^tDSX
» D \
W
PRO^
Policy Assessment for the Review of the Ozone
National Ambient Air Quality Standards
-------
-------
EP A-452/R-20-001
May 2020
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
Health and Environmental Impacts Division
Research Triangle Park, NC
-------
DISCLAIMER
This document has been prepared by staff in the U.S. Environmental Protection Agency's
Office of Air Quality Planning and Standards. Any findings and conclusions are those of the
authors and do not necessarily reflect the views of the Agency. This document does not represent
and should not be construed to represent any Agency determination or policy. Mention of trade
names or commercial products does not constitute endorsement or recommendation for use.
Questions or comments related to this document should be addressed to Dr. Deirdre
Murphy, U.S. Environmental Protection Agency, Office of Air Quality Planning and Standards,
C504-06, Research Triangle Park, North Carolina 27711 (email: murphy.deirdre@epa.gov).
-------
TABLE OF CONTENTS
1 INTRODUCTION 1-1
1.1 Purpose 1-1
1.2 Legislative Requirements 1-3
1.3 History of the O3 NAAQS, Reviews and Decisions 1-5
1.4 Current O3 NAAQ S Review 1-11
References 1-15
2 AIR QUALITY 2-1
2.1 O3 and Photochemical Oxidants in the Atmosphere 2-1
2.2 Sources and Emissions of O3 Precursors 2-4
2.3 Ambient Air Monitoring and Data Handling Conventions 2-10
2.3.1 Ambient Air Monitoring Requirements and Monitoring Networks 2-10
2.3.2 Data Handling Conventions and Computations for Determining Whether the
Standards are Met 2-14
2.4 O3 in Ambient Air 2-15
2.4.1 Concentrations Across the U.S 2-15
2.4.2 Trends in U.S. 03 Concentrations 2-16
2.4.3 Diurnal Patterns 2-20
2.4.4 Seasonal Patterns 2-23
2.4.5 Variation in Recent Daily Maximum 1-hour Concentrations 2-25
2.5 Background O3 2-27
2.5.1 Summary of U.S. Background 03 Sources 2-28
2.5.1.1 Stratosphere 2-31
2.5.1.2 Biogenic VOC 2-32
2.5.1.3 Wildland Fires 2-33
2.5.1.4 Lightning Nitrogen Oxides 2-33
2.5.1.5 Natural and Agricultural Soil NOx 2-34
2.5.1.6 Post-Industrial Methane 2-35
2.5.1.7 International Anthropogenic Emissions 2-36
2.5.2 Approach for Quantifying U.S. Background Ozone 2-37
2.5.2.1 Methodology: USB Attribution 2-38
2.5.2.2 Methodology: Strengths, Limitations and Uncertainties 2-40
1
-------
2.5.3 Estimates of USB and Contributions to USB in 2016 2-42
2.5.3.1 Spatial Characterization of O3 Contributions 2-43
2.5.3.2 Seasonal and Geographic Variations in Ozone Contributions 2-45
2.5.3.3 Ozone Source Contributions as a function of Total Ozone
Concentration 2-52
2.5.3.4 Predicted USB Seasonal Mean and USB on Peak O3 Days 2-58
2.5.4 Summary of USB 2-64
References 2-68
3 REVIEW OF THE PRIMARY STANDARD 3-1
3.1 Background on the Current Standard 3-1
3.1.1 Considerations Regarding Adequacy of the Prior Standard 3-5
3.1.2 Considerations for the Revised Standard 3-10
3.1.2.1 Indicator 3-10
3.1.2.2 Averaging time 3-11
3.1.2.3 Form 3-12
3.1.2.4 Level 3-13
3.2 General Approach and Key Issues in This Review 3-17
3.3 Health Effects Evidence 3-21
3.3.1 Nature of Effects 3-21
3.3.1.1 Respiratory Effects 3-23
3.3.1.2 Other Effects 3-28
3.3.2 Public Health Implications and At-risk Populations 3-30
3.3.3 Exposure Concentrations Associated with Effects 3-38
3.3.4 Uncertainties in the Health Effects Evidence 3-47
3.4 Exposure and Risk Information 3-49
3.4.1 Conceptual Model and Assessment Approach 3-50
3.4.2 Population Exposure and Risk Estimates for Air Quality Just Meeting the
Current Standard 3-61
3.4.3 Population Exposure and Risk Estimates for Additional Air Quality
Scenarios 3-66
3.4.4 Key Uncertainties 3-70
3.4.5 Public Health Implications 3-76
3.5 Key Considerations Regarding the Current Primary Standard 3-80
3.5.1 Evidence-based Considerations 3-81
11
-------
3.5.2 Exposure/risk-based Considerations 3-84
3.5.3 CASAC Advice 3-88
3.5.4 Conclusions on the Primary Standard 3-89
3.6 Key Uncertainties and Areas for Future Research 3-100
References 3-103
4 REVIEW OF THE SECONDARY STANDARD 4-1
4.1 Background on the Current Standard 4-1
4.1.1 Considerations Regarding Adequacy of the Prior Standard 4-4
4.1.2 Considerations for the Revised Standard 4-6
4.2 General Approach and Key Issues in this Review 4-13
4.3 Welfare Effects Evidence 4-16
4.3.1 Nature of Effects 4-16
4.3.2 Public Welfare Implications 4-24
4.3.3 Exposures Associated with Effects 4-33
4.3.4 Key Uncertainties 4-49
4.4 Exposure and Air Quality Information 4-55
4.4.1 Influence of Form and Averaging Time of Current Standard on W126 Index . 4-58
4.4.2 Environmental Exposures in Terms of W126 Index 4-62
4.4.3 Limitations and Uncertainties 4-67
4.5 Key Considerations Regarding the Current Secondary Standard 4-68
4.5.1 Evidence and Exposure/Risk-based Considerations 4-68
4.5.2 CASAC advice 4-93
4.5.3 Conclusions 4-95
4.6 Key Uncertainties and Areas for Future Research 4-106
References 4-108
iii
-------
APPENDICES
APPENDIX 2A. ADDITIONAL DETAILS ON DATA ANALYSIS PRESENTED IN PA
SECTION 2.4
APPENDIX 2B. ADDITIONAL DETAILS ON BACKGROUND OZONE MODELING AND
ANALYSIS
APPENDIX 3 A. DETAILS ON CONTROLLED HUMAN EXPOSURE STUDIES
APPENDIX 3B. AIR QUALITY INFORMATION FOR LOCATIONS OF EPIDEMIOLOGIC
STUDIES OF RESPIRATORY EFFECTS
APPENDIX 3C. AIR QUALITY DATA USED IN POPULATION EXPOSURE AND RISK
ANALYSES
APPENDIX 3D. EXPOSURE AND RISK ANALYSIS FOR THE OZONE NAAQS REVIEW
APPENDIX 4A. EXPOSURE-RESPONSE FUNCTIONS FOR 11 TREE SPECIES AND TEN
CROPS
APPENDIX 4B. U.S. DISTRIBUTION OF 11 TREE SPECIES
APPENDIX 4C. VISIBLE FOLIAR INJURY SCORES AT U.S. FOREST SERVICE
BIOSITES (2006-2010)
APPENDIX 4D. ANALYSIS OF THE W126 03 EXPOSURE INDEX AT U.S. AMBIENT
AIR MONITORING SITES
APPENDIX 4E. OZONE WELFARE EFFECTS AND RELATED ECOSYSTEM SERVICES
AND PUBLIC WELFARE ASPECTS
iv
-------
TABLE OF TABLES
Table 2-1. Simulation names and descriptions for hemispheric-scale and regional-scale
simulations 2-39
Table 2-2. Expressions used to calculate contributions from specific sources 2-40
Table 2-3. Predicted USB for U.S. and U.S. regions based on averages for all U.S. grid
cells 2-62
Table 2-4. Predicted USB for high elevation locations (>1500 m) 2-63
Table 2-5. Predicted USB for locations within 100 km of Mexico or Canada Border 2-63
Table 2-6. Predicted USB for low-elevation (<1500 m) that are 100 km or farther from the
border 2-64
Table 3-1. National prevalence of asthma, 2017 3-36
Table 3-2. Summary of 6.6-hour controlled human exposure study-findings, healthy
adults 3-43
Table 3-3. Percent and number of simulated children and children with asthma estimated to
experience at least one or more days per year with a daily maximum 7-hour
average exposure at or above indicated concentration while breathing at an
elevated rate in areas just meeting the current standard 3-64
Table 3-4. Percent of simulated children and children with asthma estimated to experience at
least one or more days per year with a lung function decrement at or above 10, 15
or 20% while breathing at an elevated rate in areas just meeting the current
standard 3-66
Table 3-5. Percent and number of simulated children and children with asthma estimated
to experience one or more days per year with a daily maximum 7-hour average
exposure at or above indicated concentration while breathing at an elevated
rate - additional air quality scenarios 3-69
Table 3-6. Percent of risk estimated for air quality just meeting the current standard in three
study areas using the E-R function approach on days where the daily maximum
7-hour average concentration is below specified values 3-75
Table 3-7. Percent of risk estimated for air quality just meeting the current standard in three
study areas using the MSS model approach on days where the daily maximum
7-hour average concentration is below specified values 3-75
Table 3-8. Comparison of current assessment and 2014 HREA (all study areas) for percent
of children estimated to experience at least one, or two, days with an exposure
at or above benchmarks while at moderate or greater exertion 3-88
Table 4-1. Distribution of 3-year average seasonal W126 index for sites in Class I areas and
across U.S. that meet the current standards and for those that do not 4-66
v
-------
TABLE OF FIGURES
Figure 2-1. U.S. O3 precursor emissions by sector: A) NOx; B) CO; C) VOCs; D) CH4 2-6
Figure 2-2. U.S. anthropogenic O3 precursor emission trends for: A) NOx; B) CO;
C) VOCs; and D) CI I . 2-7
Figure 2-3. U.S. county-level CO emissions density estimates (tons/year/mi2) for 2014 2-8
Figure 2-4. U.S. county-level NOx emissions density estimates (tons/year/mi2) for 2014. ...2-9
Figure 2-5. U.S. county-level VOC emissions density estimates (tons/year/mi2) for 2014....2-9
Figure 2-6. Current O3 monitoring seasons in the U.S 2-12
Figure 2-7. Map of U.S. ambient air O3 monitoring sites reporting data to the EPA during
the 2016-2018 period 2-14
Figure 2-8. O3 design values in ppb for the 2016-2018 period 2-16
Figure 2-9. Trends in O3 design values based on data from 2000-2002 through
2016-2018 2-17
Figure 2-10. National trend in annual 4th highest MDA8 values, 1980 to 2018 2-18
Figure 2-11. National trend in annual 4th highest MDA8 concentrations and O3 design values
in ppb, 2000 to 2018 2-18
Figure 2-12. Regional trends in median annual 4th highest MDA8 concentrations,
2000 to 2018 2-19
Figure 2-13. Diurnal patterns in hourly O3 concentrations at selected monitoring sites: A) an
urban site in Los Angeles; B) a downwind suburban site in Los Angeles; C) a low
elevation rural site in New Hampshire; and D) a high elevation rural site in New
Hampshire 2-22
Figure 2-14. Seasonal patterns in MDA8 O3 concentrations at selected monitoring sites
(2015-2017): A) an urban site in Baltimore, MD; B) an urban site in Baton
Rouge, LA; C) a rural site in Colorado; and D) a site in Utah experiencing
high wintertime O3 2-24
Figure 2-15. Boxplots showing the distribution of MDA1 concentrations (2016-2018), binned
according to each site's 2016-2018 design value 2-26
Figure 2-16. Number of days in 2016-2018 at each monitoring site with a MDA1
concentration greater than or equal to 120 ppb compared to its 8-hour design
value in ppb 2-26
Figure 2-17. National trend in the annual 2nd highest MDA1 O3 concentration,
2000 to 2018 2-27
Figure 2-18. Conceptual models for 03 sources: (a) in the U.S., and (b) at a single
location 2-30
vi
-------
Figure 2-19. Predicted MDA8 total O3 concentration (top left), Natural (top right),
International (bottom left), and USA (bottom right) contributions in spring
(March, April, May) 2-44
Figure 2-20. Predicted MDA8 total O3 concentration (top left), Natural (top right),
International (bottom left), and USA (bottom right) contributions in summer
(June, July, Aug) 2-45
Figure 2-21. Predicted contribution of International sources as a function of distance from
Mexico/Canada (left) and at "interior" locations (excluding border areas) by
elevation (right) 2-47
Figure 2-22. Grid cell assignments to East, West, High Elevation, Near Border, and Near and
High (i.e., both High Elevation and Near Border) 2-48
Figure 2-23. Annual time series of regional average predicted MDA8 total O3 concentration
and contributions of each source (see legend) for the West (top), and the East
(bottom) 2-49
Figure 2-24. Annual time series of regional urban area-weighted average predicted MDA8
total O3 concentration and contributions of each source (see legend) for the
High-elevation West (top), near-border West (middle), and Low/Interior West
(bottom) 2-51
Figure 2-25. Predicted contribution of Natural as a function of predicted total (Base)
MDA8 O3 concentration in the West and East 2-53
Figure 2-26. Predicted contribution of International as a function of predicted total (Base)
MDA8 O3 concentration in the West and East 2-54
Figure 2-27. Predicted contribution of USA as a function of predicted total (Base) MDA8
O3 concentration in the West and East. Sloped lines show percent contribution
as a quick reference 2-54
Figure 2-28. Annual time series of regional average predicted MDA8 O3 and contributions
of each source to predicted MDA8 total O3 (see legend) in the West (top) and
East (bottom) including only those grid-cell days with MDA8 greater than
70 ppb 2-56
Figure 2-29. Annual time series of regional average predicted MDA8 O3 and contributions of
each source to predicted MDA8 O3 (see legend) in the high-elevation West (top),
in the near-border West (middle), and in the Low/Interior West weighted toward
urban areas (bottom) including only those grid-cell days with MDA8 O3 greater
than 70 ppb 2-57
Figure 2-30. Map of predicted USB contributions by O3 season for spring average (top left),
summer average (top right), top 10 predicted total O3 days (center left),
4th highest total O3 simulated day (center right), and all days with total O3
greater than 70 ppb (bottom left), along with a map of the number of days with
total O3 above 70 ppb (bottom right) 2-60
Figure 3-1. Overview of general approach for review of the primary O3 standard 3-20
vii
-------
Figure 3-2. Group mean Cte-induced reduction in FEV1 from controlled human exposure
studies of healthy adults exposed for 6.6 hours with quasi-continuous
exercise 3-39
Figure 3-3. Conceptual model for exposure-based risk assessment 3-51
Figure 3-4. Analysis approach for exposure-based risk analyses 3-52
Figure 4-1. Overview of general approach for review of the secondary O3 standard 4-15
Figure 4-2. Potential effects of O3 on the public welfare 4-32
Figure 4-3. Established RBL functions for seedlings of 11 tree species 4-39
Figure 4-4. Established RYL functions for 10 crops 4-39
Figure 4-5. Distribution of nonzero BI scores at USFS biosites (normal soil moisture)
grouped by assigned W126 index estimates 4-46
Figure 4-6. W126 index at monitoring sites with valid design values (2016-2018
average) 4-57
Figure 4-7. Relationship between the W126 index and design values for the current
standard (2016-2018). The W126 is analyzed in terms of averages across the 3-
year design value period (left) and annual values (right) 4-59
Figure 4-8. Relationship between trends in the W126 index and trends in design values
across a 19-year period (2000-2019). The W126 is analyzed in terms of averages
across the 3-year design value period (left) and annual values (right) 4-61
Figure 4-9. Analytical approach for characterizing vegetation exposure 4-63
viii
-------
1 INTRODUCTION
This document, Policy Assessment for the Review of the Ozone National Ambient Air
Quality Standards (hereafter referred to as the PA), presents the policy assessment for the U.S.
Environmental Protection Agency's (EPA's) current review of the ozone (O3) national ambient
air quality standards (NAAQS).1 The overall plan for this review was presented in the Integrated
Review Plan for the Ozone National Ambient Air Quality Standards (IRP; [U.S. EPA, 2019]).
The IRP also identified key policy-relevant issues to be addressed in this review and discussed in
the main documents that generally inform NAAQS reviews, including an Integrated Science
Assessment (ISA), and a Policy Assessment (PA).
This document is organized into four chapters. Chapter 1 presents introductory
information on the purpose of the PA, legislative requirements for reviews of the NAAQS, an
overview of the history of the O3 NAAQS, including background information on prior reviews,
and a summary of the progress to date for the current review. Chapter 2 provides an overview of
how photochemical oxidants, including O3, are formed in the atmosphere, along with current
information on sources and emissions of important precursor chemicals. Chapter 2 also
summarizes key aspects of the ambient air monitoring requirements, and current O3 air quality,
including estimates of O3 resulting from natural sources and anthropogenic sources outside the
U.S. Chapters 3 and 4 focus on policy-relevant aspects of the currently available health and
welfare effects evidence and exposure/risk information, identifying and summarizing key
considerations related to this review of the primary (health-based) and secondary (welfare-based)
standard, respectively.
1.1 PURPOSE
The PA, when final, presents an evaluation, for consideration by the EPA Administrator,
of the policy implications of the currently available scientific information, assessed in the ISA,
any quantitative air quality, exposure or risk analyses based on the ISA findings, and related
limitations and uncertainties. Ultimately, a final decision on the O3 NAAQS will reflect the
judgments of the Administrator. The role of the PA is to help "bridge the gap" between the
Agency's scientific assessment and quantitative technical analyses, and the judgments required
of the Administrator in determining whether it is appropriate to retain or revise the O3 NAAQS.
1 This review focuses on the presence in ambient air of photochemical oxidants, a group of gaseous compounds of
which ozone (the indicator for the current standards) is the most prevalent in the atmosphere and the one for
which there is a veiy large, well-established evidence base of its health and welfare effects. The standards that are
the focus of this review were set in 2015 (80 FR 65292, October 26, 2015) and are referred to in this document as
the "current" or "existing" standards.
1-1
-------
In evaluating the question of adequacy of the current standards and whether it may be
appropriate to consider alternative standards, the PA focuses on information that is most
pertinent to evaluating the standards and their basic elements: indicator, averaging time, form,
and level.2 These elements, which together serve to define each standard, must be considered
collectively in evaluating the public health and public welfare protection the standards afford.
The development of the PA is also intended to facilitate advice to the Agency and
recommendations to the Administrator from an independent scientific review committee, the
Clean Air Scientific Advisory Committee (CASAC), as provided for in the Clean Air Act
(CAA). As discussed below in section 1.2, the CASAC is to advise on subjects including the
Agency's assessment of the relevant scientific information and on the adequacy of the current
standards, and to make recommendations as to any revisions of the standards that may be
appropriate. The EPA generally makes available to the CASAC and the public one or more drafts
of the PA for CASAC review and public comment.
In this PA, we take into account the available scientific information, as assessed in the
Integrated Science Assessment for Ozone and Related Photochemical Oxidants (ISA [U.S. EPA,
2020]) and additional policy-relevant quantitative air quality, exposure and risk analyses.3 Thus,
the PA is based on the final ISA and the evaluation and conclusions in this document have also
been informed by the advice received from the CASAC in its reviews of the draft PA and draft
IRP, and also by public comment received thus far in the review.
The PA is designed to assist the Administrator in considering the currently available
scientific and risk information and formulating judgments regarding the standards. Accordingly,
the PA will inform the Administrator's decision in this review. Beyond informing the
Administrator and facilitating the advice and recommendations of the CASAC, the PA is also
intended to be a useful reference to all parties interested in the review of the O3 NAAQS. In
these roles, it is intended to serve as a source of policy-relevant information that supports the
Agency's review of the O3 NAAQS, and it is written to be understandable to a broad audience.
2 The indicator defines the chemical species or mixture to be measured in the ambient air for the purpose of
determining whether an area attains the standard. The averaging time defines the period over which air quality
measurements are to be averaged or otherwise analyzed. The form of a standard defines the air quality statistic
that is to be compared to the level of the standard in determining whether an area attains the standard. For
example, the form of the annual NAAQS for fine particulate matter is the average of annual mean concentrations
for three consecutive years, while the form of the 8-hour NAAQS for carbon monoxide is the second-highest 8-
hour average in a year. The level of the standard defines the air quality concentration used for that purpose.
3 The terms "staff," "we" and "our" throughout this document refer to the staff in the EPA's Office of Air Quality
Planning and Standards (OAQPS).
1-2
-------
1.2 LEGISLATIVE REQUIREMENTS
Two sections of the CAA govern the establishment and revision of the NAAQS. Section
108 (42 U.S.C. 7408) directs the Administrator to identify and list certain air pollutants and then
to issue air quality criteria for those pollutants. The Administrator is to list those pollutants
"emissions of which, in his judgment, cause or contribute to air pollution which may reasonably
be anticipated to endanger public health or welfare"; "the presence of which in the ambient air
results from numerous or diverse mobile or stationary sources"; and for which he "plans to issue
air quality criteria...." (42 U.S.C. § 7408(a)(1)). Air quality criteria are intended to "accurately
reflect the latest scientific knowledge useful in indicating the kind and extent of all identifiable
effects on public health or welfare which may be expected from the presence of [a] pollutant in
the ambient air...." (42 U.S.C. § 7408(a)(2)).
Section 109 [42 U.S.C. 7409] directs the Administrator to propose and promulgate
"primary" and "secondary" NAAQS for pollutants for which air quality criteria are issued [42
U.S.C. § 7409(a)], Section 109(b)(1) defines primary standards as ones "the attainment and
maintenance of which in the judgment of the Administrator, based on such criteria and allowing
an adequate margin of safety, are requisite to protect the public health."4 Under section
109(b)(2), a secondary standard must "specify a level of air quality the attainment and
maintenance of which, in the judgment of the Administrator, based on such criteria, is requisite
to protect the public welfare from any known or anticipated adverse effects associated with the
presence of [the] pollutant in the ambient air."5
In setting primary and secondary standards that are "requisite" to protect public health
and welfare, respectively, as provided in section 109(b), the EPA's task is to establish standards
that are neither more nor less stringent than necessary. In so doing, the EPA may not consider the
costs of implementing the standards. See generally, Whitman v. American Trucking Ass'ns, 531
U.S. 457, 465-472, 475-76 (2001). Likewise, "[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], cert, denied, 455 U.S. 1034
[1982]; accord Murray Energy Corp. v. EPA, 936 F.3d 597, 623-24 [D.C. Cir. 2019]). At the
same time, courts have clarified the EPA may consider "relative proximity to peak background
4 The legislative history of section 109 indicates that a primary standard is to be set at "the maximum permissible
ambient air level. . . which will protect the health of any [sensitive] group of the population," and that for this
purpose "reference should be made to a representative sample of persons comprising the sensitive group rather
than to a single person in such a group." S. Rep. No. 91-1196, 91st Cong., 2d Sess. 10 (1970).
5 Under CAA section 302(h) (42 U.S.C. § 7602(h)), effects on welfare include, but are not limited to, "effects on
soils, water, crops, vegetation, manmade materials, animals, wildlife, weather, visibility, and climate, damage to
and deterioration of property, and hazards to transportation, as well as effects on economic values and on personal
comfort and well-being."
1-3
-------
... concentrations" as a factor in deciding how to revise the NAAQS in the context of
considering standard levels within the range of reasonable values supported by the air quality
criteria and judgments of the Administrator (.American Trucking Ass'ns, v. EPA, 283 F.3d 355,
379 [D.C. Cir. 2002], hereafter referred to as "ATA III').
The requirement that primary standards provide an adequate margin of safety was
intended to address uncertainties associated with inconclusive scientific and technical
information available at the time of standard setting. It was also intended to provide a reasonable
degree of protection against hazards that research has not yet identified. See Lead Industries
Ass'n v. EPA, 647 F.2d 1130, 1154 (D.C. Cir 1980), cert, denied, 449 U.S. 1042 (1980);
American Petroleum Institute v. Costle, 665 F.2d at 1186; Coalition of Battery Recyclers Ass'n v.
EPA, 604 F.3d 613, 617-18 (D.C. Cir. 2010); Mississippi v. EPA, 744 F.3d 1334, 1353 (D.C. Cir.
2013). Both kinds of uncertainties are components of the risk associated with pollution at levels
below those at which human health effects can be said to occur with reasonable scientific
certainty. Thus, in selecting primary standards that include an adequate margin of safety, the
Administrator is seeking not only to prevent pollution levels that have been demonstrated to be
harmful but also to prevent lower pollutant levels that may pose an unacceptable risk of harm,
even if the risk is not precisely identified as to nature or degree. The CAA does not require the
Administrator to establish a primary NAAQS at a zero-risk level or at background concentration
levels (see Lead Industries v. EPA, 647 F.2d at 1156 n.51, Mississippi v. EPA, 744 F.3d at 1351),
but rather at a level that reduces risk sufficiently so as to protect public health with an adequate
margin of safety.
In addressing the requirement for an adequate margin of safety, the EPA considers such
factors as the nature and severity of the health effects involved, the size of the sensitive
population(s), and the kind and degree of uncertainties. The selection of any particular approach
to providing an adequate margin of safety is a policy choice left specifically to the
Administrator's judgment. See Lead Industries Ass'n v. EPA, 647 F.2d at 1161-62; Mississippi v.
EPA, 744 F.3d at 1353.
Section 109(d) (1) of the Act requires periodic review and, if appropriate, revision of
existing air quality criteria to reflect advances in scientific knowledge on the effects of the
pollutant on public health and welfare. Under the same provision, the EPA is also to periodically
review and, if appropriate, revise the NAAQS, based on the revised air quality criteria.6
Section 109(d) (2) addresses the appointment and advisory functions of an independent
scientific review committee. Section 109(d) (2) (A) requires the Administrator to appoint this
6 This section of the Act requires the Administrator to complete these reviews and make any revisions that may be
appropriate "at five-year intervals."
1-4
-------
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 of the EPA's Science
Advisory Board. A number of other advisory functions are also identified for the committee by
section 109(d)(2)(C), which reads:
Such committee shall also (i) advise the Administrator of areas in which
additional knowledge is required to appraise the adequacy and basis of existing,
new, or revised national ambient air quality standards, (ii) describe the research
efforts necessary to provide the required information, (iii) advise the
Administrator on the relative contribution to air pollution concentrations of
natural as well as anthropogenic activity, and (iv) advise the Administrator of any
adverse public health, welfare, social, economic, or energy effects which may
result from various strategies for attainment and maintenance of such national
ambient air quality standards.
As previously noted, the Supreme Court has held that section 109(b) "unambiguously bars cost
considerations from the NAAQS-setting process" (Whitman v. American Trucking Ass'ns, 531
U.S. 457, 471 [2001]). Accordingly, while some of the issues listed in section 109(d)(2)(C) as
those on which Congress has directed the CASAC to advise the Administrator, are ones that are
relevant to the standard setting process, others are not. Issues that are not relevant to standard
setting may be relevant to implementation of the NAAQS once they are established.7
1.3 HISTORY OF THE 03 NAAQS, REVIEWS AND DECISIONS
Primary and secondary NAAQS were first established for photochemical oxidants in
1971 (36 FR 8186, April 30, 1971) based on the air quality criteria developed in 1970 (U.S.
7 Because some of these issues are not relevant to standard setting, some aspects of CASAC advice may not be
relevant to EPA's process of setting primary and secondary standards that are requisite to protect public health
and welfare. Indeed, were the EPA to consider costs of implementation when reviewing and revising the
standards "it would be grounds for vacating the NAAQS" (Whitman v. American Trucking Ass'ns, 531 U.S. 457,
471 n.4 [2001]). At the same time, the CAA directs CASAC to provide advice on "any adverse public health,
welfare, social, economic, or energy effects which may result from various strategies for attainment and
maintenance" of the NAAQS to the Administrator under section 109(d) (2) (C) (iv). In Whitman, the Court
clarified that most of that advice would be relevant to implementation but not standard setting, as it "enable[s] the
Administrator to assist the States in carrying out their statutory role as primary implementers of the NAAQS" {id.
at 470 [emphasis in original]). However, the Court also noted that CASAC's "advice concerning certain aspects
of 'adverse public health ... effects' from various attainment strategies is unquestionably pertinent" to the
NAAQS rulemaking record and relevant to the standard setting process {id. at 470 n. 2).
1-5
-------
DHEW, 1970; 35 FR 4768, March 19, 1970). The EPA set both primary and secondary standards
at 0.08 parts per million (ppm), as a 1-hour average of total photochemical oxidants, not to be
exceeded more than one hour per year based on the scientific information in the 1970 air quality
criteria document (AQCD). Since that time, the EPA has reviewed the air quality criteria and
standards a number of times, with the most recent review being completed in 2015.
The EPA initiated the first periodic review of the NAAQS for photochemical oxidants in
1977. Based on the 1978 AQCD (U.S. EPA, 1978), the EPA published proposed revisions to the
original NAAQS in 1978 (43 FR 26962, June 22, 1978) and final revisions in 1979 (44 FR 8202,
February 8, 1979). At that time, the EPA changed the indicator from photochemical oxidants to
O3, revised the level of the primary and secondary standards from 0.08 to 0.12 ppm and revised
the form of both standards from a deterministic (i.e., not to be exceeded more than one hour per
year) to a statistical form. With these changes, attainment of the standards was defined to occur
when the average number of days per calendar year (across a 3-year period) with maximum
hourly average O3 concentration greater than 0.12 ppm equaled one or less (44 FR 8202,
February 8, 1979; 43 FR 26962, June 22, 1978).
Following the EPA's decision in the 1979 review, several petitioners sought judicial
review. Among those, the city of Houston challenged the Administrator's decision arguing that
the standard was arbitrary and capricious because natural O3 concentrations and other physical
phenomena in the Houston area made the standard unattainable in that area. The U.S. Court of
Appeals for the District of Columbia Circuit (D.C. Circuit) rejected this argument, holding (as
noted in section 1.1 above) that attainability and technological feasibility are not relevant
considerations in the promulgation of the NAAQS (.American Petroleum Institute v. Costle, 665
F.2d at 1185). The court also noted that the EPA need not tailor the NAAQS to fit each region or
locale, pointing out that Congress was aware of the difficulty in meeting standards in some
locations and had addressed this difficulty through various compliance related provisions in the
CAA (id. at 1184-86).
The next periodic reviews of the criteria and standards for O3 and other photochemical
oxidants began in 1982 and 1983, respectively (47 FR 11561, March 17, 1982; 48 FR 38009,
August 22, 1983). The EPA subsequently published the 1986 AQCD (U.S. EPA, 1986) and the
1989 Staff Paper (U.S. EPA, 1989). Following publication of the 1986 AQCD, a number of
scientific abstracts and articles were published that appeared to be of sufficient importance
concerning potential health and welfare effects of O3 to warrant preparation of a supplement to
the 1986 AQCD (U.S. EPA, 1992). In August of 1992, the EPA proposed to retain the existing
primary and secondary standards based on the health and welfare effects information contained
in the 1986 AQCD and its 1992 Supplement (57 FR 35542, August 10, 1992). In March 1993,
1-6
-------
the EPA announced its decision to conclude this review by affirming its proposed decision to
retain the standards, without revision (58 FR 13008, March 9, 1993).
In the 1992 notice of its proposed decision in that review, the EPA announced its
intention to proceed as rapidly as possible with the next review of the air quality criteria and
standards for O3 and other photochemical oxidants in light of emerging evidence of health effects
related to 6- to 8-hour O3 exposures (57 FR 35542, August 10, 1992). The EPA subsequently
published the AQCD and Staff Paper for that next review (U.S. EPA, 1996). In December 1996,
the EPA proposed revisions to both the primary and secondary standards (61 FR 65716,
December 13, 1996). With regard to the primary standard, the EPA proposed to replace the then-
existing 1-hour primary standard with an 8-hour standard set at a level of 0.08 ppm (equivalent
to 0.084 ppm based on the proposed data handling convention) as a 3-year average of the annual
third-highest daily maximum 8-hour concentration. The EPA proposed to revise the secondary
standard either by setting it identical to the proposed new primary standard or by setting it as a
new seasonal standard using a cumulative form. The EPA completed this review in 1997 by
setting the primary standard at a level of 0.08 ppm, based on the annual fourth-highest daily
maximum 8-hour average concentration, averaged over three years, and setting the secondary
standard identical to the revised primary standard (62 FR 38856, July 18, 1997).
On May 14, 1999, in response to challenges by industry and others to the EPA's 1997
decision, the D.C. Circuit remanded the O3 NAAQS to the EPA, finding that section 109 of the
CAA, as interpreted by the EPA, effected an unconstitutional delegation of legislative authority
(American Trucking Ass'ns v. EPA, 175 F.3d 1027, 1034-1040 [D.C. Cir. 1999]). In addition, the
court directed that, in responding to the remand, the EPA should consider the potential beneficial
health effects of O3 pollution in shielding the public from the effects of solar ultraviolet (UV)
radiation, as well as adverse health effects (id. at 1051-53). In 1999, the EPA sought panel
rehearing and for rehearing en banc on several issues related to that decision. The court granted
the request for panel rehearing in part and denied it in part but declined to review its ruling with
regard to the potential beneficial effects of O3 pollution (American Trucking Ass'ns v. EPA,195
F.3d 4, 10 [D.C Cir., 1999]). On January 27, 2000, the EPA petitioned the U.S. Supreme Court
for certiorari on the constitutional issue (and two other issues) but did not request review of the
ruling regarding the potential beneficial health effects of O3. On February 27, 2001, the U.S.
Supreme Court unanimously reversed the judgment of the D.C. Circuit on the constitutional
issue (Whitman v. American Trucking Ass'ns, 531 U.S. 457, 472-74 [2001], [holding that section
109 of the CAA does not delegate legislative power to the EPA in contravention of the
Constitution]). The Court remanded the case to the D.C. Circuit to consider challenges to the O3
NAAQS that had not been addressed by that court's earlier decisions. On March 26, 2002, the
D.C. Circuit issued its final decision on the remand, finding the 1997 O3 NAAQS to be "neither
1-7
-------
arbitrary nor capricious," and so denying the remaining petitions for review. See ATA III, 283
F.3d at 379.
