in the analyses with regard to simulating the averting behavior (e.g., prevalence and
duration of the behavior), given the lack of actual activity pattern data that explicitly incorporated this

-------
Based on the available information, including the limited analyses in the 2014 HREA, we
find it reasonable to conclude that overall averting behavior will likely have little effect on
the range of exposures generated from the APEX simulations.

With regard to the commenter's statement regarding correlations of asthma prevalence and
race, the PA provides appreciable discussion (e.g., PA, Appendix 3D, sections 3D.2.2.2 and
3D.3.4). As explained in the PA, there are personal attributes other than those used to stratify
asthma prevalence for use in the exposure and risk analysis that have been shown to
influence asthma prevalence, such as race, ethnicity, obesity, smoking, health insurance, and
activity level (e.g. Zahran & Bailey, 2013). The set of variables chosen to stratify asthma
prevalence for use in the exposure and risk analysis (i.e., age, sex, and family income level)
was based on maximizing the potential range in asthma prevalence variability, maximizing
the number of survey respondents comprising a representative subset study group, and
having the ability to link the set of attributes to variables within the U.S. Census population
demographic data sets. Of the additional influential factors identified here, race is perhaps the
only attribute common to both the prevalence and population data sets that could be an
important influential factor and was not directly used to calculate asthma prevalence.
However, the use of race in calculating asthma prevalence, either alone or in combination
with family income level, would further stratify the NHIS analytical data set and appreciably
reduce the number of individuals of specific age, sex, race, and family income level,
potentially reducing the confidence in calculated asthma prevalence based on having so few
data in a given stratification. Because family income level already strongly influences asthma
prevalence across all races and stratifies the NHIS data into only two subgroups (i.e., above
or below the poverty threshold) in comparison to the larger number of subgroups a race
variable might yield, family income was chosen as the next most important variable beyond
age and sex to rely on for weighting the spatial distribution of asthma prevalence.

With regard to the comment regarding the CASAC recommendation concerning quantitative
uncertainty analyses, as an initial matter, we agree that the PA presents estimates for each
study area (by simulated at-risk population) and note that those estimates are described in the
PA as simply that (i.e., they are not implied to represent uncertainty). With regard to the
uncertainty characterization provided in the PA, it does not include a completely quantitative
uncertainty analysis. We did not have the data necessary to complete a full probabilistic
uncertainty analysis (that would have required, for example, use of expert elicitation to
derive confidence levels for all key inputs). Rather, the PA includes a full and rigorous
analysis of key sources of uncertainty including discussion of the potential nature and
magnitude of the impact of individual sources of uncertainty on the estimates generated,
following an approach that is employed in exposure/risk analyses conducted for NAAQS
reviews. The mainly qualitative approach used in this and other REAs, also informed by
quantitative sensitivity analyses, is described by WHO (2008). This approach identifies key

type of behavioral response (2014 HREA, p. 5-33). As noted in the 2014 HREA, "because most activity
diaries are limited to a single day and the survey participants were not directly asked if they altered their
daily activities in response to a high air pollution event, we do not know if any diary day represents the
activities of an individual who averted. Thus it is entirely possible that the 'no averting' simulation
includes, to an unknown extent, individuals who spent less time outdoors than would have occurred if
absolutely no individuals averted." (2014 HREA, p. 5-33).

15

-------
aspects of the assessment approach that may contribute to uncertainty in the exposure and
risk estimates and provided the rationale for the inclusion of such aspects (PA, Appendix 3D,
sections 3D.2.9.2 and 3D.3.4). The output of the uncertainty characterization is a summary
that describes, for each identified source of uncertainty, the magnitude of the impact and the
direction of influence the uncertainty may have on the exposure and risk results. The
uncertainty characterization in the PA additionally draws on uncertainties associated with
APEX exposure modeling that have been previously characterized in the REAs for nitrogen
dioxide (NO2), carbon monoxide (CO), and sulfur dioxide (SO2) conducted for recent
primary NAAQS reviews, along with other pollutant-specific issues (U.S. EPA, 2010, 2017b,
2018a), all complementary to quantitative uncertainty characterizations conducted for the
2007 O3 exposure assessment by Langstaff (2007). Conclusions drawn from each of these
characterizations have been considered in the current analyses in light of new information,
data, tools, and approaches used in this exposure and risk analysis. Thus, the PA includes full
discussion of uncertainties and an extensive table describing key uncertainties (PA pages 3D-
144 to 3D-158). Further, the characterization in the PA is also informed by an array of
quantitative sensitivity analyses. As described in the PA, a number of additional quantitative
analyses aimed at informing our characterization of uncertainty were added in consideration
of the CASAC advice. For example, in consideration of CASAC recommendations and
public comments, the PA exposure/risk analysis includes presentations reflecting further
analyses, investigations, and/or clarifications of the available data with regard to a number of
areas (listed below).

•	Analyses of data on outdoor activity by different population groups including those
identified as at risk in this review (e.g., children with asthma and older adults) during
times of day when O3 may be elevated (PA, section 3D.2.5.3);

•	Estimates for the comparison-to-benchmarks analysis additionally summarized in
light of the estimates from the last review (PA, section 3D.3.2.4);

•	Evaluation of risk characterization uncertainty related to its representation of
population groups having health conditions other than asthma, of older adults, and of
outdoor workers (PA, section 3D.3.4.1);

•	Evaluation of uncertainty in estimates for people with asthma that may be associated
with method for identifying individuals with asthma (PA, section 3D.3.4.1);

•	Evaluation of uncertainty with the E-R function and risk estimates (PA, section
3D.3.4.1);

•	Analyses investigating the sensitivity of the MSS model outputs to the value assigned
the individual variability parameter, and to low-level ventilation rates, as well as
overall model uncertainty in the MSS model (PA, section 3D.3.4.1).

(7) Comment: In claiming there to be problems with the EPA's exposure assessment, some
commenters state that the EPA did not adequately address a CASAC comment regarding
performance evaluations conducted for the air quality modeling in each of the study areas.
The CASAC letter on the draft PA stated that "it would be useful" for the EPA to conduct
additional model performance evaluations for the O3 precursors, NOx and VOC, suggesting
that if the precursor concentrations did not match the observations, the HDDM sensitivities
"may not be" accurate (Cox, 2020, Consensus Responses to Charge Questions p. 11).

16

-------
Response: Model evaluations presented in the PA focused on the O3 concentrations predicted
for the three air quality scenarios assessed primarily because the population exposure
assessment focuses on O3 concentrations (and does not utilize NOx and VOC
concentrations). In the model-based rollback, predicted model concentrations were not used
at all,9 and so an evaluation of predicted NOx or VOC concentrations individually would
have limited relevance to judgements related to confidence in the rollback outcomes. Since
the HDDM sensitivities reflect the model characterization of O3 production chemistry
derived from first principles, the relationship between NOx and VOC in the atmosphere
along with the representation of local meteorological conditions is what matters for the
predicted chemistry, rather than absolute concentrations of either species. Additionally,
ambient air monitoring data for VOC concentrations are not uniformly available, with very
few locations that include measurements are available for both NOx and VOC compounds. A
performance evaluation for both precursors would therefore only be possible for a very small
number of sites, thus making it also less useful or relevant for consideration across the set of
study areas assessed.

(8) Comment: In support of their disagreement with the exposure and risk analysis conclusions
related to the 70 ppb benchmark, one commenter claims that the "sample size" of the
exposure and risk analysis is limited ("extremely small") due to the focus on children in the
study areas breathing at an elevated rate. In so doing, they state that of the full simulated
populations for the eight study areas, about 2.3% are children with asthma (pointing to Table
3D-25 of the PA), and claim, based on an assumption that 79% of those children spend seven
hours outdoors at moderate or greater exertion, that 1.82% of the full simulated populations
would be children with asthma that are outdoors at exertion for seven hours a day.

Response: To the extent that the commenter intends to disagree with the premise of the
comparison to benchmarks analysis that the pertinent exposures for comparison are those
while at elevated exertion, we disagree. As discussed in the PA and the ISA, the controlled
human exposure studies on which the benchmarks are based evaluate effects of exposures
experienced while subjects are engaged in quasi-continuous exercise (6.6 hours including six
50-minute periods of moderate or greater exertion) (ISA, Appendix 3, section 3.1.4.1; PA,
sections 3.3.3 and 3.4). Further, to the extent that the commenter is claiming that the
exposure assessment only estimates exposure for a subset of the identified population in a
study area, we also disagree. In fact, the exposure assessment estimates population exposures
for simulated individuals in eight study areas. The "sample size" consists of 60,000 simulated
children in each of the eight study areas with the number of individuals estimated to be
exposed weighted to reflect the actual population residing within the census tracts that
comprise each respective study area. Thus, of the over 8.2 million children simulated, 10.8%
were children asthma which is somewhat higher than the actual national prevalence of 7.9%
children with asthma (PA, Table 3-1, appendix 3D, Table 3D-25). Lastly, the commenter
implies that the exposure assessment only considers exposures occurring outdoors. This is

9 The model-based adjustment approach employed to develop the air quality scenarios for the exposure
and risk analyses is described in detail in Appendix 3C of the PA (e.g., see PA, Appendix 3C, section
3C.5). As part of the methodology used, photochemical modeling results are not used in an absolute sense
but instead are applied to modulate ambient air measurements, thus tying estimated O3 distributions to
measured values (PA, Appendix 3C, section 3C.5.2).

17

-------
incorrect. As described in the PA, the exposure assessment focuses on exposures to O3 that
originated in ambient air, and such exposures include microenvironments that are both indoor
and outdoor locations (PA, section 3.4.1 and Appendix 3D, section 3D.2.6).

B. Environmental Exposure Analyses

This section addresses significant comments on quantitative environmental exposure

related analyses that are not addressed in the final decision notice.

(1) Comment: One commenter identifies what they describe as flaws in example analyses
included in the PA that use biomass measurements from the multiyear study by King et al.
(2005) to estimate above-ground aspen biomass over a multiyear period using the established
E-R function for aspen with a constant single-year W126 index, e.g., of 17 ppm-hrs, or with
varying annual W126 index values (10, 17 and 24 ppm-hrs) for which the 3-year average is
17 ppm-hrs (PA, Appendix 4A, section 4A.3). The first of the two criticisms concerning the
example analyses concerns application of the RBL (for each W126 index value) to an annual
growth increment derived from the "control" trees (represented in the analysis by the trees in
King et al (2005) exposed to unadjusted ambient air). The commenter states that this assumes
the effect of O3 in one year to have no effect on tree growth in a subsequent one, yet this
ignores that smaller trees (presumably including an O3 exposed tree compared to a control)
grow less than larger trees (Binkley et al., 2013) and ignores that this can explain long-term
negative cumulative effects of O3 on tree growth (Talhelm et al., 2014). The second criticism
of the analysis was the EPA's use of aspen, for which the E-R relationship is relatively linear.
The commenter claims that it does not reflect species with a more pronounced curve, such as
sugar maple. Lastly, the commenter claims that a statement regarding multiyear studies in the
PA summary of limitations for this analysis was "boldly incorrect."

Response: While the example analysis in the PA critiqued by the commenter was not, itself,
recognized as a consideration in the Administrator proposed decision on the secondary
standard (85 FR 49907-49913, August 14, 2020), or in his final decision (NFA, section
III.B.4), the EPA has considered the points raised by this commenter.

With regard to the first point, the EPA agrees with the general concept explained by the
commenter, essentially that an 03-exposed tree would have a smaller annual incremental
growth than a control tree due to its being of smaller size as a result of the O3 exposure in
prior year.10 We note, however, that the comparison in this analysis was between two 03-
exposed scenarios, and the point raised by the commenter has little impact on that
comparison in the PA example. This is illustrated in a revised version of the PA example,
provided in Appendix B of this document, in which the annual biomass increment is
estimated using a function of prior year biomass (derived from the measurements for the
control aspen in King et al. [2005]). As the commenter suggested, the difference of the O3-

10 As recognized in the ISA, "[i]t is well established that exposure duration influences the degree of plant
response and that ozone effects on plants are cumulative," "effects are clearly demonstrated to be related
to the cumulative exposure over the growing season for crops and herbaceous plant species," and "[f|or
long-lived plants, such as trees, exposures occur over multiple seasons and years" (ISA, p. ES-19). These
are reasons identified for the focus on using cumulative indices of exposure to assess vegetation exposure
(ISA, p. ES-19).

18

-------
exposed scenario from the control scenario is larger (than in the PA Appendix 4A example).
The difference between estimates for the different patterns of O3 exposure (the focus of the
example analysis), however, is still relatively small (a fraction of a percent). Further, we note
that the varying W126 values in the simple example include a W126 index value of 24 ppm-
hrs every third year. Yet the frequency of such a value for air quality meeting the current
standard is quite rare, as can be seen from the air quality analyses in the PA, which show that
across the period from 2000 through 2018 for even just a subset of sites meeting the current
standard, i.e., those with design value closest to 70 ppb (66-70), the 99th percentile is below
20 ppm-hrs (PA, Appendix 4D, Figure 4D-8). Focusing just on Class I areas (sites meeting
the current standard), there are no more than 15 occurrences of a single-year W126 index
value above 19 ppm-hrs in the entire time period (2000-2018) and all of those occurrences
date prior to 2013 (FR 85 49904, August 14, 2020). Thus, the varying scenario in the
example analyses generally represents much higher cumulative exposures that those
occurring across the U.S. at sites that meet the current standard, with a much greater
frequency of higher W126 index values. Specifically, they include as one of the three years
of air quality, a magnitude of W126 index that has been quite rarely observed in areas that
meet the current standard since 2000.

With regard to the second point, the EPA recognizes that the shape of the E-R relationship
may have the potential to influence the difference in the comparison analyzed. We note,
however, that the large change in slope of the relationship for sugar maple (the only species
for which such a pattern is seen in the 11 E-R functions) occurs at or above the highest W126
index values observed at U.S. ambient air monitoring sites (as well as the highest value used
in the PA example analysis, 24 ppm-hrs), as summarized above (PA, Appendix 4A, Figure
4A-1). For W126 index values at or below 19 ppm-hrs, the slope for sugar maple is much
smaller and RBL estimates associated with its E-R function are appreciably lower than those
for the aspen (PA, Appendix 4A, Figure 4A-1).11 Thus, while the shape of the sugar maple
differs substantially from the aspen at higher exposure conditions, there are much more
limited differences at exposures more common in U.S. areas where the current standard is
met. Accordingly, any difference that a sugar maple example might indicate from the aspen
example would not be significant. Further, sugar maple occurs in the northeast and upper
midwest of the U.S., areas with among the relatively lower W126 index levels across the
U.S. (PA, Appendix 4D, Figure 4D-2). Across the other 10 species for which there are
established E-R functions, nine of them have generally linear E-R relationships. The tenth,
black cherry, has a slope for the E-R function that presents the opposite pattern to that of red
maple (the slope of the curve slowly declines with increasing W126 index).

11 At sites and time periods during 2000 through 2018 in which the current standard was met, even
focusing on just the design values closest to the current standard (e.g., 66 -70 ppb), the sugar maple RBL
estimated for the 75th percentile is less than 1% (PA, Appendix 4A, Table 4A-4 and Appendix 4D, Figure
4D-8). It is also of note that the lowest nonzero O3 exposure represented in the dataset from which the
sugar maple E-R function is derived has been reported as a SUM06 of 25.2 ppm-hrs (a value that the 2007
Staff Paper indicated may be similar to a W126 index of 21 ppm-hrs (PA, Appendix 4A, Table 4A-6;
U.S. EPA 2007). This lack of an exposure representative of exposure levels more commonly occurring in
U.S. areas that meet the current standard contributes appreciable uncertainty to interpretations of sugar
maple RBL for exposure for air quality meeting the current standard.

-------
Lastly, in characterizing the EPA's statement regarding multiyear studies "boldly incorrect,"
the commenter has taken the statement out of the context intended by the EPA. The sentence
referenced by the commenter is as follows: "However, datasets of tree growth across
multiple-year periods such as that available for aspen in the study by King et al. (2005) are
not prevalent" (PA, p. 4A-23). The commenter apparently assumes the statement concerned
any multi-year tree growth study, of which the EPA agrees there are many. In the statement
cited, however, the EPA was referring to multiyear studies of growth that investigated the
impact of O3 as in the case of the study by King et al (2005) that the EPA's sentence cites.

IV. Legal, Administrative, and Procedural Issues and Misplaced Comments

A number of comments were received that addressed a wide range of issues including

legal, administrative, and procedural issues, as well as issues that are not germane to the review

of the NAAQS. Many legal issues are addressed generally throughout the notice of final action.

Specific responses to other comments are presented below

(1) Comment: Some commenters conveyed the view that the process followed by the EPA in this
review of the O3 NAAQS has been conducted in compliance with the applicable substantive
and procedural requirements under the Act and relevant case law. In support of this view,
these commenters state that the CASAC fulfills the CAA criteria for the committee, that the
CAA-required air quality criteria are provided in the ISA, which they note was reviewed by
the CASAC (which provided the Administrator with written advice on the draft document),
and that the CASAC provided its advice to the Administrator on the NAAQS. These
commenters additionally cite the EPA actions that fulfilled requirements for a public docket,
a published proposed decision that explains the basis for associated judgments and
consideration of CASAC advice, and a public comment period of 48 days (that also included
two days of public hearings with oral testimony from the public). The commenters also
recognize the expedited nature of the current review which they find appropriate, stating that
the EPA has acted appropriately and lawfully to ensure timely completion of the review. In
response to questions raised during the review, these commenters note that the Act does not
require that the EPA issue a final ISA prior to a draft PA, further noting that the CASAC
written comments on the draft ISA and PA were available to the EPA, as well as the final
ISA, prior to its completion of the final PA. The commenters also note that the Act does not
require the EPA to release multiple review drafts of a document, or of a separate
exposure/risk document (noting that this was also not done in the NO2 primary NAAQS
review).

Response: We agree with this comment that this review of the O3 NAAQS complied with the
applicable legal requirements, including for reasons identified in the comment, as well as for
reasons described elsewhere in this Response to Comments document and in the Federal
Register notice describing the final decision in this review. We further note that, as described
in Administrator Pruitt's memo (Pruitt, 2018), the EPA recognizes the importance of
completing NAAQS reviews in a timely, efficient, and transparent manner in order to ensure
a full evaluation of whether any revisions to the standards are necessary to provide requisite
protection of public health and welfare, while remaining cognizant of the statutory time
frames for completion such reviews.

20

-------
(2) Comment: Some commenters criticized the process that the EPA followed in this NAAQS
review, stating that the EPA had failed to provide any explanation (reasoned or not) for
process changes and state that when the EPA made changes to the process in 2006, it
provided an opportunity for CASAC and others to provide input on the changes. Some
commenters claim that the EPA has failed to conduct the necessary scientific review to
reflect the latest scientific knowledge that Congress intended (or that it was impossible to
ascertain based on a lack of explanation for the selection of the CASAC and the consultant
pool). Some commenters emphasize that particularly given the CAA specified role filled by
the CASAC, and the concerns they raise with regard to CASAC, the process followed in the
current review was flawed, arbitrary and capricious, and unlawful and that accordingly the
proposed decision suffers from fatal procedural errors and is contrary to law. Some
commenters assert that these flaws mean that the Administrator should not defer to CASAC
advice and that they call into question any decision that relies on CASAC review.

With regard to the CASAC and its role in the review, some commenters claim that CASAC
was not lawfully constituted. Such commenters assert that the CASAC is unqualified for
scientific advice on NAAQS or the expertise contemplated by its charter, that it lacked or did
not have appropriate access to relevant scientific expertise (e.g., epidemiology, toxicology,
clinical experience, exposure assessment, plant and climate impacts); and that it did not
provide "valid expert advice." Such commenters additionally claim that CASAC members
were appointed based on geographic location and government affiliation versus scientific
expertise, and that the chair appears to lack impartiality. Some commenters claimed that the
CASAC is at odds with the CAA-specified role that the EPA describes the CASAC as filling
and that it lacked critical expertise because the EPA failed to form an expert O3 review panel,
such that the EPA could not ensure that the NAAQS accurately reflects the latest scientific
knowledge, as required under the CAA. Some commenters state that the process of allowing
CASAC to seek written input from a pool of consultants did not provide an adequate
substitute for the O3 review panel, citing similar concerns about lack of expertise in the group
of consultants, as well as concerns that the process of exchanging information through
writing does not replace the process that CASAC and expert review panels have used in past
reviews. Some commenters further claim that the process followed in the current NAAQS
review fails to meet requirements for federal advisory committees and federal peer review
guidance, citing the EPA's Peer Review Handbook and the Office of Management and
Budget's 2004 Peer Review Bulletin, and asserting CASAC to lack appropriate expertise and
balance contemplated by these documents. These commenters also assert that the EPA barred
from CASAC nongovernment recipients of the EPA funding (a requirement they state was
later overruled by court decisions and for which the EPA provided no rationale), which they
further state resulted in a lack of balance of scientific perspectives on the panel. Further,
some commenters suggested that the change in membership in the chartered CASAC
conflicts with established Science Advisory Board (SAB) criteria for advisory panels, which
call for "continuity of knowledge." Some commenters also stated that CASAC overlooked or
was not informed of important aspects of the review, such as thorough review of the form of
the standard, consideration of lower primary standard levels with regard to margin of safety,
and potential conflict between draft PA and Murray decision.

Some commenters also criticize the approach the EPA followed for preparing the ISA, PA,
and exposure and risk analyses. These commenters variously state that what they describe as

21

-------
simultaneous preparation of draft the ISA and PA, and presentation of the exposure/risk
analyses within the PA (versus in a separate document) results in insufficient time for
CASAC and public review, as well as a "commingling" of science and policy. The
commenters claim that presenting the exposure/risk analyses in the PA document
"eliminates" CASAC and public review of these analyses, and that development of the PA
prior to CASAC review of the ISA or risk and exposure analyses contributes to CASAC and
the EPA reviewing policy considerations in the PA based on unreviewed scientific analyses,
stating the view that this results in severe limitations on CASAC's ability to advise on the
NAAQS. Additionally, some commenters state that the lack of second draft documents
handicaps accuracy and contributes to the PA not reflecting the best science. Some
commenters point out that CASAC requested a second draft of the PA, but the EPA did not
provide one.

Some commenters also raise concerns regarding public participation and review of
documents prepared during the course of the review, claiming among other things that it did
not meet CAA requirements for public participation. In support of this view, commenters
variously cite the overlapping public review periods for the draft ISA and PA, a lack of
second drafts of these documents, and the brevity of the public comment period for the
proposal, claiming that these placed burdens on public review (especially for respiratory
health professionals who may have less time for review during the current pandemic).

Response: We disagree with these commenters that the current review process, including
aspects involving the CASAC review, public participation, and development of scientifically
sound documents in support of the decision, is flawed, inadequate, arbitrary, capricious, or
contrary to law.