Specifically, in ATA III, the D.C. Circuit upheld the EPA's decision on the 1997 O3
standard as the product of reasoned decision making. With regard to the primary standard, the
court made clear that the most important support for the EPA's decision to revise the standard
was the health evidence of insufficient protection afforded by the then-existing standard ("the
record [is] replete with references to studies demonstrating the inadequacies of the old one-hour
standard"), as well as extensive information supporting the change to an 8-hour averaging time
(id. at 378). The court further upheld the EPA's decision not to select a more stringent level for
the primary standard noting "the absence of any [emphasis in original] human clinical studies at
ozone concentrations below 0.08 [ppm]" which supported the EPA's conclusion that "the most
serious health effects of ozone are 'less certain' at low concentrations, providing an eminently
rational reason to set the primary standard at a somewhat higher level, at least until additional
studies become available" (id. at 379, internal citations omitted). The court also pointed to the
significant weight that the EPA properly placed on the advice it received from the CASAC (id. at
379). In addition, the court noted that "although relative proximity to peak background ozone
concentrations did not, in itself, necessitate a level of 0.08 [ppm], EPA could consider that factor
when choosing among the three alternative levels" (id. at 379).
Coincident with the continued litigation of the other issues, the EPA responded to the
court's 1999 remand to consider the potential beneficial health effects of O3 pollution in
shielding the public from effects of UV radiation (66 FR 57268, Nov. 14, 2001; 68 FR 614,
January 6, 2003). The EPA provisionally determined that the information linking changes in
patterns of ground-level O3 concentrations to changes in relevant patterns of exposures to UV
radiation of concern (UV-B) to public health was too uncertain, at that time, to warrant any
relaxation in 1997 O3 NAAQS. The EPA also expressed the view that any plausible changes in
UV-B radiation exposures from changes in patterns of ground-level O3 concentrations would
likely be very small from a public health perspective. In view of these findings, the EPA
proposed to leave the 1997 primary standard unchanged (66 FR 57268, Nov. 14, 2001). After
considering public comment on the proposed decision, the EPA published its final response to
this remand in 2003, re-affirming the 8-hour primary standard set in 1997 (68 FR 614, January 6,
2003).
The EPA initiated the fourth periodic review of the air quality criteria and standards for
O3 and other photochemical oxidants with a call for information in September 2000 (65 FR
57810, September 26, 2000). In 2007, the EPA proposed to revise the level of the primary
standard within a range of 0.075 to 0.070 ppm (72 FR 37818, July 11, 2007). The EPA proposed
to revise the secondary standard either by setting it identical to the proposed new primary
1-8
-------
standard or by setting it as a new seasonal standard using a cumulative form. Documents
supporting these proposed decisions included the 2006 AQCD (U.S. EPA, 2006) and 2007 Staff
Paper (U.S. EPA, 2007) and related technical support documents. The EPA completed the
review in March 2008 by revising the levels of both the primary and secondary standards from
0.08 ppm to 0.075 ppm while retaining the other elements of the prior standards (73 FR 16436,
March 27, 2008).
In May 2008, state, public health, environmental, and industry petitioners filed suit
challenging the EPA's final decision on the 2008 O3 standards. On September 16, 2009, the EPA
announced its intention to reconsider the 2008 O3 standards,8 and initiated a rulemaking to do so.
At the EPA's request, the court held the consolidated cases in abeyance pending the EPA's
reconsideration of the 2008 decision.
In January 2010, the EPA issued a notice of proposed rulemaking to reconsider the 2008
final decision (75 FR 2938, January 19, 2010). In that notice, the EPA proposed that further
revisions of the primary and secondary standards were necessary to provide a requisite level of
protection to public health and welfare. The EPA proposed to revise the level of the primary
standard from 0.075 ppm to a level within the range of 0.060 to 0.070 ppm, and to revise the
secondary standard to one with a cumulative, seasonal form. At the EPA's request, the CASAC
reviewed the proposed rule at a public teleconference on January 25, 2010 and provided
additional advice in early 2011 (Samet, 2010, Samet, 2011). Later that year, in view of the need
for further consideration and the fact that the Agency's next periodic review of the O3 NAAQS
required under CAA section 109 had already begun (as announced on September 29, 2008),9 the
EPA decided to consolidate the reconsideration with its statutorily required periodic review.10
In light of the EPA's decision to consolidate the reconsideration with the current review,
the D.C. Circuit proceeded with the litigation on the 2008 O3 NAAQS decision. On July 23,
2013, the court upheld the EPA's 2008 primary standard, but remanded the 2008 secondary
standard to the EPA (.Mississippi v. EPA, 744 F.3d 1334 [D.C. Cir. 2013]). With respect to the
primary standard, the court first rejected arguments that the EPA should not have lowered the
level of the existing primary standard, holding that the EPA reasonably determined that the
existing primary standard was not requisite to protect public health with an adequate margin of
safety, and consequently required revision. The court went on to reject arguments that the EPA
should have adopted a more stringent primary standard. With respect to the secondary standard,
8 The press release of this announcement is available at:
https://archive.epa.gOv/epapages/newsroom_archive/newsreleases/85f90b7711acb0c88525763300617d0d.html.
9 The Call for Information initiating the new review was announced in the Federal Register (73 FR 56581,
September 29, 2008).
10 This rulemaking, completed in 2015, concluded the reconsideration process.
1-9
-------
the court held that the EPA's explanation for the setting of the secondary standard identical to the
revised 8-hour primary standard was inadequate under the CAA because the EPA had not
adequately explained how that standard provided the required public welfare protection.
At the time of the court's decision, the EPA had already completed significant portions of
its next statutorily required periodic review of the O3 NAAQS. This review had been formally
initiated in 2008 with a call for information in the Federal Register (73 FR 56581, September 29,
2008). In late 2014, based on the ISA, Risk and Exposure Assessments (REAs) for health and
welfare, and PA11 developed for this review, the EPA proposed to revise the 2008 primary and
secondary standards by reducing the level of both standards to within the range of 0.070 to 0.065
ppm (79 FR 75234, December 17, 2014).
The EPA's final decision in this review was published in October 2015, establishing the
now-current standards (80 FR 65292, October 26, 2015). In this decision, based on consideration
of the health effects evidence on respiratory effects of O3 in at-risk populations, the EPA revised
the primary standard from a level of 0.075 ppm to a level of 0.070 ppm, while retaining all the
other elements of the standard (80 FR 65292, October 26, 2015). The EPA's decision on the
level for the standard was based on the weight of the scientific evidence and quantitative
exposure/risk information. The level of the secondary standard was also revised from 0.075 ppm
to 0.070 ppm based on the scientific evidence of O3 effects on welfare, particularly the evidence
of O3 impacts on vegetation, and quantitative analyses available in the review.12 The other
elements of the standard were retained. This decision on the secondary standard also
incorporated the EPA's response to the D.C. Circuit's remand of the 2008 secondary standard in
Mississippi v. EPA, 744 F.3d 1344 (D.C. Cir. 2013). The 2015 revisions to the NAAQS were
accompanied by revisions to the data handling procedures, and the ambient air monitoring
requirements13 (80 FR 65292, October 26, 2015).14
After publication of the final rule, a number of industry groups, environmental and health
organizations, and certain states filed petitions for judicial review in the D.C. Circuit. The
industry and state petitioners argued that the revised standards were too stringent, while the
11 The final versions of these documents, released in August 2014, were developed with consideration of the
comments and recommendations from the CAS AC, as well as comments from the public on the draft documents
(Frey, 2014a, Frey, 2014b, Frey, 2014c, U.S. EPA, 2014a, U.S. EPA, 2014b, U.S. EPA, 2014c).
12 These standards, set in 2015, are specified at 40 CFR 50.19.
13 The current federal regulatory measurement methods for 03 are specified in 40 CFR 50, Appendix D and 40 CFR
part 53. Consideration of ambient air measurements with regard to judging attainment of the standards set in
2015 is specified in 40 CFR 50, Appendix U. The O3 monitoring network requirements are specified in 40 CFR
58.
14 This decision additionally announced revisions to the exceptional events scheduling provisions, as well as changes
to the air quality index and the regulations for the prevention of significant deterioration permitting program.
1-10
-------
environmental and health petitioners argued that the revised standards were not stringent enough
to protect public health and welfare as the Act requires. On August 23, 2019, the court issued an
opinion that denied all the petitions for review with respect to the 2015 primary standard while
also concluding that the EPA had not provided a sufficient rationale for aspects of its decision on
the 2015 secondary standard and remanding that standard to the EPA {Murray Energy Corp. v.
EPA, 936 F.3d 597 [D.C. Cir. 2019]).
In the August 2019 decision, the court additionally addressed arguments regarding
considerations of background O3 concentrations, and socioeconomic and energy impacts. With
regard to the former, the court rejected the argument that the EPA was required to take
background O3 concentrations into account when setting the NAAQS, holding that the text of
CAA section 109(b) precluded this interpretation because it would mean that if background O3
levels in any part of the country exceeded the level of O3 that is requisite to protect public health,
the EPA would be obliged to set the standard at the higher nonprotective level (id. at 622-23).
Thus, the court concluded that the EPA did not act unlawfully or arbitrarily or capriciously in
setting the 2015 NAAQS without regard for background O3 (id. at 624). Additionally, the court
denied arguments that the EPA was required to consider adverse economic, social, and energy
impacts in determining whether a revision of the NAAQS was "appropriate" under section
109(d)(1) of the CAA (id. at 621-22). The court reasoned that consideration of such impacts was
precluded by Whitman's holding that the CAA "unambiguously bars cost considerations from the
NAAQS-setting process" (531 U.S. at 471, summarized in section 1.2 above). Further, the court
explained that section 109(d) (2) (C)'s requirement that CASAC advise the EPA "of any adverse
public health, welfare, social, economic, or energy effects which may result from various
strategies for attainment and maintenance" of revised NAAQS had no bearing on whether costs
are to be considered in setting the NAAQS (Murray Energy Corp. v. EPA, 936 F.3d at 622).
Rather, as described in Whitman and discussed further in section 1.2 above, most of that advice
would be relevant to implementation but not standard setting (id.).
1.4 CURRENT 03 NAAQS REVIEW
In May 2018, the Administrator directed his Assistant Administrators to initiate this
review of the O3 NAAQS (Pruitt, 2018). In conveying this direction, the Administrator further
directed the EPA staff to expedite the review, implementing an accelerated schedule aimed at
completion of the review within the statutorily required period (Pruitt, 2018). Accordingly, the
EPA took immediate steps to proceed with the review. In June 2018, the EPA announced the
initiation of the current periodic review of the air quality criteria for photochemical oxidants and
the O3 NAAQS and issued a call for information in the Federal Register (83 FR 29785, June 26,
2018). Two types of information were called for: information regarding significant new O3
1-11
-------
research to be considered for the ISA for the review, and policy-relevant issues for consideration
in this NAAQS review. Based in part on the information received in response to the call for
information, the EPA developed a draft IRP which was made available for consultation with the
CASAC and for public comment (83 FR 55163, November 2, 2018; 83 FR 55528, November 6,
2018). Comments from the CASAC (Cox, 2018) and the public were considered in preparing the
final IRP (U.S. EPA, 2019).
Under the plan outlined in the IRP and consistent with revisions to the process identified
by the administrator in his 2018 memo directing initiation of the review, the current review of
the O3 NAAQS is progressing on an accelerated schedule (Pruitt, 2018). The EPA is
incorporating a number of efficiencies in various aspects of the review process, as summarized in
the IRP, to support completion within the statutorily required period (Pruitt, 2018). As one
example of such an efficiency, rather than produce two separate documents, the exposure and
risk analyses for the primary standard are included as an appendix in the PA, along with a
number of other technical appendices. The draft PA (including these analyses as appendices) was
reviewed by the CASAC and made available for public comment while the draft ISA was also
being reviewed by the CASAC and was available for public comment (84 FR 50836, September
26, 2019; 84 FR 58711, November 1, 2019).15 The CASAC was assisted in its review by a pool
of consultants with expertise in a number of fields (84 FR 38625, August 7, 2019). The approach
employed by the CASAC in utilizing outside technical expertise represents an additional
modification of the process from past reviews. Rather than join with some or all of the CASAC
members in a CASAC review panel as has been common in other NAAQS reviews in the past, in
this O3 NAAQS review (and also in the recent CASAC review of the PA for the particulate
matter NAAQS), the consultants comprised a pool of expertise that CASAC members drew on
through the use of specific questions, posed in writing prior to the public meeting, regarding
aspects of the documents being reviewed, obtaining subject matter expertise for its document
review in a focused, efficient and transparent manner.
The CASAC discussed its review of both the draft ISA and the draft PA over three days
at a public meeting in December 2019 (84 FR 58713, November 1, 2019).16 The CASAC
discussed its draft letters describing its advice and comments on the documents in a public
15 The draft ISA and draft PA were released for public comment and CASAC review on September 26, 2019 and
October 31, 2019, respectively. The charges for the CASAC review summarized the overarching context for the
document review (including reference to Pruitt [2018], and the CASAC's role under section 109(d)(2)(C) of the
Act), as well as specific charge questions for review of each of the documents.
16 While simultaneous reviews of first drafts of both documents has not been usual in past reviews, there have been
occurrences of the CASAC review of a draft PA (or draft REA when process involved policy assessment being
included within the REA document) simultaneous with review of a second (or later) draft ISA (e.g., 73 FR 19835,
April 11, 2008; 73 FR 34739, June 18, 2008; 77 FR 64335, October 19, 2020; 78 FR 938, January 7, 2013).
1-12
-------
teleconference in early February 2020 (85 FR 4656; January 27, 2020). The letters to the
Administrator conveying the CASAC advice and comments on the draft PA and draft ISA were
released later that month (Cox, 2020a, Cox, 2020b).
The letters from the CASAC and public comment on the draft ISA and draft PA have
informed completion of the final documents and further inform development of the
Administrator's proposed decision in the review. Comments from the CASAC on the draft ISA
have been considered by the EPA and led to a number of revisions in developing the final
document. The CASAC review and the EPA's consideration of CASAC comments are described
in Appendix 10, section 10.4.5 of the final ISA. As noted by Administrator Wheeler noted, in his
response to the CASAC letter conveying its review, "for those comments and recommendations
that are more significant or cross-cutting and which were not fully addressed, the Agency will
develop a plan to incorporate these changes into future Ozone ISAs as well as ISAs for other
criteria pollutant reviews." The ISA was completed and made available to the public in April
2020 (85 FR 21849, April 20, 2020).
The CASAC comments additionally provided a number of comments intended to
improve the PA. For example, it recommended that the process followed in the current review,
including its distinctions from prior reviews, be clearly summarized, as has been done in the
presentation in this section of the PA. Further, the CASAC and public comment also provided
comments on improving the clarity and other aspects of the presentations of air quality
information in Chapter 2, the scientific evidence of health and welfare effects in Chapters 3 and
4, and the quantitative exposure and risk analyses, presented in detail in Appendices 3C and 3D.
These comments have been considered in completing these sections of this document. For
example, the summary of the health effects evidence has been strengthened, consistent with the
final ISA, including further recognition of evidence of Ch-related inflammatory response and
susceptibility of people with asthma. Additions to the quantitative exposure and risk analyses are
summarized in Appendix 3D, section 3D.1. And, additional data presentations on O3 precursors
trends and regional emissions patterns have been added to Chapter 2.
The CASAC advice to the Administrator regarding the O3 standards has also been
described and considered in this document. Advice on the primary standard is summarized in
section 3.5.3 and considered in the conclusions discussed in section 3.5.4. For the secondary
standard, the CASAC advice is summarized in section 4.5.2. and considered in the PA
conclusions discussed in section 4.5.3.
The current timeline for the review of the standards projects a proposed decision near the
middle of 2020. Materials upon which this proposed decision is based, including the documents
1-13
-------
described above, will be made available to the public in the docket for the review.17 Following a
public comment period on the proposed decision, a final decision in the review is projected for
late in 2020.
17 The docket for the current O3 NAAQS review is identified as EPA-HQ-OAR-2018-0279. This docket has
incorporated the ISA docket (EPA-HQ-ORD-2018-0274) by reference. Both dockets are publicly accessible at
www.regulations.gov.
1-14
-------
REFERENCES
Cox, LA. (2018). Letter from Dr. Louis Anthony Cox, Jr., Chair, Clean Air Scientific Advisory
Committee, to Acting Administrator Andrew R. Wheeler, Re: Consultation on the EPA's
Integrated Review Plan for the Review of the Ozone. December 10, 2018. EPA-CASAC-
19-001. Office of the Administrator, Science Advisory Board U.S. EPA HQ, Washington
DC. Available at:
https://yosemite.epa.gov/sab/sabproduct.nsf/LookupWebReportsLastMonthCASAC/A286
A0F0151DC8238525835F007D348A/$File/EPA-CASAC-19-001.pdf.
Cox, LA. (2020a). Letter from Louis Anthony Cox, Jr., Chair, Clean Air Scientific Advisory
Committee, to Administrator Andrew R. Wheeler. Re:CASAC Review of the EPA's
Integrated Science Assessment for Ozone and Related Photochemical Oxidants (External
Review Draft - September 2019). February 19, 2020. EPA-CASAC-20-002. Office of the
Adminstrator, Science Advisory Board Washington, DC Availbale at:
https://yosemite.epa.gov/sab/sabproduct.nsf/264cbl227d55e02c85257402007446a4/F22
8E5D4D848BBED85258515006354D0/$File/EPA-CASA C-20-002.pdf.
Cox, LA. (2020b). Letter from Louis Anthony Cox, Jr., Chair, Clean Air Scientific Advisory
Committee, to Administrator Andrew R. Wheeler. Re:CASAC Review of the EPA's
Policy Assessment for the Review of the Ozone National Ambient Air Quality Standards
(External Review 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/264cbl227d55e02c85257402007446a4/4713
D217BC07103485258515006359BA/$File/EPA-CASAC-20-003.pdf.
Frey, HC. (2014a). Letter from Dr. H. Christopher Frey, Chair, Clean Air Scientific Advisory
Committee, to Administrator Gina McCarthy. Re: CASAC Review of the EPA's Welfare
Risk and Exposure Assessment for Ozone (Second External Review Draft). June 18,
2014. EPA-CASAC-14-003. Office of the Administrator, Science Advisory Board
Washington, DC. Available at:
httpY/nepis. epa.gov/Exe/ZyPDF. cgi?Dockey=Pl 00JMSY.PDF.
Frey, HC. (2014b). Letter from Dr. H. Christopher Frey, Chair, Clean Air Scientific Advisory
Committee to Honorable Gina McCarthy, Administrator, US EPA. Re: CASAC Review
of the EPA's Second Draft Policy Assessment for the Review of the Ozone National
Ambient Air Quality Standards. June 26, 2014. EPA-CASAC-14-004. Office of the
Administrator, Science Advisory Board Washington, DC. Available at:
https.Y/nepis. epa.gov/Exe/ZyPURL. cgi?Dockey=Pl 00JR6F. txt.
Frey, HC. (2014c). Letter from Dr. H. Christopher Frey, Chair, Clean Air Scientific Advisory
Committee, to Administrator Gina McCarthy. Re: Health Risk and Exposure Assessment
for Ozone (Second External Review Draft - February 2014) EPA-CASAC-14-005. Office
of the Administrator, Science Advisory Board Washington, DC. Available at:
https.Y/nepis. epa.gov/Exe/ZyPURL. cgi?Dockey=Pl 00JR8I. txt.
1-15
-------
Pruitt, E. (2018). Memorandum from E. Scott Pruitt, Administrator, U.S. EPA to Assistant
Administrators. Back-to-Basics Process for Reviewing National Ambient Air Quality
Standards. May 9, 2018. Office of the Administrator U.S. EPA HQ, Washington DC.
Available at: https://www. epa.gov/criteria-air-pollutants/back-basics-process-reviewing-
national-ambient-air-quality-standards.
Samet, JM. (2010). Letter from Jonathan Samet, Chair, Clean Air Scientific Advisory
Committee, to Administrator Lisa Jackson. Re: CASAC Review of EPA's Proposed
Ozone National Ambient Air Quality Standard (Federal Register, Vol. 75, Nov. 11,
January 19, 2010).. February 19, 2010. EPA-CASAC-10-007. Office of the
Administrator, Science Advisory Board U.S. EPA HQ, Washington DC. Available at:
https://nepis.epa.gov/Exe/ZyPURL.cgi?Dockey=P10072Tl.txt.
Samet, JM. (2011). Letter from Jonathan Samet, Chair, Clean Air Scientific Advisory
Committee, to Administrator Lisa Jackson. Re: CASAC Response to Charge Questions
on the Reconsideration of the 2008 Ozone National Ambient Air Quality Standards..
March 30, 2011. EPA-CASAC-11-004. Office of the Administrator, Science Advisory
Board U.S. EPA HQ, Washington DC. Available at:
https://yosemite.epa.gov/sab/sabproduct.nsf/368203f97al5308a852574ba005bbd01/F08
BEB48C1139E2A8525785E006909AC/$File/EPA-CASAC-l l-004-unsigned+.pdf.
U.S. DHEW (1970). Air Quality Criteria for Photochemical Oxidants. National Air Pollution
Control Administration Washington, DC. U.S. DHEW. publication no. AP-63. NTIS,
Springfield, VA; PB-190262/BA.
U.S. EPA (1978). Air Quality Criteria for Ozone and Other Photochemical Oxidants
Environmental Criteria and Assessment Office. Research Triangle Park, NC. EPA-600/8-
78-004. April 1978. Available at:
https.Y/nepis. epa.gov/Exe/ZyPURL. cgi?Dockey=200089CW. txt.
U.S. EPA (1986). Air Quality Criteria for Ozone and Other Photochemical Oxidants (Volume I -
V). Environmental Criteria and Assessment Office. Research Triangle Park, NC. U.S.
EPA. EPA-600/8-84-020aF, EPA-600/8-84-020bF, EPA-600/8-84-020cF, EPA-600/8-
84-020dF, EPA-600/8-84-020eF. August 1986. Available at:
https.Y/nepis. epa.gov/Exe/ZyPURL. cgi?Dockey=30001D3J. txt
https.Y/nepis. epa.gov/Exe/ZyPURL. cgi?Dockey=30001DA V. txt
https.Y/nepis. epa.gov/Exe/ZyPURL. cgi?Dockey=30001DNN. txt
https.Y/nepis. epa.gov/Exe/ZyPURL. cgi?Dockey=30001EOF. txt
https.Y/nepis. epa.gov/Exe/ZyPURL. cgi?Dockey=30001E9R. txt.
U.S. EPA (1989). Review of the National Ambient Air Quality Standards for Ozone: Policy
Assessment of Scientific and Technical Information. OAQPS Staff Paper. Office of Air
Quality Planning and Standards. Research Triangle Park, NC U.S. EPA.
U.S. EPA (1992). Summary of Selected New Information on Effects of Ozone on Health and
Vegetation: Supplement to 1986 Air Quality Criteria for Ozone and Other Photochemical
1-16
-------
Oxidants. Office of Research and Development. Washington, DC. U.S. EPA. EPA/600/8-
88/105F.
U.S. EPA (1996). Air Quality Criteria for Ozone and Related Photochemical Oxidants. Volume I
- III. Office of Research and Development Research Triangle Park, NC. U.S. EPA. EPA-
600/P-93-004aF, EPA-600/P-93-004bF, EPA-600/P-93-004cF. July 1996. Available at:
https://nepis. epa.gov/Exe/ZyPURL. cgi?Dockey=300026GN. txt
https://nepis. epa.gov/Exe/ZyPURL. cgi?Dockey=300026SH. txt
https.Y/nepis. epa.gov/Exe/ZyPURL. cgi?Dockey=l0004RHL. txt.
U.S. EPA (2006). Air Quality Criteria for Ozone and Related Photochemical Oxidants (Volume I
- III). Office of Research and Development U.S. EPA. EPA-600/R-05-004aF, EPA-
600/R-05-004bF, EPA-600/R-05-004cF February 2006. Available at:
https.Y/cfpub. epa.gov/ncea/risk/recordisplay. cfm ?deid=l49923.
U.S. EPA (2007). Review of the National Ambient Air Quality Standards for Ozone: Policy
Assessment of Scientific and Technical Information: OAQPS Staff Paper. Office of Air
Quality Planning and Standards. Research Triangle Park, NC. U.S. EPA. EPA-452/R-07-
003. January 2007. Available at:
https.Y/nepis. epa.gov/Exe/ZyPURL.cgi?Dockey=Pl0083VX. txt.
U.S. EPA (2014a). Policy Assessment for the Review of National Ambient Air Quality
Standards for Ozone (Final Report). Office of Air Quality Planning and Standards, Health
and Environmental Impacts Divison. Research Triangle Park, NC. U.S. EPA. EPA-
452/R-14-006 August 2014. Available at:
https.Y/nepis. epa.gov/Exe/ZyPDF. cgi?Dockey=Pl 00KCZ5. txt.
U.S. EPA (2014b). Welfare Risk and Exposure Assessment for Ozone (Final).. Office of Air
Quality Planning and Standards. Research Triangle Park, NC. U.S. EPA. EPA-452/P-14-
005a August 2014. Available at:
https.Y/nepis. epa.gov/Exe/ZyPURL. cgi?Dockey=Pl 00KB9D. txt.
U.S. EPA (2014c). Health Risk and Exposure Assessment for Ozone. (Final Report). Office of
Air Quality Planning and Standards. Research Triangle Park, NC. U.S. EPA. EPA-452/R-
14-004a. August 2014. Available at:
https.Y/nepis. epa.gov/Exe/ZyPURL. cgi?Dockey=Pl OOKBUF. txt.
U.S. EPA (2019). Integrated Review Plan for the Ozone National Ambient Air Quality
Standards. Office of Air Quality Planning and Standards. Research Triangle Park, NC.
U.S. EPA. EPA-452/R-19-002. Available at:
httpsY/www.epa.gov/sites/production/files/2019-08/documents/o3-irp-aug27-
2019_final.pdf.
U.S. EPA (2020). 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.Y/www.epa.gov/isa/integrated-
science-assessment-isa-ozone-and-related-photochemical-oxidants.
1-17
-------
2 AIR QUALITY
This chapter begins with an overview of O3 and other photochemical oxidants in the
atmosphere (section 2.1). Subsequent sections summarize the sources and emissions of O3
precursors (section 2.2), ambient air monitoring and data handling conventions for determining
whether the standards are met (section 2.3), O3 concentrations measured in the U.S. ambient air
(section 2.4), and available evidence and information related to background O3 in the U.S.
(section 2.5). These focus primarily on tropospheric O3 and surface-level concentrations
occurring in ambient air1.
2.1 03 AND PHOTOCHEMICAL OXIDANTS IN THE ATMOSPHERE
O3 is one of a group of photochemical oxidants formed in the troposphere2 by
photochemical reactions of precursor gases in the presence of sunlight (ISA, Appendix 1, section
l.l)3 and is generally not directly emitted from specific sources. Tropospheric O3 and other
oxidants, such as peroxyacetyl nitrate (PAN) and hydrogen peroxide, form in polluted areas by
atmospheric reactions involving two main classes of precursor pollutants: volatile organic
compounds (VOCs) and nitrogen oxides (NOx). This occurs especially during the summer, as a
result of the photolysis of primary pollutants such as nitrogen dioxide (NO2). The reaction is
disrupted by the presence of VOCs, the radical that results from methane (CH4) oxidation; or a
reaction between carbon monoxide (CO) and the hydroxyl radical (OH) in the atmosphere. Thus,
the substances NOx, VOC, CH4 and CO are considered to be the primary precursors of
tropospheric O3. The formation of O3, other oxidants and oxidation products from these
precursors is a complex, nonlinear function of many factors including (1) the intensity and
spectral distribution of sunlight; (2) atmospheric mixing; (3) concentrations of precursors in the
ambient air and the rates of chemical reactions of these precursors; and (4) processing on cloud
and aerosol particles (ISA, Appendix 1, section 1.4; 2013 ISA, section 3.2).
Rather than varying directly with emissions of its precursors, O3 changes in a nonlinear
fashion with the concentrations of its precursors (2013 ISA, section 3.2.4). Emissions of NOx
lead to both the formation and destruction of O3, depending on the local quantities of NOx,
VOCs, radicals, and sunlight. O3 chemistry is often described in terms of which precursors most
1 Ambient air means that portion of the atmosphere, external to buildings, to which the general public has access
(see 40 CFR 50.1(e)).
2 Ozone also occurs in the stratosphere, where it serves the beneficial role of absorbing the sun's harmful ultraviolet
radiation and preventing the majority of this radiation from reaching the Earth's surface.
3 The only other appreciable source of O3 to the troposphere is transport from the stratosphere, as described in
section 2.5.1.1 below.
2-1
-------
directly impact formation rates. A NOx-limited regime indicates that O3 concentrations will
decrease in response to decreases in ambient NOx concentrations and vice-versa. These
conditions tend to occur when NOx concentrations are generally low compared to VOC
concentrations and during warm, sunny conditions when NOx photochemistry is relatively fast.
NOx-limited conditions are more common during daylight hours, in the summertime, in
suburban and rural areas, and in portions of the country with high biogenic VOC emissions like
the Southeast. In contrast, NOx-saturated conditions (also referred to as VOC-limited or radical-
limited) indicate that O3 will increase as a result of NOx reductions but will decrease as a result
of VOC reductions (2013 ISA, section 3.2; 2006 AQCD, chapter 2). NOx-saturated conditions
occur at times when and at locations with lower levels of available sunlight, resulting in slower
photochemical formation of O3, and when NOx concentrations are in excess compared to VOC
concentrations. NOx-saturated conditions are more common during nighttime hours, in the
wintertime, and in densely populated urban areas or industrial plumes. These varied relationships
between precursor emissions and O3 chemistry result in localized areas in which O3
concentrations are suppressed compared to surrounding areas, but which contain NO2 that
contributes to subsequent O3 formation further downwind (2013 ISA, section 3.2.4).
Consequently, O3 response to reductions in NOx emissions is complex and may include
decreases in O3 concentrations at some times and locations and increases in O3 concentrations at
other times and locations. Over the past decade, there have been substantial decreases in NOx
emissions in the U.S. (see Figure 2-2) and many locations have transitioned from NOx-saturated
to NOx-limited (Jin et al., 2017) during times of year that are conducive to O3 formation
(generally summer). As these NOx emissions reductions have occurred, lower O3 concentrations
have generally increased while the higher O3 concentrations have generally decreased, resulting
in a compressed O3 distribution, relative to historical conditions (ISA, Appendix 1, section 1.7).
Prior to 1979, the indicator for the NAAQS for photochemical oxidants was total
photochemical oxidants (36 FR 8186, April 30, 1971). Early ambient air monitoring indicated
similarities between O3 measurements and the photochemical oxidant measurements, as well as
reduced precision and accuracy of the latter (U.S. EPA, 1978). To address these issues, the EPA
established O3 as the indicator for the NAAQS for photochemical oxidants in 1979 (44 FR 8202,
February 8, 1979), and it is currently the only photochemical oxidant other than nitrogen dioxide
that is routinely monitored in a national ambient air monitoring network.
O3 is present not only in polluted urban atmospheres, but throughout the troposphere,
even in remote areas of the globe. The same basic processes involving sunlight-driven reactions
of NOx, VOCs, and CO contribute to O3 formation throughout the troposphere. These processes
also lead to the formation of other photochemical products, such as PAN, HNO3, and H2SO4, and
2-2
-------
to other gaseous compounds, such as HCHO and other carbonyl compounds, as well as a number
of particulate compounds (ISA, Appendix 1, section 1.4; 2013 ISA, section 3.2).
As mentioned above, the formation of O3 from precursor emissions is also affected by
meteorological parameters such as the intensity of sunlight and atmospheric mixing (2013 ISA,
section 3.2). Major episodes of high O3 concentrations in the eastern U.S. are often associated
with slow-moving high-pressure systems which can persist for several days. High pressure
systems during the warmer seasons are associated with the sinking of air, resulting in warm,
generally cloudless skies, with light winds. The sinking of air results in the development of
stable conditions near the surface which inhibit or reduce the vertical mixing of O3 precursors,
concentrating them near the surface. Photochemical activity involving these precursors is
enhanced because of higher temperatures and the availability of sunlight during the warmer
seasons. In the eastern U.S., concentrations of O3 and other photochemical oxidants are
determined by meteorological and chemical processes extending typically over areas of several
hundred thousand square kilometers. Therefore, O3 episodes are often regarded as regional in
nature, although more localized episodes often occur in some areas, largely the result of local
pollution sources during summer, e.g., Houston, TX (2013 ISA, section 2.2.1; Webster et al.,
2007). In addition, in some parts of the U.S. (e.g., Los Angeles, CA), mountain barriers limit O3
dispersion and result in a higher frequency and duration of days with elevated O3 concentrations
(2013 ISA, section 3.2).
More recently, high O3 concentrations of up to 150 parts per billion (ppb)4 have been
measured during the wintertime in two western U.S. mountain basins (ISA, Appendix 1, section
1.4.1). Wintertime mountain basin O3 episodes occur on cold winter days with low wind speeds,
clear skies, substantial snow cover, extremely shallow boundary layers driven by strong
temperature inversions, and substantial precursor emissions activity from the oil and gas sector.
The results of recent modeling studies suggest that photolysis of VOCs provides the source of
reactive chemical species (radicals) needed to initiate the chemistry driving these wintertime O3
episodes. This mechanism is markedly different from the chemistry driving summertime O3
formation, which is initiated with the photolysis of NO2 followed by the formation of the OH
radicals (ISA, Appendix 1, section 1.4.1).
O3 concentrations in a region are affected both by local formation and by transport of O3
and its precursors from upwind areas. O3 transport occurs on many spatial scales including local
transport within urban areas, regional transport over large regions of the U.S., and long-range
transport which may also include international transport. In addition, O3 can be transferred into
4 Although the standards are specified in ppm (e.g., as described in Chapter 1), the units, ppb, are commonly used in
describing O3 concentrations throughout this document, with 0.070 ppm being equivalent to 70 ppb.
2-3
-------
the troposphere from the stratosphere, which is rich in naturally occurring O3, through
stratosphere-troposphere exchange (STE). These intrusions usually occur behind cold fronts,
bringing stratospheric air with them and typically affect O3 concentrations in higher elevation
areas (e.g. > 1500 m) more than areas at lower elevations, as discussed in section 2.5.3.2 (ISA,
Appendix 1, section 1.3.2.1; 2013 ISA, section 3.4.1.1).
2.2 SOURCES AND EMISSIONS OF 03 PRECURSORS
Sources of emissions of O3 precursor compounds can be divided into anthropogenic and
natural source categories, with natural sources further divided into emissions from biological
processes of living organisms (e.g., plants, microbes, and animals) and emissions from chemical
or physical processes (e.g., biomass burning, lightning, and geogenic sources). Anthropogenic
emissions associated with combustion processes, including mobile sources and power plants,
account for the majority of U.S. NOx and CO emissions (Figure 2-1 and Figure 2-2). Emissions
of these chemicals from mobile sources have declined appreciably since 2002 (Figure 2-2).
Anthropogenic sources are also important for VOC emissions, though in some locations and
times of the year (e.g., southern states during summer) the majority of VOC emissions come
from vegetation (2013 ISA, section 3.2.1)5. In practice, the distinction between natural and
anthropogenic sources is often unclear, as human activities directly or indirectly affect emissions
from what would have been considered natural sources during the preindustrial era. Thus,
precursor emissions from plants, animals, and wildfires could be considered either natural or
anthropogenic, depending on whether emissions result from agricultural practices, forest
management practices, lightning strikes, or other types of events. Additional challenges are
presented because much O3 results from reactions between anthropogenic and natural precursors
(ISA, Appendix 1, section 1.8.1.2).
The National Emissions Inventory (NEI) is a comprehensive and detailed estimate of air
emissions of criteria pollutants, precursors to criteria pollutants, and hazardous air pollutants
from air emissions sources (U.S. EPA, 2018c). The NEI is released every three years based
primarily upon data provided by State, Local, and Tribal air agencies for sources in their
jurisdictions and supplemented by data developed by the US EPA. The NEI is built using the
EPA's Emissions Inventory System (EIS) first to collect the data from State, Local, and Tribal
air agencies and then to blend that data with other data sources.6
5 It should be noted that the definition of VOCs used in this section does not include CH4 because it is excluded
from the EPA's regulatory definition of VOCs in 40 CFR 51.100 (s). More information about this regulatory
definition of VOCs is available at https://www.epa.gov/indoor-air-quality-iaq/technical-overview-volatile-organic-
compounds.
6 More details are available from: https://www.epa.gov/enviro/nei-overview.
2-4
-------
Accuracy in an emissions inventory reflects the extent to which the inventory represents
the actual emissions that occurred. Anthropogenic emissions of air pollutants result from a
variety of sources such as power plants, industrial sources, motor vehicles and agriculture. The
emissions from any individual source typically varies in both time and space. For the thousands
of sources that make up the NEI, there is uncertainty in one or both of these factors. For some
sources, such as power plants, direct emission measurements enable the emission factors derived
from them to be more certain than sources without such direct measurements. However, it is not
practically possible to directly monitor each of the emission sources individually and, therefore,
emission inventories necessarily contain assumptions, interpolation and extrapolation from a
limited set of sample data (U.S. EPA, 2018c).
2-5
-------
A) NOx (14,366 kTon/yr)
Commercial
Marine
Vessels
9%
On-Road
Diesel
Heavy Duty
Vehicles
15%
On-Road non-Diesel
Light Duty Vehicles
17%
Non-Road Biogenics -
Equipment- Vegetation and Soil
Diesel 6%
8%
Locomotives
5%
Oil & Gas
Production
5%
Industrial
Boilers, ICEs
Natural Gas
4%
Non-Road
Equipment
- Gasoline
Residential-Natural 2%
Gas Combustion
2%
Electric
Generation
Combustion
Other
17%
B) CO (72,353 kTons/yr)
On-Road non-
Diesel Light
Duty Vehicles
31%
Non- Road
Equipment-
Gasoline
16%
Other
11%
w
Wildfires
15%
Prescribed
Fires
12%
Biogenics-
fuel Comb-
Vegetation
Residential-Woocf
and Soil
3%
9%
C) VOCs (55,630 kTon/yr)
Vegetation and Soil (Biogenics)
69%
Other
8%
Non-Road
Equipment
Gasoline
3%
9P
Wildfires
Prescribed
Oil & Gas
Production
6%
D) CH4 (26,298 kTon/yr)
Other
Petroleum Systems 3%
6%
Coal Mining
8%
Landfills
16%
Agriculture-
Animal
Husbandry
36%
Consumer &
Commercial
Solvent Use
|On-Road non-
Diesel Light
Duty Vehicles
3%
Natural Gas
Systems
25%
Source: Based on Figure 1-2 of ISA Appendix 1. Sources are the 2014 National Emissions Inventory, version 2 (U.S. EPA,
2018c) for panels A-C, and the Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2016 (U.S. EPA, 2018b) for panel
D. Categories contributing less than 2% each have been summed and are represented by the "other" category.
Figure 2-1. U.S. O3 precursor emissions by sector: A) NOx; B) CO; C) VOCs; D) CHt.
2-6
-------
Solvent
Off-Highway
Equipment
Fuel
Combustion
EGUs
Agriculture • Animal Husbandry
Natural Gas Systems
Landfills
Coal Mining
Petroleum Systems
Other
D) CH,
10000
6000
2000
0
Agriculture - Animal Husbandry
"O
0 8000 Natural Gas Systems
Landfills
c 4000 oth„
O omer Coal Mining
Petroleum
2002 2005 2008 2011 2014 2016
Year
A)
NOx
12000
10000
8000
6000
4000
2000
0
B)
o
o
CO
60000
50000
40000
30000
20000
10000
0
C)
o
>
2002 2005 2008 2011 2014 2017
Inventory Year
VOCs
4500
4000
3500
3000
2500
2000
1500
1000
500
0
Highway
Vehicles
Petroleum
& Related
Industries
Wildfires
2002 2005 2008 2011 2014 2017
Inventory Year
Highway Vehicles
Highway Vehicles
2002 2005 2008 20112014 2017
Inventory Year
Off Highway
Equipment
Off-Highway
Equipment
Wildfires
Leeend: NOx. CO, VOCs
Petroleum & Related industries
1 Fuel Combustion - EGUs
¦ Other Industrial Processes
Miscellaneous (w/o Wildfires)
— Storage and Transport
Chemical & Allied Product MFG
Waste Disposal & Recycling
Fuel Combustion - Industrial
¦ Highway Vehicles
Solvent
Off-Highway Equipment
Metals Processing
1 Wildfires
Fuel Combustion - Other
Source: Based on Figure 1-3 of ISA Appendix 1. Sources are the EPA's Emissions Inventory System (EIS) for panels A-C, and
the Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2016 (U.S. EPA, 2018a) for panel D. Estimates for 2017
come from air pollutant emissions trends estimates available on the EPA's website (https://www.epa.gov/air-emissions-
imentories/air-pollutant-emissions-trends-data). Categories contributing less than 2% each have been summed and are
represented by the "other" category. Sources shown generate 90% or more of the estimated U.S. anthropogenic emissions.
Figure 2-2. U.S. anthropogenic O3 precursor emission trends for: A) NOx; B) CO; C)
VOCs; and D) CH4.
2-7
-------
Figure 2-3, Figure 2-4 and Figure 2-5 show county-level estimates of U.S. emissions
densities (in tons/year/mi2) for CO, NOx, and VOCs, respectively. In general, CO and NOx
emissions tend to be highest in urban areas which typically have the most anthropogenic sources,
however, CO emissions may be higher in some rural areas due to fires, and similarly NOx
emissions may be higher in some rural areas due to sources such as electricity generation, oil and
gas extraction, and traffic along major highways. While there are some significant anthropogenic
sources of VOC emissions in urban areas, in rural areas the vast majority of VOC emissions
come from plants and trees (biogenics), particularly in the southeastern U.S. In other areas of the
U.S., such as the Great Plains region and parts of the inter-mountain west, areas with higher
levels of VOC emissions are largely due to oil and gas extraction (U.S. EPA, 2018c).
It should be noted that O3 levels in a given area are impacted by both local emissions that
form O3 in the area as well as remote emissions that form O3 which is then transported into the
area. Biogenic VOC emissions that lead to O?, formation may vary greatly depending on the type
and amount of vegetation, which is generally much lower in urban areas than in rural areas.
However, biogenic VOC emissions that are upwind of an urban area can have a significant
impact on urban O3 levels. Thus, while the county-level maps shown in Figure 2-3, Figure 2-4
and Figure 2-5 illustrate the variability in precursor emissions in the U.S., it is not sufficient to
look only at the patterns in local emissions when considering the impact on O3 concentrations.
Total Carbon Monoxide Emissions Density (tons/year/miA2)
O 0-9(788) ~ 10-19(918) ~ 20-49(958) ® 50-99(335) ¦ 100-5,198(221)
Source: 2014 National Emissions Inventory, version 2 (US EPA, 2018c; data downloaded from
https://edap.epa.gov/public/extensions/nei_report_2014/dashboard.html)
Figure 2-3. U.S. county-level CO emissions density estimates (tons/year/mi2) for 2014.
2-8
-------
Total Nitrogen Oxides Emissions Density (tons/year/miA2)
~ 2-4(1165) H 5-9 (603) H 10-19 (294)
Source: 2014 National Emissions Inventory, version 2 (US EPA, 2018c; data downloaded from
https://edap.epa.gov/public/extensions/nei_report_2014/dashboard.html)
Figure 2-4. U.S. county-level NOx emissions density estimates (tons/year/mi2) for 2014.
Total Volatile Organic Compounds Emissions Density (tons/year/miA2)
n 10-19(910) ~ 20-49(1327) Q 50-99(125) ¦
Source: 2014 National Emissions Inventory, version 2 (US EPA, 2018c; data downloaded from
https://edap. epa. gov/public/extensions/nei_report_2014/dashboard. htm!)
Figure 2-5. U.S. county-level VOC emissions density estimates (tons/year/mi2) for 2014.
2-9
-------
2.3 AMBIENT AIR MONITORING AND DATA HANDLING
CONVENTIONS
2.3.1 Ambient Air Monitoring Requirements and Monitoring Networks
State and local environmental agencies operate O3 monitors at state or local air monitoring
stations (SLAMS) as part of the SLAMS network. The requirements for the SLAMS network
depend on the population and most recent O3 design values7 in the area. The minimum number
of O3 monitors required in a metropolitan statistical area (MSA) ranges from zero for areas with
a population less than 350,000 and no recent history of an O3 design value greater than 85
percent of the level of the standard, to four for areas with a population greater than 10 million
and an O3 design value greater than 85 percent of the standard level.8 Within an O3 monitoring
network, at least one site for each MSA must be designed to record the maximum concentration
for that particular metropolitan area. Siting criteria for SLAMS includes horizontal and vertical
inlet probe placement; spacing from minor sources, obstructions, trees, and roadways; inlet probe
material; and sample residence times.9 Adherence to these criteria ensures uniform collection and
comparability of O3 data. Since the highest O3 concentrations tend to be associated with a
particular season for various locations, the EPA requires O3 monitoring during specific O3
monitoring seasons (shown in Figure 2-6) which vary by state from five months (May to
September in Oregon and Washington) to all twelve months (in 11 states), with the most
common season being March to October (in 27 states).10
Most of the state, local, and tribal air monitoring stations that report data to the EPA use
ultraviolet Federal Equivalent Methods (FEMs). The Federal Reference Method (FRM) was
revised in 2015 to include a new chemiluminescence by nitric oxide (NO-CL) method. The
previous ethylene (ET-CL) method, while still included in the CFR as an acceptable method, is
no longer used due to lack of availability and safety concerns with ethylene.11 The NO-CL
method is beginning to be implemented in the SLAMS network.
7 A design value is a statistic that summarizes the air quality data for a given area in terms of the indicator, averaging
time, and form of the standard. Design values can be compared to the level of the standard and are typically used
to designate areas as meeting or not meeting the standard and assess progress towards meeting the NAAQS.
8 The SLAMS minimum monitoring requirements to meet the O3 design criteria are specified in 40 CFR Part 58,
Appendix D. The minimum O3 monitoring network requirements for urban areas are listed in Table D-2 of
Appendix D to 40 CFR Part 58 (accessible at https://www.ecfr.gov).
9 The probe and monitoring path siting criteria for ambient air quality monitoring are specified in 40 CFR, Part 58,
Appendix E.
10 The required O3 monitoring seasons for each state are listed in Table D-3 of Appendix D to 40 CFR Part 58.
11 The current FRM for 03 (established in 2015) is a chemiluminescence method, which is fully described in
Appendix D to 40 CFR Part 50.
2-10
-------
Ambient air quality data and associated quality assurance (QA) data are reported to the
EPA via the Air Quality System (AQS). Data are reported quarterly and must be submitted to
AQS within 90 days after the end of the quarterly reporting period. Each monitoring agency is
required to certify data that is submitted to AQS from the previous year. The data are certified,
taking into consideration any QA findings, and a data certification letter is sent to the EPA
Regional Administrator. Data must be certified by May 1st of the following year. Data collected
by FRM or FEM monitors that meet the QA requirements must be certified.12 To provide
decision makers with an assessment of data quality, the EPA's QA group derives estimates of
both precision and bias for O3 and the other gaseous criteria pollutants from quality control (QC)
checks using calibration gas, performed at each site by the monitoring agency. The data quality
goal for precision and bias is 7 percent.13
12 Quality assurance requirements for monitors used in evaluations of the NAAQS are provided in 40 CFR Part 58,
Appendix A.
13 Annual summary reports of precision and bias can be obtained for each monitoring site at
https://www.epa.gov/outdoor-air-quality-data/single-point-precision-and-bias-report.
2-11
-------
I 9
OR
5-9
CA
1-12
* Mi
x
4-10
Alaska
WA
4-9
5 9
ME
MT
NO
3-9
4-9
Wl
SO
NY
MA "\ >
t §~>
4-9
3-10
3-10
\VY
I WIS \ J 3-10
1-9
p* >5
310 /
7^'tO.OE
/ T&g
3-10 S.J'
VA -4'
HE
3-10
OH
3-10
KV
3-10
3-10
UT
1 12
CO
'3-10
f 7' KY
1-12
.3-10
KS
MO
3-10
3-10
NC
3-10
OK
AZ
AR
MM
3-11
1-12
3-11
1-12
Al_
GA
MS
3-11
3-10
3-10
3-10
3-10
A 1 1
M2
Puerto Rico &
Virgin Islands
1-12
1-12
Hawaii
1-12
3-9:
NH
MA
3-10:
MD
DC
Figure 2-6. Current O3 monitoring seasons in the U.S. Numbers in each state indicate the months of the year the state is required
to monitor for O3 (e.g., 3-10 means O3 monitoring is required from March through October).
2-12
-------
In 2018, there were over 1,300 federal, state, local, and tribal ambient air monitors
reporting O3 concentrations to the EPA. Figure 2-7 shows the locations of such monitoring sites
that reported data to the EPA at any time during the 2016-2018 period. Nearly 80% of this
network are SLAMS monitors operated by state and local governments to meet regulatory
requirements and provide air quality information to public health agencies; these sites are largely
focused on urban and suburban areas.
Two important subsets of SLAMS sites separately make up the National Core (NCore)
multi-pollutant monitoring network and the Photochemical Assessment Monitoring Stations
(PAMS) network. Each state is required to have at least one NCore station, and O3 monitors at
NCore sites are required to operate year-round. At each NCore site located in a MSA with a
population of 1 million or more (based on the most recent census), a PAMS network site is
required.14 At a minimum, monitoring sites in the PAMS network are required to measure certain
O3 precursors during the months of June, July and August, although some precursor monitoring
may be required for longer periods of time to improve the usefulness of data collected during an
area's O3 season (U.S. EPA, 2018a).
In addition to reporting O3 concentrations, the NCore and PAMS networks provide data
on O3 precursor chemicals. The NCore sites feature co-located measurements of chemical
species such as nitrogen oxide and total reactive nitrogen, along with various meteorological
measurements. The additional data collected at the PAMS sites include measurements of NOx,
and a target set of VOCs. The enhanced monitoring at sites in these two networks informs our
understanding of local O3 formation.
While the SLAMS network has a largely urban and population-based focus, there are
monitoring sites in other networks that can be used to track compliance with the NAAQS in rural
areas. For example, the Clean Air Status and Trends Network (CASTNET) monitors are located
in rural areas. There were 77 CASTNET sites operating in 2018, with most of the sites in the
eastern U.S. being operated by the EPA, and most of the sites in the western U.S. being operated
by the National Park Service (NPS). Finally, there are also a number of Special Purpose
Monitoring Stations (SPMs), which are not required but are often operated by air agencies for
short periods of time (less than 3 years) to collect data for human health and welfare studies, as
well as other types of monitoring sites, including monitors operated by tribes and industrial
sources. The SPMs are typically not used to assess compliance with the NAAQS.15
14 The requirements for PAMS, which were most recently updated in 2015, is fully described in section 5 of
Appendix D to 40 CFR Part 58.
15 However, SPMs that use federal reference or equivalent methods, meet all applicable requirements in 40 CFR Part
58, and operate continuously for at least 3 years may be used to assess compliance with the NAAQS.
2-13
-------
. • *%
:r£
>WWJ
V
• •
• SLAMS (979) • CASTNET (77) • NCORE/PAMS (129) • SPM/OTHER (210)
Figure 2-7. Map of U.S. ambient air O3 monitoring sites reporting data to the EPA during
the 2016-2018 period.
2.3.2 Data Handling Conventions and Computations for Determining Whether the
Standards are Met
To assess whether a monitoring site or geographic area (usually a county or urban area)
meets or exceeds a NAAQS, the monitoring data are analyzed consistent with the established
regulatory requirements for the handling of monitoring data for the purposes of deriving a design
value. A design value summarizes ambient air concentrations for an area in terms of the
indicator, averaging time and form for a given standard such that its comparison to the level of
the standard indicates whether the area meets or exceeds the standard. The procedures for
calculating design values for the current O3 NAAQS (established in 2015) are detailed in
Appendix U to 40 CFR Part 50 and are summarized below.
Hourly average O3 concentrations at the monitoring sites used for assessing whether an
area meets or exceeds the NAAQS are required to be reported in ppm to the third decimal place,
with additional digits truncated, consistent with the typical measurement precision associated
with most O3 monitoring instruments. Monitored hourly O3 concentrations flagged by the States
as having been affected by an exceptional event, having been the subject of a demonstration, and
having received concurrence from the appropriate EPA Regional Office, are excluded from
2-14
-------
design value calculations consistent with 40 CFR 50.14.16 The hourly concentrations are used to
compute moving 8-hour averages, which are stored in the first hour of each 8-hour period (e.g.,
the 8-hour average for the 7:00 AM to 3:00 PM period is stored in the 7:00 AM hour), and digits
to the right of the third decimal place are truncated. Each 8-hour average is considered valid if 6
or more hourly concentrations are available for the 8-hour period.
Next, the daily maximum 8-hour average (MDA8) concentration for each day is
identified as the highest of the 17 consecutive, valid 8-hour average concentrations beginning at
7:00 AM and ending at 11:00 PM (which includes hourly O3 concentrations from the subsequent
day). MDA8 values are considered valid if at least 13 valid 8-hour averages are available for the
day, or if the MDA8 value is greater than the level of the NAAQS. Finally, the O3 design value is
calculated as the 3-year average of the annual 4th highest MDA8 value17. An O3 design value less
than or equal to the level of the NAAQS is considered to be valid if valid MDA8 values are
available for at least 90% of the days in the O3 monitoring season (as defined for each state and
shown in Figure 2-6) on average over the 3 years, with a minimum of 75% data completeness in
any individual year. Design values greater than the level of the NAAQS are always considered to
be valid.
An O3 monitoring site meets the NAAQS if it has a valid design value less than or equal
to the level of the standard, and it exceeds the NAAQS if it has a design value greater than the
level of the standard. A geographic area meets the NAAQS if all ambient air monitoring sites in
the area have valid design values meeting the standard. Conversely, if one or more monitoring
sites has a design value exceeding the standard, then the area exceeds the NAAQS.
2.4 03 IN AMBIENT AIR
2.4.1 Concentrations Across the U.S.
Figure 2-8 below shows a map of the O3 design values at U.S. ambient air monitoring
sites based on data from the 2016-2018 period. From the figure it is apparent that many
monitoring sites have design values exceeding the current NAAQS, and that most of these sites
are located in or near urban areas. The highest design values are located in California, Texas,
along the shoreline of Lake Michigan, and near large urban areas in the northeastern and western
U.S. There are also high design values associated with wintertime O3 in the Uinta Basin in Utah.
16 A variety of resources and guidance documents related to identification and consideration of exceptional events in
design value calculations are available at https://www.epa.gov/air-quality-analysis/final-2016-exceptional-events-
rule-supporting-guidance-documents-updated-faqs.
17 Design values are reported in ppm to the third decimal place, with additional digits truncated. This truncation step
also applies to the initially calculated 8-hour average concentrations (Appendix 2 A, section 2A.1).
2-15
-------
The lowest design values are located in the north central region of the U.S., rural parts of New
England and the southeastern U.S., and along the Pacific Ocean, including Alaska and Hawaii.
• 41-60 ppb (157 sites) • 66 -70 ppb (346 sites) • 76 - 111 ppb (109 sites)
O 61 - 65 ppb (346 sites) © 71 - 75 ppb (164 sites)
Figure 2-8. Os design values in ppb for the 2016-2018 period.
2.4.2 Trends in U.S. O3 Concentrations
Figure 2-9 shows a map of the site-level trends in the O3 design values at U.S. monitoring
sites having complete data18 from 2000-2002 through 2016-2018. The trends were computed
using the Thiel Sen estimator (Sen, 1968; Thiel, 1950), and tests for significance were computed
using the Mann-Kendall test (Kendall, 1948; Mann, 1945). From this figure it is apparent that
design values have decreased significantly over most of the eastern U.S. during this period.
These decreases are in part due to EPA programs such as the Clean Air Interstate Rule and the
Cross-State Air Pollution Rule with the goal of achieving broad, regional reductions in
summertime NOx emissions, as well as mobile emission reductions from federal motor vehicle
emissions and fuel standards and local controls resulting from implementation of the existing O3
standards. Other areas of the country have also experienced decreases in design values, most
notably in California and near urban areas in the intermountain west.
18 The data completeness criteria for Figure 2-8 through Figure 2-14 are listed in Table 2A-1 of Appendix 2A.
2-16
-------
~ Decreasing > 1 ppb/yr (364 sites) ° No Significant Trend (37 sites)
v Decreasing < 1 ppb/yr (223 sites) A increasing < 1 ppb/yr (5 sites)
Figure 2 9. Trends in O3 design values based on data from 2000-2002 through 2016-2018.
Figure 2-10 shows the national trend in the annual 4th highest MDA8 values based on 196
ambient air monitoring sites with complete data from 1980 to 2018. This figure shows that, on
average, there has been a 31% decrease in U.S. annual 4th highest MDA8 levels since 1980.
Since relatively few sites have been monitoring continuously since 1980, Figure 2-11 shows the
national trend in the annual 4th highest MDA8 values and the design values based on the 870
monitoring sites with complete data from 2000 to 2018. The U.S. median annual 4th highest
MDA8 values decreased by 25% nationally from 2002 (88 ppb) to 2013 (66 ppb), with some
variability among individual years in this period which can generally be attributed to changes in
meteorological conditions. Similarly, the U.S. median design value decreased by 20% from
2000-2002 (84 ppb) to 2013-2015 (67 ppb). However, the trend in the annual 4th highest MDA8
concentrations has been relatively flat since 2013, and the design values have been relatively
constant since 2015. In general, the design value metric is more stable and therefore better
reflects long-term changes in Os than the annual 4th highest MDA8 metric.
2-17
-------
Ozorie Air Quality. 1980 - 2018
(Annual 4th Maximum of Daily Max 8-Hour Average)
National Trend based on 196 Sites
015
e
CL
Q.
010-
£Z
o
J—
ra
A—-
c
National standard
$ 0 05-
ooo
99999999999999999999000 0 00000000 0 000000
8888888888999999999900000000001111 1 1111
0123456789Q1234567890123456789012345678
1980 to 2018 : 31 % decrease in National Average
Source: EPA's Air Trends website (https://www.epa.gov/air-trends/ozone-trends/).
Figure 2-10. National trend in annual 4th highest MDA8 values, 1980 to 2018. The white
center line is the average while the filled area represents the range between the 10lil
and 90th percentiles. The dotted line is the level of the standard.
0
/
\
10th/90th Percentile Annual Value
Median Annual Value
/
\
\
—r
\
11
10th/90th Percentile Design Value
Median Design Value
0
^ \
\
\
t-urreni iianaara
(7)
S
Q_
\
\
/
/
-
*s
\
\
d)
3
CCS 0
/
^/
\ \
\ v
V
\
~
1*
if
~
\
%
> 00 -
X
o
-------
Figure 2-12 shows regional trends in the median annual 4th highest MDA8 values for the
9 National Oceanic and Atmospheric Administration (NOAA) climate regions19 based on
ambient air monitoring sites with complete O3 monitoring data for 2000-2018. The five eastern
U.S. regions (Central, East North Central, Northeast, Southeast, South) have all shown decreases
of at least 10 ppb in median annual 4th highest MDA8 values since the early 2000's, with the
Southeast region in particular showing the largest decrease of over 20 ppb. On the other hand,
the median annual 4th highest MDA8 values have changed by less than 10 ppb in each of the four
western U.S. regions (Northwest, Southwest, West, West North Central). The large increase in
the Northwest region in 2017 and 2018 is largely due to the influence of wildfires.
100
90
80
70
60
Regional Trends in Annual 4th Highest Daily Maximum 8-hour Ozone (ppb)
Figure 2-12. Regional trends in median annual 4th highest MDA8 concentrations, 2000 to
2018.
Trends presented in this section have focused on annual 4th high VI DAS concentrations
and design values. Additional information from the published literature has examined trends in
0
T—
CM
CO
LO
CD
r-
CO
CD
O
T—
CM
CO
LO
CD
co
0
O
O
O
O
O
O
O
O
O
x—
Y-
r—
T—
—
t—
T—
—
x—
0
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
0
O
C\J
CM
CM
CM
CM
CM
CM
CM
CM
CM
CM
CM
CM
CM
CM
CM
CM
CM
CM
NOAA Climate Region (# sites)
Central (156) Southeast (168)
East North Central (58) Southwest (60)
Northeast (143) West (151)'
Northwest (17) West North Central (13)
South (104)
19 These regions are defined per Karl and Koss (1984) as illustrated in Appendix 2B, Figure 2B-1.
2-19
-------
MDA8 concentrations across the distribution of high and low O3 days. Simon et al., 2015) found
that, similar to results presented in this section for DVs and annual 4th high MDA8
concentrations, the 95th percentile of summertime MDA8 concentrations decreased significantly
at most sites across the U.S. between 1998 and 2013. In contrast, trends over that time period for
the 5th percentile, median and mean of MDA8 varied with location and time of year. Similarly,
Lefohn et al. (2017) reported that between 1980 and 2014 there was a compression of the
distribution of measured hourly O3 values with extremely high and extremely low concentrations
becoming less common. As a result, O3 metrics impacted by high hourly O3 concentrations, such
as the annual 4th highest MDA8 value, decreased at most US sites across this period.
Concurrently, metrics that are impacted by averaging longer time periods of hourly O3
measurements, such as the 6-month (April-September) average of daytime (8am-7pm) O3
concentrations, were more varied with only about half of the sites exhibiting decreases in this
metric and most other sites exhibiting no trend (Lefohn et al., 2017).
2.4.3 Diurnal Patterns
Tropospheric O3 concentrations in most locations exhibit a diurnal pattern due to the
photochemical reactions that drive formation and destruction of O3 molecules. Figure 2-13
shows boxplots of O3 concentrations in ambient air, by hour of the day for four monitoring sites
that represent diurnal patterns commonly observed in the U.S. The top panels show diurnal
patterns, based on available data from 2015-2017, at urban (panel A) and downwind suburban
(panel B) monitoring sites in the Los Angeles metropolitan area. Both sites generally experience
their highest O3 concentrations during the early afternoon hours, and their lowest concentrations
during the early morning hours, as is typical of most urban and suburban areas in the U.S.
However, higher levels of NOx emissions near the urban site may suppress O3 formation
throughout the day and increase the O3 titration rate at night, resulting in lower O3 concentrations
than those typically observed at the downwind site.
Rural areas generally experience lower O3 concentrations than urban and suburban areas,
with less pronounced diurnal patterns. However, elevation and transport also play a larger role in
influencing concentrations in rural areas than in urban areas. The bottom panels in Figure 2-13
show diurnal patterns at low elevation (panel C) and high elevation (panel D) rural monitoring
sites in New Hampshire. The low elevation site experiences O3 concentrations that are 10-20 ppb
lower, on average, than the high elevation site. The low elevation site experiences a slight diurnal
pattern similar to that seen at the urban and suburban sites (generally related to photochemical O3
formation that increases concentrations in the late morning and afternoon), while the high
elevation site does not appear to experience any sort of diurnal pattern in O3 concentrations. The
lack of a diurnal pattern observed at this site is typical of high elevation rural sites throughout the
2-20
-------
U.S., suggesting that observed O3 concentrations are primarily driven by transport from upwind
areas rather than being formed from local precursor emissions. The presence of higher peak O3
concentrations at the high elevation site than the low elevation site at all hours of the day
indicates that the high elevation site may be influenced by transport from the free troposphere to
a greater extent than the low elevation site.
2-21
-------
AQS Site ID: 06-037-1302 Site Name: Compton
Q-100
Q-100 -
N 50
o
n 50 -
o
AQS Site ID: 06-037-0016 Site Name: Glendora
6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23
Hour (Local Standard Time)
s ID: 33-007-4002 Site Name: CAMP DODGE, GREENS GRANT
HQ-
S'
Q-100
Q-
£ 90-
S 70
2 «H
40
ttttYY
I
8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23
Hour (Local Standard Time)
: i
0
150
5 4 5 6 7 8 9 10 11 12 13 14 15
Hour (Local Standard Time)
AQS Site ID: 33-007-4001 Site Name: MT.
16 17 18 19 20
WASHINGTON
140 -
130
120-
110-
2T
Q-100-
Q.
| 90-
"5
£ !
0)
£ 70-|
^ 60
n 50
O
• , • : •
* 4-1 i «
BSBfflaaQ
PP0000B
000000
—i—t-
9 10 11 12 13 14 15 16 17 18 19 20 21 22 23
Hour (Local Standard Time)
Figure 2-13. Diurnal patterns in hourly O3 concentrations at selected monitoring sites: A) an urban site in Los Angeles; B) a
downwind suburban site in Los Angeles; C) a low elevation rural site in New Hampshire; and D) a high elevation
rural site in New Hampshire.
2-22
-------
2.4.4 Seasonal Patterns
Tropospheric O3 concentrations also tend to experience seasonal patterns due to seasonal
changes in meteorological conditions and the length and intensity of daylight. High O3
concentrations are most commonly observed on hot, sunny, and stagnant days during the spring
and summer. Figure 2-14 shows boxplots of MDA8 O3 concentrations by month of the year for
four monitoring sites that represent different kinds of seasonal patterns commonly observed in
the U.S. This figure is based on data from 2015-2017. Panel A shows the seasonal pattern for an
urban site in Baltimore, MD, which reflects the typical seasonal pattern observed at many urban
and suburban monitoring sites across the U.S. The highest O3 concentrations are observed during
May to September, when the days are the longest and solar radiation is strongest.
Panel B shows the seasonal pattern for an urban site in Baton Rouge, LA. In parts of the
southeastern U.S., the highest O3 concentrations are often observed in April and May due to the
onset of warm temperatures combined with abundant emissions of biogenic VOCs at the start of
the growing season. This is often followed by lower concentrations during the summer months,
which is associated with high humidity levels that tend to suppress O3 formation in the region
(Camalier et al., 2007). Some areas, particularly in the states bordering the Gulf of Mexico, may
experience a second peak in O3 concentrations in September and October.
Panel C shows the seasonal pattern for a high elevation rural site in Colorado. The
highest O3 concentrations in rural areas are typically observed in the spring. This can be due to
several factors, including those mentioned previously, and additionally, long-range transport
from Asia is most prevalent at this time of year. Stratospheric Tropospheric Exchange (STE)
events, which most often affect high elevation areas in the western U.S., are also most common
during the spring.
Finally, Panel D shows the seasonal pattern for a monitoring site in Utah where high
wintertime O3 concentrations were observed. Over the past decade, high O3 concentrations have
been observed in two mountain basins in the western U.S. during the winter months (December
to March). These wintertime O3 episodes require a unique set of conditions, including a shallow
inversion layer, snow cover, calm or light winds, and pervasive local NOx and VOC emissions
(in these cases, from oil and gas extraction). These conditions are relatively uncommon, and
elevated wintertime O3 levels may not occur in some years.
2-23
-------
AOS Site ID: 24-005-3001 Site Name: Essex
AOS Site ID: 22-033-0009 Site Name: Capitol
120 -
110
2-100-
Q.
a.
| 90'
03
§ 80-|
£=
o
o
70"
c
o
M
° 60
5
o
sz
^ 50"
20-
10-
40 1—
= 30-|—~LZ
T
T
—i 1 1 1 1 1 1 1 1 1 1—
Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov
AOS Site ID: 08-051-9991 Site Name: Gothic
Dec
120 -
110
-100-
90-
® 80-
70'
60"
50-
40"
f 30H
~
20-
10-
—I 1 1 1 1 1 1 1 1 1 1 1—
Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec
120
110
2100-1
Q.
Q.
| 90'
TO
§ 80-
c
o
O
«> 70-
C
o
N
° 60'
50-
5 40-
€ 30-
20'
10"
B
Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec
AOS Site ID: 49-047-2003 Site Name: Ouray
120 -
110 -
100-
'! i'
Q_
C
O
03
90-
C
O
80-
O
O
<1>
/»-
C
N
O
60-
_C
00
50-
b
_>
h
4U -
(O
>
"to
30-
L5
20-
10-
—i 1 1 1 1 1 1 1 1 1 1 1—
Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec
Figure 2-14. Seasonal patterns in MDA8 Os concentrations at selected monitoring sites
(2015-2017): A) an urban site in Baltimore, Ml); B) an urban site in Baton
Rouge, LA; C) a rural site in Colorado; and D) a site in Utah experiencing
high wintertime O3.