The process by which the CASAC was established was proper and met all applicable CAA
and Federal Advisory Committee Act (FACA) requirements, and was consistent with the
EPA policy and procedure as outlined in the CASAC Charter. Section 109(d)(2)(A) of the
CAA 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 CASAC. The seven-member
chartered CASAC meets these statutory requirements.12

12 The list of chartered CASAC members, along with their bio sketches, is available at:

https://yosemite.epa.gov/sab/sabpeople.nsf/WebExternalCommitteeRosters?OpenView&committee=CAS

AC&secondname=Clean%20Air%20Scientific%20Advisory%20Committee%20

22

-------
With regard to commenters' claims that the CASAC lacks necessary scientific expertise, thus
jeopardizing its ability to provide the EPA with the required scientific advice or the EPA's
ability to rely on CASAC's advice, we first note the CAA requirements for the independent
scientific review committee whose role is currently filled by the CASAC. Section 108(a)(2)
of the CAA directs the EPA to prepare "air quality criteria" that accurately reflect the latest
scientific knowledge regarding all identifiable effects on public health and welfare that may
result from the presence of the criteria pollutant in the atmosphere. Section 109(d)(2) of the
CAA directs the EPA to appoint an independent scientific review committee that shall
conduct a review of the air quality 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. Section 109(d)(2)(A)
provides that the review committee 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." The CAA does not further identify specific
areas of experience or scientific expertise to be represented on the review committee, and it
would be unreasonable to expect all the individual disciplines from the wide array of
atmospheric, health, and welfare sciences to be explicitly represented on the committee itself.
To address this expertise concern as expressed by the CASAC, in this review the EPA made
available a pool of consultants with expertise in a number of scientific fields germane to the
O3 NAAQS (84 FR 38625, August 7, 2019). The approach employed for the CASAC to
utilize outside technical expertise represents a modification of the process used in past
reviews. Rather than join with some or all of the CASAC members in a pollutant-specific
review panel as has been common in other NAAQS reviews in the past, in this review, the
consultants comprised a pool of expertise that CASAC members drew on through the use of
specific questions, posed in writing prior to their public meeting on the review, regarding
aspects of the documents being reviewed. This allowed the CASAC to obtain subject matter
expertise for its document review in a focused, efficient, and transparent manner.

The pool of expert consultants utilized by the CASAC was assembled using a public process
beginning with a Federal Register notice requesting the nomination of scientists from a broad
range of 15 disciplines, with demonstrated expertise and research in the field of air pollution
related to PM and O3, including epidemiology (84 FR 38625, August 7, 2019). From among
the nominees, the Administrator selected scientists, all with advanced scientific degrees
(PhDs), many active in academia, and reflecting a broad range of areas of expertise,
including aerosol science and atmospheric chemistry, predictive and causal modeling method
development and application, human health risk assessment, quantitative risk analyses,
environmental exposures, toxicology, statistics, environmental and genetic epidemiology,
and human clinical studies in respiratory effects. Therefore, the Agency does not agree with
commenters who contend that the current CASAC review process lacked the appropriate
breadth or balance of expertise. Further, the EPA notes that the decision to provide
supplemental expertise to the chartered CASAC, whether with an additional panel of experts
as in previous NAAQS reviews or via a publicly nominated pool of experts as in this review,
goes beyond the requirements of the Act.

As outlined in the EPA's call for nominees to the consultant pool, the EPA was resolved to
provide complete public transparency with respect to the consultant pool's input to the
independent scientific review and established a process whereby all communications between

23

-------
the chartered CASAC and the individual consultants was accomplished via written requests
and responses. Contrary to claims of these commenters, which include a concern that having
the process involve CASAC raising an issue before the consultant pool could address it could
have led to deficiencies in the review documents being overlooked (a concern for which the
commenters provided no evidence), the EPA believes that this process allowed for focused
and transparent information exchange on the subjects where CASAC members most needed
additional expertise. Certainly, there is ample evidence of robust information exchange
between the chartered CASAC and individual consultants in which CASAC members pose
questions and, in addition to responding to the questions, consultants provide comments on a
wide range of issues regarding the EPA documents. For example, Appendix D of the CASAC
February 19, 2020 letter providing its review of the draft ISA, and Appendix C of the
CASAC February 19, 2020 letter providing its review of the draft PA, contain 126 and 53
pages, respectively, of CASAC questions and consultant answers that document substantive
technical and scientific interpretations. The CASAC questions and consultant feedback were
presented on the EPA webpage prior to the public meeting of the CASAC. In sum, while not
explicitly required by the CAA, the EPA believes that the pool of consultants established for
this review provided CASAC access to additional expertise beyond that of the seven
chartered members and helped contribute to a rigorous and thorough scientific review that
satisfied the CAA requirements.

While some commenters referred to an EPA directive concerning membership of the EPA
advisory committees, claiming that it prohibited leading research scientists from CASAC
membership which resulted in a lack of balance on the Committee, we note that the grants
portion of the directive was never applied to exclude any person from membership on the
CASAC. With respect to the recent court decisions that address the EPA's directive
concerning membership of advisory committees, those decisions were issued after the
CASAC membership was finalized. Further, although the CASAC membership was finalized
prior to the court decisions, the only court decision to vacate the section of the directive
pertaining to EPA grants does not require the EPA to reopen the composition of any advisory
committees. See Nat. Res. Def. Council, Inc. v. U.S. Envtl. Prot. Agency, No. 19CV5174
(DLC), 2020 WL 2769491, at *1 (S.D.N.Y. Apr. 15, 2020).

Further, in light of the above discussion, we disagree with these commenters that the CASAC
composition and the expertise brought to bear in the CASAC reviews of the draft ISA and
PA were contrary to federal peer review guidance. While the two documents cited by the
commenters do not constitute statutory requirements, we note that, as described in the
paragraphs above, the factors with which commenters expressed concern — expertise and
balance — were of major importance in selection of committee members (in addition to the
CAA specific representation requirements). The CASAC review process addressed these
factors in its provision of external scientific review while additionally providing the Agency
with a streamlined review process that facilitated Agency attention to statutory deadlines.
Regarding the Administrator's decision to increase state, tribal, and local government
participation and enhance geographic diversity, the Administrator has discretion to consider
such criteria in staffing the EPA's advisory committees, and doing so is not inconsistent with
any requirement of the CAA or FACA. Further, the EPA does not agree with commenters'
claim that the EPA considered such criteria instead of scientific expertise. Rather, the EPA
considered expertise and geographic diversity and governmental affiliation in addition to

-------
scientific expertise. With respect to the comment concerning continuity of knowledge, one
CASAC member has prior experience with a NAAQS review thus providing a degree of
continuity of knowledge. Further, the "continuity of knowledge" criteria referred to by
commenters is contained in a report issued by the Science Advisory Board Staff Office and
describes non-binding criteria used to develop a list of potential candidates.

We also disagree with the comments suggesting that the EPA has not provided a rationale for
the changes in process in this review. Changes can be made to these procedures without
public comment. The EPA has made changes to the NAAQS review process on multiple
occasions such that it has evolved in multiple ways since the Agency began conducting
reviews back in the 1970's while continuing to meet the CAA requirements for such reviews.
In addition to these changes in the larger framework for NAAQS reviews, each review has
also had its own unique characteristics, related to considerations such as the specific nature
of the pollutant and the standard(s) being evaluated. While the scientific review process here
may have differed in format from previous O3 NAAQS reviews, it still fully comported with
all relevant CAA requirements, including in sections 108 and 109 of the Act, and reflected a
thorough review of the latest scientific knowledge relevant in reviewing the air quality
criteria for O3 and related photochemical oxidants and the adequacy of the existing standards.

Prior to initiating this review, the EPA released to the public and posted on its website
Administrator Pruitt's memo (Pruitt, 2018), which explains the rationale for the process
changes the EPA incorporated into this particular NAAQS review. Two of the key reform
principles specifically centered on the CAA statutory requirements in Section 109(d) for both
meeting the five-year review deadlines and properly establishing the functions of the
independent scientific review committee. Emphasizing these goals is well within the
Administrator's discretion. The memo announced publicly that the EPA would be identifying
ways in which the review process could be streamlined, including development of robust
initial draft versions of the ISA and PA that could allow the CASAC and the public to
comment significantly on the review documents. While some of these aspects differed from
how prior reviews had been structured, the specific processes used in those reviews were not
binding or required by the statute. The comments fail to establish that the process changes
made for this review, or the manner in which they were made, were inconsistent with any
statutory requirement or that the process changes invalidate the final decision or limit the
EPA's ability to rely on the advice that the duly constituted CASAC has provided in this
review.

The EPA disagrees that the current review process did not provide sufficient time for public
and CASAC review documents prepared in the review. As summarized in section I.D of the
Federal Register notice describing the final decision in this review, the draft ISA and draft
PA were made available to the public and the CASAC 69 days and 34 days, respectively,
prior to being discussed by the CASAC at a public meeting on December 4, 5 and 6, 2019
(84 FR 50836, September 26, 2019; 84 FR 58713, November 1, 2019). They were
subsequently discussed by the CASAC at a public teleconference on February 11 and 12,
2020 (85 FR 4656, January 27, 2020). As described in the public notice announcing the
meeting and teleconference, members of the public had the opportunity to provide written
comments in advance of, and oral comments during, the meeting and teleconference. Further,
the Agency publicly announced the availability of the draft ISA and draft PA for public

25

-------
review over periods of 68 days (9/26/2019 to 12/2/2019) and 46 days (11/01/2019 to
12/16/2019), respectively, specifying the submission of comments to the respective dockets
(84 FR 50836, September 26, 2019; 84 FR 58711, November 1, 2019). Comments from the
public and the advice of the CASAC were considered by the EPA in preparing the final ISA,
and also in the development of the draft PA. Further, the public comment period for the
proposal provided 48 days for the submission of written comments and two days of public
hearings for the provision of oral comments. This process met the legal requirements for
CASAC review and public participation on NAAQS decisions. Comments received on the
proposal have been considered in reaching the final decisions. Thus, we disagree with these
commenters that the process adopted in the current O3 NAAQS review has not allowed
sufficient opportunity for public input and CASAC advice.

Lastly, with regard to concerns raised on the development of support documents in the
review, as directed by Administrator Pruitt's memo, the EPA determined to streamline the
NAAQS review process to expedite its completion in an efficient and transparent manner
(Pruitt, 2018). The CAA leaves the Administrator considerable discretion as to what
documents and analyses are useful in informing a particular review of the air quality criteria
and existing standards and how to structure the review; it does not establish a particular
sequential order for the documents prepared during a review, nor does it require completion
of multiple drafts of any review document. Several recent NAAQS reviews have established
precedent for single drafts of the PA (e.g., the lead and sulfur dioxide NAAQS reviews
completed in 2016 and 2018, respectively). Additionally, it is not uncommon, as implied by
commenters, for past reviews to have included CASAC review of draft PAs before a final
ISA had been released, nor to have the same public meeting include CASAC's review of a
draft ISA, a draft REA, and a draft PA, as occurred in the 2015 O3 review (77 FR 46755,
August 6, 2012). Further, regarding the inclusion of the exposure and risk analyses in the PA
rather than being presented in a separate document, we disagree with commenters that that
the current NAAQS review process necessitates a separate REA. Consistent with the recently
completed particulate matter NAAQS review, detailed and complete quantitative analyses,
including those focused on human exposure and health risk, environmental exposure, and
various welfare effects are included in the PA. In this O3 NAAQS review, the health-related
analyses and associated policy-relevant considerations are described in Chapter 3 of the final
PA, while appendices 3C and 3D provide comprehensive details on the data, tools, and
methods employed, specific results and sensitivity, uncertainty, and variability analyses. The
welfare-related analyses and associated policy-relevant considerations are described in
Chapter 4 of the final PA, while appendices 4C and 4D provide comprehensive details on
data, tools, and methods employed, specific results and summaries of limitations and
uncertainties. The CASAC provided advice on all of these quantitative analyses, as did
several public commenters, when reviewing the draft PA. In some past NAAQS reviews, the
risk and exposure analyses have been included in the PA (most recently in the nitrogen
dioxide NAAQS review completed in 2017),13 and as noted above, there have been reviews

13 Prior to embarking on these steps in the review, that were described in the draft IRP, we noted
comments received from individual CASAC members during consultation on the draft IRP, which did not
generate consensus CASAC advice. Such comments did not take consider the Agency's past experiences
with such an approach in prior reviews or raise issues that had not been accounted for in those prior
reviews. Thus, after consideration of past experience, the efficiency objectives for the process objectives

26

-------
in which the CASAC and public review of a draft REA were coincident with review of a
draft PA. Therefore, we disagree with the commenter that a separate REA is required to
allow for a technically rigorous analysis of risk, or that inclusion of the risk and exposure
analyses in the PA improperly commingles policy and scientific considerations or prevents
transparent and informed review by the CASAC and the public. Similarly, we disagree with
the comments that review of the draft PA before the ISA had been finalized limited the
CASAC's ability to provide advice on the NAAQS. CASAC considered the draft ISA and
draft PA each in turn and provided separate advice on each document. In addition, the EPA
completed the ISA before it completed the PA, so it was able to consider the final scientific
assessment, which reflected consideration of the CASAC's advice, in finalizing the PA.

With regard to claims by some commenters stated that the CASAC overlooked or was not
informed of important aspects of the review, such as thorough review of the form of the
standard, consideration of lower primary standard levels with regard to margin of safety, and
potential conflict between draft PA and Murray Energy decision, we note that, as in past
reviews, the PA discuss these topics. Although this might be implied by these comments,
PAs do not always discuss specific alternative standards, as illustrated by the PAs for the
recent Pb, NO2, and SO2 NAAQS reviews (U.S. EPA, 2014b, 2017b, 2018a). As in those
cases, the O3 PA discussed policy-relevant aspects of the current evidence and quantitative
analyses in discussing key considerations for the Agency's consideration in its review of the
existing standards, in addition to describing the CAA requirements for NAAQS reviews and
the approaches employed in past reviews. Further, the Agency solicited input from the public
for consideration by the CASAC in its review of the PA and such concerns could have been
raised at that time (84 FR 58713, November 1, 2019).

(2) Comment: One commenter states that the EPA was required to and failed to conduct a
consultation with the U.S. Fish and Wildlife Service and the National Marine Fisheries
Service under Section 7(a)(2) of the Endangered Species Act, stating that without such
consultation the EPA cannot assure that any final standard is not likely to jeopardize
continued existence of endangered or threatened species or result in the destruction or
adverse modification of critical habitat. The commenter further states that Section 7
"consultation" is required under the ESA for "any action [that] may affect listed species or
critical habitat" to "insure that any action authorized, funded, or carried out by such
agency... is not likely to jeopardize the continued existence of any endangered species or
threatened species or result in the adverse modification of habitat of such species...
determined.. .to be critical...." The commenter asserts that agency "action" is broadly
defined in the ESA's implementing regulations at 50 CFR 402.02 to include: "(b) the
promulgation of regulations; ... or (d) actions directly or indirectly causing modifications to
the land, water, or air." The comment states that the EPA's review of the NAAQS is "an
activity carried out by a federal agency in the United States which directly and indirectly
causes modifications" to the land, water or air.

The commenter also notes that ESA regulations at 50 CFR 402.03 provide that Section 7 of
the ESA applies to all actions in which there is discretionary Federal involvement or control

described in the Pruitt (2018) memo, as well as any public or CASAC member comments on the issue, we
implemented the approach.

27

-------
and asserts that the EPA has discretion to consider impacts to listed species in its review of
both the primary and secondary NAAQS. With respect to the secondary NAAQS, the
commenter states that the CAA both empowers and mandates the EPA to exercise its
discretion to consider impacts to listed species and critical habitat in reviewing the secondary
NAAQS because the secondary NAAQS is designed to protect the "public welfare," which is
defined to include effects on soil, water, crops, vegetation, animals, wildlife, weather,
visibility, and climate. The commenter further claims that review of the primary standard
allows for consideration of impacts to listed species and critical habitat because the primary
NAAQS is designed to protect the public health because the health and vigor of human
societies and the integrity and wildness of the natural environment are closely linked, and
many people suffer significant long-term stress from species going extinct and their critical
habitat being adversely modified. The comment further asserts that biodiversity is a
foundation of human health because it allows humans to secure life-sustaining goods and
services, while the loss of biodiversity is a threat to public health, for example through
threats to food security, increased prevalence of infectious disease, and because biodiversity
can provide important resources for traditional practices, medical research, and drug
development, as well as reducing risks from climate disasters and supporting recovery and
adaptation efforts.

The commenter additionally cites the ISA's description of O3 impacts on a variety of plant
and animal species, asserting that O3 can directly harm several listed species and critical
habitat for several listed species, as well as claiming that O3 can harm listed species through
its effects on plant-insect signaling, herbivore growth and reproduction, and climate change.
The commenter also cites additional studies that it claims shows that O3 can adversely affect
the growth and flowering of plants, alter species composition and richness, change water flux
regulation, pollination efficiency, plant pathogen development, and functioning
belowground, including nutrient cycling and carbon pools. As further support for their view,
the commenter further claims that the EPA must consider co-benefits of a more stringent O3
NAAQS for listed species and habitats, as a more stringent O3 NAAQS would lead to
reductions in emissions of NOx, mercury, fine particulate matter, VOCs, and greenhouse
gases. The comment further states that such consideration of co-benefits is consistent with
OMB Circular A-4.

Response: The EPA disagrees with the commenter's assertion that this review is "an activity
carried out by a federal agency in the United States which directly and indirectly causes
modifications" to the land, water or air, and therefore the EPA was required to consult under
Section 7(a)(2) of the ESA. Even assuming that the ESA consultation requirement could
apply to a decision to revise the NAAQS, the EPA does not agree that leaving the NAAQS
unaltered triggers the requirement to consult under the ESA. Leaving the NAAQS unchanged
does not authorize or carry out any "action" under the statutory terms of the ESA.14 Both the
Code of Federal Regulations and the status quo regarding the NAAQS are entirely
undisturbed. The EPA is not taking any affirmative action. Moreover, leaving the NAAQS
unaltered will not require the EPA to make new air quality designations, nor will it require

14 Section 7(a)(2) of the ESA only applies to "action authorized, funded, or carried out" by a federal agency.

28

-------
States or authorized tribes to undertake new planning or control efforts or to change air
quality.

Similarly, even if the EPA's review decision on the O3 NAAQS were found to be an "action"
for ESA purposes, the EPA's decision to leave the O3 NAAQS unaltered causes no change to
the status quo for air quality and regulatory requirements, and thus has no effect on species or
their habitat.

Additionally, the EPA does not believe it is necessary or appropriate to consider O3 impacts
on species and habitats as part of the review of the primary O3 NAAQS. To the extent the
commenter is suggesting the primary standard should be set to protect species and habitats,
the EPA believes that would be inconsistent with the text and structure of the Clean Air Act.
Section 109 of the CAA requires the EPA to establish primary standards to protect public
health (see section 109(b)(1)) and secondary standards to protect public welfare (see section
109(b)(2)). Under Section 109(b)(1) and Whitman v. Am. Trucking Associations, 531 U.S.
457 (2001), the EPA sets primary standards that are requisite to protect public health,
allowing an adequate margin of safety. The EPA does not have discretion to set a different
primary standard than the one the Administrator judges is required under Section 109(b)(1)
to protect public health in order to protect species and habitats.

Moreover, even if the EPA's review of the NAAQS were an "action" that was anticipated to
have some effect on a listed species or habitats, consultation under Section 7(a)(2) of the
ESA would not be required, because the statute leaves the EPA no discretion to set a
NAAQS more stringent than is "requisite" to protect public health and welfare. See 50 CFR
402.03. The Supreme Court adopted the Solicitor General's definition of "requisite" to mean
"sufficient, but not more than necessary." Whitman v. Am. Trucking Associations, 531 U.S.
457, 473 (2001). In other words, the Administrator must select a standard that is "not lower
or higher than necessary" to achieve the CAA's statutory objectives. Id. at 476. This leaves
the Administrator no discretion to weigh factors that do not bear on public health and
welfare.

As to the primary standard, the EPA must set a standard that is "requisite to protect the public
health" with "an adequate margin of safety." CAA § 109(b)(1). The EPA would have no
discretion to modify the requisite standard based on consideration of factors that are not
related to health. To the extent the commenter is arguing that effects on species also have
effects on people, e.g., because the commenter knows of people who place great value on the
continued existence of species, the commenter is describing an effect that cannot be
considered without assessing the effects on species, which is done in the review of the
secondary standard, not the primary review. Thus, any such effects on people are beyond the
scope of this review of the primary standard. Even as to the secondary standard, the
Administrator has no discretion to adapt the welfare standard to be more protective of listed
species and habitats than is necessary "to protect the public welfare from any known or
anticipated adverse effects associated with the presence of such air pollutant in the ambient
air." CAA §109(b)(2). To be sure, the Clean Air Act says that "[a]ll language referring to
effects on welfare includes, but is not limited to, effects on . . . vegetation [and] animals"
among other factors. CAA §302(h). The Administrator must therefore set a secondary
standard which protects against known or anticipated effects on plants and animals to the

29

-------
extent they constitute adverse effects on public welfare. But the NAAQS are not zero risk
standards, see, e.g., Murray Energy Corp. v. EPA, 936 F.3d 597, 610 (D.C. Cir. 2019), and
not every adverse effect on welfare constitutes an adverse effect on public welfare.

The EPA may only set a standard that is requisite (neither more nor less stringent than
necessary) to protect against adverse effects on public welfare. The universe of "effects"
which are to be addressed under the ESA can be considerably broader than the "adverse
effects on public welfare" which are the proper subject of a secondary NAAQS. In particular,
consultation under ESA § 7(a)(2), when applicable, requires the agency to "insure" that the
agency's action "is not likely to jeopardize the continued existence of any endangered
species or threatened species or result in the destruction or adverse modification of [critical]
habitat of such species." ESA § 7(a)(2). The EPA lacks the discretion to set a secondary
NAAQS that goes beyond the proper scope and stringency required by the Clean Air Act.