2-24
-------
2.4.5 Variation in Recent Daily Maximum 1-hour Concentrations
To provide a characterization of recent O3 concentrations in the U.S. for periods shorter
than 8 hours, this section presents recent O3 monitoring data in terms of daily maximum 1-hour
average (MDA1) concentrations, and their variation across monitoring sites that vary with regard
to design values for the current O3 standard.
Figure 2-15 shows boxplots of MDA1 values at U.S. monitoring sites based on 2016-
2018 data stratified by each site's 8-hour O3 design value. The boxes representing the 25th
percentile, median, and 75th percentile MDA1 values increase slightly with higher design values.
The range (min/max) of observed MDA1 values does not appear to change much, except for the
presence of higher MDA1 values up to around 160 ppb for the rightmost bin which includes only
sites that exceed the current standards. The boxplots show that there are only a small number of
MDA1 values above 120 ppb for sites that meet the current standards.
Figure 2-16 shows a scatter plot of the number of days at each monitoring site that have a
MDA1 value of 120 ppb or greater based on 2016-2018 data compared to the site's 2016-2018
design value. According to the figure, a small proportion of O3 monitoring sites in the U.S.
observe MDA1 values at or above 120 ppb more than once per year, but these sites all exceed the
current 8-hour standard. There are no sites that were meeting the current standards based on
2016-2018 data that had MDA1 values above 120 ppb more than twice over the same 3-year
period (Appendix 2A, Table 2A-2).
Figure 2-17 shows the national trend in the annual 2nd highest MDA1 O3 concentration,
which was the metric used to track progress towards meeting the 1-hour O3 NAAQS, originally
set in 1979 and later replaced by the current 8-hour metric in 1997 (62 FR 38856, July 18,
1997).20 The monitoring sites represented in Figure 2-17 are the 861 sites with complete data
from 2000 to 2018 (as summarized in Appendix 2A, Section 2A.2). The shapes of the trend lines
in Figure 2-17 are similar to those shown for the annual 4th highest MDA8 values in Figure 2-11.
The national median annual 2nd highest MDA1 value decreased by 27% from 2002 (105 ppb) to
2013 (77 ppb), which is comparable to the decrease observed in the national median annual 4th
highest MDA8 value (25%) during the same period.
20 The 1-hour O3 standards were formally revoked in 2005 (70 FR 44470, August 3, 2005).
2-25
-------
200
160-
c 120 -
80-
40-
<60 61-70 71-84 >84
8-hour 03 Design Value (ppb)
Figure 2-15. Boxplots showing the distribution of MDA1 concentrations (2016-2018),
binned according to each site's 2016-2018 design value.
50-
40-
15 30-
>
a- 20
ra
Q
10
Number of Monitors
• 1
•
i-6
4-9
10 or more
*
•
~
•
*
«
*
*
•
•
*
*
m
•
*
«
~ •
*
~
«
w «•
*
~
*
40
50
70 80 90 100
8-hour Ozone Design Value (ppb)
110
120
Figure 2-16. Number of days in 2016-2018 at each monitoring site with a MDA1
concentration greater than or equal to 120 ppb compared to its 8-hour design
value in ppb.
2-26
-------
jQ
140
CL
CL
C
O
130
>
90
<0
Q
CO
O)
JO
80
O)
X
"O
c
CM
70
To
3
c
c
<
60
Ox-c\Jcorfmtor^aDCDOv-cMco^rmtor-ao
OOOOOOOOOOr-r-t-r*T-r- i-r-r-
ooooooooooooooooooo
CnJCMCMCMCvICNCNCNCNCMCMCnJCMCNCMCMCMCMCM
Figure 2-17. National trend in the annual 2nd highest MDA1 O3 concentration, 2000 to
2018. The solid blue line represents the median value, dotted blue lines represent
the 25th and 75th percentile values, and the light blue shaded area represents the
range from the 10th to the 90th percentile values.
2.5 BACKGROUND 03
There are a number of definitions of background Os used in various contexts that differ
by the specific emissions sources and/or natural processes the definition includes (e.g., see ISA,
Appendix 1, section 1.2.2). In this review, as in past reviews, the EPA generally characterizes Os
concentrations that would exist in the absence of U.S. anthropogenic emissions as U.S.
background (USB). An alternative phrasing for USB is the O3 concentrations created collectively
from global natural sources and from anthropogenic sources existing outside of the U.S. Such a
definition helps distinguish the O3 that can be controlled by precursor emissions reductions
within the U.S. from O3 originating from global natural and foreign precursor sources that cannot
be controlled by U.S. regulations (ISA, section 1.2.2).
2-27
-------
Because monitors cannot distinguish the origins of the O3 they measure,21 photochemical
grid models have been widely used to estimate the contribution of background sources to
observed surface O3 concentrations. This section summarizes results of a state-of-the-science
modeling analysis that the EPA performed for this assessment to estimate the magnitude of
present-day USB and its various components. Conceptually, these USB estimates represent O3
concentrations that occur as a result of global natural sources (or processes, see section 2.5.1 for
more details) and those anthropogenic sources existing outside the U.S., i.e., the O3
concentrations that would occur in the absence of any U.S. anthropogenic O3 precursor
emissions. Modeling results summarized in this section include average estimates of MDA8
USB concentrations for several temporal periods including seasons. Average USB estimates are
also presented for days on which the total model-predicted MDA8 O3 concentration was greater
than either 60 ppb or 70 ppb, and for the days on which the 4th-highest MDA8 O3 concentration
was predicted to occur. Additionally, this modeling analysis investigated the contributions to
USB of some specific groups of sources, such as international anthropogenic sources, and how
those contributions vary by season and by location.
The section is organized as follows. Section 2.5.1 provides an overview of the various
sources that contribute to USB, including currently available information on the magnitude,
seasonal variability, and spatial variability of their contributions to USB. Section 2.5.2
summarizes the methodology for the modeling analyses used to quantify USB and component
contributions. More detailed information about the modeling methodology is presented in
Appendix 2B. Section 2.5.3 summarizes USB estimates using methodology described in section
2.5.2, including estimates specific to certain subgroups of sources. Section 2.5.4 summarizes key
findings of the analyses.
2.5.1 Summary of U.S. Background O3 Sources
Jaffe et al. (2018) most recently reviewed the literature on sources that contribute to USB.
While the term "background" may imply a low concentration well-mixed22 environment,
background sources can create well-defined plumes and/or contribute to the well-mixed
environment. The USB definition, which is based on sources, includes both the well-mixed
environment and more well-defined plumes. Figure 2-18a (adapted from Jaffe et al. (2018))
21 Ozone concentrations that do not include contributions from U.S. anthropogenic emissions cannot be determined
exclusively from O3 measurements because even relatively remote monitoring sites in U.S. receive transport of
U.S. anthropogenic O3 from other locations.
22 We use the term "well-mixed" here to refer to conditions when the contributions from various types of sources are
mixed due to chemistry or physical processes to the point where it is not possible to discern the contribution to O3
from each individual source.
2-28
-------
illustrates sources of USB O3 (blue) and US anthropogenic sources of O3 (yellow). Figure 2-18b
shows two theoretical examples where background sources contribute to the total ground-level
O3. The first example (Ex 1) highlights a typical monitoring site with lower USB, and the second
example (Ex 2) presents a scenario in which USB is a large contributor. Both examples
oversimplify methane, which has both natural and anthropogenic and both domestic and foreign
contributions. Source contributions to USB vary in space and time, and the stacked bar plot in
this figure oversimplifies the complex relationship between USB and total O3. Even so, USB
sources can broadly be discussed as global natural sources (see sections 2.5.1.1 to 2.5.1.6) and
international anthropogenic sources (see section 2.5.1.7). In the simplest interpretation, the
natural sources are background regardless of where they occur, or which definition of
background is being used (e.g., USB or natural background23). By contrast, anthropogenic
sources are only considered as background when they are not from sources within the focus area.
However, this paradigm is complicated by the fact that many sources of O3 precursors are the
result of interactions between human and natural systems (for instance forest management
practices can impact both biogenic VOC emissions from trees and wildfires). In the context of
USB, anthropogenic background is synonymous with O3 originating from international
anthropogenic emission sources. The relative contribution of international and natural
background sources can vary dramatically from place to place and are most notably larger at
locations near borders (international) or high elevation (natural). At non-border locations and
many border locations, the natural background is usually the dominant background source.
23 Natural background is the O3 that would exist in the absence of anthropogenic emission sources.
2-29
-------
Stratospheric Intrusion , >'.
j Lightning ¦
Foreign Pollution
/ -|1 .**? ["Wildfire Impacts
if"* 1. ¦ ;j 1
>, in rn -
[ Agricultural Emissions
Recirculated
Domestic Pollution
Interstate Transport,
I Export of ^
[Domestic Pollution
Local Photochemical
Pollution
| Prescribed Burning1
Biogenic Emissions
(a) U.S. O3 sources shown with yellow boxes or arrows represent domestic sources.
Sources shown with blue boxes or arrows represent USB sources. Note that locations for
each process are not specific to any one region. The base map shows satellite-observed
tropospheric NO2 columns for 2014 from the Ozone Monitoring Instrument (OMI) onboard
the NASA Aura satellite (Credit: NASA Goddard's Scientific Visualization Studio/T.
Schindler). NO2 column amounts are relative with red colors showing highest values,
followed by yellow then blue. We use the OMI NO2 columns as a proxy to show local O3
precursor emission sources, (b) The bar chart shows two theoretical examples of USB Ql
contributions combine with domestic sources to produce elevated O3 at a specific location
on any given day. Each source varies daily and there are also nonlinear interactions
between USB O3 sources and anthropogenic sources that can further add to O3
formation, e.g., wildfires and urban anthropogenic emissions (e.g., Singh et al., 2012).
Minor adaptation from DOI: https://doi.org/10.1525/elernenta.309.f1
(b)
Figure 2-18. Conceptual models for 03 sources: (a) in the U.S., and (b) at a single location.
The natural and anthropogenic sources of background O3 vary by location and by season.
Emissions from anthropogenic sources largely occur in the same areas year after year. Natural
sources of O3 and precursors, on the other hand, vary both in magnitude and in location from day
to day and year to year. As a result, certain types of natural sources may have large Os
contributions measured at a monitor at one point in time but not at other times. The combination
of varying proximity and magnitude means that natural sources can contribute to background in
the form of localized plumes of elevated Os that contribute to O3 at monitoring sites on an
episodic basis In the absence of locally well- defined plumes, global natural and international
anthropogenic sources are constantly contributing to the well-mixed background.
80
70
US Anthro
60
50
_o
1
Intl Anthro
a. 40
O-
1!
¦ Wildfires
30
i
Lightning
20
¦
Methane
10
0
I
1
- Biogenics
¦
Stratospheric
i-H
(N
X
X
LU
LU
2-30
-------
USB varies by location and by season due to both the nature of sources and the loss
processes. The nature of emission sources leads to seasonal and spatial patterns that will be
described further below. The contribution of these sources is modulated by transport patterns that
interact with deposition and chemical losses. For illustration, two emission sources of identical
magnitudes may have different contributions if one emits near the surface in summer and the
other emits in the free troposphere in spring. Warmer moister air in the summer at the surface
enhances O3 chemistry losses and deposition of O3 to the surface increases losses further. In
contrast, cooler, drier temperatures in the spring and free troposphere lengthen O3 lifetimes and
faster winds in the free troposphere enable longer transport. The seasonality of temperature and
transport patterns gives O3 USB a distinct seasonal cycle that results from both sinks and
sources.
The sections below summarize the state of the science estimates of USB contributions.
Each source type is described with respect to its seasonality as well as its local vs well-mixed
contribution potential. Jaffe et al. (2018) reviewed contributions of various sources to USB O3
from modeling studies and the references therein are used to illustrate the range of O3
contributions from each source. The literature-based estimate ranges provide context to the
estimates of USB that are reported in section 2.5.3.
2.5.1.1 Stratosphere
The only direct source of O3 to the troposphere with appreciable contributions to O3
concentrations is STE (other sources are indirect via precursors). STE occurs when stratospheric
air, which is relatively rich in O3, is transported across the tropopause where it enhances
tropospheric concentrations. Most STE events create enhancements that do not immediately
reach the surface. Instead, STE-enhanced O3 mixes into the free troposphere where it is
dispersed. In cases when the transported air reaches the surface before enough dispersion occurs,
it creates a localized plume of O3 referred to as a Stratospheric Ozone Intrusion (SOI). The total
stratospheric contribution includes both the well-mixed contribution from the distant stratosphere
exchanges as well as any localized SOI plume.
The total global O3 flux from the stratosphere to the troposphere is estimated at 510±90
teragrams per year (Tg/y) compared to 4620±600 Tg/y (post-2000 literature in Table 2 in Wu et
al., 2007) produced within the troposphere. The majority of the earth's surface is outside the U.S.
and only STE that take place over the U.S. are likely to create a large magnitude local
enhancement at a U.S. monitor.24 A SOI that occurs outside the U.S. would likely be dispersed
24 Recently methods have been developed for identifying and estimating SOIs that have clear localized contributions
to O3 concentrations with the potential to contribute to standards' exceedances. These are described in documents
available at: https://www.epa.gov/air-quality-analysis/guidance-preparation-exceptional-events-demonstrations-
stratospheric-ozone.
2-31
-------
into the well-mixed background and reduced through chemical loss and deposition before it
reaches many monitors.
Modeling and observational studies show that SOI can episodically contribute large
amounts of O3 at a subset of U.S. monitors, but stratospheric mixing more frequently contributes
smaller quantities of O3. Modeling studies focused on seasons with frequent SOI find median
total stratospheric contributions to MDA8 are 10-22 ppb in the West and 3-13 ppb in the East
with episodic contributions up to 40 ppb mostly in the West (Table S2, Jaffe et al., 2018).
Because these studies focus on the most active season, these medians are expected to be upper
bounds for the annual average. Further, SOI are most common in the spring when MDA8 O3
concentrations above 70 ppb are less common (ISA, section 1.3.2).
2.5.1.2 Biogenic VOC
Biogenic VOCs are the quintessential "natural" source of O3 precursors. At global scales,
biogenic sources are the largest contributor to VOCs - even though local anthropogenic sources
of highly reactive VOCs can be very important in some areas. VOCs are also an important
source of carbon monoxide. Biogenic VOCs are emitted by various types of vegetation and
emissions peak in summer which is also when O3 production is fast and O3 lifetimes are short.
The large abundance of biogenic VOCs leads to NOx-limited O3 production in most of
the world. That is, concentrations of biogenic VOCs are in excess with respect to concentrations
of NOx; therefore O3 production is controlled by the availability of NOx. The methodologies25
typically used by the air quality community estimate contribution based on sensitivity of O3
production. As a result, the sensitivity-based contribution estimate of biogenic VOC sources to
O3 shows relatively small contributions considering the large amount of emissions.
Estimates of biogenic VOC contributions in the literature are generally small compared to
NOx. For example, Lapina et al. (2014) found that North American Background (NAB)26 for
W12627 O3 was relatively insensitive to VOC (10.8% of NAB sensitivity) compared to NOx
(79.8% of NAB sensitivity). This well-known global-scale sensitivity to NOx would not exist if
concentrations of biogenic VOCs were a broadly limiting factor. Even though background O3 is
not particularly sensitive to small changes in the biogenic VOC, natural sources of VOCs are a
critical component of all background O3 estimates.
25 Source apportionment techniques and derivative-normalization techniques use sensitivity to attribute
concentrations to sources. When a concentration is insensitive to VOC sources, the contribution estimate solely
from that source of VOC will be zero.
26 North American Background is analogous to USB; but NAB is generally characterized as the O3 concentrations
that would exist in the absence of North American anthropogenic emissions.
27 W126 is a daytime weighted average concentration where higher concentrations are given greater weight based on
a sigmoidal curve (see Chapter 4).
2-32
-------
2.5.1.3 Wildland Fires
Fires emit a complex mixture of nitrogen oxides, nitrogen reservoir species (e.g., PANs),
and VOCs that are all precursors to O3. In the northern hemisphere, the fire season generally
starts in spring and extends into fall with the specific timing varying widely by region. Fires also
exhibit significant year to year variability, with emissions varying by an order of magnitude
between high and low fire years in some places (van der Werf et al., 2017). While smoke from
fires affects most of the contiguous U.S. at some point during the year, the fire season in the
western U.S. occurs primarily late in the summer. Fires across western states and parts of Canada
can contribute both to regional background and episodic surface O3 enhancements (McClure and
Jaffe, 2018) .2S
Ozone production in fire plumes depends on a range of factors including the type of fuel
combusted, plume age, and interactions with other air masses (e.g. urban plumes) (Jaffe and
Wigder, 2012). While some studies have estimated wildfire O3 contributions to seasonal mean
O3 of up to several ppb during high fire years in the Western U.S. (Jaffe et al., 2018), O3
production from individual fires varies substantially (Akagi et al., 2013). Several studies have
shown that locations near large fires can even experience suppressed O3 formation, perhaps due
to titration from fresh NO emissions and/or reduced solar radiation resulting from high aerosol
concentrations (McClure and Jaffe, 2018;Buysse et al., 2019). Large variability in O3 precursor
emissions from fires combined with complex in-plume dynamics and chemistry make accurately
quantifying O3 production from fires extremely difficult at both regional and local scales.29
New data from recent and upcoming field and aircraft campaigns30 are expected to
provide new insights that expand current understanding of contributions from fires to O3
concentrations in the U.S., both in the context of regional background concentrations and
production during individual fire episodes.
2.5.1.4 Lightning Nitrogen Oxides
Lightning is an indirect natural O3 precursor source. Lightning produces NOx from
molecular nitrogen and oxygen, similar to traditional combustion processes. Because NOx is the
28 Fires may occur on wildlands naturally or accidentally, or fires may be planned (prescribed) for various purposes
and set intentionally. In the USB modeling work described in section 2.5.2.1 below, emissions associated with
prescribed fires are categorized as anthropogenic emissions and are not included in estimating USB.
29 Recently methods have been developed for identifying and estimating wild or prescribed fire contributions to 03
concentrations with the potential to contribute to standards' exceedances. These are described in documents
available at https://www. epa.gov/air-quality-analysis/final-2016-exceptional-events-rule-supporting-guidance-
documents-updated-faqs.
30 Western Wildfire Experiment for Cloud Chemistry, Aerosol Absorption and Nitrogen (WE-CAN,
https://www.eol.ucar.edu/field_projects/we-can) in 2018 and Fire Influence on Regional to Global Environments
and Air Quality (FIREX-AQ, https://www.esrl.noaa.gov/csd/projects/firex-aq/) in 2019.
2-33
-------
globally limiting precursor for O3 production and lightning emits where there are few other
sources, O3 production is quite sensitive to this source. Over the U.S., lightning NOx (LNOx)
emissions peak in summer with convective activity and are characterized as having high
interannual variability (Murray, 2016). Allen et al. (2012) showed that the majority of LNOx is
emitted in the free troposphere (i.e, troposphere above the planetary boundary layer). Thus,
LNOx is produced in a NOx-limited environment where any O3 formed as a result will be
efficiently transported and loss pathways are limited.
The total NOx created by lightning is highly uncertain (Murray, 2016). Murray (2016)
discusses the uncertainty in NO yield per flash rate and the role of large spatial gradients in the
yield. The effect of such uncertainties is evident in the range of global lightning emissions
(std/mean=0.4). Murray (2016) also discusses the uncertainty in the vertical distribution of NO
production and post-production redistribution.
Jaffe et al. (2018) reviewed contributions from lightning to surface USB O3 based on
modeling studies using various flash rate yields, which shows large single day contributions to
modeled MDA8 O3 (up to 46 ppb, Murray, 2016) and smaller contributions to annual means (1-6
ppb) and seasonal means (6-10 ppb). Lapina et al. (2014) showed that, in their modeling, W126
had a 15% contribution from lightning NOx over the U.S.31 A 15% contribution is consistent
with the annual and seasonal mean contributions to MDA8 reported by Zhang et al. (2014) and
Murray (2016). Lapina et al. (2014) also noted that 40% of the lightning NOx sensitivity comes
from lightning strikes outside the U.S. The findings from these studies highlight the primary
importance of lightning NOx as a contributor to the well-mixed background concentrations
(Murray, 2016).
2.5.1.5 Natural and Agricultural Soil NOx
Nitrogen oxides from soils are a naturally occurring source that is enhanced by
anthropogenic activity. Truly natural soil NOx is created as a byproduct of nitrogen fixation in
natural environments. The fixation and byproduct release are affected by flora composition,
nitrogen availability, and environmental conditions (e.g., humidity). Human activity affects the
amount and location of soil NOx emissions by changing land cover and by increasing the
availability of nitrogen for fixation though the application of fertilizer to crop lands or additions
31 The numbers shown in this report are derived from reported values in Lapina et al. (2014) which showed
sensitivity of W126 to anthropogenic NOx sources was 58% (of that, 80% US; 9% CAN; 4% MEX) and natural
NOx sources was 25%. The remaining 17% was attributed natural isoprene (1.3%), VOCs/CO from fires (Fig 9:
-3%) and international VOC/CO (Fig 9: -14%). So non-North American anthropogenic NOx (58% * 7% non-NA
= 4%) and natural NOx (25%) create a total NAB NOx sensitivity of 29% and total NAB sensitivity of 35% (29%
/ 79.8%). Of the total sensitivity (parentheses contain percent of NAB NOx sensitivity, see Fig 12), lightning was
15% (52.9%), soil NOx was 8% (28.2%), fire NOx was 1% (4.3%) and international anthropogenic NOx was 4%
(14.5%).
2-34
-------
of nitrogen via deposition of emissions from other sources. The effect of human land cover
alteration is readily apparent in soil NOx emission measurements. Steinkamp and Lawrence
(2011), highlight that soils in pristine natural ecosystems emit more NOx compared to similar
ecosystems that have been disturbed by human activity. At the same time, human managed crop
lands emit more than natural ecosystems (pristine or disturbed) environments because of the
applied fertilizer.
Soil NOx clearly has both anthropogenic and natural sources, but these are rarely
separated in the literature. First, Hudman et al., 2012 estimate that the majority (-80%) of soil
NOx emissions are currently attributed to land surfaces without considering active fertilization or
deposition of anthropogenic nitrogen. Second, the emissions and attribution are relatively
uncertain. Finally, anthropogenic soil NOx is associated with agricultural ammonia application
that is not directly regulated in the United States. As a result, the attribution of soil NOx as a
"background" source is imperfect. In this assessment, no distinction is made between natural and
fertilizer-enhanced soil NOx and instead we include both within "natural sources."
Hudman et al. (2012) estimated the global soil NOx emissions at 10.7 TgN/y. As noted
above, soil NOx emissions are linked to nitrogen availability in the soil, which is increased by
anthropogenic activities. Hudman et al. (2012) attributed 1.8 TgN/y to anthropogenic soil
fertilization and 0.5 TgN/y to atmospheric deposition. Like lightning, most soil NOx emissions
occur outside of the U.S. Unlike lightning, soil NOx has a smaller long-range transport
component because it is emitted at the surface. For example, Lapina et al. (2014) calculated that
W126 had an 8% sensitivity to soil NOx (see footnote 26) and noted that a small fraction (only
7%) was from emissions outside the U.S. The more local sensitivity is likely due to the emission
height and spatial distribution of soil NOx.
2.5.1.6 Post-Industrial Methane
Like other VOCs, CH4 is a hydrocarbon that can form O3 in the presence of NOx and
sunlight. While some atmospheric methane is emitted naturally from wetlands, wildfires,
geogenic sources, and insects, significant global methane enhancements following the industrial
revolution are clearly associated with increased emissions from anthropogenic fossil fuel
combustion (Pachauri et al., 2015). Other human activities such as livestock cultivation, landfills
and land use modification (e.g., rice paddies) also release methane. More recently, changing
climate conditions have led to increased emissions from natural sources (e.g., permafrost
melting) in some areas (Reay et al., 2018), although the exact magnitude of these effects on
global methane concentrations, and consequently O3 in the U.S., over longer time scales remains
uncertain.
2-35
-------
Due to its long atmospheric lifetime (-10 years), methane is well-mixed at seasonal and
annual time scales. As a result, isolating contributions to atmospheric methane concentrations
from individual geographic areas or specific emission sectors is very difficult (Turner et al.,
2017). However, sensitivity simulations with chemical transport models can be used to assess the
overall influence of global methane concentrations on regional O3 budgets. For example, Lin et
al. (2017) used the GFDL-AM3 chemistry-climate model to estimate that increasing global
methane concentrations contributed -20% to background MDA8 O3 trends during boreal spring
and summer at several western U.S. sites during the period 1988 to 2012. In general, post-
industrial anthropogenic methane is estimated to contribute -5 ppb to surface O3 in the U.S., an
estimate that primarily comes from modeling studies (Jaffe et al., 2018 and references therein).
A major limitation with existing model-based estimates of the influence of global
methane on current U.S. O3 concentrations is our limited understanding of historical methane
emissions. The U.S. and the rest of the world's anthropogenic methane emissions have not been
tracked quantitatively in detail until relatively recently. As a result, the pre-industrial methane
concentration is relatively unconstrained. Further, post-industrial methane can be attributed to
direct emissions and emissions from natural sources (e.g., permafrost). Many modeling studies,
including this one, do not explicitly track methane sources and sinks, further complicating
attribution in an air quality context. Therefore, the post-industrial methane contribution is
difficult to quantitatively attribute. The post-industrial enhancement of methane is clearly related
to emissions and human activity, which includes both foreign and domestic contribution.
2.5.1.7 International Anthropogenic Emissions
International anthropogenic emissions are the only anthropogenic contribution to USB.
For the purposes of discussion, NOx and VOCs will be discussed separately from methane
(methane is covered in section 2.5.1.6). NOx and VOC emission estimates from outside the U.S.
are derived from international collaborative efforts like the Hemispheric Transport of Air
Pollutants (HTAP) task force of the United Nations Economic Commission for Europe. HTAP
harmonized national emission databases from individual countries with global estimates that
cover areas without their own estimates. Collecting and harmonizing these emission datasets
requires coordination and technical expertise, which recently occurred twice (HTAP Phase I and
HTAP Phase II) and may occur again soon. Global estimates that incorporate national
information are available (e.g., Community Emissions Data System and Emissions Database for
Global Atmospheric Research), but do not always have as much participation from individual
countries. This is particularly important because individual countries are most aware of
regulations and controls that have been promulgated within their borders.
2-36
-------
International anthropogenic sources of O3 include emissions within the borders of other
countries (e.g., onroad sources, power plants, etc.) as well as sources in international waters and
air space. Sources within the borders of other countries can be easily attributed to those countries
using geographical bounds based on emission source location. Some studies (e.g., Lin et al.,
2014), however, have done more complex analyses to spatially attribute emissions globally based
on the consumption of produced goods. For the purposes of this document, international
emissions are attributed based on the emission source location. Using emission source location,
maritime shipping and aircraft sources require more artificial distinctions. Typically, aircraft
takeoff and landing are assigned completely to the country where it occurs. Aircraft cruising
emissions are attributed based on geographic boundaries. This assumes that both inbound and
outbound flights change source type (domestic/international) when they cross a border.
2.5.2 Approach for Quantifying U.S. Background Ozone
Updating USB estimates is motivated by interannual variability, trends in international
anthropogenic emissions, and continual improvements in simulating processes affecting USB.
USB sources are expected to vary from year to year because natural emissions vary in response
to meteorology (e.g., temperature) and long-range transport patterns alter the efficiency of
transport from long-range USB sources (Lin et al., 2015). In addition, the scientific
characterization of background emission sources continues to evolve. As a result, we provide an
updated assessment of USB for 2016 using the latest stable version of the Community Multiscale
Air Quality (CMAQ) model applied at hemispheric to regional scales.
This assessment uses a firmly source-oriented definition of USB based on modeling. The
source composition of a model estimate can be quantified using tagging techniques or by
sensitivity analysis. By contrast, the source composition of measured O3 is difficult to isolate. In
most areas at most times, measured O3 concentrations are the result of contributions from a
variety of anthropogenic and non-anthropogenic sources. Measurements from locations
sometimes suggested to be representative of USB often have contributions from U.S.
anthropogenic sources. As a result, some researchers have filtered measurements to focus on
times when US contributions are minimized (e.g., based on wind direction or other indicators).
The measurement filtering approach is based on conceptual or quantitative models of source
contributions as a function of wind direction or another environmental variable. After correction,
the degree of contamination is minimized but not precisely known. Recently, urban
measurements have been paired with simplistic statistical models to estimate background
(Parrish et al., 2017). However, Jaffe et al. (2018) concluded that statistical adjustment cannot be
directly interpreted as "background" - even though the estimate is useful for bounding simulated
background. Due to the complications of quantifying background based on ambient air
2-37
-------
measurements, the sources that contribute to background are most clearly defined using an air
quality model. Using separate nomenclature (baseline: monitors; background: models) helps to
clearly delineate between these approaches that each have their strengths and weaknesses.
This section of the PA quantifies O3 from sources using a sensitivity approach. The
multiscale system is applied to predict total O3 and then applied multiple times to predict O3
without U.S. anthropogenic emission sources. The difference between total O3 and O3 without
the U.S. anthropogenic emissions is used to characterize the USB.
2.5.2.1 Methodology: USB Attribution
This assessment attributes O3 to USB sources using one of several available techniques.
Jaffe et al. (2018) reviewed the methods for identifying USB contributions. The methodologies
reviewed range in complexity from simply turning off U.S. anthropogenic (or specific sources)
emissions, to normalizing derivatives from instrumented models, to complex tagging techniques
(e.g., CAMx OSAT, APCA, or Grewe, 20 1 3).32 This analysis follows the zero-out approach for
simplicity of interpretation and consistency with previous EPA analyses. In urban areas, this
approach will estimate higher natural and USB contributions when NOx titration is present. The
estimate, therefore, is an estimate of what concentrations could be without US anthropogenic
emissions and not the fraction of observed O3 that is USB.
This analysis is designed to specifically separately quantify (Mrom global natural,
international anthropogenic, and U.S. anthropogenic sources. The precursors that this analysis
focuses on are NOx and VOC because they have a response on timescales relevant to the
NAAQS planning schedules (i.e., not methane). Table 2-1 lists simulations and the sources they
exclude at the various spatial scales modeled (i.e., hemispheric - 108 km resolution, regional -
36 km resolution and regional - 12 km resolution). For international shipping and aviation, the
U.S. domain is either included (ZROW) or excluded (ZUSA). These simulations form the basis
for estimating the contributions of USB and its components. Given the long atmospheric lifetime
and attributability to U.S. sources, methane is not separately identified nor is it perturbed in any
simulations. This has the effect of attributing methane to natural processes, which are a
background source.
32 For a discussion of methods and the effect on estimates, see (Jaffe et al., 2018).
2-38
-------
Table 2-1. Simulation names and descriptions for hemispheric-scale and regional-scale
simulations.
Simulation Description
Performed at Hemispheric A and RegionalB Scales
BASE All emission sectors are included
ZUSA All U.S. anthropogenic emissions are removed including prescribed fires.c
ZROW All international anthropogenic emissions are removed including prescribed fires where
possible.
ZANTH All anthropogenic emissions are removed including prescribed fires.
Performed at Hemispheric Scale only
ZCHN All Chinese anthropogenic emissions are removed.
ZIND All India anthropogenic emissions are removed.
ZSHIP Zero all near-U.S. commercial marine vessel category 3 and all global shipping.
ZFIRE Zero all fire emissions (agricultural, prescribed, and wild).
A Hemispheric-scale simulations use 108 km grid cells defined on a polar stereographic projection.
B Regional-scale simulations use a nested 36 km and 12km simulation on a lambert conformal projection.
c Emissions estimated to be associated with intentionally set fires ("prescribed fires") are grouped with anthropogenic fires.
Table 2-2 describes the calculations that are used to derive contributions. It is important
to note that contributions are not strictly additive. Large NOx sources can create non-linear
conditions that decrease O3 concentrations due to titration which is most relevant at night and in
the winter. In some cases, removing a source only increases the efficiency of other sources. In
that case, some anthropogenic contribution exists unless all anthropogenic sources are removed.
This residual anthropogenic contribution occurs in the model for both International and U.S.
sources. The results presented in this section focus on Base, USB, International, Natural
contributions. Some components of International and Natural were separately analyzed.
Canada/Mexico are separately quantified at both hemispheric and regional scales. The India,
China, Fire, and shipping contributions are analyzed only at the hemispheric scale and are
presented in Appendix 2B. The analyses in Appendix 2B support the interpretation in the
discussion below.
2-39
-------
Table 2-2. Expressions used to calculate contributions from specific sources.
Label
Name
Description
Expression
BASE
Total
Total Concentration
BASE
USB
USB
U.S. Background
ZUSA
USA
USA
U.S. Contribution
BASE-ZUSA
Intl
International
Rest of the World Contribution
BASE-ZROW
Natural
Natural
Natural Contribution
ZANTH
Res-Anth
Anthropogenic contribution that is
not attributed directly to either the
U.S. or International due to non-
linear chemistry
BASE -ZANTH -1
itl - USA
IND
India
India Contribution
BASE-ZIND
CHN
China
China Contribution
BASE -ZCHN
Ship
Ship
Ship Contribution
BASE-ZSHIP
FIRE
Fire
Global fire contributions
BASE-ZFIRE
2.5.2.2 Methodology: Strengths, Limitations and Uncertainties
The model was evaluated to assess the accuracy of predictions and infer possible biases
in USB estimates. Evaluations included comparison to satellite retrievals, O3 sondes33,
CASTNET monitors, and AQS monitors. Results were also qualitatively compared to the
Tropospheric Ozone Assessment Report (TOAR) database, which has global O3 observations
that have been well characterized34 but only extends through 2014. The evaluation of the
hemispheric simulation that provides boundary conditions to the 36 km model simulation relies
heavily upon the satellites, O3 sondes and CASTNET monitors. Since the satellite data can be
used to provide concentration estimates in areas without surface monitors, these data are
particularly useful for evaluating O3 column totals in the hemispheric modeling. The sonde data
provide a means to evaluate predictions aloft which are important for understanding model
performance of long-range transport. The regional evaluation analysis focuses on data measured
at CASTNET and AQS monitors.35 Evaluation using the AQS monitors provides information on
how the model performs at urban/suburban O3, which may exhibit large space/time gradients in
O3 concentration. CASTNET data are included in the evaluation of both the hemispheric and
33 03 sondes are balloon-borne instruments that ascend through the atmosphere taking 03 and meteorological
measurements. For more information, see https://www.esrl.noaa.gov/gmd/ozwv/ozsondes/
34 The TOAR database includes 03 globally where each monitor has been consistently characterized as urban or
rural. The global observations have been processed for several metrics (MDA8, W126, etc) and gridded to 2-
degree by 2-degree global fields for easy comparison to large-scale models.