Furthermore, the EPA disagrees that it is necessary or appropriate to consider in this review
the co-benefits of reductions in pollutants other than O3, such as mercury, that might result
from establishing a more stringent O3 NAAQS. To the extent the commenter is suggesting
that either the primary or secondary standard should be set to protect against effects from
pollutants other than O3 or related photochemical oxidants, the EPA believes that would be
inconsistent with the text and structure of the Clean Air Act. Under section 109(a)(1) and (2)
of the CAA, NAAQS are to be promulgated for air pollutants for which air quality criteria
have been issued. Section 109(b) of the CAA requires the EPA to establish primary standards
to protect public health (see section 109(b)(1)) and secondary standards to protect public
welfare (see section 109(b)(2)), and provides that both the primary and secondary standards
are to be based on the air quality criteria. Section 108(a)(2) of the CAA requires that the air
quality criteria for an air pollutant "shall 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 expectedfrom the presence of such pollutant in the ambient air" (emphasis
added). Under section 109(d), the EPA is required to periodically review the air quality
criteria and standards, consistent with sections 108 and 109(b). As described in the Integrated
Review Plan, in this review the EPA is revising the air quality criteria and NAAQS for
photochemical oxidants including O3 (U.S. EPA, 2019a). Accordingly, the 2020 ISA
"review[ed] and synthesize[d] the air quality criteria for the health and welfare effects of
ozone and related photochemical oxidants in ambient air" (ISA, IS-2). Accordingly, this
review is appropriately focused on the health and welfare effects of O3, and whether any
revisions to the existing standards would be appropriate under section 109 to provide
additional protection from O3 effects. Cf. WildEarth Guardians v. EPA, 759 F.3d 1196, 1209
(10th Cir. 2014) ("Just as the ESA consultation requirement cannot be invoked by
characterizing agency nonaction as action, it cannot be invoked by trying to piggyback
nonaction on an agency action by claiming that the nonaction is really part of some broader
action. When an agency action has clearly defined boundaries, we must respect those
boundaries and not describe inaction outside those boundaries as merely a component of the
agency action."). The EPA does not have discretion to set a different primary or secondary
standard than the one the Administrator judges is required under section 109(b)(1) or (2), to
protect public health or welfare from effects associated with other pollutants, for which air
quality criteria have not been issued.

30

-------
(3) Comment: Some commenters state that the proposed decision had not adequately considered
environmental justice and equity concerns. More specifically, some commenters variously
state that the EPA's consideration of at-risk groups in NAAQS decisions does not fulfill its
obligations under EO 12898; that the EPA's finding that the action does not have
disproportionately high and adverse human health or environmental impacts on minority,
low-income, or indigenous populations is unfounded; and, that the EPA must undertake a
review of cumulative impacts and set the O3 standard using a precautionary approach. Some
of these commenters assert that given disproportionate distribution of some diseases and
health outcomes (e.g., asthma, associated emergency department visits, cardiovascular
disease, and associated deaths), that the commenters consider to also be related to O3, the
EPA should and has not considered how a lower standard level would benefit these Black
and Native American communities. Other commenters state that "fact that the status quo is
inequitable is not a rational reason to forego meaningful analysis of the inequity of the status
quo or to avoid redressing these inequities." In support of their view that EJ concerns were
not adequately considered, some commenters state that tribal populations are more
vulnerable than the general population to O3 and socioeconomic-related impacts.

Response: As described in section I.A of the NFA, the NAAQS must protect public health
with an adequate margin of safety, including for sensitive groups (or populations) as well as
the general populace. Minority populations, low-income populations, and/or indigenous
peoples are often such sensitive populations. The EPA agrees that the NAAQS should be set
in a precautionary fashion because the margin of safety is intended to provide a reasonable
degree of protection against hazards that research has not yet identified, and does not assume
that the status quo is adequate, but disagrees that the EPA is required to undertake a review
of cumulative impacts of other pollutants. The primary O3 standard is a nationally uniform
standard which in the Administrator's judgment is requisite to protect public health,
including the health of sensitive groups (also termed at-risk populations), with an adequate
margin of safety, from the effects which may be expected from the presence of O3 in the
ambient air, consistent with CAA requirements. Thus, the CAA requires the NAAQS to be
"requisite," i.e., neither more nor less stringent than necessary, to provide protection, with an
adequate margin of safety, for sensitive groups.

In making its determination regarding the requisite protection of at-risk populations, as
discussed in section II. A.2.b of the final action, in other comment responses, and as
summarized in section IX.K of final action, the EPA expressly considered the available
information regarding O3 exposure and health effects among sensitive populations, including
low income and minority populations. The ISA and PA for this review, which include
identification of populations at risk from 03-related health effects, are available in the docket,
EPA-HQ-OAR-2018-0279. In accordance with E.O. 12898, the EPA has considered
whether the decision may have disproportionate negative impacts on minority populations,
low-income populations, or indigenous peoples. This decision retaining the existing primary
O3 standard, without revision, is not expected to have disproportionate negative impacts on
minority or low-income populations. Rather, the EPA expects that actions taken to bring all
areas of the U.S. into compliance with this standard will reduce health risks in the areas
subject to the highest ambient air concentrations of O3. We further note that to the extent that
areas currently not complying with the current standard are disproportionately populated by
Black or low-income populations, as implied by commenters, improvements in air quality to

31

-------
come into compliance with the current standard will reduce Cb-related health risks in these
populations.

To the extent any of the commenters is suggesting E.O. 12898 requires additional
quantitative analysis or assessment of environmental justice issues, or that the standard
should be revised to be more stringent than necessary to protect the health of sensitive and
other groups with an adequate margin of safety, the EPA disagrees. This action retains the
existing O3 NAAQS. States have primary response for implementing the NAAQS and
implementation plans are beyond the scope of this action. However, the EPA notes that
recipients of EPA financial assistance must comply with all federal nondiscrimination
statutes that together prohibit discrimination on the bases of race, color, national origin
(including limited-English proficiency), disability, sex, and age. These laws include: Title VI
of the Civil Rights Act of 1964; Section 504 of the Rehabilitation Act of 1973; Section 13 of
the Federal Water Pollution Control Act Amendments of 1972; Title IX of the Education Act
Amendments of 1972; and the Age Discrimination Act of 1975. The EPA's Office of Civil
Rights (OCR) is responsible for carrying out compliance with these federal
nondiscrimination statutes and does so through a variety of means, including: complaint
investigation; agency-initiated compliance reviews; pre-grant award assurances and audits;
and technical assistance and outreach activities. Anyone who believes that any of the federal
nondiscrimination laws enforced by OCR have been violated by a recipient of EPA financial
assistance may file an administrative complaint with the EPA's OCR.

The commenter has provided no evidence that the current O3 standards result in a
disproportionate impact on Native Americans and Alaska natives with asthma or that revision
would not, and the EPA is not aware of such evidence. Further, contrary to the comments,
and consistent with the assessment conducted in each NAAQS review, the EPA has
evaluated the available evidence with regard to populations that may be at greater risk of O3
health effects than the general population. That assessment, described in the ISA, identified
people with asthma as an at-risk population but did not identify native Americans as an at-
risk population. Likewise, the ISA concluded that available evidence is not adequate to
conclude an increased risk status based solely on racial, ethnic, or income variables, or pre-
existing cardiovascular disease or diabetes, alone (ISA, section IS.4.4). The term "at-risk
populations" is used to recognize populations that have a greater likelihood of experiencing
03-related health effects (sometimes referred to as sensitive groups). Thus, the EPA expects
that a standard providing protection for populations identified as at risk also provides
protection for other groups. Finally, to the extent that these comments are premised on the
commenters' view that the current O3 standards do not provide adequate public health or
public welfare protection, the EPA disagrees for the reasons described elsewhere in this
Response to Comments document and the NFA describing the EPA's final decision in this
review.

(4) Comment: Some commenters express the view that the EPA should have engaged in

consultation with Tribes and should not have claimed there are no tribal implications of the
proposed action. In support of this view, the commenters state that American Indian or
American Native (AI/AN) populations in America suffer disproportionately from health
discrepancies that leave them more vulnerable to impacts from pollution than the general
public. The commenters state that the EPA ignored these implications and has not offered

32

-------
Tribal consultation on the proposal, directly disregarding EO 13175. The commenters further
state that only Tribes (not the EPA) can determine whether Tribes are impacted by this
action, claiming immeasurable cost to Tribes from the loss or diminishment of culturally
significant species due to the impacts of O3 or other pollutants and noting the benefits to an
entire Tribe and individuals from being able to exercise hunting, fishing, and gathering rights
as guaranteed by treaties with the United States.

Response: Executive Order 13175, entitled "Consultation and Coordination with Indian
Tribal Governments" (65 FR 67249, November 9, 2000), directs federal agencies to develop
an accountable process to ensure "meaningful and timely input by tribal officials in the
development of regulatory policies that have tribal implications." It provides that '"policies
that have tribal implications' refers to regulations, legislative comments or proposed
legislation, and other policy statements or actions that have substantial direct effects on one
or more Indian tribes, on the relationship between the Federal Government and Indian tribes,
or on the distribution of power and responsibilities between the Federal Government and
Indian tribes." This NAAQS decision retains existing national standards to address the health
and welfare effects of O3, providing protection for sensitive groups from adverse effects to
public health, with an adequate margin of safety, and protection of public welfare from
known or anticipated adverse effects.15 There are no changes to regulations by this action,
and Tribes are not obligated to implement these standards, or to conduct monitoring or adopt
monitoring requirements, such that no direct requirements are placed on Tribes by this
action. This action does not have tribal implications, as specified in Executive Order 13175,
and is therefore not subject to the Executive Order. Even if this action was determined to
have tribal implications within the meaning of the Executive Order, it will neither impose
substantial direct compliance costs on tribal governments, nor preempt tribal law, and
therefore consultation under the Executive Order was not required.

While consultation was not required under the EO, the EPA undertook a number of outreach
activities to inform the Tribal community about the O3 NAAQS review. Specifically, on
three occasions, prior to and subsequent to signature on the proposed decision, we
participated in National Tribal Air Association (NTAA)/EPA Air Policy calls to describe the
status, current and future steps in the O3 NAAQS review, including with regard to
development of review documents and associated opportunities for public review. These
NTAA calls occurred on December 12, 2019, May 28, 2020, and July 30, 2020. Additionally,
on the date the Administrator signed the proposed O3 NAAQS decision and the EPA placed a
copy on its website, July 14, 2020, we sent an email to the Tribal community and the NTAA
about the proposal. We received no requests for further informational meetings or for
consultations.

Further, during the public comment period, we received comments on the proposed action
from seven tribes and three tribal organizations. These comments were considered in
reaching a final decision on the existing primary and secondary standards in this review, and
all significant comments are addressed in the preamble to the final action or in this RTC.

15 The comment related to a potential for Tribes to be increased risk of O3 health effects is addressed in
the NFA.

33

-------
(5) Comment: Some commenters expressed the view that the EPA's approach for the exposure
and risk analyses is inconsistent with CAA "which promises air in which people can engage
in their normal range of activity free from adverse effects." The commenter goes on to claim
that the EPA has failed to establish that the demographic variables such as age and gender
provide for appropriate predictions of time spent outdoors with elevated ventilation rates.
They further assert that the APEX modeling is flawed and systematically underestimates the
likelihood of multiple exposures of concern for simulated individuals.

Response: The premise of this comment seems to be that the Agency's use of an exposure
assessment to estimate population exposures to O3 in ambient air that may occur when the
current standard is met does not take into account the "normal range of activity" in which
people engage. However, as described in detail in the PA, the exposure model is intended to
do just that. Briefly, the APEX model calculates the exposure time-series for a user-specified
duration and number of individuals (PA, Appendix 3D, section 3D.2). Collectively and by
design, these simulated individuals are intended to be a representative random sample of the
population in the chosen study area. To this end, demographic data from the decennial census
are used so that appropriate model sampling probabilities can be derived considering
personal attributes such as age and sex and used to properly weigh the distribution of
individuals in any given geographical area. For each simulated person, the following general
steps are performed: (1) Select personal attribute variables and choose values to characterize
the simulated individual (e.g., age, sex, body weight, disease status); (2) Construct an activity
event sequence (a minute-by-minute time-series) by selecting a sequence of appropriate daily
activity diaries for the simulated individual (using demographic and other influential
variables); (3) Calculate the pollutant concentrations in the microenvironments that simulated
individuals visit; and (4) Calculate the simulated individual's exposure, and simultaneously,
their breathing rate for each exposure event and summarize for the selected exposure metric
(PA, Appendix 3D, section 3D.2). A simulated individual's complete time-series of
exposures (i.e., exposure profile), representing intra-individual variability in exposures, is
combined with the exposure profiles for all simulated individuals in each study area and
summarized to generate the population distribution of exposures, representing inter-
individual variability in exposures. The overarching goal of the exposure and risk analysis is
to account for the most significant factors contributing to inhalation exposure and risk, i.e.,
the temporal and spatial distribution of people and pollutant concentrations throughout the
study area and among the microenvironments (PA, Appendix 3D, section 3D.2).