35 In the discussion here in section 2.5, the data for CASTNET sites are referred to as "CASTNET data" and data for
all other sites in AQS are referred to as "AQS data" (even though data for many, if not all, CASTNET monitors
are stored in AQS).
2-40
-------
regional models since monitoring sites in this network are intended to represent O3
concentrations across broad areas of the U.S.
The evaluation using sonde data shows that the hemispheric model predictions of O3 are
generally within 20% of the corresponding measurements throughout much of the free
troposphere. Near the tropopause, there is a low bias in the model that is most pronounced in the
spring. The low bias at the tropopause likely suggests an underestimate of stratospheric
exchange. Mean bias drops to below 20% in the middle troposphere (600-300 hPa). The low-bias
in the free troposphere may stem from underestimation of spring time stratospheric contribution
in some regions.
The acceptability of model performance was judged for the 2016 CMAQ O3 performance
results considering the range of performance found in recent regional O3 model applications
(NRC, 2002, Phillips et al.. 2008, Simon et al.. 2012, U.S. EPA, 2009, U.S. EPA, 2018d). The
model performance results, as described in this document, demonstrate the predictions from the
2016 modeling platform closely replicate the corresponding observed concentrations in terms of
the magnitude, temporal fluctuations, and spatial differences for 8-hour daily maximum O3. At
CASTNET sites, the model performance is similarly good, but has a distinct seasonal pattern
(see Appendix 2B.3). The normalized mean bias increases from a low-bias in boreal Winter
(West: -16%; East: -14%) to relatively neutral in boreal Fall (West: 0%: East: 7%). These results
are consistent with the free troposphere bias seen in the comparison of model predictions to
sonde data. Despite the conceptual consistency, the low-bias in winter at CASTNET sites is also
influenced by local sources. For example, the Uintah Basin monitors have extremely high winter
observations that are underpredicted by the model. These are most likely due to underestimation
of O3 formed from precursors emitted by local sources as well as the need for finer resolution
meteorological inputs to capture cold pool conditions that characterize these events.36
Model predictions have historically shown poor performance for capturing the impacts
from O3 of wildfires and stratospheric intrusions. Wildfire contributions have been overpredicted
by models (Baker et al., 2016, Baker et al., 2018). Model predictions of O3 from stratospheric
intrusions have ranged from underestimated to overestimated (e.g., Emery et al., 2012). Models
are not expected to perform well in capturing the contributions from wildfires and stratospheric
intrusions without a focused effort on properly characterizing the physical properties of
individual events.
36 The DIN431 CASTNET monitor, among others, is in the Uintah basin where wintertime O3 can be caused by
snow-cover enhanced photolysis combined with light VOC emissions from the oil and gas production, (see
Ahmadov et al., 2015).
2-41
-------
This analysis uses an emission inventory with known issues in the fire inventory. The
" 2016fe" inventory had double counting of some grassland fires.37 To minimize the effects of
double counting, a filter is applied to the data to remove large episodic natural influences
including fires. The filter removes days where natural contributions deviate from the mean for
that grid cell by whichever is higher: 20 ppb or twice the standard deviation for that grid cell.
Using this approach, 0.1% of grid cell days were removed -- 71% of grid cells have no days
removed and fewer than 5% have more than 1% removed. Of the days that were removed, fewer
than 21% had MDA8 concentrations above 70 ppb.
This study does not directly quantify USB uncertainty. Jaffe et al. (2018) highlight that
uncertainties in USB and USB component estimates come from multi-model comparisons.
Dolwick et al., 2015) showed that multi-model estimates converged when applying bias
correction, indicating that differences in USB estimates are correlated with model performance.
No bias correction has been applied here, so in a limited manner bias in ambient predictions can
help set expectations for bias in USB. Based on hemispheric model evaluation, the stratospheric
component in spring is likely underestimated leading to a USB low bias in spring. As a single
estimate, this study relies upon the literature based ±10 ppb for seasonal means and higher for
individual days (Jaffe et al., 2018). Further, differences between models that share
parameterizations may not fully quantify underlying uncertainty and the year-to-year variability
complicates comparing model simulations done for different years.
2.5.3 Estimates of USB and Contributions to USB in 2016
Background O3 is known to vary seasonally, spatially, and with elevation (as discussed in
section 2.5.1, above). Seasonal variations are related to temporal changes in both sources and
sinks. Spatial variations are related to differential transport patterns and the proximity to sources
of background O3. Elevation is important in determining USB because it relates to the proximity
to the free troposphere. In addition, the seasonality and spatial relationships of USB and USA
contributions are not always aligned. As a result, USB can be highest on days with lower total
O3. For these reasons, estimates of USB and USB components (i.e., Natural and International)
contributions developed from the current modeling are summarized spatially, over time, and as a
function of total O3.
All analyses of USB and components focus on model predictions over land within the
U.S. The U.S. and adjoining areas are represented in the modeling using grid cells. Only grid
37 More information related to this issue is available on the fire working group wiki page
http://views, cira. colostate. edu/wiki/wiki/9175#July-12-2018.
2-42
-------
cells in the U.S. are included in this analysis.38 Grid cells with water as the dominant land use
(e.g., lake or ocean) were simply excluded from analysis to acknowledge the potential bias of
total O3 over water bodies (U.S. EPA, 2018). The USB estimates provided here are all in terms
of a metric, MDA8, closely related to the form of the current O3 standards, and do not directly
apply to other metrics.
Section 2.5.3.1 characterizes the spatial variation of model-predicted MDA8 O3
concentrations and contributions using maps of seasonal averages. Section 2.5.3.2 characterizes
the time variation of the predicted MDA8 O3 and contributions using time series of spatial
averages. Section 2.5.3.3 characterizes the relationship between predicted USB components and
predicted total O3. Section 2.5.3.4 summarizes USB predictions across regions and seasons.
2.5.3.1 Spatial Characterization of O3 Contributions
Figure 2-19 and Figure 2-20 provide seasonally aggregated maps that show the spatial
distribution of total model-predicted MDA8 O3 and contributions from natural, international, and
U.S. anthropogenic sources across the U.S.
Figure 2-19 shows predicted MDA8 values for the 12 km domain averaged for spring
months (March, April, and May) for total O3 and contributions from Natural, International, and
USA. Natural is a relatively large contributor to total O3 in spring with a relatively small range of
values (ratio max:min = 2). International contributes less with a larger range (ratio max:min = 3).
There are spatial gradients primarily along parts of the Mexico border, and an overarching
general West-East gradient. The USA contribution, even in spring, has the largest variation (ratio
max:min > 20) with enhancements in some urban areas.
38 Modeling grid cells are assigned to the U.S. based on the grid cell centers. For grid cells whose area covers the
U.S. and an adjoining area, the grid cell is only assigned to the U.S. if the fraction of anthropogenic NOx
emissions contributed by the U.S. is greater than 80%. This is designed to remove grid cells from the analysis
when the model cannot differentiate the border.
2-43
-------
Figure 2-19. Predicted MDA8 total O3 concentration (top left), Natural (top right),
International (bottom left), and USA (bottom right) contributions in spring
(March, April, May). Each panel displays the simple spatial average and range
(min, max) in ppb in the lower left-hand corner of the panel.
Figure 2-20 shows the same type of information for the summer (June, July, August). The
summer total concentrations are higher than spring due to increases in USA and Natural
contributions. The international contribution spatial gradients have increased (reflecting shorter
O3 lifetimes), so that the maximum International contribution at the border is higher and the
average contribution is lower compared to spring. Similarly, the West-East gradient of Natural,
International, and USA contributions is enhanced in the summer. In addition, the USA
contributions show distinct gradients in urban areas. Figure 2-20 highlights the increasingly near-
border or high-elevation influence of international contribution during the summer when O3
concentrations are most likely to violate the NAAQS.
2-44
Natural: 21 ppb
(14. 28 ppb)
v-
USA: 10 ppb
(1. 21 ppb)
-------
Natural: 22 ppb
(11, 34 ppb)
V
Intl: 6 ppb
(2. 23 ppb)
il
USA: 14 ppb
(2. 51 ppb)
Figure 2-20. Predicted MDA8 total O3 concentration (top left), Natural (top right),
International (bottom left), and USA (bottom right) contributions in summer
(June, July, Aug). Each contribution has the spatial average and range (min, max)
in ppb in the lower left-hand corner of the panel.
2.5.3.2 Seasonal and Geographic Variations in Ozone Contributions
Seasonal and geographic variations are an important part of background O3. The
geographic variation helps us to understand where USB contributes appreciably to ():;
concentrations. The seasonal variation is particularly important as it determines whether high
USB and MDA8 concentrations above 70 ppb are likely to occur at the same time. This section
begins by characterizing the dependencies of predictions for different USB components 011
season and geography to define regions for further analysis. These dependencies are used to
define regions for subsequent time series analysis.
Seasonal dependence: Comparing Figure 2-19 and Figure 2-20 highlights the seasonal
differences in the predicted contributions from Natural, International, and USA sources. Between
spring and summer, the International contribution decreases by 33%; the USA contribution
increases by 40%; and the contribution from Natural sources shows a relatively small increase of
5%. The differences in contributions between the spring and summer are due to a complex
relationship between Os production, O3 lifetime, and therefore transport efficiency. Cooler drier
conditions increase the lifetime of O3 in winter/spring compared to summer/fall (Liu et al.,
1987). As a result, winter and spring have more efficient transport of O3 compared to summer
2-45
-------
and fall. Summer and fall, however, have warmer weather that promotes higher local O3
production rates. Thus, summer and fall have locally fast O3 production and relatively inefficient
transport, which combined increase the relative contribution of proximate sources.
Border dependence: In the summer, model-predicted gradients of International O3 at the
borders are most obvious. As previously discussed, summer temperatures increase O3 production
rates and decrease O3 lifetimes. As a result, areas with locally high O3 are evident near the border
in southern California and the Big Bend and lower Rio Grande areas of Texas. These local
enhancements generally occur within tens of kilometers from the border due to the short O3
lifetime in summer as noted above.
Topography dependence: High elevation monitors are closer to the free troposphere; in
fact, at certain times of day and locations, the surface can sample free tropospheric air (Jaffe et
al., 2018). Complex topography can also enhance downward transport - for example, free
tropospheric air can "downwash" on the lee-side of high elevation mountains. Sites on the lee-
side can then be affected by this large-scale downwash. High elevation sites or sites influenced
by enhanced vertical transport may show higher contributions from more distant sources.
Combined Seasonal and Geographic Dependence: The simultaneous effects of
topography, proximity to international borders, and seasonal variations are highlighted by
Hovmoller diagrams (Figure 2-21). The Hovmoller diagram shows the average concentration as
a function of month (y-axis) and distance-to-border or elevation (x-axis). Due to the higher
magnitude of estimates of USB sources in the West than the East (Figure 2-19 and Figure 2-20),
the effects of distance and elevation are shown for the West. For the purposes of this analysis, we
use the 97W longitude line as a convenient way to separate the West from the East. The figures
show average estimated values and should not be used to estimate the international contribution
at any specific location. In addition, there are distinct gradients within the 100 m resolution of
the distance-to-border bins. For instance, the 0-100 km from the border grid cell values represent
a spatial average such that the locations directly adjacent to the border have Mexican
contributions higher than that average and the locations 100 km from the border have Mexican
contributions lower than that average.
Figure 2-21 shows that proximity to the border with Canada or Mexico is a good
indicator of the role of international contributions on USB predictions. In the spring, the average
international contribution can be as much as 12.4 ppb within 100 km of the border (62 miles). In
the early spring, large contributions persist further from the border because of the longer O3
lifetimes. Near the borders the contributions also have much higher variability, both from day-to-
day and between locations on the border. The contribution from international sources drops
notably in the summer months when O3 concentrations are highest. The day-to-day variability is
associated with the variations in wind direction, while the location variability is associated with
2-46
-------
the proximity to an international population center. International contributions are highest in
near-border areas of the U.S. where there are emissions sources on the other side of the border.
To isolate the effect of elevation alone, Figure 2-21 shows the predicted international
contributions as a function of elevation after excluding border areas. In the spring, higher
international contributions are seen at all elevations. The international contribution at all
elevations decreases in summer compared to spring, but to lower contributions at lower elevation
and mostly slowly for the very high elevations (> 1500 m). This is consistent with findings from
Zhang et al. (2011) who used this elevation as a threshold.
Mean 03 8HRMAX West of 97W Atl 12US2
T 13
12 ¦
11 •
| -11 10 •
9
-9 £ g 8
* >7.
2 0 '
0 £ 6
7 2 §
7 1= £ 5
c
4 -
-5 3-
2 ¦
1 -
3
Mean 03 8HRMAX West of 97W notnearbord 12US2
0 200 400 600 800 1000
distance from MEXCAN border (0. 1094 km)
500 1000 1500
Efevation M2. 3660 mj
Figure 2-21. Predicted contribution of International sources as a function of distance from
Mexico/Canada (left) and at "interior" locations (excluding border areas) by
elevation (right).
Timeseries Analysis: The maps in Figure 2-19 and Figure 2-20 and the Hovmoller plots in
Figure 2-21 highlight the impact of season and location on predicted Os and contributions. To
further characterize the temporal variations in contributions, the contribution data are averaged
over West and East regions individually using 97W as a dividing line. The coarse "all-cells"
averaging of the data from individual grid cells ignores the major features of the relationship
between the sources and receptors on a sub-regional basis. For example, there are more grid cells
with high urban density and high anthropogenic NOx in the East, so the USA contribution will
be higher in the East. Similarly, there are more high elevation areas in the West, so transported
O3 from outside the U.S. will be higher there. Within the West, however, there are also urban
areas that have both high predicted contributions from international transport and anthropogenic
emissions in the U.S. An analysis using "all-cells" will highlight the general characteristics of the
region. To highlight the within region variability in the West, we also include analyses that focus
on urban cells at high-elevation, near borders, and elsewhere. Figure 2-22 shows regions (West
2-47
-------
and East) with high-elevation and near border areas and urban areas highlighted by contours. As
can be seen, all the high-elevation areas and Mexico/U.S. border are assigned to the West, the
Canada/U.S. border extends across both East and West, and there are no high-elevation areas in
the East.
Near and High
Near Border
High Elevation
- West
- East
- Excluded
Figure 2-22. Grid cell assignments to East, West, High Elevation, Near Border, and Near
and High (i.e., both High Elevation and Near Border). The purple outlines
highlight grid cells with 20% or greater urban land use. Near Border areas are
in both the West and East, while High Elevation areas are exclusively in the West.
Areas matching colors denoted East and West, are thus the Low Elevation/Interior
areas.
Figure 2-23 shows the time series of regional average (C) MDA8 O:; and O3 contributions
over the year for the West and East at "all-cells," calculated using equation 2-1,
r-_E£C£ Equation 2-1
Nx
where,
Nx = number of grid cells (x) included
Cx = concentration at each grid cell location (a)
The temporal pattern in the regional average clearly shows that the seasonality of MDA8
predictions for each total O3 component varies by region. The natural contribution has a single
maximum in late summer in the West, whereas, in the East there is evidence of two peaks— the
largest in late Spring and a second peak in early Fall. The somewhat lower MDA8 Os in summer
2-48
-------
in the East requires further analysis but may be related to the lack of lightning emissions within
the regional domain. The seasonality international contribution predictions is more similar
between the two regions. The international contributions in both the West and East are greatest in
Spring, but the contribution in the West is larger both at its peak and its trough, compared to the
East. The total international contribution and the separately analyzed long-distance components
(e.g., China, India, international shipping) peak in spring when O3 lifetimes favor long-range
transport (see Appendix 2B, Figure 2B-29). However, the Canada/Mexico component of
international contributions peaks in summer because of the relative proximity to the U.S.
receptors. The predicted USA contribution increases in the summer for both the West and the
East, but the USA contribution in the West is smaller than in the East. As mentioned previously,
this "all cells" average is disproportionately rural in the West. The following analysis looks
further at the different types of land in the West, including urban areas that are more
representative of population centers that behave differently than the "all cells" analysis.
C West 97W 12km All >0 ppb Natural Res-Anth ¦¦ Intl USA
60
2016-01 2016-03 2016-05 2016-07 2016-09 2016-11 2017-01
C East 97W 12km All >0 ppb wm Natural ¦ Res-Anth ¦ Intl USA
60
2016-01 2016-03 2016-05 2016-07 2016-09 2016-11 2017-01
Figure 2-23. Annual time series of regional average predicted MDA8 total O3 concentration
and contributions of each source (see legend) for the West (top), and the East
2-49
-------
(bottom). Natural is global natural sources, Intl is international anthropogenic
sources, USA is U.S. anthropogenic sources, and Res-Anth is the residual
anthropogenic (see Table 2-2 for further descriptions).
Figure 2-24 shows the predicted contributions to total O3 in the West split into three
parts: the highest elevation areas, the near border areas, and Low/Interior areas with a weighted
average focusing on urban areas. Each of these subsets is illustrated in Figure 2-22, which shows
high elevation areas (exclusively in the West), near border areas (along the U.S./Mexico and
U.S./Canada borders), and dense urban areas. The Low/Interior areas are neither high elevation
nor near border. In each subset of cells, the purple outlines show the areas whose urban land use
is highest. The effect on O3 contributions of the relative amount of urban land use can be
illustrated by computing an urban area weighted average contribution (Cu), calculated using
equation 2-2.
where,
A% is the urban area in the grid cell x
The urban area weighted average gives a larger weight to data in those urban areas that have
dense emission sources (e.g., mobile). The urban area weighted average shows higher
contribution from USA while Natural and International are lower compared to Figure 2-23. The
differences between urban-weighted and non-weighted contributions are smaller in the East (not
shown) than in the West (compare Figure 2-23 top and Figure 2-24 bottom). Compared to the
West, the East has a larger fraction of land use that is urban (see Figure 2-22), which explains
this difference. Thus, the non-weighted regional average contributions in the East includes the
effects of urban areas much more so than the West. The seasonality of International is also
different between the highest elevation areas, near border areas, and urbanized areas. At
low/interior and at high-elevation sites, the simulated International contribution peaks earlier in
the year than at border sites. This earlier season peak is consistent with seasonality of O3 lifetime
necessary for long-range transport and a smaller contribution of long-distance sources (India,
China, and Ships, see Appendix 2B, Figure 2B-30). At near-border sites, the seasonal cycle of
predicted USB contributions from Canada/Mexico and from long-range transport combine to
create a maximum later in the spring or early summer that is dominated by Canada/Mexico
contributions (see Appendix 2B, Figure 2B-30, middle panel).
Equation 2-2
2-50
-------
2016-07
2016-09
2016-11
2016 11
C" West 97W 12km > 1500m >0 ppb Hi Natural Res-Anth HI Intl USA
x
<
£
C£
X
CO
< 40
S
CC
co, 30
C0 West 97W 12km Low/Interior >0 ppb Natural Res-Anth Hi Intl USA
3 40
s
5 30
t
Cw West 97W 12km MX/CAN < 100km >0 ppb Hi Natural Hi Res-Anth Intl Hi USA
o
2016-01
2016-03
2016-05
2016-07
2016-09
2016-11
2017-01
Figure 2-24. Annual time series of regional urban area-weigh ted average predicted MDA8
total O3 concentration and contributions of each source (see legend) for the
High-elevation West (top), near-border West (middle), and Low/Interior West
(bottom). Natural is global natural sources, Intl is international anthropogenic
sources, USA is U.S. anthropogenic sources, and Res-Anth is the residual
anthropogenic (see Table 2-2 for further descriptions).
2-51
-------
2.5.3.3 Ozone Source Contributions as a function of Total Ozone Concentration
Background contributions are also known to vary as a function of total O3. To illustrate
the relationship, specialized scatter density plots were created to show the contributions as a
function of total O3. Unlike the rest of this section, the scatter density plots do not apply the
episodic natural filter described in section 2.5.2. Thus, episodic natural contributions including
double counted fires are included in these presentations, and the effect of large events may be
overestimated.39 In the scatter density plots (Figure 2-25 through Figure 2-27), each pixel
represents a 5 ppb O3 bin. In a traditional scatter density plot, the pixel color would represent the
proportion of all points that fall within that pixel. However, in Figure 2-25 through Figure 2-27
the color represents the fraction of grid-cell-days within each 5 ppb total O3 bin (i.e., the x-axis)
that have a particular model-predicted contribution value (i.e., the y-axis). Brighter colors show
where the most frequent model-predicted contribution (y-axis: Natural or International) lies
within each 5-ppb bin of total O3 value (x-axis). As a reference, percent contribution lines are
overlaid on the plots to help contextualize the results.
Figure 2-25 shows the simulated daily Natural contribution as a function of total MDA8
concentration in the West and East for the whole year. In both regions the majority of total O3
concentrations are under 40-50 ppb. At these low concentrations, the natural contribution
correlates well with total O3 and frequently contributes half of the total O3. At low
concentrations, natural contributions estimated by a zero-out approach can be larger than 100%
of the total prediction. This is a result of NOx-titration by local anthropogenic emissions, which
reduces O3 concentrations and is a well-known non-linearity of O3 chemistry. Thus, removing
the local NOx source increases prediction concentrations. At higher concentrations, Figure 2-25
shows that predicted natural contributions in both regions have a bimodal distribution (or a fork
in frequency of contributions). The lower mode represents a plateau of natural contributions with
increasing total O3, which represents enhancement by anthropogenic sources. The upper mode
represents instances where natural contributions are correlated with total predicted O3. In the
West, the lower mode is less dominant than the East. This suggests, at least in the modeling, that
there are more frequent model-predicted contributions from wildfires and/or stratospheric
intrusions in the West. Wildfire emissions are known to be overestimated in this emission
inventory and their contribution to O3 concentrations are also often overestimated by CMAQ
predictions. As a result, these predictions of very high natural contributions should be interpreted
39 When episodic natural events contribute to elevated O3 concentrations documented in air quality monitoring data
to such an extent that they result in a regulatorily significant exceedance or violation of the NAAQS, they can be
addressed via the Exceptional Events Rule (40 CFR 50.14).
2-52
-------
qualitatively as simply indicating that such contributions can be appreciable, rather than as
providing accurate and precise quantitative predictions.
rr 1-0
n 10
200
¦ 0.1
OB
150
150
0 6 5
a. 100
o. 100
04
0-2
0.2
0.0
East 97W
West 97W
50 100
Base ppb
150
200
0.0
50 100
Base ppb
200
0.2
0.0
Figure 2-25. Predicted contribution of Natural as a function of predicted total (Base)
MDA8 O3 concentration in the West and East. Sloped lines show percent
contribution as a quick reference. The number of cells in each column is
identified using the probability density function above the plot, which is on a log
scale that highlights infrequent high concentrations.
Figure 2-26 shows the predicted contribution in the West and East from international
anthropogenic sources. Unlike natural contributions, there is very little correlation between
international anthropogenic and total O3. There are rare large model-predicted contributions,
which are more frequent in the West than in the East and rarely contribute more than 50% total
Os in either region. There are also negative contributions (up to -15 ppb), which arise from non-
linearities in chemistiy. The largest negative contribution predictions are along the Mexico
border. These can either be NOx-titration events or cases where chemistiy associated with
international NOx-sources remove precursors that would otherwise enhance ();; from U.S.
sources. Negative international contributions tend to occur at relatively low total O3
concentrations.
2-53
-------
West 97W
r to
•o.i
50 100
Base ppb
East 97W
50 100
Base ppb
Figure 2-26. Predicted contribution of International as a function of predicted total (Base)
MDA8 O3 concentration in the West and East. Sloped lines show percent
contribution as a quick reference. The number of cells in each column is
identified using the probability density function above the plot, which is on a log
scale that highlights infrequent high concentrations.
Figure 2-27 illustrates the relationship between predictions of U.S. anthropogenic sources
and total O3. Above 50 ppb, the predicted contribution from USA increases with total O3 in both
the West and the East. The relationship is stronger in the East, than the West, where near border
contributions, fire contributions, and stratospheric exchange are smaller. Even so, the higher total
O3 in the West has a similar association of larger USA contributions at larger concentrations.
This is consistent with previous findings (Henderson et al., 2012; U.S. EPA, 2014).
oc
-------
Another way of looking at the contributions is to restrict the time series to grid cells
where the concentration is above a threshold. Restricting to grid cells with high concentrations
implicitly weights the results toward urban areas where these high concentrations occur most
frequently. Figure 2-28 shows the seasonal and regional variation of USB (International
Anthropogenic and Natural) and USA (anthropogenic only) sources on high O3 days (MDA8
>70 ppb). The largest magnitude differences between sources in the East and West come from
contributions predicted for Natural and USA sources. Recall that the West contains all the high-
elevation areas (>1500 m) and the full length of the U.S./Mexican border. Figure 2-29 includes
time series for high elevation, near Mexico border, and low-elevation interior areas separately.
Compared to the East, the low/interior sites in the West have 9 ppb larger contribution from
Natural and 2 ppb more from International. Compared to low/interior sites in the West, the high-
elevation sites have 7 ppb larger contributions from Natural and 4 ppb more from International.
For border areas, the International contribution is 13 ppb greater than in Low/Interior sites. As
previously noted, there are large gradients of predicted international contributions even within
the border areas, such that some locations within the 100 km of the border are predicted to
receive larger international contributions while others are predicted to receive substantially
smaller international contributions than noted above.
2-55
-------
C West 97W 12km All >70 ppb
Natural
C East 97W 12 km All >70 ppb
Natural
2016-01 2016-03 2016-05 2016-07 2016-09 2016-11 2017-01
Figure 2 28. Annual time series of regional average predicted MDA8 O3 and contributions
of each source to predicted MDA8 total O3 (see legend) in the West (top) and
East (bottom) including only those grid-cell days with MDA8 greater than 70
ppb. Natural is global natural sources, Intl is international anthropogenic sources,
USA is U.S. anthropogenic sources, and Res-Anth is the residual anthropogenic
(see Table 2-2 for further descriptions).
2-56
-------
C West 97W 12km >1500m >70 ppb
Natural
Res-Anth
Intl
USA
so
70
60
< 50
a.
5 40
30
E 20
10
04.
2016-01
2016-03
2016-05
2016-07
2016-09
2016-11
2017-01
C West 97W 12km MX/CAN < 100km >70 ppb
Natural
Res-Anth
Intl
USA
80
70
£ 60
x
< 50
cc
5 40
?30
E 20
10
0
2016-01
2016-03
2016-05
2016-07
2016-09
2016-11
C" West 97W 12km Low/Interior >70 ppb
Natural
Res-Anth
so
70
a 60
X
I 50
a.
x
00
40
30
2 20
10
0
2016-01
mm
2017-01
2016-03
2016-05
2016-07
2016-09
2016-11
2017-01
Figure 2-29. Annual time series of regional average predicted MDA8 O3 and contributions
of each source to predicted MDA8 O3 (see legend) in the high-elevation West
(top), in the near-border West (middle), and in the Low/Interior West
weighted toward urban areas (bottom) including only those grid-cell days with
MDA8 O3 greater than 70 ppb. Natural is global natural sources, Intl is
international anthropogenic sources, USA is U.S. anthropogenic sources, and Res-
Anth is the residual anthropogenic (see Table 2-2 for further descriptions).
2-57
-------
2.5.3.4 Predicted USB Seasonal Mean and USB on Peak O3 Days
The analyses above describe the contributions from the components of USB to MDA8 O3
over seasons and days. Jaffe et al. (2018) concluded that model predictions of seasonal means
have more certainty than individual daily or episodic estimates of USB. However, from a policy
perspective, it is also useful to understand the USB contributions for various regulatory-relevant
metrics. In addition to reporting predicted USB using a seasonal average metric, we also examine
predicted USB (1) on days with the highest predicted MDA8 total O3 concentrations (top 10
days); (2) on days predicted to have the 4th highest MDA8 total O3 concentrations in the year;
and, (3) on days when predicted MDA8 for total O3 is above 60 ppb or above 70 ppb.
Figure 2-30 shows USB predicted by a single simulation with U.S. anthropogenic
emissions zeroed-out. Similar to what was found for the seasonal average metric, the effect of
topography and proximity to borders are readily evident for predicted MDA8 USB on the top 10
days and the 4th highest days. The differences in seasonal average contributions between the East
and West are also evident with the top 10 days metric and 4th highest day metric. The speckled
nature of the USB plot for the 4th highest day is due to the day or even season on which the 4th
high is predicted to occur, which varies from grid cell to grid cell. The season in which the 4th
highest day occurs influences the expected contribution from long-range international transport.
The average USB contributions for the top 10 days exhibit a smoother spatial pattern because
there is a tendency for high days to be grouped seasonally, even if the 4th highest is not. Because
the USB contribution varies by season, the predicted USB contribution on the predicted 4th
highest day is quite sensitive to model bias because bias may change the season on which the 4th
highest predicted day occurs.
It is also important to highlight that areas with high predicted USB contributions do not
always coincide with areas where MDA8 total O3 concentrations are predicted to be above 70
ppb. On the 10 highest predicted MDA8 O3 days, predicted USB is relatively constant over large
areas (see Figure 2-30 middle left). Within these areas of relatively constant USB, Figure 2-30
shows that the locations having model-predicted MDA8 concentrations above 70 ppb are
generally in or near urban areas (Figure 2-30 lower right).
The USB contribution predicted in urban areas on the predicted top 10 days tends to be
lower than in surrounding rural areas. This is due to the temporal anti-correlation of local
contribution with natural and international contributions. In urban areas, MDA8 total O3
concentrations above 70 ppb tend to occur in summer and fall when anthropogenic sources result
in locally high increments of O3. Also during these seasons, long-range transport is limited and
USB from intercontinental transport is at its lowest. As a result, the predicted top 10 and 4th
highest concentration days in urban areas tend to have lower predicted USB contributions than
do such days in rural parts of the region even though rural areas have lower MDA8 O3. As a
2-58
-------
result, the areas with predicted top 10 days having MDA8 total O3 above 70 ppb tend to have
lower percentage USB contributions than the surrounding areas.
Predicted USB contributions can be large on top 10 days near populated U.S./Mexico
border areas. In near-border areas with large anthropogenic emissions, international transport can
make a large contribution. For example, across the 4th highest days predicted for every grid cell
in this model simulation, the highest predicted MDA8 USB is 80 ppb (at a location immediately
adjacent to the border). Given the uncertainties associated with such single value predictions,
averaged predictions are important to consider. Compared to the maximum USB on the 4th high,
the maximum USB is 10 ppb lower for the average of top 10 days (Figure 2-30, middle left
panel) and 11 ppb lower the average of days with MDA8 above 70 ppb (Figure 2-30, lower left
panel). The very high USB values associated with international anthropogenic emissions are very
near the U.S./Mexico border and, to the extent that associated areas have been designated
nonattainment for the NAAQS, these areas may qualify under Clean Air Act section 179B, titled
"International border areas," for specified regulatory relief upon submission of a satisfactory
demonstration.
2-59
-------
USB: 32 ppb
{26, 45 ppb)
jA_
30
USB: 37 ppb
(14, 70 ppb)
L\_
Figure 2-30. Map of predicted USB contributions by O3 season for spring average (top left),
summer average (top right), top 10 predicted total O3 days (center left), 4th
highest total O3 simulated day (center right), and all days with total O3 greater
than 70 ppb (bottom left), along with a map of the number of days with total
O3 above 70 ppb (bottom right). Each contribution has the spatial average and
range (m in, max) in the lower left-hand corner of the panel.
>70 ppb: 0 days
(0, 90 days)
USB: 33 ppb
{10, 69 ppb)
2-60
-------
The maps in Figure 2-30 provide a detailed spatial representation of predicted USB but
may imply more precision than can be expected from a modeling system. For example, the
maximum USB on predicted fourth highest day reaches 80 ppb near the Mexico border. The
largest USB at nearby monitoring sites was 71 ppb.40 The observed 4th highs at those monitors
occurred in late February and early March, while the predicted 4th highs occurred in summer.
After selecting the 4th highs based on the observations and applying bias correction
proportionally to contributions, the new USB at these locations is 51 and 63 ppb. The USB
values for any given grid cell may be biased due to local features of topography, meteorology,
emissions bias, or model construct.
To complement the spatially resolved data and reduce bias associated with individual
daily model predictions, we also spatially aggregate the data by NOAA climate region. The
predicted USB values by climate region are provided in Table 2-3 to Table 2-6. Similar to the
figures, the tables separately quantify all grid cells (Table 2-3), high elevation (>1500 m) areas
(Table 2-4), near border areas (Table 2-5), and low-elevation (<1500 m) interior areas (Table 2-
6). These tables show the spatial averages of USB within each climate region for the annual
average, seasonal averages, averages of days when MDA8 O3 is greater than 60 or 70 ppb,
averages of each grid cell's top 10-days, and each cell's 4th highest day. Note that top 10-day
average and 4th high day for each grid cell may be from different times of the year compared to
the neighboring grid cells. As a result, grid cells with highest O3 driven by transport in the Spring
are being mixed with grid cells with highest O3 driven by local formation. Applying these
averages to interpret observations must, therefore, be done in the full context of time, space, and
concentration range.
40 Monitor 06-025-1003 measured 4th maximum value was 74 ppb on March 1, 2016. Monitor 06-073-1011
measured 4th maximum was 75 ppb on February 28, 2016. Predicted USB on predicted 4th high at both locations
was 71 ppb without bias correction in July and August.
2-61
-------
Table 2-3. Predicted USB for U.S. and U.S. regions based on averages for all U.S. grid
cells.
Regions*
Mean MDA8 for Seasons or Year
Mean MDA8 of Values in Subset
Annual
4th highest
MDA8
DJFB
MAMC
JJAD
SONE
ANNF
>60ppb
>70ppb
Top10
U.S.
26
32
31
29
30
38
33
37
37
West
28
35
36
32
33
47
43
44
44
East
24
29
24
25
26
28
27
28
28
NW
27
33
33
32
31
43
32
41
41
W
30
34
38
34
34
47
43
46
47
WNC
24
33
36
30
31
48
44
43
44
SW
31
38
39
35
36
51
48
49
49
S
27
33
26
27
28
34
29
33
33
ENC
21
30
28
26
26
31
34
32
33
C
24
30
25
26
26
28
28
28
28
SE
25
28
20
24
24
25
22
25
25
NE
25
29
27
27
27
29
26
28
27
A U.S.continental U.S, West= >97 degrees West longitude, East= <97 degrees West longitude, NW=Northwest, W=West,
WNC=WestNorthCentral, SW=Southwest, S=South, ENC=EastNorthCentral, C=Central, SE=Southeast, and NE=Northeast.
B Season defined as December, January and February.
c Season defined as March, April and May.
D Season defined as June, July and August.
E Season defined as September, October and November.
F Annual mean.
2-62
-------
Table 2-4. Predicted USB for high elevation locations (>1500 m).
Mean M
DA8 for Seasons or Year
Mean MDA8 of values in subset
Annual
4th highest
MDA8
Regions*
DJFB
MAMC
JJAD
SONE
ANNF
>60ppb
>70ppb
ToplO
U.S.
31
37
40
35
35
52
49
49
50
West
31
37
40
35
35
52
49
49
50
East
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
NW
29
35
38
33
34
52
42
47
48
W
32
36
42
36
36
53
47
51
52
WNC
28
35
39
34
34
52
48
48
49
SW
32
38
39
35
36
51
50
50
50
S
35
43
36
35
37
55
59
52
53
ENC
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
C
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
SE
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
NE
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
AU.S.=continental U.S, West= >97 degrees West longitude, East= <97 degrees West longitude, NW=Northwest, W=West,
WNC=WestNorthCentral, SW=Southwest, S=South, ENC=EastNorthCentral, C=Central, SE=Southeast, and NE=Northeast.
B Season defined as December, January and February.
c Season defined as March, April and May.
D Season defined as June, July and August.
E Season defined as September, October and November.
F Annual mean.
Table 2-5. Predicted USB for locations within 100 km of Mexico or Canada Border.
Regions*
Mean MDA8 for Seasons or Year
Mean MDA8 of values in subset
Annual
4th highest
MDA8
DJFB
MAMC
JJAD
SONE
ANNF
>60ppb
>70ppb
ToplO
U.S.
26
34
32
30
30
45
43
40
40
West
28
36
34
32
32
51
56
45
45
East
22
29
28
27
27
33
34
31
31
NW
27
32
30
31
30
46
N/A
38
38
W
30
35
41
36
36
46
51
51
51
WNC
21
33
34
29
29
49
N/A
42
42
SW
32
40
36
35
36
53
55
49
50
S
32
41
33
32
34
52
63
48
49
ENC
20
29
28
26
26
32
35
32
32
C
24
30
29
28
28
31
30
31
32
SE
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
NE
24
29
28
27
27
34
41
30
30
AU.S.=continental U.S, West= >97 degrees West longitude, East= <97 degrees West longitude, NW=Northwest, W=West,
WNC=WestNorthCentral, SW=Southwest, S=South, ENC=EastNorthCentral, C=Central, SE=Southeast, and NE=Northeast.
B Season defined as December, January and February.
c Season defined as March, April and May.
D Season defined as June, July and August.
E Season defined as September, October and November.
F Annual mean.
2-63
-------
Table 2-6. Predicted USB for low-elevation (<1500 m) that are 100 km or farther from
the border.
Mean MDA8
for Seasons or Year
Mean MDA8 of values in subset
Annual 4th
highest
MDA8
Regions*
DJFB
MAMC
JJAD
SONE
ANNF
>60ppb
>70ppb
ToplO
U.S.
25
31
28
28
28
33
30
34
34
West
27
34
34
31
31
43
39
41
41
East
24
29
24
25
26
27
27
28
28
NW
27
32
31
31
30
37
32
38
38
W
29
32
35
33
32
42
41
42
42
WNC
23
33
36
29
30
44
42
41
42
SW
29
37
38
33
34
49
43
47
47
S
26
32
26
27
28
32
26
32
32
ENC
21
30
28
26
26
31
33
32
33
C
24
30
25
26
26
28
28
28
28
SE
25
28
20
24
24
25
22
25
25
NE
25
29
26
27
27
28
25
27
26
AU.S.=continental U.S, West= >97 degrees West longitude, East= <97 degrees West longitude, NW=Northwest, W=West,
WNC=WestNorthCentral, SW=Southwest, S=South, ENC=EastNorthCentral, C=Central, SE=Southeast, and NE=Northeast.
B Season defined as December, January and February.
c Season defined as March, April and May.
D Season defined as June, July and August.
E Season defined as September, October and November.
F Annual mean.
2.5.4 Summary of USB
Background O3 results from a variety of sources, each of which has its own temporal
pattern and spatial distribution. The location and timing of these sources impacts O3 production,
dispersion and loss and thus different background O3 sources have unique seasonality and spatial
patterns. The analysis presented here provides updated model-based estimates of magnitude,
seasonality and spatial patterns of background O3 contributions. The analysis separately
characterizes the estimated magnitude and spatial/temporal patterns of MDA8 O3 from three
sources: natural, international anthropogenic, and USA anthropogenic.
The current analysis indicates that natural and USA O3 contributions peak during the
traditional O3 season (May through September), while long-range intercontinental transport of
international O3 (i.e. contributions from China, India etc.) peaks in the spring (February through
May). The contributions from Canada/Mexico at near-border locations are associated with
relatively short-range transport and the seasonality peaks during May through September, similar
to USA anthropogenic O3. The influence of Canada/Mexico, however, is indicated by the model
predictions to have a stronger spatial gradient in summer, so Canada/Mexico contributions are
most evident near the border. Of the three categories of contributions, the USA anthropogenic is
best correlated with total O3 at concentrations above 40-50 ppb in both the West and the East
2-64
-------
suggesting that US anthropogenic emissions are usually the driving cause of high O3 events in
the US. This is largely explained by temporal patterns of background O3 influences in relation to
typical high O3 events. There can be exceptions to this rule that are generally associated natural
contributions at high-elevation, during fires events, or at near-border sites.
This modeling analysis indicates the relationship between predicted international and
USA anthropogenic contributions depend upon the international sources and the location. Long-
range transport and USA anthropogenic contributions tend peak at different times of the year, so
the contribution of international is often at its minimum when local sources are the driving factor
for high total O3 during the May through September O3 season. Even in cases where O3 formed
from international anthropogenic emissions does coincide seasonally with high O3 periods, the
impact of those sources can have large spatial variation. For example, O3 formed from
anthropogenic emissions in Canada and Mexico can peak in late spring or early summer when
total O3 is high. During this time-period, there is a strong spatial variability not shown in the
regional mean. As a result, specific days at specific locations may experience larger or smaller
contributions from cross-border transport on an episodic basis that is not well characterized by
average seasonal contributions. Another example of spatial heterogeneity is exemplified by
wintertime O3 events associated with emissions from local oil and gas production in the
Intermountain West. Even though these episodes can occur as early in the year as February,
international emissions may not contribute to them substantially. The conditions associated with
these events result in decoupling of the local air masses from the upper atmosphere, essentially
isolating air in the mountain valleys from the atmosphere above and reducing the influence of
long-range transport compared to other winter and early spring days. As a result, these unique
wintertime O3 episodes may have little relative influence from international emissions despite
occurring at a time of year when long-range transport from Asia is efficient. This highlights the
need to perform location specific analysis rather than relying on regional averages.
In addition to seasonal patterns, the ISA highlights interannual patterns in background O3
as well as long-term trends (ISA, section IS.2.2.1). Natural emissions and international transport
are highly impacted by meteorological patterns which vary from year to year. One key ISA
finding is that decreasing East Asian NOx emissions starting around 2010, which would suggest
decreasing contributions from East Asia in the future if those trends continue, and therefore
decreasing spring USB.
Assessments of background O3 in the last review reported regional variation in
background O3 (2013 ISA; 2014 PA). Consistent with those assessments, modeling presented
here predicts that USB is higher in the West than in the East. In this analysis, we found that on
high O3 days (greater than 70 ppb) the West-East differences are largely associated with
international contributions in near-border areas and natural contributions at high-elevation
2-65
-------
locations. The Natural component of USB exhibits the largest magnitude difference between the
West and East. International contributions from intercontinental transport (e.g., Asia) are most
important at high elevations in the West, while international contributions from Canadian and
Mexican sources are most pronounced immediately adjacent to the borders.
The modeling performed for this assessment does not differentiate between natural
sources of ozone. For this analysis we did not attempt to separately quantify the contributions
from individual Natural sources (e.g., lightning, soil, fires, stratosphere) or to address exceptional
events beyond basic screening to remove veiy large fire plumes. Literature-based emissions
estimates and photochemical modeling studies can help to inform the likely contributors to
natural. In the northern hemisphere, the natural NOx sources with the largest emissions estimates
are lightning (9.4 megatonN/yr), soils (5.5 megatonN/yr), and wildland fires (-2.2
megatonN/yr). Because NOx is the limiting precursor at hemispheric scales, the emissions
estimates suggest that lightning and soils are most likely the largest contributors to Natural O3,
except when impacted by specific fire episodes. As noted by Lapina et al. (2014), a large
contribution from lightning may be the result of lightning strikes outside the U.S. while the
contribution from soil NOx tends to be largest from emissions within the U.S. The distant
lightning source is likely to have its effect as part of the well-mixed background. The local soil
NOx emissions have a clear seasonal cycle and are known to have large local contributions. The
relative effect at any specific site would require further analysis.
The overall findings of this assessment are consistent with the 2014 PA, with the EPA's
Background Ozone whitepaper (U.S. EPA, 2015), and with the peer reviewed literature (e.g.,
Jaffe et al. 2018). The definition of USB is also consistent with the assessment in the 2014 PA
and includes global natural and international anthropogenic emission sources (NOx and VOC).
Specific findings from the current analysis are summarized as:
• USB has important spatial variation that is related to geography, topography, and
international borders. The spatial variation is influenced by seasonal variation with long-
range international transport contributions peaking in the spring while US anthropogenic
contributions peak in summer.
• The West has higher predicted USB concentrations than the East, which includes higher
contributions from International and Natural sources. Within the West, high-elevation
and near-border areas stand out as having particularly high USB. The high-elevation
areas have more International and Natural contributions than low-interior areas in the
same region. The near-border areas in the West can have substantially more international
contribution than other parts of the West.
• The USA contributions that drive predicted MDA8 total O3 concentrations above 70 ppb
are predicted to typically peak in summer. In this typical case, the predicted USB is
overwhelmingly from Natural sources. The most notable exception to the typical case is
reflected by predictions for an area near the Mexico border where the modeling indicates
2-66
-------
that a combination of Natural and Canada/Mexico contributions can lead to predicted
MDA8 USB concentrations 60-80 ppb, on specific days, which is consistent with the
previous O3 PA (Section 2.4).41
• Predicted international contributions, in most places, are lowest during the season with the
most frequent occurrence of MDA8 concentrations above 70 ppb. Except for the near-
border areas, the International contribution requires long-distance transport that is most
efficient in Spring.
• Days for which MDA8 total O3 concentrations are predicted to be above 70 ppb tend to
have a substantially higher model-predicted USA (anthropogenic) contribution than other
days in both the West and the East.
41 Uncertainties associated with such model predictions for individual days are recognized in section 2.5.3.4 above,
along with observations of how they may differ from measurements at monitoring locations in the same area. It is
also important to note that the modeling analyses presented here do not provide estimates of design values, which
are derived from monitoring data (collected over three years) and used to assess exceedances of the O3 standards.
Additionally, as noted earlier, where such exceedances occur and are shown to be caused by USB, regulations for
exceptional events may pertain.
2-67
-------
REFERENCES
Ahmadov, R, McKeen, S, Trainer, M, Banta, R, Brewer, A, Brown, S, Edwards, PM, de Gouw,
JA, Frost, GJ, Gilman, J, Helmig, D, Johnson, B, Karion, A, Koss, A, Langford, A,
Lerner, B, Olson, J, Oltmans, S, Peischl, J, Petron, G, Pichugina, Y, Roberts, JM,
Ryerson, T, Schnell, R, Senff, C, Sweeney, C, Thompson, C, Veres, PR, Warneke, C,
Wild, R, Williams, EJ, Yuan, B and Zamora, R (2015). Understanding high wintertime
ozone pollution events in an oil- and natural gas-producing region of the western US.
Atmos Chem Phys 15(1): 411-429.
Akagi, SK, Yokelson, RJ, Burling, IR, Meinardi, S, Simpson, I, Blake, DR, McMeeking, GR,
Sullivan, A, Lee, T, Kreidenweis, S, Urbanski, S, Reardon, J, Griffith, DWT, Johnson, TJ
and Weise, DR (2013). Measurements of reactive trace gases and variable 03; formation
rates in some South Carolina biomass burning plumes. Atmos Chem Phys 13(3): 1141-
1165.
Allen, DJ, Pickering, KE, Pinder, RW, Henderson, BH, Appel, KW and Prados, A (2012).
Impact of lightning-NO on eastern United States photochemistry during the summer of
2006 as determined using the CMAQ model. Atmos Chem Phys 12(4): 1737-1758.
Baker, KR, Woody, MC, Tonnesen, GS, Hutzell, W, Pye, HOT, Beaver, MR, Pouliot, G and
Pierce, T (2016). Contribution of regional-scale fire events to ozone and PM2.5 air
quality estimated by photochemical modeling approaches. Atmos Environ 140: 539-554.
Baker, KR, Woody, MC, Valin, L, Szykman, J, Yates, EL, Iraci, LT, Choi, HD, Soja, AJ,
Koplitz, SN, Zhou, L, Campuzano-Jost, P, Jimenez, JL and Hair, JW (2018).
Photochemical model evaluation of 2013 California wild fire air quality impacts using
surface, aircraft, and satellite data. Sci Total Environ 637-638: 1137-1149.
Buysse, CE, Kaulfus, A, Nair, U and Jaffe, DA (2019). Relationships between Particulate
Matter, Ozone, and Nitrogen Oxides during Urban Smoke Events in the Western US.
Environ Sci Technol 53(21): 12519-12528.
Camalier, L, Cox, W and Dolwick, P (2007). The effects of meteorology and their use in
assessing ozone trends. Atmos Environ 41: 7127-7137.
Dolwick, P, Akhtar, F, Baker, KR, Possiel, N, Simon, H and Tonnesen, G (2015). Comparison of
background ozone estimates over the western United States based on two separate model
methodologies. Atmos Environ 109: 282-296.
Emery, C, Jung, J, Downey, N, Johnson, J, Jimenez, M, Yarvvood, G and Morris, R (2012).
Regional and global modeling estimates of policy relevant background ozone over the
United States. Atmos Environ 47: 206-217.
Grewe, V (2013). A generalized tagging method. Geosci Model Dev 6(1): 247-253.
2-68
-------
Henderson, BH, Possiel, N, Akhtar, F and Simon, H. (2012). Regional and Seasonal Analysis of
North American Background Ozone Estimates from Two Studies. 08/15/2012. U.S. EPA
Research Triangle Park, NC.
Hudman, RC, Moore, NE, Mebust, AK, Martin, RV, Russell, AR, Valin, LC and Cohen, RC
(2012). Steps towards a mechanistic model of global soil nitric oxide emissions:
implementation and space based-constraints. Atmos Chem Phys 12(16): 7779-7795.
Jaffe, DA, Cooper, OR, Fiore, AM, Henderson, BH, Tonneson, GS, Russell, AG, Henze, DK,
Langford, AO, Lin, M and Moore, T (2018). Scientific assessment of background ozone
over the U.S.: Implications for air quality management. Elem Sci Anth 6(1): 56.
Jaffe, DA and Wigder, NL (2012). Ozone production from wildfires: A critical review. Atmos
Environ 51: 1-10.
Jin, X, Fiore, AM, Murray, LT, Valin, LC, Lamsal, LN, Duncan, B, Folkert Boersma, K, De
Smedt, I, Abad, GG, Chance, K and Tonnesen, GS (2017). Evaluating a Space-Based
Indicator of Surface Ozone-NO *-VOC Sensitivity Over Midlatitude Source Regions and
Application to Decadal Trends: Space-Based Indicator of O 3 Sensitivity. Journal of
Geophysical Research: Atmospheres.
Lapina, K, Henze, DK, Milford, JB, Huang, M, Lin, M, Fiore, AM, Carmichael, G, Pfister, GG
and Bowman, K (2014). Assessment of source contributions to seasonal vegetative
exposure to ozone in the US. JOURNAL OF GEOPHYSICAL RESEARCH-
ATMOSPHERES 119(1): 324-340.
Lefohn, AS, Malley, CS, Simon, H, Wells, B, Xu, X, Zhang, L and Wang, T (2017). Responses
of human health and vegetation exposure metrics to changes in ozone concentration
distributions in the European Union, United States, and China. Atmos Environ 152: 123-
145.
Lin, J-T, Martin, RV, Boersma, KF, Sneep, M, Stammes, P, Spurr, R, Wang, P, Van Roozendael,
M, Clemer, K and Irie, H (2014). Retrieving tropospheric nitrogen dioxide from the
Ozone Monitoring Instrument: effects of aerosols, surface reflectance anisotropy, and
vertical profile of nitrogen dioxide. Atmos Chem Phys 14(3): 1441-1461.
Lin, M, Fiore, AM, Horowitz, LW, Langford, AO, Oltmans, SJ, Tarasick, D and Rieder, HE
(2015). Climate variability modulates western US ozone air quality in spring via deep
stratospheric intrusions. Nature Communications 6(1): 7105.
Lin, M, Horowitz, LW, Payton, R, Fiore, AM and Tonnesen, G (2017). US surface ozone trends
and extremes from 1980 to 2014: quantifying the roles of rising Asian emissions,
domestic controls, wildfires, and climate. Atmos Chem Phys 17(4): 2943-2970.
Liu, SC, Trainer, M, Fehsenfeld, FC, Parrish, DD, Williams, EJ, Fahey, DW, Hiibler, G and
Murphy, PC (1987). Ozone production in the rural troposphere and the implications for
regional and global ozone distributions. Journal Of Geophysical Research-Atmospheres
92 (D4).
2-69
-------
McClure, CD and Jaffe, DA (2018). Investigation of high ozone events due to wildfire smoke in
an urban area. Atmos Environ 194: 146-157.
Murray, LT (2016). Lightning NO x and Impacts on Air Quality. Curr Pollut Rep 2(2): 115-133.
NRC (2002). National Research Council Committee on Estimating the Health-Risk-Reduction
Benefits of Proposed Air Pollution Regulations. National Academies Press (US).
Washington (DC).
Pachauri, RK, Mayer, L and and Intergovernment Panel on Climate Change (2015). Climate
Change 2014: Synthesis Report. Contribution of Working Groups I, II and III to the Fifth
Assessment Report of the Intergovernmental Panel on Climate Change. IPCC. Geneva,
Switzerland. https://epic. awi. de/id/eprint/37530/
Parrish, DD, Young, LM, Newman, MH, Aikin, KC and Ryerson, TB (2017). Ozone Design
Values in Southern California's Air Basins: Temporal Evolution and U.S. Background
Contribution: Southern California Ozone Design Values. Journal of Geophysical
Research: Atmospheres 122(20): 11,166-111,182.
Phillips, S, Wang, K, Jang, C, Possiel, N, Strum, M and Fox, T (2008). Evaluation of 2002
Multi-pollutant Platform: Air Toxics, Ozone, and Particulate Matter. 7th Annual CMAS
Conference.
Reay, DS, Smith, P, Christensen, TR, James, RH and Clark, H (2018). Methane and Global
Environmental Change. Annu Rev Environ Resour 43(1): 165-192.
Simon, H, Baker, KR and Phillips, S (2012). Compilation and interpretation of photochemical
model performance statistics published between 2006 and 2012. Atmos Environ 61: 124-
139.
Simon, H, Reff, A, Wells, B, Xing, J and Frank, N (2015). Ozone trends across the United States
over a period of decreasing NOx and VOC emissions. Environ Sci Technol 49(1): 186-
195.
Steinkamp, J and Lawrence, MG (2011). Improvement and evaluation of simulated global
biogenic soil NO emissions in an AC-GCM. Atmos Chem Phys 11(12): 6063-6082.
Turner, AJ, Frankenberg, C, Wennberg, PO and Jacob, DJ (2017). Ambiguity in the causes for
decadal trends in atmospheric methane and hydroxyl. Proc Natl Acad Sci USA 114(21):
5367-5372.
U.S. EPA (1978). Air Quality Criteria for Ozone and Other Photochemical Oxidants
Environmental Criteria and Assessment Office. Research Triangle Park, NC. EPA-600/8-
78-004. April 1978. Available at:
https.Y/nepis. epa.gov/Exe/ZyPURL. cgi?Dockey=200089CW. txt.
U.S. EPA (2009). Technical Support Document for the Proposal to Designate an Emissions
Control Area for Nitrogen Oxides, Sulfur Oxides, and Particulate Matter. U.S.
2-70
-------
Environmental Protection Agency. Research Triangle Park, NC. U.S. EPA. EPA-420-R-
007. Available at: http://www.epa.gov/otaq/regs/nonroad/marine/ci/420r09007.pdf.
U.S. EPA (2014). Policy Assessment for the Review of the Ozone National Ambient Air Quality
Standards,. U.S. Environmental Protection Agency. Research Triangle Park, NC. U.S.
EPA. EPA-452/R-14-006. Available from:
https.Y/nepis. epa.gov/Exe/ZyPDF. cgi?Dockey=Pl 00KCZ5. txt.
U.S. EPA (2015). Implementation of the 2015 Primary Ozone NAAQS: Issues Associated with
Background Ozone White Paper for Discussion. U.S. Environmental Protection Agency.
Research Triangle Park, NC. U.S. EPA. https://www.epa.gov/sites/production/files/2016-
03/documents/whitepaper-bgo3-final.pdf.
U.S. EPA (2018a). Air Quality Modeling Technical Support Document for the Updated 2023
Projected Ozone Design Values. Office of Air Quality Planning and Standards United
States Environmental Protection Agency. RTP, NC. U.S. EPA. 83FR65878.
https://www. epa.gov/sites/production/files/2018-
06/documents/aq_modelingtsd_updated_2023_modeling_o3_dvs.pdf.
U.S. EPA (2018b). Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2016. Office
of Air Quality Planning and Standards. Research Triangle Park, North Carolina. U.S.
EPA. EPA 430-R-18-003 Available at: https://www.epa.gov/sites/production/files/2018-
01/documents/2018_ complete_report.pdf.
U.S. EPA (2018c). 2014 National Emissions Inventory, version 2 Technical Support Document.
Office of Air Quality Planning and Standards. Research Triangle Park, North Carolina.
U.S. EPA. Available: https://www.epa.gov/sites/production/files/2018-
07/documents/nei2014v2_tsd_05jul2018.pdf.
U.S. EPA (2018d). Modeling Guidance for Demonstrating Attainment of Air Quality Goals for
Ozone, PM2.5, and Regional Haze. Office of Air Quality Planning and Standards.
Research Triangle Park, NC. U.S. EPA. EPA 454/R-18-009. Available at:
https://www3.epa.gov/ttn/scram/guidance/guide/03-PM-RH-Modeling_Guidance-
2018.pdf.
van der Werf, GR, Randerson, JT, Giglio, L, van Leeuwen, TT, Chen, Y, Rogers, BM, Mu, M,
van Marie, MJE, Morton, DC, Collatz, GJ, Yokelson, RJ and Kasibhatla, PS (2017).
Global fire emissions estimates during 1997-2016. Earth Syst Sci Data 9(2): 697-720.
Wu, S, Mickley, LJ, Jacob, DJ, Logan, JA, Yantosca, RM and Rind, D (2007). Why are there
large differences between models in global budgets of tropospheric ozone? J Geophys
Res 112 (D 5).
Zhang, L, Jacob, DJ, Downey, NV, Wood, DA, Blewitt, D, Carouge, CC, van Donkelaar, A,
Jones, DBA, Murray, LT and Wang, Y (2011). Improved estimate of the policy-relevant
background ozone in the United States using the GEOS-Chem global model with 1/2° x
2/3° horizontal resolution over North America. Atmos Environ 45(37): 6769-6776.
2-71
-------
Zhang, L, Jacob, DJ, Yue, X, Downey, NV, Wood, DA and Blewitt, D (2014). Sources
contributing to background surface ozone in the US Intermountain West. Atmos Chem
Phys 14(11): 5295-5309.
2-72
-------
3 REVIEW OF THE PRIMARY STANDARD
This chapter presents and evaluates the policy implications of the key aspects of the
currently available scientific and technical information pertaining to this review of the O3
primary standard. In so doing, the chapter presents key aspects of the current evidence of the
health effects of O3, as documented in the ISA, with support from the prior ISA and AQCDs, and
associated public health implications. It also presents key aspects of updated quantitative risk and
exposure analyses conducted for this review, as detailed in the appendices associated with this
chapter. Together this information provides the basis for our evaluation of the current scientific
information regarding health effects of O3 in ambient air and the potential for effects to occur
under air quality conditions associated with the existing standard (or any alternatives
considered), as well as the associated implications for public health. Our evaluation is framed
around key policy-relevant questions derived from the IRP (IRP, section 3.1.1), and also takes
into account conclusions reached in the last review. In this way we identify key policy-relevant
considerations and summary conclusions regarding the public health protection provided by the
current standard for the Administrator's consideration in this review of the primary O3 standard.
Within this chapter, background information on the current standard, including
considerations in its establishment in the last review, is summarized in section 3.1. The general
approach for considering the currently available information in this review, including policy-
relevant questions identified to frame our policy evaluation, is summarized in section 3.2. Key
aspects of the currently available health effects evidence and associated public health
implications and uncertainties are addressed in section 3.3, and the current air quality and
exposure information, with associated uncertainties is addressed in section 3.4. Section 3.5
summarizes the key evidence- and air quality or exposure-based considerations identified in our
evaluation, as well as the advice and recommendations received from the CASAC during its
review of the draft PA and public comments received on the draft document, and also presents
associated summary conclusions of this analysis. Key remaining uncertainties and areas for
future research are identified in section 3.6.
3.1 BACKGROUND ON THE CURRENT STANDARD
The current primary standard was set in 2015 based on the scientific evidence and
quantitative exposure and risk analyses available at that time, and on the Administrator's
judgments regarding the available scientific evidence, the appropriate degree of public health
protection for the revised standard, and the available exposure and risk information regarding the
exposures and risk that may be allowed by such a standard (80 FR 65292, October 26, 2015).
3-1
-------
The 2015 decision revised the level of the primary standard from 0.075 to 0.070 ppm,1 in
conjunction with retaining the then-current indicator (O3), averaging time (eight hours), and form
(annual fourth-highest daily maximum 8-hour average concentration, averaged across three
consecutive years). This action provided increased protection for at-risk populations,2 such as
children and people with asthma, against an array of adverse health effects. The 2015 decision
drew upon the available scientific evidence assessed in the 2013 ISA, the exposure and risk
information presented and assessed in the 2014 health REA (HREA), the consideration of that
evidence and information in the 2014 PA, the advice and recommendations of the CAS AC, and
public comments on the proposed decision (79 FR 75234, December 17, 2014).
The health effects evidence base available in the 2015 review included extensive
evidence from previous reviews as well as the evidence that had emerged since the prior review
had been completed in 2008. This evidence base, spanning several decades, documents the
causal relationship between exposure to O3 and a broad range of respiratory effects (2013 ISA, p.
1-14). Such effects range from small, reversible changes in pulmonary function and pulmonary
inflammation (documented in controlled human exposure studies involving exposures ranging
from 1 to 8 hours) to more serious health outcomes such as emergency department visits and
hospital admissions, which have been associated with ambient air concentrations of O3 in
epidemiologic studies (2013 ISA, section 6.2). In addition to extensive controlled human
exposure and epidemiologic studies, the evidence base includes experimental animal studies that
provide insight into potential modes of action for these effects, contributing to the coherence and
robust nature of the evidence. Based on this evidence, the 2013 ISA concluded there to be a
causal relationship between short-term O3 exposures and respiratory effects, and also concluded
that the relationship between longer-term exposure and respiratory effects was likely to be causal
(2013 ISA, p. 1-14).3
1 Although ppm are the units in which the level of the standard is defined, the units, ppb, are more commonly used
throughout this PA for greater consistency with their use in the more recent literature. The level of the current
primary standard, 0.070 ppm, is equivalent to 70 ppb.
2 As used here and similarly throughout the document, the term population refers to persons having a quality or
characteristic in common, such as, and including, a specific pre-existing illness or a specific age or lifestage. A
lifestage refers to a distinguishable time frame in an individual's life characterized by unique and relatively stable
behavioral and/or physiological characteristics that are associated with development and growth. Identifying at-
risk populations includes consideration of intrinsic (e.g., genetic or developmental aspects) or acquired (e.g.,
disease or smoking status) factors that increase the risk of health effects occurring with exposure to sulfur oxides
as well as extrinsic, nonbiological factors, such as those related to socioeconomic status, reduced access to health
care, or exposure.
3 The 2013 ISA also concluded there likely to be a causal relationship between short-term exposure and mortality, as
well as short-term exposure and cardiovascular effects, including related mortality, and that the evidence was
suggestive of causal relationships between long-term O3 exposures and total mortality, cardiovascular effects and
3-2
-------
With regard to the short-term respiratory effects that were the primary focus of the 2015
decision, the controlled human exposure studies were recognized to provide the most certain
evidence indicating the occurrence of health effects in humans following specific O3 exposures
(80 FR 65343, October 26, 2015; 2014 PA, section 3.4). These studies additionally illustrate the
role of ventilation rate4 and exposure duration in eliciting responses to O3 exposure at the lowest
studied concentrations. The exposure concentrations eliciting a given level of response in
subjects at rest are higher than those eliciting a response in subjects exposed while at elevated
ventilation, such as while exercising (2013 ISA, section 6.2.1.1).5 Further, while the study
subjects in the vast majority of the controlled human exposure studies (and in all of these studies
conducted at the lowest exposures) are healthy adults, the 2013 ISA identified several groups,
including children and adults with asthma, as being at increased risk of Ch-related effects. In
light of this finding with regard to children and adults with asthma, the exposure-based analyses
in the HREA included these population groups (U.S. EPA, 2014, hereafter 2014 HREA, p. 3-14).
The exposure and risk information available in the 2015 review included exposure and
risk estimates for air quality conditions just meeting the then-existing standard, and also for air
quality conditions just meeting potential alternative standards. Estimates were derived for two
exposure-based analyses, as well as for an analysis based on epidemiologic study associations.
The first of the exposure-based analyses involved comparison of population exposure estimates
at elevated exertion to exposure benchmark concentrations (exposures of concern).6 These
benchmark concentrations are based on exposure concentrations from controlled human
reproductive and developmental effects, and between short-term and long-term O3 exposure and nervous system
effects (2013 ISA, section 2.5.2).
4 Ventilation rate is a specific technical term referring to breathing rate in terms of volume of air taken into the body
per unit of time. The units for ventilation rate (Ve) are usually liters (L) per minute (min). Another related term is
equivalent ventilation rate (EVR), which refers to Ve normalized by a person's body surface area in square meters
(m2). Accordingly, the units for EVR are generally L/min-m2. For different activities, a person will experience
different levels of exertion and different ventilation rates.
5 In the controlled human exposure studies, the magnitude or severity of the respiratory effects induced by 03 is
influenced by ventilation rate and exposure duration, as well as exposure concentration, with physical activity
increasing ventilation and potential for effects. In studies of generally healthy young adults exposed while at rest
for 2 hours, 500 ppb is the lowest concentration eliciting a statistically significant 03-induced reduction in group
mean lung function measures, while a much lower concentration produces a statistically significant response in
lung function when the ventilation rate of the group of study subjects is sufficiently increased with exercise (2013
ISA, section 6.2.1.1). For example, the lowest exposure concentration examined that elicited a statistically
significant 03-induced reduction in group mean lung function in an exposure of 2 hours or less was 120 ppb in a
1-hour exposure of trained cyclists who maintained a high exertion level throughout the exposure period (2013
ISA, section 6.2.1.1; Gong et al., 1986) after 2-hour exposure (heavy intermittent exercise) ofyoung healthy
adults (2013 ISA, section 6.2.1.1; McDonnell et al., 1983).
6 The benchmark concentrations to which exposure concentrations experienced while at moderate or greater exertion
were compared were 60, 70 and 80 ppb.
3-3
-------
exposure studies in which lung function changes and other effects were measured in healthy,
young adult volunteers exposed to O3 while engaging in quasi-continuous moderate physical
activity for a defined period (generally 6.6 hours).7 The second exposure-based analysis provided
population risk estimates of the occurrence of days with Ch-attributable lung function reductions
of varying magnitudes by using the exposure-response (E-R) information in the form of E-R
functions or other quantitative descriptions of biological processes.8 In the epidemiologic study-
based analysis, risk estimates were also derived from ambient air concentrations using
concentration-response (C-R) functions derived from epidemiologic studies. These latter
estimates were given less weight by the Administrator in her decision on the standard in light of
conclusions reached in the PA and the HREA, which reflected lower confidence in these
estimates (80 FR 65316-17, October 26, 2015).
The 2014 HREA developed exposure-based estimates for several population groups
including all children and all adults. The type of exposure-based estimates that involved
comparison of exposures to benchmarks was also derived for children with asthma and adults
with asthma. The estimates of percentages of all children with exposures at or above benchmarks
were virtually indistinguishable from the corresponding estimates for children with asthma.9
When considered in terms of the number of children (rather than percentages of the child
populations), the estimates for all children were much higher than those for children with asthma,
with the magnitude of the differences varying based on asthma prevalence in each study area
(2014 HREA, sections 5.3.2, 5.4.1.5 and section 5F-1). The estimates for percent of children
experiencing an exposure at or above the benchmarks were higher than percent of adults due to
the greater time that children spend outdoors and engaged in activities at elevated exertion (2014
HREA, section 5.3.2). Thus, consideration of the exposure-based results in the 2015 decision
focused on the results for all children and children with asthma.
In weighing the 2013 ISA conclusions with regard to the health effects evidence and
making judgments regarding the public health significance of the quantitative estimates of
exposures and risks allowed by the then-existing standard and potential alternative standards
considered, as well as judgments regarding margin of safety, the Administrator considered the
7 The studies given primary focus were those for which O3 exposures occurred over the course of 6.6 hours during
which the subjects engaged in six 50-minute exercise periods separated by 10-minute rest periods, with a 35-
minute lunch period occurring after the third hour (e.g., Folinsbee et al., 1988 and Schelegle et al., 2009).