The goal in addressing variability in this exposure and risk analysis is to ensure that the
estimates of exposure and risk reflect the variability of O3 concentrations in ambient air,
population characteristics, associated O3 exposures, physiological characteristics of simulated
individuals, and potential health risk across the study areas and for the simulated at-risk
populations (PA, Appendix 3D, section 3D.2.9.1). The APEX model is designed to account
for variability in the model input data, including the physiological variables that are
important inputs to determining exertion levels and associated ventilation rates. The resulting
collection of probabilistically sampled individuals represents the variability of the target
population, and by accounting for several types of variability, including demographic,
physiological, and human behavior, APEX is able to represent much of the variability in the
exposure and risk estimates. For example, variability may arise from differences in the
population residing within census tracts (e.g., age distribution) and the activities that may

34

-------
affect population exposure to O3 (e.g., time spent outdoors, performing moderate or greater
exertion level activities outdoors). The range of exposure and associated risk estimates are
intended to reflect such sources of variability, although we note that the range of values
obtained reflects the input parameters, algorithms, and modeling system used, and may not
necessarily reflect the complete range of the true exposure or risk values (PA, Appendix 3D,
section 3D.2.9.1).

Further, the commenters appear to also fundamentally object to the EPA's consideration of
exposure estimates in reaching conclusions on the primary O3 standard. The EPA disagrees
with these commenters' conclusions regarding the appropriateness of considering exposure
estimates, and notes thatNAAQS must be "requisite" (i.e., "sufficient, but not more than
necessary" (Whitman, 531 U.S. at 473)) to protect the "public health" ("the health of the
public" (Whitman, 531 U.S. at 465)). Estimating exposure patterns based on extensive
available data16 is a reasonable means of ascertaining that standards are neither under- nor
over-protective, and that standards address issues of public health rather than health issues
pertaining only to isolated individuals. Behavior patterns are critical in assessing whether
ambient air concentrations of O3 may pose a public health risk.17 Exposures to the
concentrations of O3 that occur in ambient air have only been shown to result in potentially
adverse effects if the ventilation rates of people in the exposed populations are raised to a
sufficient degree (e.g., through physical exertion) (ISA, Appendix 3).18 Ignoring whether
such elevated ventilation rates are actually occurring, as advocated by these commenters,
would not provide an accurate assessment of whether the public health is at risk. Indeed, a
standard established without regard to behavior of the public would likely lead to a standard
which is more stringent than necessary to protect the public health.

While setting the primary O3 standard based only on ambient concentrations, without
consideration of activity patterns and ventilation rates, would likely result in a standard that
is over-protective, the EPA also concludes that setting a standard based on the assumption
that people will adjust their activities to avoid exposures on high-pollution days would likely
result in a standard that is under-protective. The exposure analysis does not make this latter
assumption. The time-location-activity diaries that provided the basis for exposure estimates
reflect actual variability in human activities. While some diary days may reflect individuals
spending less time outdoors than would be typical for them, it is similarly likely that some
days reflect individuals spending more time outdoors than would be typical. Considering the

16	The CHAD database used in the exposure assessment contains data for nearly 180,000 individual diary
days, and includes time-location-activity patterns for individuals of both sexes across a wide range of ages
(PA, Appendix 3D, section 3D.2.5.5.1).

17	As the EPA explained in the last review: "The activity pattern of individuals is an important
determinant of their exposure. Variation in O3 concentrations among various microenvironments means
that the amount of time spent in each location, as well as the level of activity, will influence an
individual's exposure to ambient O3. Activity patterns vary both among and within individuals, resulting
in corresponding variations in exposure across a population and over time" (80 FR 65312, October 26,
2015). This continues to be true in the current review.

18	For healthy young adults exposed at rest for 2 hours, 500 ppb is the lowest O3 concentration reported to
produce a statistically significant 03-induced group mean FEVi decrement (PA, section 3.3.3; 2013 ISA,
p. 6-5).

35

-------
actual variability in time-location-activity patterns is at the least a permissible way of
identifying standards that are neither over- nor under-protective.19

Further, the EPA sees nothing in the CAA that prohibits consideration of the O3 exposures
that could result in effects of public health concern. While a number of judicial opinions have
upheld the EPA's decisions in other NAAQS reviews to place little weight on particular risk
or exposure analyses (i.e., because of scientific uncertainties in those analyses), none of these
opinions have suggested that such analyses are irrelevant because actual exposure patterns do
not matter. See, e.g. Mississippi v. EPA, 744 F.3d 1334, 1352-53 (D.C. Cir. 2013); American
Trucking Ass 'ns, v. EPA, 283 F.3d 355, 373-74 (D.C. Cir. 2002). Further, in upholding the
EPA's decision on the 2015 primary O3 NAAQS, the court noted that the Administrator had
considered activity patterns in the exposed population "because adverse health responses to
ozone exposure are critically dependent on breathing rates" and concluded that this "use of
the exposure assessment was rational." See Murray Energy Corp. v. EPA, 936 F.3d 597, 610-
11 (D.C. Cir. 2019).

Therefore, because behavior patterns are critical in assessing whether ambient concentrations
of O3 may pose a public health risk, the EPA disagrees with the views expressed by these
commenters objecting to the consideration of O3 exposures in reaching decisions on the
primary O3 standard.

(6)	Comment: One commenter expresses the view that the EPA failed to consider how action
will undermine implementation of Chesapeake Bay total maximum daily load (TMDL)
which they note relies on NAAQS for achieving atmospheric N deposition allocation.

Response: The CAA requirements for NAAQS are that the standards provide the requisite
protection of public health and welfare. The CAA does not, as implied by this comment,
require the NAAQS to be used as a tool for implementing TMDLs established under section
303(d) of the Clean Water Act. Further, the NAAQS pollutant addressed in this review, O3
and related photochemical oxidants, are not identified in the TMDL referenced by the
commenter. The pollutants identified in the TMDL as interfering with attainment of water
quality standards for the Chesapeake Bay are nitrogen, phosphorus, and sediment. To the
extent atmospheric deposition of nitrogen has implications for the health of the Chesapeake
Bay, and related impacts on the public welfare, such a consideration may be germane to the
secondary NAAQS for oxides of nitrogen or for particulate matter, standards, which are
currently being considered in the separate ongoing review of the secondary standards for
oxides of nitrogen, oxides of sulfur and PM. Initial plans for that review are described in
publicly available planning documents for the review (U.S. EPA, 2017a, 2018b) accessible
on the EPA's website (https://www.epa.gov/naaqs/nitrogen-dioxide-no2-and-sulfur-dioxide-
so2-secondary-air-quality-standards)

(7)	Comment: In describing their support for the proposal to retain the existing NAAQS, some
commenters identify proximity to background of the air quality under the current standard.

19 See Mississippi, 744 F. 3d at 1343 ("[determining what is 'requisite' to protect the 'public health' with
an 'adequate' margin of safety may indeed require a contextual assessment of acceptable risk. See
Whitman, 531 U.S. at 494-95 (Breyer, J. concurring...))."

36

-------
These commenters cite comments provided to the CASAC by the pool of expert consultants,
a news article, and EPA research. In the commenter's view, the expert consultant comments
suggest that the EPA's analysis in the PA implies that natural sources can result in a design
value above 70 ppb in some areas and claim that meeting the current 70 ppb NAAQS is
impossible in some areas through domestic precursor emissions controls alone. The
commenter also cites a COVID-related news article to highlight Phoenix background levels
and cites research on lightning emissions. The commenter claims that such information
implies USB could prevent attainment of the current NAAQS and accordingly of a more
stringent one.

Response: We note that the part of the CASAC consultants' comments referring to the ability
for some areas to meet the current NAAQS through implementation of domestic precursor
emissions controls is not making the statement that the commenter claims. Rather, the
CASAC consultant is asking a seemingly rhetorical question implying non-achievability of a
70 ppb NAAQS. A USB value of 80 ppb on a 4th highest day in a single model year (an
observation from the EPA's modeling analysis in the PA that the consultant cites and that the
commenter emphasizes) does not suggest the standard is not attainable. First, a design value
is made of a multi-year average of 4th highs. Second, the prediction with the maximum USB
value is an unmonitored site but appears to be high-biased based on surrounding monitors.
Third, there are mechanisms within the Clean Air Act to address USB during
implementation. Sqq Murray Energy, 936 F.3d at 623.

In support of their view regarding contributions from background, the commenter points to
excerpts from a news article about Phoenix Arizona during COVID lock down and highlights
"transported and natural emissions resulted in NOx levels within the city that were 12.5
percent higher than the previous year." We note that the article includes many references to
USB contributions: highlighting California as a transport source, stagnation that would
emphasize local sources, and high temperatures that would favor local production. The
12.5% annual difference cited is from a narrow time window (9-days) with no attempt to
account for meteorological variability. The article also references a 6% decrease over a
slightly longer time period (March 17 to May 4). Further, "Most of the days of the shutdown
period prior to the hot weather had below average pollution compared to 2019 and especially
compared to a 10-year average." The excerpts referenced from the article, therefore, do not
account for factors other than changes in U.S. emissions that impacted O3 concentrations
during 2019 and 2020, and consequently do not provide any direct information on USB
levels in Phoenix.

The commenter additionally highlights research showing that lightning NOx is
approximately a "30 percent of total NOx ... in certain areas." It is worth noting that the area
selected for analysis is, by design, a predominantly rural area with very little anthropogenic
emissions. Further, the lightning emissions occur above the planetary boundary layer, so it is
unclear to what extent they impact surface-level air pollution. Thus, simply highlighting the
large fraction of emissions, as the commenter does, does not convey a true representation of
their O3 contribution to concentrations at non-attainment monitors.

The EPA notes that natural, international, and US domestic emission sources contribute to
O3. The commenter provides no evidence that USB (i.e., O3 concentrations that would exist

-------
in the absence of U.S. anthropogenic emissions) would prevent attainment of the proposed
standard. The attainability of the O3 NAAQS depends on the standard averaging time, level,
form, and implementation tools. Furthermore, "[attainability 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]; accord
Murray Energy Corp., 936 F.3d at 623-24).

(8) Comment: One commenter states that the EPA should help states attain existing standards
versus, as they described it, rushing, the current NAAQS review. In support of this position,
the commenter presented several examples related to implementation programs where they
find a need for action.

Response: As described in section I.A of the final action, this action is being taken pursuant
to CAA section 109(d)(1) and relevant case law. Under CAA section 109(d)(1) the EPA has
the obligation to periodically review the air quality criteria and the existing NAAQS and
make such revisions as may be appropriate. Accordingly, the scope of this action is to satisfy
that obligation; it is not to address concerns related to implementation of the existing
standards. Thus, comments related to the implementation of the existing O3 standards are
outside the scope of this action and require no further response. State and federal O3 control
programs, such as those discussed in section I.D of the final action, may provide an
opportunity for such implementation concerns to be addressed.