Responses after O3 exposure were compared to those after filtered air exposure.
8 The E-R information and quantitative models derived from it are based on controlled human exposure studies.
9 This reflects use of the same time-location-activity diary pool to construct each simulated individual's time-activity
series, which is based on the similarities observed in the available diary data with regard to time spent outdoors
and exertion levels (2014 HREA, sections 5.3.2 and 5.4.1.5).
3-4
-------
currently available information and commonly accepted guidelines or criteria within the public
health community, including statements of the American Thoracic Society (ATS), an
organization of respiratory disease specialists,10 advice from the CASAC and public comments.
In so doing, she recognized that the determination of what constitutes an adequate margin of
safety is expressly left to the judgment of the EPA Administrator (Lead Industries Ass'n v. EPA,
647 F.2d 1130, 1161-62 [D.C. Cir 1980], cert, denied, 449 U.S. 1042 [1980]; Mississippi v. EPA,
744 F.3d 1334, 1353 [D.C. Cir. 2013]). In NAAQS reviews generally, evaluations of how
particular primary standards address the requirement to provide an adequate margin of safety
include consideration of such factors as the nature and severity of the health effects, the size of
the sensitive population(s) at risk, and the kind and degree of the uncertainties present.
Consistent with past practice and long-standing judicial precedent, the Administrator took the
need for an adequate margin of safety into account as an integral part of her decision-making.
In the 2015 decision, the Administrator first addressed the adequacy of protection
provided by the then-existing primary standard and decided that the standard should be revised.
Considerations related to that decision are summarized in section 3.1.1 below. The
considerations and decisions on the revisions to the then-existing standard in order to provide the
requisite protection under the Act, including an adequate margin of safety, are summarized in
section 3.1.2.
3.1.1 Considerations Regarding Adequacy of the Prior Standard
The Administrator's conclusion regarding the adequacy of the primary standard that
existed at the time of the last review was based on careful consideration of the available
evidence, analyses and conclusions contained in the 2013 ISA, including information newly
available in the review; the quantitative exposure and risk analyses in the 2014 HREA; the
information, evaluations, considerations and conclusions presented in the 2014 PA; advice from
the CASAC; and public comments. Key considerations informing the Administrator's decision
that the then-current standard should be revised are summarized below.
The Administrator gave primary consideration to the evidence of respiratory effects from
controlled human exposure studies, including those newly available in the review, and for which
the exposure concentrations were at the lower end of those studied (80 FR 65343, October 26,
2015). This emphasis was consistent with CASAC comments on the strength of this evidence
(Frey, 2014, p. 5). In placing weight on these studies, the Administrator took note of the variety
of respiratory effects reported from the studies of healthy adults engaged in six 50-minute
periods of moderate exertion within a 6.6-hour exposure to O3 concentrations of 60 ppb and
10 With regard to commonly accepted guidelines or criteria within the public health community, the 2014 PA
considered statements issued by the ATS that had also been considered in prior reviews (ATS, 2000; ATS, 1985).
3-5
-------
higher. The broadest range of effects (lung function decrements, respiratory symptoms, airway
inflammation, airway hyperresponsiveness, and impaired lung host defense) have been studied
and reported following such 6.6-hour exposures to 80 ppb O3 or higher, and the most severe
respiratory effects have been reported at these concentrations. The lowest exposure concentration
in such studies for which a combination of statistically significant reduction in lung function and
increase in respiratory symptoms was reported was 72 ppb,11 while reduced lung function and
increased pulmonary inflammation were reported following such exposures to O3 concentrations
as low as 60 ppb. In considering these findings, the Administrator noted that the combination of
03-induced lung function decrements and respiratory symptoms met ATS criteria for an adverse
response.12 She additionally noted the CASAC comments on this point and also its caution that
these study findings were for healthy adults and thus indicated the potential for such effects in
some groups of people, such as people with asthma, at lower exposure concentrations (Frey,
2014, pp. 5-6). In light of this, the Administrator concluded that "the controlled human exposure
studies indicate that adverse effects are likely to occur following exposures to O3 concentrations
below the [75 ppb] level of the [then-current] standard" (80 FR 65343, October 26, 2015).
The 2013 ISA indicated that the pattern of effects observed across the range of exposures
assessed in the controlled human exposure studies, increasing with severity at higher exposures,
is coherent with (i.e., reasonably related to) the health outcomes reported to be associated with
ambient air concentrations in epidemiologic studies (e.g., respiratory-related hospital admissions,
emergency department visits). With regard to the available epidemiologic studies, while analyses
of O3 air quality in the 2014 PA indicated that most O3 epidemiologic studies reported health
effect associations with O3 concentrations in ambient air that violated the then-current (75 ppb)
standard, the Administrator took particular note of a study that reported associations between
short-term O3 concentrations and asthma emergency department visits in children and adults in a
U.S. location that would have met the then-current standard over the entire 5-year study period
11 For the 70 ppb target exposure, Schelegle et al. (2009) reported, based on O3 measurements during the six 50-
minute exercise periods, that the mean O3 concentration during the exercise portion of the study protocol was 72
ppb. Based on the measurements for the six exercise periods, the time weighted average concentration across the
full 6.6-hour exposure was 73 ppb (Schelegle et al., 2009).
12 The most recent statement from the ATS available at the time of the 2015 decision stated that" [i]n drawing the
distinction between adverse and nonadverse reversible effects, this committee recommended that reversible loss
of lung function in combination with the presence of symptoms should be considered as adverse" (ATS, 2000).
3-6
-------
(80 FR 65344, October 26, 2015; Mar and Koenig, 2009).13 14 While uncertainties15 limited the
extent to which the Administrator based her conclusions on air quality in locations of multicity
epidemiologic studies, she additionally noted some support from several multicity studies of
morbidity or mortality in which the majority of study locations would have met the then-current
standard (80 FR 65344, October 26, 2015; 2014 PA, section 3.1.4.2). Accordingly, looking
across the body of epidemiologic evidence, the Administrator reached the conclusion that
analyses of air quality in some study locations supported the occurrence of adverse 03-associated
effects at O3 concentrations in ambient air that met, or are likely to have met, the then-current
standard (80 FR 65344, October 26, 2016). Taken together, the Administrator concluded that the
scientific evidence from controlled human exposure and epidemiologic studies called into
question the adequacy of the public health protection provided by the 75 ppb standard that had
been set in 2008.
In considering the exposure and risk information, the Administrator gave particular
attention to the estimates of exposures of concern, focusing on the estimates for children, in 15
urban areas for air quality conditions just meeting the then-current standard. Consistent with the
finding that larger percentages of children than adults were estimated to experience exposures at
or above benchmarks, the Administrator focused on the results for all children and for children
with asthma, noting that the results for these two groups, in terms of percent of the population
group, are virtually indistinguishable (2014 HREA, sections 5.3.2, 5.4.1.5 and section 5F-1). In
considering these estimates, she placed the greatest weight on estimates of two or more days with
occurrences of exposures at or above the benchmarks, in light of her increased concern about the
potential for adverse responses with repeated occurrences of such exposures. In particular, she
noted that the types of effects shown to occur following exposures to O3 concentrations from 60
13 The design values in this location over the study period were at or somewhat below 75 ppb (Wells, 2012).
14 The Administrator also took note of analyses in the 2014 PA for some single-city study locations where the then-
current standard was not met during the study period (i.e., those evaluated in Silverman and Ito, 2010; Strickland
et al., 2010), finding support for the association of hospital admissions and emergency department visits with
short-term O3 on subsets of days with virtually all ambient air 03 concentrations below the level of the then-
current standard. These analyses generally focused on the range of short-term concentrations for which the
confidence intervals for the concentration-response relationship were tightest, finding these to be on many days
with O3 concentrations below the level of the then-current standard (80 FR 65344, October 26, 2015).
15 Compared to the single-city epidemiologic studies the Administrator noted additional uncertainty in interpreting
the relationships between short-term O3 air quality in individual study cities and reported O3 multicity effect
estimates. This uncertainty applied specifically to interpreting air quality analyses within the context of multicity
effect estimates for short-term O3 concentrations, where effect estimates for individual study cities are not
presented (as is the case for the key O3 studies analyzed in the PA, with the exception of the study by Stieb et al.
(2009) where none of the city-specific effect estimates for asthma emergency department visits were statistically
significant) (80 FR 65344; October 26, 2015).
3-7
-------
ppb to 80 ppb, such as inflammation, if occurring repeatedly as a result of repeated exposure,
could potentially result in more severe effects based on the ISA conclusions regarding mode of
action (80 FR 65343, 65345, October 26, 2015; 2013 ISA, section 6.2.3).16 While generally
placing the greatest weight on estimates of repeated exposures, the Administrator also considered
estimates for single exposures at or above the higher benchmarks of 70 and 80 ppb (80 FR
65345, October 26, 2015). Further, while the Administrator recognized the effects documented in
the controlled human exposure studies for exposures to 60 ppb to be less severe than those
associated with exposures to higher O3 concentrations, she also recognized there to be limitations
and uncertainties in the evidence base with regard to unstudied population groups. As a result,
she judged it appropriate for the standard, in providing an adequate margin of safety, to provide
some control of exposures at or above the 60 ppb benchmark (80 FR 65345-65346, October 26,
2015).
With regard to multiple exposures, the HREA found that under conditions just meeting
the then-current standard, fewer than 1% of children in the 15 study areas would be estimated to
experience multiple days in a year with 8-hour exposures at or above 70 ppb while at elevated
ventilation, while the percentage was as high as approximately 2% in the year and location with
the highest exposure estimates (80 FR 65345 and Table 1, October 26, 2015). Although she
expressed less concern with single occurrences, the Administrator noted that the then-current (75
ppb) standard could allow just over 3% of children to experience one or more days (i.e., at least
one day), averaged over the years of analysis, with an 8-hour exposure at or above 70 ppb (while
at moderate or greater exertion), based on the worst-case location, and up to 8% in the worst-case
year and location (80 FR 65345, October 26, 2015). She additionally noted that, that in the
worst-case year and location across the 15 study areas, the then-current standard could allow up
to about 1% of children to experience at least one day per year with 8-hour exposures at elevated
ventilation at or above 80 ppb, the highest benchmark evaluated (80 FR 65345, October 26,
2015). Additionally, while expressing less confidence in the adversity of effects observed
following exposures as low as 60 ppb (particularly single exposures), the Administrator noted
that the HREA found that under air quality conditions just meeting the then-current standard,
approximately 3 to 8% of children in the 15 urban study areas (including approximately 3 to 8%
of children with asthma), on average across the years of analysis, were estimated to experience
two or more days per year with 8-hour exposures at or above 60 ppb, while at elevated
ventilation (80 FR 65345, October 26, 2015).
16 In addition to recognizing the potential for continued inflammation to evolve into other outcomes, the 2013 ISA
also recognized that inflammation induced by a single exposure (or several exposures over the course of a
summer) can resolve entirely (2013 ISA, p. 6 76; 80 FR 65331, October 26, 2015).
3-8
-------
In considering these exposure estimates with regard to public health implications, the
Administrator concluded that the exposures and risks projected to remain upon meeting the then-
current (75 ppb) standard could reasonably be judged to be important from a public health
perspective. In particular, this conclusion was based on her judgment that it is appropriate to set a
standard that would be expected to eliminate, or almost eliminate, the occurrence of exposures,
while at moderate exertion, at or above 70 and 80 ppb (80 FR 65346, October 26, 2015). In
addition, given that the average percent of children estimated to experience two or more days
with exposures at or above the 60 ppb benchmark approaches 10% in some urban study areas (on
average across the analysis years), the Administrator concluded that the then-current standard did
not incorporate an adequate margin of safety against the potentially adverse effects that could
occur following repeated exposures at or above 60 ppb (80 FR 65345-46, October 26, 2015).
With regard to the HREA estimates of lung function risk in terms different magnitudes of
decrements in forced expiratory volume in one second (FEVi), the Administrator also gave
greater weight to estimates of multiple occurrences than to single occurrences, while additionally
noting CAS AC advice regarding uses of FEVi decrement estimates as scientifically relevant
surrogates for adverse health outcomes (Frey, 2014, p. 3). The Administrator noted that, when
averaged over the years of evaluation, the then-current (75 ppb) standard was estimated to allow
about 2 to 3% of children in the 15 urban study areas to experience two or more 03-induced lung
function decrements >15%, and to allow about 8 to 12% of children to experience two or more
03-induced lung function decrements >10% (80 FR 65346, October 26, 2015). Although she
recognized increased uncertainty in and placed less weight on both the HREA estimates for lung
function risk and the epidemiologic-study-based risk estimates, the Administrator concluded that
both types of estimates further support a conclusion that the 03-associated health effects
estimated to remain upon just meeting the then-current standard are an issue of public health
importance on a broad national scale. Thus, she concluded that O3 exposure and risk estimates,
when taken together, supported a conclusion that the exposures and health risks associated with
just meeting the then-current standard could reasonably be judged to be of public health
significance, such that the then-current standard was not sufficiently protective and did not
incorporate an adequate margin of safety.
In addition to the evidence and exposure/risk information, the Administrator also took
note of CASAC advice, which included the finding that "the current NAAQS for ozone is not
protective of human health" and the unanimous recommendation "that the Administrator revise
3-9
-------
the current primary ozone standard to protect public health" (Frey, 2014, p. 5). She further noted
similar CASAC advice in the prior 2008 review.17
In consideration of all of the above, the Administrator concluded that the then-current
primary O3 standard (with its level of 75 ppb) was not requisite to protect public health with an
adequate margin of safety, and that it should be revised to provide increased public health
protection. This decision was based on the Administrator's conclusions that the available
evidence and exposure and risk information clearly called into question the adequacy of public
health protection provided by the then-current primary standard such that it was "not appropriate,
within the meaning of section 109(d) (1) of the CAA, to retain the current standard" (80 FR
65346, October 26, 2015).
3.1.2 Considerations for the Revised Standard
The following subsections focus on the individual elements - indicator, averaging time,
form and level - for the new primary standard established in the 2015 review. While these
sections summarize the Administrator's key considerations and conclusions for each element
individually, she considered the elements collectively in evaluating the health protection afforded
by the standard, consistent with past practice.
3.1.2.1 Indicator
In considering whether O3 continued to be the most appropriate indicator for a standard
meant to provide protection against photochemical oxidants in ambient air, the Administrator
considered findings and assessments in the 2013 ISA and 2014 PA, as well as advice from the
CASAC and public comment. The 2013 ISA specifically noted that O3 is the only photochemical
oxidant (other than nitrogen dioxide) that is routinely monitored and for which a comprehensive
database exists (2013 ISA, section 3.6; 80 FR 65347, October 26, 2015). The PA additionally
noted that, since the precursor emissions that lead to the formation of O3 also generally lead to
the formation of other photochemical oxidants, measures leading to reductions in population
exposures to O3 can generally be expected to lead to reductions in other photochemical oxidants.
The CASAC indicated its view that O3 is the appropriate indicator "based on its causal or likely
causal associations with multiple adverse health outcomes and its representation of a class of
pollutants known as photochemical oxidants" (Frey, 2014c, p. ii). Based on all of these
considerations and public comments, the Administrator concluded that O3 remained the most
appropriate indicator for a standard meant to provide protection against photochemical oxidants
17 The CASAC O3 Panel for the 2008 review likewise recommended revision of the standard to one with a level
below 75 ppb. This earlier recommendation was based entirely on the evidence and information in the record for
the 2008 decision, which had been expanded in the 2015 review (Samet, 2011; Frey and Samet, 2012).
3-10
-------
in ambient air, and she retained O3 as the indicator for the primary standard (80 FR 65347,
October 26, 2015).
3.1.2.2 Averaging time
The 8-hour averaging time for the primary O3 standard was established in 1997 with the
decision to replace the then-existing 1-hour standard with an 8-hour standard (62 FR 38856, July
18, 1997). The decision in that review was based on evidence from numerous controlled human
exposure studies in healthy adults of adverse respiratory effects resulting from 6- to 8-hour
exposures, as well as quantitative analyses indicating the control provided by an 8-hour
averaging time of both 8-hour and 1-hour peak exposures and associated health risk (62 FR
38861, July 18, 1997; U.S. EPA, 1996). The 1997 decision was also consistent with advice from
the CASAC (62 FR 38861, July 18, 1997; 61 FR 65727, December 13, 1996). The EPA reached
similar conclusions in the subsequent 2008 review in which the 8-hour averaging time was
retained (73 FR 16436, March 27, 2008).
In the review completed in 2015, the Administrator again considered the averaging time
for the standard in light of both the strong evidence for Ch-associated respiratory effects
following short-term exposures and the available evidence related to effects following longer-
term exposures (80 FR 65347-50, October 26, 2015). In so doing, the Administrator noted the
substantial health effects evidence from controlled human exposure studies that demonstrate that
a wide range of respiratory effects (e.g., pulmonary function decrements, increases in respiratory
symptoms, lung inflammation, lung permeability, decreased lung host defense, and airway
hyperresponsiveness) occur in healthy adults following exposures ranging from 1 to 8 hours (80
FR 65348, October 26, 2015; 2013 ISA, section 6.2.1.1). The Administrator also noted the
strength of evidence from epidemiologic studies that evaluated a wide variety of populations
(e.g., including at-risk lifestages and populations, such as children and people with asthma,
respectively) using a number of different short-term averaging times, including the maximum 1-
hour concentration within a 24-hour period (1-hour max), the maximum 8-hour average
concentration within a 24-hour period (8-hour max), and the 24-hour average (80 FR 65348,
October 26, 2015; 2013 ISA, chapter 6). It was recognized that an 8-hour averaging time is
similar to the exposure periods evaluated in the more recent controlled human exposure studies
conducted at the lowest concentrations, and the Administrator noted that the epidemiologic
evidence alone did not provide a strong basis for distinguishing between the appropriateness of
1-hour, 8-hour and 24-hour averaging times. Thus, in consideration of the then-available health
effects information, the Administrator concluded that an 8-hour averaging time remained
appropriate for addressing health effects associated with short-term exposures to ambient air O3
(80 FR 65348, October 26, 2015).
3-11
-------
In considering the evidence related to longer-term exposures, the Administrator initially
considered the extent to which currently available evidence and exposure/risk information
suggested that a standard with an 8-hour averaging time can provide protection against
respiratory effects associated with longer-term exposures to ambient air O3. As in previous
reviews, the review completed in 2015 recognized and further evaluated changes in long-term air
quality patterns in response to attaining an 8-hour standard and the reduction in potential risk of
health effects associated with such patterns in areas meeting an 8-hour standard (80 FR 65348,
October 26, 2015).18 Furthermore, the Administrator observed that the CASAC agreed with the
choice of an 8-hour averaging time (Frey, 2014, p. ii). Therefore, based on the then-available
evidence and information discussed in detail in the 2013 ISA, 2014 HREA, and 2014 PA, along
with CASAC advice and public comments, the Administrator concluded that a standard with an
8-hour averaging time could effectively limit health effects attributable to both short- and long-
term O3 exposures and that it was appropriate to retain the 8-hour averaging time (80 FR 65350,
October 26, 2015).
3.1.2.3 Form
While giving foremost consideration to the adequacy of public health protection provided
by the combination of all elements of the standard, including the form, the Administrator placed
considerable weight on the findings from prior reviews with regard to the use of the .nth-high
metric, as described below (80 FR 65350-65352, October 26, 2015). Based on these findings and
consideration of CASAC advice, the Administrator judged it appropriate to retain the fourth-high
form, more specifically the annual fourth-highest daily maximum 8-hour O3 average
concentration, averaged over 3 years (80 FR 65352, October 26, 2015).
The concentration-based form was established in the 1997 review when it was recognized
that such a form better reflects the continuum of health effects associated with increasing O3
concentrations than an expected exceedance form, which had been the form of the standard prior
to 1997. Unlike an expected exceedance form, a concentration-based form gives proportionally
more weight to years when 8-hour O3 concentrations are well above the level of the standard
than years when 8-hour O3 concentrations are just above the level of the standard. More weight
was given to high O3 concentrations, in light of the available health evidence that indicated a
continuum of effects associated with exposures to varying concentrations of O3, and because the
extent to which public health is affected by exposure to O3 in ambient air is related to the actual
magnitude of the O3 concentration, not just whether the concentration is above a specified level.
18 Analyses described in detail in the 2014 HREA suggested that reductions in O3 precursors emissions in order to
meet a standard with an 8-hour averaging time, coupled with the appropriate form and level, would be expected to
reduce O3 concentrations in terms of the metrics reported in epidemiologic studies to be associated with
respiratory morbidity and mortality (80 FR 65348, October 26, 2015).
3-12
-------
With regard to a specific concentration-based form, the fourth-highest daily maximum was
selected in 1997, recognizing that a less restrictive form (e.g., fifth highest) would allow a larger
percentage of sites to experience O3 peaks above the level of the standard, and would allow more
days on which the level of the standard may be exceeded when the site attains the standard (62
FR 38868-38873, July 18, 1997), and there was not an basis identified for selection of a more
restrictive form (62 FR 38856, July 18, 1997).
In the subsequent 2008 review, the EPA considered the potential value of a percentile-
based form, recognizing that such a statistic is useful for comparing datasets of varying length
because it samples approximately the same place in the distribution of air quality values, whether
the dataset is several months or several years long (73 FR 16474, March 27, 2008). However, the
EPA concluded that, because of the differing lengths of the monitoring season for O3 across the
U.S., a percentile-based statistic would not be effective in ensuring the same degree of public
health protection across the country. Specifically, a percentile-based form would allow more
days with higher air quality values (i.e., higher O3 concentrations) in locations with longer O3
seasons relative to locations with shorter O3 seasons. Thus, the EPA concluded in the 2008
review that a form based on the ifh-highest maximum O3 concentration would more effectively
ensure that people who live in areas with different length O3 seasons received the same degree of
public health protection (73 FR 16474-75, March 27, 2008). The importance of a form that
provides stability to ongoing control programs was also recognized (73 FR 16474, March 27,
2008). In the case of O3, for example, it was noted that it was important to have a form that
provides stability and insulation from the impacts of extreme meteorological events that are
conducive to O3 occurrence. Such events could have the effect of reducing public health
protection, to the extent they result in frequent shifts in and out of attainment due to
meteorological conditions because such frequent shifting could disrupt an area's ongoing
implementation plans and associated control programs (73 FR 16475, March 27, 2008).
In the 2015 review, the Administrator continued to recognize the considerations
supporting the decisions in 1997 and 2008, and additionally noted recent CASAC advice in
which the CASAC indicated that the O3 standard should be based on the fourth-highest, daily
maximum 8-hour average value (averaged over 3 years), by stating that this form "provides
health protection while allowing for atypical meteorological conditions that can lead to
abnormally high ambient ozone concentrations which, in turn, provides programmatic stability"
(Frey, 2014, p. 6; 80 FR 65352, October 26, 2015).
3.1.2.4 Level
The Administrator's decision to set the level of the revised primary O3 standard at 70 ppb
built upon her conclusion (summarized in section 3.1.1 above) that the overall body of scientific
3-13
-------
evidence and exposure/risk information called into question the adequacy of the public health
protection afforded by the then-current standard, particularly for at-risk populations and
lifestages (80 FR 65362, October 26, 2015). In her decision on level, the Administrator placed
the greatest weight on the results of controlled human exposure studies and on quantitative
analyses based on information from these studies, particularly analyses of O3 exposures of
concern. The Administrator viewed the results of the lung function risk assessment, analyses of
O3 air quality in locations of epidemiologic studies, and epidemiologic-study-based quantitative
health risk assessment as providing information in support of her decision to revise the then-
current standard, but of less utility for selecting a particular standard level among a range of
options (80 FR 65362, October 26, 2015).
In placing weight on information from controlled human exposure studies and analyses
based on information from these studies, the Administrator noted that controlled human exposure
studies provide the most certain evidence indicating the occurrence of health effects in humans
following specific O3 exposures, noting in particular that the effects reported in the controlled
human exposure studies are due solely to O3 exposures, and are not complicated by the presence
of co-occurring pollutants or pollutant mixtures (as is the case in epidemiologic studies). The
Administrator's emphasis on the information from the controlled human exposure studies was
consistent with the CASAC's advice and interpretation of the scientific evidence (80 FR 65362,
October 26, 2015; Frey, 2014). In this regard, the Administrator recognized that: (1) the largest
respiratory effects, and the broadest range of effects, have been studied and reported following
exposures to 80 ppb O3 or higher (i.e., decreased lung function, increased airway inflammation,
increased respiratory symptoms, airway hyperresponsiveness, and decreased lung host defense);
(2) exposures to O3 concentrations somewhat above 70 ppb have been shown to both decrease
lung function and to result in respiratory symptoms; and (3) exposures to O3 concentrations as
low as 60 ppb have been shown to decrease lung function and to increase airway inflammation
(80 FR 65363, October 26, 2015).
The Administrator considered both ATS recommendations and CASAC advice to inform
her judgments on the potential adversity to public health of effects reported in controlled human
exposure studies (80 FR 65363, October 26, 2015). In so doing, the Administrator concluded that
the evidence from controlled human exposure studies provided strong support for the conclusion
that a revised standard with a level of 70 ppb is requisite to protect public health with an
adequate margin of safety. This conclusion was based, in part, on the fact that such a standard
level would be well below the O3 exposure concentration documented to result in the widest
range of respiratory effects (i.e., 80 ppb), and below the lowest O3 exposure concentration shown
to result in the adverse combination of lung function decrements and respiratory symptoms (80
FR 65363, October 26, 2015).
3-14
-------
In considering the degree of protection provided by a revised primary O3 standard, the
Administrator considered the extent to which that standard would be expected to limit population
exposures to the broad range of O3 exposures shown to result in health effects (80 FR 65363,
October 26, 2015). In considering the exposure estimates from the HREA, the Administrator
focused on the estimates of two or more exposures of concern in order to provide a health-
protective approach to take into account the potential for repeated occurrences of exposures that
could result in adverse effects. In so doing, she placed the most emphasis on setting a standard
that appropriately limits repeated occurrences of exposures at or above the 70 and 80 ppb
benchmarks, while at elevated ventilation. She noted that a revised standard with a level of 70
ppb was estimated to eliminate the occurrence of two or more days with exposures at or above
80 ppb and to virtually eliminate the occurrence of two or more days with exposures at or above
70 ppb for all children and children with asthma, even in the worst-case year and location
evaluated.19 Given the considerable protection provided against repeated exposures of concern
for all benchmarks evaluated in the HREA, the Administrator judged that a standard with a level
of 70 ppb incorporated a margin of safety against the adverse Ch-induced effects shown to occur
in the controlled human exposure studies (80 FR 65364, October 26, 2015).20
While she was less confident that adverse effects would occur following exposures to O3
concentrations as low as 60 ppb,21 as discussed above, the Administrator judged it to also be
appropriate to consider estimates of exposures (while at moderate or greater exertion) for the 60
ppb benchmark (80 FR 65363-64, October 26, 2015). In so doing, she recognized that while
CASAC advice regarding the potential adversity of effects observed in studies of 60 ppb was less
definitive than for effects observed at the next higher concentration studied, the CASAC did
clearly advise the EPA to consider the extent to which a revised standard is estimated to limit the
effects observed in studies of 60 ppb exposures (80 FR 65364, October 26, 2015; Frey, 2014).
The Administrator's consideration of exposures at or above the 60 ppb benchmark, and
particularly consideration of multiple occurrences of such exposures, was primarily in the
19 Under conditions just meeting an alternative standard with a level of 70 ppb across the 15 urban study areas, the
estimate for two or more days with exposures at or above 70 ppb was 0.4% of children, in the worst year and
worst area (80 FR 65313, Table 1, October 26, 2015).
20 In so judging, she noted that the CASAC had recognized the choice of a standard level within the range it
recommended based on the scientific evidence (which is inclusive of 70 ppb) to be a policy judgment (80 FR
65355, October 26, 2015; Frey, 2014).
21 The Administrator was "notably less confident in the adversity to public health of the respiratory effects that have
been observed following exposures to O3 concentrations as low as 60 ppb," based on her consideration of the
ATS recommendation on judging adversity from transient lung function decrements alone, the uncertainty in the
potential for such decrements to increase the risk of other, more serious respiratory effects in a population (per
ATS recommendations on population-level risk), and the less clear CASAC advice regarding potential adversity
of effects at 60 ppb compared to higher concentrations studied (80 FR 65363, October 26, 2015).
3-15
-------
context of considering the extent to which the health protection provided by a revised standard
included a margin of safety against the occurrence of adverse Ch-induced effects (80 FR 65334-
35, 65364, October 26, 2015). In this context, the Administrator noted that a revised standard
with a level of 70 ppb was estimated to protect the vast majority of children in urban study areas
(i.e., about 96% to more than 99% of children in individual areas) from experiencing two or
more days with exposures at or above 60 ppb (while at moderate or greater exertion). Compared
to the estimates for the then-current standard (with its level of 75 ppb), this represented a
reduction of more than 60%. Given the considerable protection provided against repeated
exposures of concern for all of the benchmarks evaluated, including the 60 ppb benchmark, the
Administrator judged that a standard with a level of 70 ppb would incorporate a margin of safety
against the adverse Ch-induced effects shown to occur following exposures (while at moderate or
greater exertion) to a somewhat higher concentration. The Administrator also judged the HREA
results for one or more exposures at or above 60 ppb to provide further support for her somewhat
broader conclusion that "a standard with a level of 70 ppb would incorporate an adequate margin
of safety against the occurrence of O3 exposures that can result in effects that are adverse to
public health" (80 FR 65364, October 26, 2015).22
While placing limited weight on the lung function risk estimates,23 epidemiologic
evidence24 and quantitative estimates based on information from the epidemiologic studies, the
Administrator additionally considered that information in the context of her consideration of a
22 While the Administrator was less concerned about single occurrences of O3 exposures of concern, especially for
the 60 ppb benchmark, she judged that estimates of one or more exposures of concern can provide further insight
into the margin of safety provided by a revised standard. In this regard, she noted that "a standard with a level of
70 ppb is estimated to (1) virtually eliminate all occurrences of exposures of concern at or above 80 ppb; (2)
protect the vast majority of children in urban study areas from experiencing any exposures of concern at or above
70 ppb (i.e., > about 99%, based on mean estimates; Table 1); and (3) to achieve substantial reductions, compared
to the then-current standard, in the occurrence of one or more exposures of concern at or above 60 ppb (i.e., about
a 50% reduction; Table 1)" (80 FR 65364, October 26, 2015).
23 The Administrator noted important uncertainties in using lung function risk estimates as a basis for considering
the occurrence of adverse effects in the population (also recognized in the prior review) that limited her reliance
on these estimates to distinguish between the appropriateness of the health protection afforded by a standard level
of 70 ppb versus lower levels (80 FR 65364, October 26, 2015). These uncertainties related to (1) the ATS
recommendation that "a small, transient loss of lung function, by itself, should not automatically be designated as
adverse" (ATS, 2000); (2) uncertainty in the extent to which a transient population-level decrease in FEVi would
increase the risk of other, more serious respiratory effects in that population (i.e., per ATS recommendations on
population-level risk); and (3) that the C ASAC did not advise considering a standard that would be estimated to
eliminate 03-induced lung function decrements >10 or 15% (Frey, 2014); 80 FR 65364, October 26, 2015).
24 While the Administrator concluded that analyses of air quality in single-city epidemiologic studies support a level
at least as low as 70 ppb, based on a study (Mar and Koenig, 2009) reporting health effect associations in a
location that met the then-current standard over the entire study period but that would have violated a revised
standard with a level of 70 ppb, she further judged that they are of more limited utility for distinguishing between
the appropriateness of the health protection estimated for a standard level of 70 ppb and the protection estimated
for lower levels (80 FR 65364, October 26, 2015).
3-16
-------
standard with a level of 70 ppb. For example, she judged that a standard with a level of 70 ppb
would be expected to result in important reductions in the population-level risk of Ch-induced
lung function decrements in children, including children with asthma (80 FR 65364, October 26,
2015). With regard to the epidemiologic evidence, the Administrator noted that a revised
standard with a level of 70 ppb would provide additional public health protection, beyond that
provided by the then-current standard, against the clearly adverse effects analyzed in
epidemiologic studies (80 FR 65364, October 26, 2015). With regard to the epidemiology-based
risk estimates, the Administrator judged that a revised standard with a level of 70 ppb will result
in meaningful reductions in the mortality and respiratory morbidity risk that is associated with
short- or long-term concentrations of O3 in ambient air (80 FR 65365, October 26, 2015).
In summary, given her consideration of the evidence, exposure and risk information,
advice from the CASAC, and public comments, the Administrator judged a primary standard of
70 ppb in terms of the 3-year average of annual fourth-highest daily maximum 8-hour average
O3 concentrations to be requisite to protect public health, including the health of at-risk
populations, with an adequate margin of safety (80 FR 65365, October 26, 2015).
3.2 GENERAL APPROACH AND KEY ISSUES IN THIS REVIEW
As is the case for all such reviews, this review of the primary O3 standard is most
fundamentally based on using the Agency's assessment of the current scientific evidence and
associated quantitative analyses to inform the Administrator's judgments regarding a primary
standard that is requisite to protect public health with an adequate margin of safety. The
approach for this review builds on the substantial assessments and evaluations performed over
the course of the last review, taking into account the more recent scientific information and air
quality data now available to inform our understanding of the key-policy relevant issues in this
review.
The evaluations in the PA of the scientific assessments in the ISA augmented by the
quantitative risk and exposure analyses25 are intended to inform the Administrator's public health
policy judgments and conclusions, including his decisions as to whether to retain or revise the
primary O3 standard. The PA evaluations consider the potential implications of various aspects
of the scientific evidence, the exposure/risk-based information, and the associated uncertainties
and limitations. In so doing, the approach for this PA involves evaluating the scientific and
technical information to address a series of key policy-relevant questions using both evidence-
25 The overarching purpose of the quantitative exposure and risk analyses is to inform the Administrator's
conclusions on the public health protection afforded by the current primary standard. An important focus is the
assessment, based on current tools and information, of the potential for exposures and risks beyond those
indicated by the information available at the time the standard was established.