(9) Comment: One commenter suggest that the EPA establish exceptional events for wildfires
without individual state petitions. Another comment recommends that the EPA "encourage
states" to invoke exceptional event rules which the commenter claims will assure use of
appropriate design values for determination of O3 NAAQS attainment and proper
development of attainment programs.

Response: As described in section I.A of the final action, this action is being taken pursuant
to CAA section 109(d)(1) and relevant case law. Under CAA section 109(d)(1) the EPA has
the obligation to periodically review the air quality criteria and the existing NAAQS and
make such revisions as may be appropriate. Accordingly, the scope of this action is to satisfy
that obligation; it is not to address concerns related to implementation of the existing
standards or the exceptional event provisions. Thus, this comment is outside the scope of this
action and requires no further response. State and federal O3 control programs, such as those
discussed in section I.D of the final action, may provide an opportunity for such
implementation concerns to be addressed.

(10) Comment: One commenter suggests that in making a decision to retain the current standard,
the EPA should give "additional recognition" to the impact of international emissions.

Response: While the EPA agrees that international transport can be a significant factor
influencing O3 concentrations, as described in Chapter 2 and Appendix 2B of the PA, it
disagrees that this is a relevant consideration in the Administrator's decision to retain the
current NAAQS in this review. With regard to implementation of the NAAQS and
consideration of international contributions, we note proposed guidance, available at:

-------
https://www.epa.gov/ground-level-ozone-pollution/international-transport-air-pollution,
regarding a "179B demonstration".

(11) Comment: Some commenters express the view that revision to more restrictive NAAQS
would be premature given that some states are still working to implement the 2008 and/or the
2015 NAAQS.

Response: As described in section I.A of the NFA, this action is being taken pursuant to
CAA section 109(d)(1) and relevant case law. Under CAA section 109(d)(1) the EPA has the
obligation to periodically review the air quality criteria and the existing NAAQS and make
such revisions as may be appropriate. Accordingly, the scope of this action is to satisfy that
obligation; it is not to address concerns related to implementation of the existing standards or
the exceptional event provisions. Thus, this comment is outside the scope of this action and
requires no further response. State and federal O3 control programs, such as those discussed
in section I.D of the final action, may provide an opportunity for such implementation
concerns to be addressed.

(12) Comment: One commenter raises a concern regarding the approach used by the EPA to
characterize and estimate USB, quoting the ISA with regard to differences in results via
different approaches. This commenter highlights multiple methodologies to estimate USB.

Response: The EPA recognizes that estimates of USB are uncertain and that there are
multiple methodologies of estimation (PA, section 2.5.2.1; Jaffe et al., 2018). The EPA has
extensive experience in considering USB, both in summarizing estimates drawn from the
current research (Henderson et al., 2012; U.S. EPA, 2014a), and in developing and
comparing estimates using alternate methodologies (Dolwick et al., 2015). That experience
informed the methodological choices made in this review.

Though methods can give substantially different results, these differences do not indicate that
one approach is more appropriate than another. For example, apportionment-based USB may
be substantially lower than zero-out USB when a monitor experiences NOx-titration of O3. If
local US-anthropogenic NOx emissions were controlled, however, the USB would be
expected to increase, and the O3 from USB sources would be conceptually more aligned with
zero-out methods. Both zero-out and apportionment-based USB are subject to model biases
that can be treated with bias correction. When biases are due to local emission sources,
proportional bias correction would incorrectly adjust both USA and USB. The source of bias
is often unknown, however, so proportional bias correction provides useful constraints where
observations are available. When calculating spatially contiguous maps of USB, bias
correction requires a technique like extended Voronoi Neighbor Averaging. This type of bias
correction can create obviously unrealistic results, and expert analysis and judgment is
necessary to identify problematic corrections. The EPA considered the potential differences
from methods in the context of the goals and national-scale of the Policy Assessment and
decided to use zero-out for this assessment.

We note that the PA analysis of USB, which is part of the characterization of current air
quality, does not play a role the Administrator's decision in this review.

-------
V. References

Bell, ML, Zanobetti, A and Dominici F (2014). Who Is More Affected by Ozone Pollution? A
Systematic Review and Meta-Analysis. Am J Epi 180(1): 15-28.

Binkley, D, Campoe, OC, Gspaltl, M and Forrester, DI (2013). Light absorption and use

efficiency in forests: Why patterns differ for trees and stands. Forest Ecol Manag 288: 5-
13.

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.

Office of the Adminstrator, Science Advisory Board Washington, DC Available at:
https://yosemite. epa.gov/sab/sabproduct.nsf/264cb122 7d55e02c85257402007446a4/4 713
D217BC07103485258515006359BA/$File/EPA-CASAC-20-003.pdf.

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.

Henderson, B, Possiel, N, Akhtar, F and Simon, H (2012). Regional and Seasonal Analysis of
North American Background Ozone Estimates from Two Studies. August 15, 2012
Available at: https://www. epa.gov/naaqs/ozone-o3-standards-other-technical-documents-
review-completed-2015

Jaffe, DA, Cooper, OR, Fiore, AM, Henderson, BH, Tonnesen, 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: 56.

Johnson, TR, Langstaff, JE, Graham, S, Fujita, EM and Campbell, DE (2018). A multipollutant
evaluation of APEX using microenvironmental ozone, carbon monoxide, and particulate
matter (PM2.5) concentrations measured in Los Angeles by the exposure classification
project. Cogent Environ Sci 4(1): 1453022.

King, J.S, 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. New Phytologist 168(3): 623-636.

Langstaff, J (2007). Memorandum to Ozone NAAQS Review Docket (EPA-HQ-OAR-2005-

0172). Analysis of Uncertainty in Ozone Population Exposure Modeling. Docket ID No.
EPA-HQ-OAR-2005-0172-0174.

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.

40

-------
Available at: https://www.epa.gov/criteria-air-pollutants/back-basics-process-reviewing-
national-ambient-air-quality-standards.

Talhelm, AF, Pregitzer, KS, Kubiske, ME, Zak, DR, Campany, CE, Burton, AJ, Dickson, RE,
Hendrey, GR, Isebrands, JG, Lewin, KF, Nagy, J and Karnosky, DF (2014). Elevated
carbon dioxide and ozone alter productivity and ecosystem carbon content in northern
temperate forests. Glob Chang Biol 20(8): 2492-2504.

U.S. EPA (2006). Air Quality Criteria for Ozone and Related Photochemical Oxidants (Volumes
I-III). EPA-600/R-05-004aF, EPA-600/R-05-004bF and EPA-600/R-05-004cF. U.S.
Environmental Protection Agency. Washington, DC. Available at:
http.V/www.epa.gov/ttn/naaqs/standards/ozone/s o3 cr cd.html.

U.S. EPA (2010). Quantitative Risk and Exposure Assessment for Carbon Monoxide. (Final
Report, Amended). Office of Air Quality Planning and Standards. Research Triangle
Park, NC. U.S. EPA. EPA-452/R-10-009. July 2010. Available at:
https://www.epa.gov/sites/production/files/2020-07/documents/co-rea-amended-
july2010.pdf

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). 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/ZyP URL. cgi ?Dockey =P 100KB UF. txl.

U.S. EPA (2014b). Policy Assessment for the Review of the Lead National Ambient Air Quality
Standards. Office of Air Quality Planning and Standards. Research Triangle Park, NC.
U.S. EPA. EPA -452/R-14-001. May 2014. Available at:
https://www3.epa.gov/ttn/naaqs/standards/pb/data/140501 pa ph fin.pdf

U.S. EPA (2015). Preamble To The Integrated Science Assessments (ISA). U.S. Environmental
Protection Agency. Washington, DC. EPA/600/R-15/067. November 2015. Available at:
https ://cfpub. epa.gov/ncea/isa/recordisplay. cfm ?deid=310244

U.S. EPA. (2017a). Integrated Review Plan for the Secondary National Ambient Air Quality

Standards for Ecological Effects of Oxides of Nitrogen, Oxides of Sulfur and Particulate
Matter. U.S. Environmental Protection Agency. Washington, DC. EPA-452/R-17-002.
January 2017. Available at:

https://www3.epa.gov/ttn/naaqs/standards/no2so2sec/data/irp-nox-sox-pm-eco.pdf

U.S. EPA (2017b). Policy Assessment for the Review of the Primary National Ambient Air

Quality Standards for Oxides of Nitrogen. Office of Air Quality Planning and Standards.
Research Triangle Park, NC. U.S. EPA. EPA-452/P-16-001. September 2016. Available

41

-------
at: https://www3. epa.gov/ttn/naaqs/standards/nox/data/2016092 7-no2-pa-external-
review-draft.pdf

U.S. EPA (2018a). Policy 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-18-002. May 2018. Available at:
https://www.epa.gov/sites/production/files/2018-05/documents/primary so2naaqs
final_pa - may 2018.pdf

U.S. EPA (2018b). 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. Research Triangle
Park, NC. U.S. EPA. EPA-452/D-18-001. August 2018. Available at:
https://www.epa.gov/sites/production/files/2018-
08/documents/reajplanJinal 080618.pdf

U.S. EPA (2019a). 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 (2019b). Policy Assessment for the Review of the Ozone National Ambient Air

Quality Standards, External Review Draft. Office of Air Quality Planning and Standards.
Research Triangle Park, NC. US Environmental Protection Agency. EPA-452/P-19-002.
October 2019. Available at: https://www.epa.gov/sites/production/files/2019-
10/documents/o3-draft_pa-oct31-2019-erd. 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.

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.

Wittig, VE, Ainsworth, EA, Naidu, SL, Karnosky, DF and Long, SP (2009). Quantifying the

impact of current and future tropospheric ozone on tree biomass, growth, physiology and
biochemistry: a quantitative meta-analysis. Glob Chang Biol 15(2): 396-424.

WHO (2008). Uncertainty and Data Quality in Exposure Assessment. The International
Programme on Chemical Safety. Geneva. WHO.

https://www.who.int/ipcs/publications/methods/harmonization/exposure assessment.pdf.

42

-------
Zahran, HS and Bailey, C (2013). Factors Associated with Asthma Prevalence among Racial and
Ethnic Groups—United States, 2009-2010 Behavioral Risk Factor Surveillance System.
J Asthma, 50(6): 583-589.

43

-------
Appendix A. List of Abbreviations and Acronyms

The following acronyms have been used for the sake of brevity in this document:

APEX

Air Pollution Exposure model

AQCD

Air Quality Criteria Document

CAA

Clean Air Act

CASAC

Clean Air Scientific Advisory Committee

CHAD

Consolidated Human Activity Database

CO

Carbon Monoxide

EPA

Environmental Protection Agency

E-R

Exposure-response

FEVi

Forced Expiratory Volume for 1 second

MSS

McDonnell- Stewart- Smith

NAAQS

National ambient air quality standards

NO2

Nitrogen dioxide

NOx

Nitrogen oxides

O3

Ozone

PA

Policy Assessment

ppm

Parts per million

ppm-hrs

Parts per million-hours

ppb

Parts per billion

RBL

Relative biomass loss

SO2

Sulfur dioxide

USB

United States background

VOCs

Volatile organic compounds

W126

Cumulative integrated exposure index with a sigmoidal weighting function

A-l

-------
Appendix B. Details of Analyses in support of Section III.B(l)

This appendix contains additional details of the response to two points raised by a
comment received on the proposal that are addressed in section III.B(l) of this document. The
comment concerned the example analysis described in the PA, Appendix 4A, section 4A.3.2. As
described there, this analysis estimates aboveground biomass for two patterns of 6-year O3
exposure of aspen: one in which the seasonal W126 index for each year equals 17 ppm-hrs, and
one in which the seasonal W126 index for each year varies (e.g., 10, 17, and 24 ppm-hrs), but the
average of each three consecutive years equals 17 ppm-hrs.

As described in the PA (PA, p. 4A-22), in this analysis, the growth rate information
(derived from King et al., 2005) was applied over six years of growth (using the yearly growth
increment, g/m2/year, for the stand). The above ground biomass of the aspen stand in each year
was compared across the exposure scenarios (Figure 4A-15; Table 4A-7). The difference
between the scenarios in total above ground biomass for the stand varied from year to year. After
the first year, this difference in the year's total above ground biomass (not to be confused with
annual growth in biomass, to which RBL is applied) was less than 2%.

In consideration of the two points raised by the commenter (summarized and addressed in
section III.B(l) above), variations on the PA calculations were examined, as presented here.

• The first of the two criticisms concerns application of the RBL (for each W126) to an
annual growth increment derived from the "control" trees (represented in the analysis
by the trees in King et al. (2005) exposed to unadjusted ambient air). The commenter
states that use of this growth rate assumes the effect of O3 in one year to have no
effect on tree growth in a subsequent one, and ignores that smaller trees (presumably
including an O3 exposed tree compared to a control) grow less than larger trees (citing
Binkley et al., 2013).

• The second criticism of the analysis was the EPA's use of aspen, for which the E-R
relationship is, like the E-R relationships for most of the other 11 species for which
functions are established (PA, Appendix 4A, Figure 4A-1), relatively linear.

Consideration of first point: The commenter did not provide a growth function for aspen; nor
did the study cited by the commenter. Therefore, to assess the impact of the first point on the
comparison of the two multiyear W126 index patterns, we derived a growth model for aspen
based on the annual measurements for that species available in the Aspen FACE data collected in
the study by King et al. (2005).20 The model is a function of the form, current year growth
increment as a function of the prior year absolute biomass. A linear model was used to represent
aspen species (r2 = 0.4137):

W126 scenario annual growth = 0.2395 * Previous Year Biomass + 215.05

20 Individual tree growth measurements from Aspen FACE (1997-2008) research, including King et al.
(2005), received from researchers (Ozone NAAQS Docket, EPA-HQ-OAR-2018-0279).

B-l

-------
This function was substituted into the analysis from the PA, Appendix 4A, section 4A.3.2. for
the annual growth increment that previously had been drawn from the control scenario biomass
measurements. Based on the recalculations resulting from this substitution, we have created an
alternate version of Table 4A-7 from the PA, Appendix 4A. This is presented in Table 1 below.
Figure 1 is an alternate version of Figure 4A-15 of Appendix 4A in the PA, that presents the
comparisons of biomass for the differing W126 scenarios (3-year vs 1-year variables) based on
the information in Table 1.

2500.0

2000.0

1500.0

1000.0

500.0

Comparison of aboveground growth (biomass) for annual W126 of 17 ppm-hrs and varying
annual W126, with 3-year average of 17 ppm-hrs

I Predicted Biomass

I W126=17,17,17, etc - biomass (g/m2)
I W126=10, 24,17, etc - biomass (g/m2)
W126= 24,17,10, etc - biomass (g/m2)
I W126= 24,10,17, etc - biomass (g/m2)
I W126= 10,17, 24 etc - biomass (g/m2)

0.0 	

yO -1997

yi

y2

y3

y4

y5

y6-2003

Figure 1. Estimated aboveground biomass of aspen with different patterns of annual seasonal
W126 index using annual growth as a function of prior year absolute biomass for trees
in the same scenario.

B-2

-------
Table 1. Comparison of total aspen above ground biomass estimated for different patterns of varying annual exposures and

constant exposure equal to 3-year average (17 ppm-hrs) using annual growth as a function of prior year absolute biomass for
trees in the same scenario.

Year

Predicted
Biomass*

Growth -

%

increase

W126=17,
biomass
(g/m2)

W126=10,
24, 17,
etc -
biomass
(g/m2)

W126=
24, 17,
10, etc -
biomass
(g/m2)

W126=
24, 10,
17, etc -
biomass
(g/m2)

W126=
10, 17,
24 etc -
biomass
(g/m2)

W126
10-17-
24 vs
17

W126
10-24-
17 vs
17

W126
24-17-
10 vs
17

W126
24-10-
17 vs
17

yO-
1997

9.1



9.1

9.1

9.1

9.1

9.1









yi

226.3

2387.14%

205.0

215.0

194.8

194.8

215.0

4.9%

4.9%

-5.0%

-5.0%

y2

495.6

118.97%

443.3

442.9

430.9

442.9

455.5

2.7%

-0.1%

-2.8%

-0.1%

y3

829.3

67.34%

733.1

732.6

732.6

732.6

732.6

-0.1%

-0.1%

-0.1%

-0.1%

y4

1243.0

49.88%

1085.4

1102.8

1066.5

1066.5

1102.8

1.6%

1.6%

-1.7%

-1.7%

y5

1755.8

41.25%

1513.8

1512.5

1490.8

1512.5

1535.0

1.4%

-0.1%

-1.5%

-0.1%

y6-2003

2391.3

36.20%

2034.8

2033.2

2033.2

2033.2

2033.2

-0.1%

-0.1%

-0.1%

-0.1%

* The value in the first row of this and other columns is the total absolute biomass measurement from King et al., 2005, Table 3 (foliage plus wood).
The subsequent rows of the first column utilize the function (above) to derive current year biomass as function of prior year biomass. In the other
columns, the annual increment derived with the function is reduced by predicted RBL for the applicable W126 index value. The W126-RBL E-R
function used is 1 -exp[-W126/109.81)12198].

B-3

-------
Consideration of second point: With regard to an E-R function for O3 RBL with a quite different
shape than that for aspen, we note the functions for two species: the sugar maple (raised by the
commenter) and the black cherry/1 As can be seen in Figure 4A-1 from Appendix 4A of the PA
(PA, Appendix 4A, section 4A.1.1) presented below, the sugar maple has a function with a slope
that changes appreciable at W126 index values above 20-25 ppm-hrs, with increasingly greater
O3 impact on growth with the W126 index value above that.

co
o

CD
O

CD

cr

o

CM
O

O
O

0	10	20	30	40

W126 (ppm-hrs)

Figure 4A-1. RBL functions for seedlings of II tree species.

50

As discussed in the PA, such W126 index levels are rare in locations that meet the current
standard. As can be seen in Figure 4D-8 from Appendix 4D of the PA (PA, Appendix 4A,
section 4D.3.2.1), The air quality analyses in the PA show that across the period from 2000
through 2018 for even just the subset of sites meeting the current standard but with design value
closest to 70 ppb (66-70), the 99th percentile is below 20 ppm-hrs (PA, Appendix 4D, Figure 4D-
8).22 Focusing just on Class I areas, there are no more than 15 occurrences of a single-year W126
index value above 19 ppm-hrs in the entire time period (200-2018), and all of those occurrences
date prior to 2013 (FR 85 49904, August 14, 2020). At W126 index levels below 20, the sugar

21	The tenth, black cherry, has a slope for the E-R function that presents the opposite pattern to that of red
maple (the slope of the curve slowly declines with increasing W126 index).

22	The established geographical range of sugar maple is predominantly in the Northeast and Upper
Midwest (PA, Appendix 4B. Table 4B-3), areas for which W126 index values are among the relatively
lower magnitudes occurring across the U.S. (PA, Appendix 4D, Figure 4D-2).

Red Maple

•	Sugar Maple

•	RedAkfer

•	Tup Poplar
Ponderosa Pine

•	While Pine

•	Loblolly Pine
Virginia Pine

•	Aspen

•	Black Cherry

•	DougtasFir

Red Mapte

•	Sugar Maple

•	Red Aider

•	Tulip Pop&i
Ponderosa Pine

•	While Pine

•	Loblolly Pine
Virginia Pine

•	Aspen

•	Slack Cherry

•	OougtasFir

10 15 20 25 30
W126 (ppm-hrs)

B-4

-------
maple slope is relatively linear and much lower than that for the aspen that was the subject of the
PA Appendix 4A example analysis.

I	r	t	t	1	1—

<60 61-65 66-70 71-75 76-80 >30
4th Max Metric Value (ppb)

Figure 4l)-8. Annual W126 index values in ppm-hrs binned by 4th max metric values based
on monitoring data for years 2000-2018. Boxes show 25th. 50th. and 75th
percentiles, whiskers extend to the 1st and 99th percentiles, and points below the l!l
percentile or above the 99th percentile are represented by dots.

Several Summary Points: Given the limited availability of controlled tree exposure data for
individual years/seasons in a multi-year exposure, as well as the simply conceptual or illustrative
nature of the analysis, there are assumptions, limitations and uncertainties inherent in the
analysis. A few key points are noted here.

•	The linear growth function derived for aspen for this example may or may not be the
correct model over the life of a tree of this species, but the impact of this uncertainty
on this analysis is unclear. The original analysis (in PA) that did not estimate annual
growth increment as a function of prior year biomass may have underestimated the
effect of the O3 exposure scenario relative to the control scenario over time, but the
amount of underestimate in this example for aspen is not highly affected by the
difference in using a 3-year average versus a variable 1-year W126.

•	The E-R function for aspen and its derivation is described in the PA, Appendix 4A,
section 4A. 1. The experimental data from which it was derived were collected from
studies of aspen seedlings. Thus, there is some uncertainty in its application in an
example describing aspen growth over six years.

B-5

-------
• Variables other than 03 that can affect growth in a given year (e.g., precipitation,
temperature, community competition) are not represented in the current analysis other
than the extent to which they affect the baseline growth rate provided by the "control"
from the aspen study by King et al. (2005).

• This example analysis includes a W126 index value of 24 ppm-hrs every third year.
Yet, the frequency of such a value is quite rare, as can be seen from the air quality
analyses in the PA, which show that across the period from 2000 through 2018 for
even just the subset of sites meeting the current standard but with design value closest
to 70 ppb (66-70 ppb), the 99th percentile is below 20 ppm-hrs (PA, Appendix 4D,
Figure 4D-8). Focusing just on Class I areas for the full period from 2000 to 2018,
there are no more than 15 occurrences of a single-year W126 index value above 19
ppm-hrs, all of which date prior to 2013 (FR 85 49904, August 14, 2020). Thus, this
example includes as one of the three years, a magnitude of W126 index that has been
quite rarely observed in areas that meet the current standard since 2000.

References

Binkley, D., Campoe, O. C., Gspaltl, M., & Forrester, D. I. 2013. Light absorption and use
efficiency in forests: Why patterns differ for trees and stands. Forest Ecology and
Management, 288, 5-13.

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.

Talhelm, A. F., Pregitzer, K. S., Kubiske, M. E., Zak, D. R., Campany, C. E., Burton, A. J., ... &
Nagy, J. (2014). Elevated carbon dioxide and ozone alter productivity and ecosystem
carbon content in northern temperate forests. Global Change Biology, 20(8), 2492-2504.

U.S. EPA (2020). Policy Assessment for the Review of National Ambient Air Quality Standards
for Ozone. Office of Air Quality Planning and Standards, Health and Environmental
Impacts Divison. Research Triangle Park, NC. U.S. EPA. EPA-452/R-20-001. May 2020.

B-6

-------