3-17
-------
and exposure/risk-based considerations. Together, consideration of the full set of evidence and
information available in this review will inform the answer to the following initial overarching
question for the review:
• Do the currently available scientific evidence and exposure/risk-based information
support or call into question the adequacy of the public health protection afforded by
the current primary O3 standard?
In reflecting on this question, we will consider the available body of scientific evidence,
assessed in the ISA and used as a basis for developing or interpreting exposure/risk analyses,
including whether it supports or calls into question the scientific conclusions reached in the last
review regarding health effects related to exposure to ambient air-related O3. Information
available in this review that may be informative to public health judgments regarding
significance or adversity of key effects will also be considered. Additionally, the currently
available exposure and risk information, whether newly developed in this review or
predominantly developed in the past and interpreted in light of current information, will be
considered, including with regard to the extent to which it may continue to support judgments
made in the last review. Further, in considering this question with regard to the primary O3
standard, as in all NAAQS reviews, we give particular attention to exposures and health risks to
at-risk populations (including at-risk lifestages) ,26
Evaluation of the available scientific evidence and exposure/risk information with regard
to this consideration of the current standard will focus on key policy-relevant issues by
addressing a series of questions including the following:
• Is there newly available evidence that indicates the importance of photochemical oxidants
other than O3 with regard to abundance in ambient air, and potential for human exposures
and health effects?
• Does the currently available scientific evidence alter our conclusions from the last review
regarding the nature of health effects attributable to human exposure to O3 from ambient air?
• Does the current evidence alter our understanding of populations that are particularly at risk
from O3 exposures?
• Does the current evidence alter our conclusions from the previous review regarding the
exposure duration and concentrations associated with health effects? To what extent does the
currently available scientific evidence indicate health effects attributable to exposures to O3
26 As used here and similarly throughout this document, the term population refers to persons having a quality or
characteristic in common, such as a specific pre-existing illness or a specific age or lifestage. Identifying at-risk
populations involves consideration of susceptibility and vulnerability. Susceptibility refers to innate (e.g., genetic
or developmental aspects) or acquired (e.g., disease or smoking status) sensitivity that increases the risk of health
effects occurring with exposure to O3. Vulnerability refers to an increased risk of 03-related health effects due to
factors such as those related to socioeconomic status, reduced access to health care or exposure.
3-18
-------
concentrations lower than previously reported and what are important uncertainties in that
evidence?
• To what extent have previously identified uncertainties in the health effects evidence been
reduced or do important uncertainties remain?
• What are the nature and magnitude of O3 exposures and associated health risks associated
with air quality conditions just meeting the current standard?
• To what extent are the estimates of exposures and risks to at-risk populations associated with
air quality conditions just meeting the current standard reasonably judged important from a
public health perspective?
• What are the important uncertainties associated with any exposure/risk estimates?
If the information available in this review suggests that revision of the current primary
standard would be appropriate to consider, the PA will also evaluate how the standard might be
revised based on the available scientific information, air quality assessments, and exposure/risk
information, and also considering what the available information indicates as to the health
protection expected to be afforded by the current or potential alternative standards. Such an
evaluation may consider the effect of revision of one or more elements of the standard (indicator,
averaging time, level and form), with the impact evaluated being based on the resulting potential
standard and all of its elements collectively. Based on such evaluations, the PA would then
identify potential alternative standards (specified in terms of indicator, averaging time, level, and
form) intended to reflect a range of alternative policy judgments as to the degree of protection
that is requisite to protect public health with an adequate margin of safety, and options for
standards expected to achieve it. The specific policy-relevant questions that frame such an
evaluation of what revision of the standard might be appropriate to consider include:
• Does the currently available information call into question the identification of O3 as the
indicator for photochemical oxidants? Is support provided for considering a different
indicator?
• Does the currently available information call into question the current averaging time? Is
support provided for considering different averaging times for the standard?
• What does the currently available information indicate with regard to a range of levels and
forms of alternative standards that may be supported and what are the uncertainties and
limitations in that information?
• What do the available analyses indicate with regard to exposure and risk associated with
specific alternative standards? What are the associated uncertainties? To what extent might
such alternatives be expected to reduce adverse impacts attributable to O3, and what are the
uncertainties in the estimated reductions?
The approach to reaching conclusions on the current primary standard and, as
appropriate, on potential alternative standards is summarized in general terms in Figure 3-1.
3-19
-------
Adequacy of Current Standard
Ex.ccsur-- 3rd R:sk-Ba:;s3 Ccnsdsraton;
>Na&ire, magnitude, and knportance of
esimaed exposures and risks associated
with juK meeSng the current standard?
>UncertainSes in the exposure and risk
esanates?
Dees currency avaflabte evidence and related
uncertainties strengthen orcal into question prior
conclusions?
¦ Evidence of heath effeas mx previously
identified?
¦ Newiysdenafed ai-risk populations?
¦ Evidence of heath effecs at tower levels or for
different exposure duraions?
¦ Uncertairtes identified in the last review
reduced or new uncertainties emerged?
Evidence-Based Considerations
NO
Consider Retaining
Current Standard
YES
Consider Potential Alternative Standards
Does
information call into
question adequacy
of current
^ standard?
Etements of Potential Aternajve Standards
^Indicator, Averaging Time, Fonm, Level
Potential Alternative Standards for Consideration
Figure 3-1. Overview of general approach for review of the primary O3 standard.
The Agency's approach in reviewing primary standards is consistent with requirements
of the provisions of the CAA related to the review of the NAAQS and with how the EPA and the
3-20
-------
courts have historically interpreted the CAA. As discussed in section 1.1 above, these provisions
require the Administrator to establish primary standards that, in the Administrator's judgment,
are requisite (i.e., neither more nor less stringent than necessary) to protect public health with an
adequate margin of safety. Consistent with the Agency's approach across all NAAQS reviews,
the approach of the PA to informing these judgments is based on a recognition that the available
health effects evidence generally reflects continuums that include ambient air exposures for
which scientists generally agree that health effects are likely to occur through lower levels at
which the likelihood and magnitude of response become increasingly uncertain. The CAA does
not require the Administrator to establish a primary standard at a zero-risk level or at background
concentration levels, but rather at a level that reduces risk sufficiently so as to protect public
health, including the health of sensitive groups,27 with an adequate margin of safety.
The decisions on the adequacy of the current primary standard and on any alternative
standards considered in a review are largely public health policy judgments made by the
Administrator. The four basic elements of the NAAQS (i.e., indicator, averaging time, form, and
level) are generally considered collectively in evaluating the health protection afforded by the
current standard, and by any alternatives considered. The Administrator's final decisions in a
review draw upon the scientific evidence for health effects, quantitative analyses of population
exposures and/or health risks, as available, and judgments about how to consider the
uncertainties and limitations that are inherent in the scientific evidence and quantitative analyses.
3.3 HEALTH EFFECTS EVIDENCE
3.3.1 Nature of Effects
The evidence base available in the current review includes decades of extensive evidence
that clearly describes the role of O3 in eliciting an array of respiratory effects and recent evidence
suggests the potential for relationships between O3 exposure and other effects. As was
established in prior reviews, the most commonly observed effects, and those for which the
evidence is strongest, are transient decrements in pulmonary function and respiratory symptoms,
such as coughing and pain on deep inspiration, as a result of short-term exposures (ISA, section
IS.4.3.1; 2013 ISA, p. 2-26). These effects are demonstrated in the large, long-standing evidence
27 More than one population group may be identified as sensitive or at-risk in a NAAQS review. Decisions on
NAAQS reflect consideration of the degree to which protection is provided for these sensitive population groups.
To the extent that any particular population group is not among the identified sensitive groups, a decision that
provides protection for the sensitive groups would be expected to also provide protection for other population
groups.
3-21
-------
base of controlled human exposure studies28 (1978 AQCD, 1986 AQCD, 1996 AQCD, 2006
AQCD, 2013 ISA, ISA). The lung function effects are also positively associated with ambient air
O3 concentrations in epidemiologic panel studies, available in past reviews, that describe these
associations for outdoor workers and children attending summer camps in the 1980s and 1990s
(2013 ISA, section 6.2.1.2; ISA, Appendix 3, section 3.1.4.1.3). The epidemiologic evidence
base additionally documents associations of O3 concentrations in ambient air with more severe
health outcomes, including asthma-related emergency department visits and hospital admissions
(2013 ISA, section 6.2.7; ISA, Appendix 3, sections 3.1.5.1 and 3.1.5.2). Extensive animal
toxicological evidence informs a detailed understanding of mechanisms underlying the
respiratory effects of short-term exposures (ISA, Appendix 3, section 3.1.11), and studies in
animal models also provide evidence for effects of longer-term O3 exposure on the developing
lung (ISA, Appendix 3, section 3.2.6).
• Does the currently available scientific evidence alter our conclusions from the last
review regarding the health effects attributable to exposure to O3?
The current evidence continues to support our prior conclusion that short-term O3
exposure causes respiratory effects. Specifically, the full body of evidence continues to support
the conclusions of a causal relationship of respiratory effects with short-term O3 exposures and a
likely causal relationship of respiratory effects with longer-term exposures (ISA, sections
IS.4.3.1 and IS.4.3.2). The current evidence base, expanded by evidence newly available in this
review, also indicates a likely causal relationship between short-term O3 exposure and metabolic
effects,29 which were not evaluated as a separate category of effects in the last review when less
evidence was available (ISA, section IS.4.3.3). The, newly available evidence is primarily from
experimental animal research. For other types of health effects, new evidence has led to different
conclusions from those reached in the prior review. Specifically, the current evidence,
particularly in light of the additional controlled human exposure studies, is less consistent than
28 The vast majority of the controlled human exposure studies (and all of the studies conducted at the lowest
exposures) involved young healthy adults as study subjects. There are also some 1-8 hr controlled human
exposure studies in older adults and adults with asthma, and there are still fewer controlled human exposure
studies in healthy children (i.e., individuals aged younger than 18 years) or children with asthma (See, for
example, Appendix 3A, Table 3A-3).
29 The term metabolic effects is used in the ISA to refer metabolic syndrome (a collection of risk factors including
high blood pressure, elevated triglycerides and low high density lipoprotein cholesterol), diabetes, metabolic
disease mortality, and indicators of metabolic syndrome that include alterations in glucose and insulin
homeostasis, peripheral inflammation, liver function, neuroendocrine signaling, and serum lipids (ISA, section
IS.4.3.3).
3-22
-------
what was previously available and less indicative of 03-induced cardiovascular effects.30 This
new evidence has altered conclusions from the last review with regard to relationships between
short-term O3 exposures and cardiovascular effects and mortality, such that likely causal
relationships are no longer concluded.31 Thus, while conclusions have changed for some effects
based on the new evidence, the conclusions reached in the last review on respiratory effects are
supported by the current evidence, and conclusions are also newly reached for an additional
category of health effects.
3.3.1.1 Respiratory Effects
As in the last review, the currently available evidence in this review supports the
conclusion of a causal relationship between short-term O3 exposure and respiratory effects (ISA,
section IS. 1.3.1). The strongest evidence for this comes from controlled human exposure studies,
also available in the last review, demonstrating 03-related respiratory effects in generally healthy
adults.32 Experimental studies in animals also document an array of respiratory effects resulting
from short-term O3 exposure and provide information related to underlying mechanisms (ISA,
Appendix 3, section 3.1). The potential for O3 exposure to elicit health outcomes more serious
than those assessed in the experimental studies, particularly for children with asthma, continues
to be indicated by the epidemiologic evidence of associations of O3 concentrations in ambient air
with increased incidence of hospital admissions and emergency department visits for an array of
health outcomes, including asthma exacerbation, COPD exacerbation, respiratory infection, and
combinations of respiratory diseases (ISA, Appendix 3, sections 3.1.5 and 3.1.6). The strongest
such evidence is for asthma-related outcomes and specifically asthma-related outcomes for
children, indicating an increased risk for people with asthma and particularly children with
asthma (ISA, Appendix 3, section 3.1.5.7).
30 As described in the ISA, "[t]he number of controlled human exposure studies showing little evidence of ozone
induced cardiovascular effects has grown substantially" and "the plausibility for a relationship between short-
term ozone exposure to cardiovascular health effects is weaker than it was in the previous review, leading to the
revised causality determination" (ISA, p. IS 43).
31 The currently available evidence for cardiovascular, reproductive and nervous system effects, as well as mortality,
is "suggestive of, but not sufficient to infer" a causal relationship with short- or long-term O3 exposures (ISA,
Table IS 1). The evidence is inadequate to infer the presence or absence of a causal relationship between long-
term O3 exposure and cancer (ISA, section IS4.3.6.6).
32 The phrases "healthy adults" or "healthy subjects" are used to distinguish from subjects with asthma or other
respiratory diseases, for which there are much fewer controlled human exposure studies. For studies of healthy
subjects "the study design generally precludes inclusion of subjects with serious health conditions," such as
individuals with severe respiratory diseases (2013 ISA, p. lx).
3-23
-------
Respiratory responses observed in human subjects exposed to O3 for periods of 8 hours or
less, while intermittently exercising, include reduced lung function,33 respiratory symptoms,
increased airway responsiveness, mild bronchoconstriction (measured as a change in specific
airway resistance [sRaw]), and pulmonary inflammation, with associated injury and oxidative
stress (ISA, Appendix 3, section 3.1.4; 2013 ISA, sections 6.2.1 through 6.2.4). The available
mechanistic evidence, discussed in greater detail in the ISA, describes pathways involving the
respiratory and nervous systems by which O3 results in pain-related respiratory symptoms and
reflex inhibition of maximal inspiration (inhaling a full, deep breath), commonly quantified by
decreases in forced vital capacity (FVC) and total lung capacity. This reflex inhibition of
inspiration combined with mild bronchoconstriction contributes to the observed decrease in
forced expiratory volume in one second (FEVi), the most common metric used to assess O3-
related pulmonary function effects. The evidence also indicates that the additionally observed
inflammatory response is correlated with mild airway obstruction, generally measured as an
increase in sRaw (ISA, Appendix 3, section 3.1.3). As described in section 3.3.3 below, the
prevalence and severity of respiratory effects in controlled human exposure studies, including
symptoms (e.g., pain on deep inspiration, shortness of breath, and cough) increases, with
increasing O3 concentration, exposure duration, and ventilation rate of exposed subjects (ISA,
Appendix 3, sections 3.1.4.1 and 3.1.4.2).
Within the evidence base from controlled human exposure studies, the majority of studies
involve healthy adult subjects (generally 18 to 35 years old), although there are studies involving
subjects with asthma, and a limited number of studies, generally of durations shorter than four
hours, involving adolescents and adults older than 50 years. A summary of salient observations
of O3 effects on lung function, based on the controlled human exposure study evidence reviewed
in the 1996 and 2006 AQCDs, and recognized in the 2013 ISA, continues to pertain to this
evidence base as it exists today (ISA, Appendix 3, section 3.1.4.1.1, p. 3-11): "(1) young healthy
adults exposed to >80 ppb O3 develop significant reversible, transient decrements in pulmonary
function and symptoms of breathing discomfort if minute ventilation (Ve) or duration of
exposure is increased sufficiently [i.e., as measured by FEVi and/or FVC]; (2) relative to young
adults, children experience similar spirometric responses but lower incidence of symptoms from
33 The measure of lung function response most commonly considered across 03 NAAQS reviews is changes in
FEVi. In considering controlled human exposure studies, an 03-induced change in FEVi is typically the
difference between the decrement observed with O3 exposure and what is generally an improvement observed
with filtered air (FA) exposure. As explained in the 2013 ISA, "[n]oting that some healthy individuals experience
small improvements while others have small decrements in FEVi following FA exposure, investigators have used
the randomized, crossover design with each subject serving as their own control (exposure to FA) to discern
relatively small effects with certainty since alternative explanations for these effects are controlled for by the
nature of the experimental design" (2013 ISA, pp. 6-4 to 6-5).
3-24
-------
O3 exposure; (3) relative to young adults, ozone-induced spirometric responses are decreased in
older individuals; (4) there is a large degree of inter-subject variability in physiologic and
symptomatic responses to O3, but responses tend to be reproducible within a given individual
over a period of several months; and (5) subjects exposed repeatedly to O3 for several days
experience an attenuation of spirometric and symptomatic responses on successive exposures,
which is lost after about a week without exposure."
The evidence is most well established with regard to the effects, reversible with the
cessation of exposure, that are associated with short-term exposures of several hours. For
example, the evidence indicates a rapid recovery from 03-induced lung function decrements
(e.g., reduced FEVi) and respiratory symptoms (2013 ISA, section 6.2.1.1). However, in some
cases, such as after exposure to higher concentrations such as 300 ppb, the recovery phase may
be slower and involve a longer time period (e.g., at least 24 hours [hrs]). Repeated daily exposure
studies at such higher concentrations also have found FEVi response to be enhanced on the
second day of exposure. This enhanced response is absent, however, with repeated exposure at
lower concentrations, perhaps as a result of a more complete recovery or less damage to
pulmonary tissues (2013 ISA, section pp. 6-13 to 6-14; Folinsbee et al., 1994).
As recognized in the last review, the persistence of other 03-induced respiratory effects,
and the potential for repeated occurrences to contribute to further effects can be an important
consideration. For example, as described in the 2013 ISA, 03-induced respiratory tract
inflammation "can have several potential outcomes: (1) inflammation induced by a single
exposure (or several exposures over the course of a summer) can resolve entirely; (2) continued
acute inflammation can evolve into a chronic inflammatory state; (3) continued inflammation can
alter the structure and function of other pulmonary tissue, leading to diseases such as fibrosis; (4)
inflammation can alter the body's host defense response to inhaled microorganisms, particularly
in potentially at-risk populations such as the very young and old; and (5) inflammation can alter
the lung's response to other agents such as allergens or toxins" (2013 ISA, p. 6-76). With regard
to 03-induced increases in airway responsiveness, the controlled human exposure study evidence
for healthy adults generally indicates a resolution within 18 to 24 hours after exposure (ISA,
Appendix 3, section 3.1.4.3.1).
The extensive evidence base for O3 health effects, compiled over several decades,
continues to indicate respiratory responses to short exposures as the most sensitive effects of O3.
Such effects are well documented in controlled human exposure studies, most of which involve
healthy adults. These studies have documented an array of respiratory effects, including reduced
lung function, respiratory symptoms, increased airway responsiveness, and inflammation, in
study subjects following 1- to 8-hour exposures, primarily while exercising. Such effects are of
increased significance to people with asthma, particularly children, who are the age group most
3-25
-------
likely to be outdoors at activity levels corresponding to those that have been associated with
respiratory effects in the human exposure studies (as recognized below in sections 3.3.2 and 3.4).
This increased significance is illustrated by the epidemiological findings of positive associations
between O3 exposure and asthma-related ED visits and hospital admissions for children with
asthma. Thus, the evidence indicates O3 exposure to increase the risk of asthma exacerbation,
and associated outcomes, in children with asthma.
The 2013 ISA and past AQCDs have also concluded that the experimental animal
evidence indicates the potential for O3 exposures to increase susceptibility to infectious diseases
through effects on defense mechanisms of the respiratory tract (2013 ISA, section 6.2.5).
Evidence regarding respiratory infections and associated effects has been augmented by a
number of epidemiologic studies reporting positive associations between short-term O3
concentrations and emergency department visits for a variety of respiratory infection endpoints
(ISA, Appendix 3, section 3.1.7.1).
Although the long-term exposure conditions that may contribute to further respiratory
effects are less well understood, the conclusion based on the current evidence base remains that
there is likely to be a causal relationship for such exposure conditions with respiratory effects
(ISA, section IS.4.3.2). Most notably, experimental studies, including with nonhuman infant
primates, have provided evidence relating O3 exposure to allergic asthma-like effects, and
epidemiologic cohort studies have reported associations of O3 concentrations in ambient air with
asthma development in children (ISA, Appendix 3, section 3.2.4.1.3 and 3.2.6). The biological
plausibility of such a role for O3 has been indicated by animal toxicological evidence on
biological mechanisms (ISA, Appendix 3, sections 3.2.3 and 3.2.4.1.2). Specifically, the animal
evidence, including the nonhuman primate studies of early life O3 exposure, indicates that such
exposures can cause "structural and functional changes that could potentially contribute to
airway obstruction and increased airway responsiveness," which are hallmarks of asthma (ISA,
Appendix 3, section 3.2.6, p. 3-113).
Overall, the respiratory effects evidence newly available in this review is generally
consistent with the evidence base in the last review (ISA, Appendix 3, section 3.1.4). A few
recent studies provide insights in previously unexamined areas, both with regard to human study
groups and animal models for different effects, while other studies confirm and provide depth to
prior findings with updated protocols and techniques (ISA, Appendix 3, sections 3.1.11 and
3.2.6). Thus, our current understanding of the respiratory effects of O3 is similar to that in the last
review.
One aspect of the evidence that has been augmented concerns pulmonary function in
adults older than 50 years of age. Previously available evidence in this age group indicated
smaller 03-related decrements in middle-aged adults (35 to 60 years) than in adults 35 years of
3-26
-------
age and younger (2006 AQCD, p. 6-23; 2013 ISA, p. 6-22; ISA, Appendix 3, section
3.1.4.1.1.2). A recent multicenter study of 55- to 70-year old subjects (average of 60 years),
conducted for a 3-hour duration involving alternating 15-minute rest and exercise periods and a
120 ppb exposure concentration, reported a statistically significant O3 FEVi response (ISA,
Appendix 3, section 3.1.4.1.1.2; Arjomandi et al., 2018). While there is not a precisely
comparable study in younger adults, the mean response for the 55- to 70-year olds, 1.2% O3-
related FEVi decrement, is lower than results for somewhat comparable exposures in adults aged
35 or younger, suggesting somewhat reduced responses to O3 exposure in this older age group
(ISA, Appendix 3, section 3.1.4.1.1.2; Arjomandi et al., 2018; Adams, 2000; Adams, 2006a).34
Such a reduced response in middle-aged and older adults compared to young adults is consistent
with conclusions in previous reviews (2013 ISA, section 6.2.1.1; 2006 AQCD, section 6.4).
The strongest evidence of 03-related health effects, as was the case in the last review,
continues to document the respiratory effects of O3 (ISA, section ES.4.1). Among the newly
available studies, there are several controlled human exposure studies that investigated lung
function effects of higher exposure concentrations (e.g., 100 to 300 ppb) in healthy individuals
younger than 35 years old, with findings generally consistent with previous studies (ISA,
Appendix 3, section 3.4.1.1.2, p. 3-17). No studies are newly available of 6.6-hour controlled
human exposures (with exercise) to concentrations below those previously studied.35 The newly
available animal toxicological studies augment the previously available information concerning
mechanisms underlying the effects documented in experimental studies. Newly available
epidemiologic studies of hospital admissions and emergency department visits for a variety of
respiratory outcomes supplement the previously available evidence with additional findings of
consistent associations with O3 concentrations across a number of study locations (ISA,
Appendix 3, sections 3.1.4.1.3, 3.1.5, 3.1.6.1.1, 3.1.7.1 and 3.1.8). These studies include a
number that report positive associations for asthma-related outcomes, as well as a few for
COPD-related outcomes. Together these studies in the current epidemiologic evidence base
34 For the same exposure concentration of 120 ppb, Adams (2006a) observed an average 3.2%, statistically
significant, 03-related FEVi decrement in young adults (average age 23 years) at the end of the third hour of an 8-
hour protocol that alternated 30 minutes of exercise and rest, with the equivalent ventilation rate (EVR) averaging
20 L/min-m2 during the exercise periods (versus 15 to 17 L/min-m2 in Arjomandi et al., 2018]). For the same
concentration with a lower EVR during exercise (17 L/min-m2), although with more exercise, Adams (2000)
observed a 4%, statistically significant, 03-related FEVi decrement in young adults (average age 22 years) after
the third hour of a 6.6-hour protocol (alternating 50 minutes exercise and 10 minutes rest).
35 As recognized in section 3.3.1.1 above, there is a newly available 3-hr study of subjects aged 55 years of age or
older that involves a slightly lower target ventilation rate for the exercise periods. The exposure concentrations
were 120 ppb and 70 ppb, only the former of which elicited a statistically significant FEVi decrement in this age
group of subjects (ISA, Appendix 3, section 3.1.4.1.1.2).
3-27
-------
continue to indicate the potential for O3 exposures to contribute to such serious health outcomes,
particularly for people with asthma.
3.3.1.2 Other Effects
As was the case for the evidence available in the last review, the currently available
evidence for health effects other than those on the respiratory system is more uncertain than that
for respiratory effects. For some of these other categories of effects, the evidence now available
has contributed to changes in conclusions reached in the last review. For example, cardiovascular
effects and mortality are no longer concluded to be likely causally related to O3 exposures based
on newly available evidence in combination with the uncertainties recognized for the evidence
available in the last review. Additionally, newly available evidence has also led to conclusions
for another category, metabolic effects, for which formal causal determinations were previously
not articulated.
The ISA finds the evidence for metabolic effects sufficient to conclude that there is likely
to be a causal relationship with short-term O3 exposures (ISA, section IS.4.3.3). The evidence of
metabolic effects of O3 comes primarily from experimental animal study findings that short-term
O3 exposure can impair glucose tolerance, increase triglyceride levels and elicit fasting
hyperglycemia and increase hepatic gluconeogenesis (ISA, Appendix 5, section 5.1.8 and Table
5-3). The exposure conditions from these studies generally involve much higher O3
concentrations than those commonly occurring in areas of the U.S. where the current standard is
met. For example, the animal studies include 4-hour concentrations of 400 to 800 ppb (ISA,
Appendix 5, Tables 5-8 and 5-10). In addition, an epidemiologic study available in the last
review has reported positive associations of multiday average O3 concentrations in ambient air
with changes in two indicators of glucose and insulin homeostasis (ISA, Appendix 5, sections
5.1.3.1.1 and 5.1.8).
The ISA additionally concludes that the evidence is suggestive of, but not sufficient to
infer, a causal relationship between long-term O3 exposures and metabolic effects (ISA, section
IS.4.3.6.2). As with metabolic effects and short-term O3, the primary evidence is from
experimental animal studies in which the exposure concentrations are appreciably higher than
those commonly occurring in the U.S. For example, the animal studies include exposures over
several weeks to concentrations of 250 ppb and higher (ISA, Appendix 5, section 5.2.3.1.1). The
somewhat limited epidemiologic evidence related to long-term O3 concentrations and metabolic
effects includes several studies reporting increased odds of being overweight or obese or having
metabolic syndrome and increased hazard ratios for diabetes incidence with increased O3
concentrations (ISA, Appendix 5, sections 5.2.3.4.1, 5.2.5 and 5.2.9, Tables 5-12 and 5-15).
3-28
-------
With regard to cardiovascular effects and total (nonaccidental) mortality and short-term
O3 exposures, the conclusions regarding the potential for a causal relationship have changed
from what they were in the last review after integrating the previously available evidence with
newly available evidence. The relationships are now characterized as suggestive of, but not
sufficient to infer, a causal relationship (ISA, Appendix 4, section 4.1.17; Appendix 6, section
6.1.8). This reflects several aspects of the current evidence base: (1) a now-larger body of
controlled human exposure studies providing evidence that is not consistent with a
cardiovascular effect in response to short-term O3 exposure; (2) a paucity of epidemiologic
evidence indicating more severe cardiovascular morbidity endpoints,36 that would be expected if
the impaired vascular and cardiac function (observed in animal toxicological studies) was the
underlying basis for cardiovascular mortality (for which epidemiologic studies have reported
some positive associations with O3); and (3) the remaining uncertainties and limitations
recognized in the 2013 ISA (e.g., lack of control for potential confounding by copollutants in
epidemiologic studies) that still remain. Although there exists consistent or generally consistent
evidence for a limited number of 03-induced cardiovascular endpoints in animal toxicological
studies and cardiovascular mortality in epidemiologic studies, there is a general lack of
coherence between these results and findings in controlled human exposure and epidemiologic
studies of cardiovascular health outcomes (ISA, section IS. 1.3.1). Related to this updated
conclusion based on the current evidence for cardiovascular effects, the evidence for short-term
O3 and mortality is also updated (ISA, Appendix 6, section 6.1.8). While there remain consistent,
positive associations between short-term O3 and total (nonaccidental), respiratory, and
cardiovascular mortality (and there are some studies reporting associations to remain after
controlling for PM10 and NO2), the full evidence base does not describe a continuum of effects
that could lead to cardiovascular mortality.37 Therefore, because cardiovascular mortality is the
largest contributor to total mortality, the relatively limited biological plausibility and coherence
within and across disciplines for cardiovascular effects (including mortality) contributes to an
accompanying change in the causality determination for total mortality (ISA, section IS.4.3.5).
Thus, the currently available evidence for cardiovascular effects and total mortality is concluded
to be suggestive of, but not sufficient to infer, a causal relationship with short-term (as well as
long-term) O3 exposures (ISA, section IS. 1.3.1).
36 These include emergency department visits and hospital admission visits for cardiovascular endpoints including
myocardial infarctions, heart failure or stroke (ISA, Appendix 6, section 6.1.8).
37 Due to findings from controlled human exposure studies examining clinical endpoints (e.g., blood pressure) that
do not indicate an O3 effect and from epidemiologic studies examining cardiovascular-related hospital admissions
and ED visits that do not find positive associations, a continuum of effects that could lead to cardiovascular
mortality is not apparent (ISA, Appendices 4 and 6).
3-29
-------
For other health effect categories, conclusions in this review are largely unchanged from
those in the last review. The available evidence for reproductive effects, as well as for effects on
the nervous system, is also concluded to be suggestive of, but not sufficient to infer, a causal
relationship (as was the case in the last review) (ISA, section IS.4.3.6). Additionally, the
evidence is inadequate to determine if a causal relationship exists between O3 exposure and
cancer (ISA, section IS.4.3.6.6).
3.3.2 Public Health Implications and At-risk Populations
The public health implications of the evidence regarding Ch-related health effects, as for
other effects, are dependent on the type and severity of the effects, as well as the size of the
population affected. Such factors are discussed here in the context of our consideration of the
health effects evidence related to O3 in ambient air. Additionally, we summarize the currently
available information related to judgments or interpretative statements developed by public
health experts, including particularly experts in respiratory health. This section also summarizes
the current information on population groups at increased risk of the effects of O3 in ambient air.
With regard to O3 in ambient air, the potential public health impacts relate most
importantly to the role of O3 in eliciting respiratory effects, the category of effects that the ISA
concludes to be causally related to O3 exposure. Controlled human exposure studies have
documented reduced lung function, respiratory symptoms, increased airway responsiveness, and
inflammation, among other effects, in healthy adults exposed while at elevated ventilation, such
as while exercising. Such effects, if of sufficient severity and in individuals with compromised
respiratory function, such as individuals with asthma, are plausibly related to emergency
department visits and hospital admissions for asthma which have been associated with ambient
air concentrations of O3 in epidemiologic studies (as summarized in section 3.3.1 above; 2013
ISA, section 6.2.7; ISA, Appendix 3, sections 3.1.5.1 and 3.1.5.2).
The clinical significance of individual responses to O3 exposure depends on the health
status of the individual, the magnitude of the changes in pulmonary function, the severity of
respiratory symptoms, and the duration of the response. With regard to pulmonary function, the
greater impact of larger decrements on affected individuals can be described. For example,
moderate effects on pulmonary function, such as transient FEVi decrements smaller than 20% or
transient respiratory symptoms, such as cough or discomfort on exercise or deep breath, would
not be expected to interfere with normal activity for most healthy individuals, while larger
effects on pulmonary function (e.g., FEVi decrements of 20% or larger lasting longer than 24
hours) and/or more severe respiratory symptoms are more likely to interfere with normal activity
for more of such individuals (e.g., 2014 PA, p. 3-53; 2006 AQCD, Table 8-2).
3-30
-------
In addition to the difference in severity or magnitude of specific effects in healthy people,
it is also important to consider aspects of the same effect with regard to its impact on population
groups that may have a pre-existing disease status. For example, the same reduction in FEVi or
increase in inflammation or airway responsiveness in a healthy group and a group with asthma
may increase the risk of a more severe effect in the group with asthma, such as asthma
exacerbation triggered by an allergen to which they may be sensitized (Cox, 2020, Responses to
Charge Questions, pp. 7-8). Duration and frequency of documented effects is also reasonably
expected to influence potential adversity and interference with normal activity. In summary,
consideration of differences and also the relative transience or persistence of such FEVi changes
and respiratory symptoms, as well as pre-existing sensitivity to effects on the respiratory system,
and other factors, are important to characterizing implications for public health effects of an air
pollutant such as O3 (ATS, 2000; Thurston et al., 2017).
Decisions made in past reviews of the O3 primary standard and associated judgments
regarding adversity or health significance of measurable physiological responses to air pollutants
have been informed by guidance, criteria or interpretative statements developed within the public
health community, including the ATS, an organization of respiratory disease specialists, as well
as the CAS AC. The ATS released its initial statement (titled Guidelines as to What Constitutes
an Adverse Respiratory Health Effect, with Special Reference to Epidemiologic Studies of Air
Pollution) in 1985 and updated it in 2000 (ATS, 1985; ATS, 2000). The ATS described its 2000
statement, considered in the last review of the O3 standard, as being intended to "provide
guidance to policy makers and others who interpret the scientific evidence on the health effects
of air pollution for the purposes of risk management" (ATS, 2000). The ATS described the
statement as not offering "strict rules or numerical criteria," but rather proposing "principles to
be used in weighing the evidence and setting boundaries," and stated that "the placement of
dividing lines should be a societal judgment" (ATS, 2000). Similarly, the most recent policy
statement by the ATS, which once again broadens its discussion of effects, responses and
biomarkers to reflect the expansion of scientific research in these areas, reiterates that concept,
conveying that it does not offer "strict rules or numerical criteria, but rather proposes
considerations to be weighed in setting boundaries between adverse and nonadverse health
effects," providing a general framework for interpreting evidence that proposes a "set of
considerations that can be applied in forming judgments" for this context (Thurston et al., 2017).
With regard to pulmonary function decrements, the earlier ATS statement concluded that
"small transient changes in forced expiratory volume in 1 s[econd] (FEVi) alone were not
necessarily adverse in healthy individuals, but should be considered adverse when accompanied
by symptoms" (ATS, 2000). The more recent ATS statement continues to support this
conclusion and also gives weight to findings of such lung function changes in the absence of
3-31
-------
respiratory symptoms in individuals with pre-existing compromised function, such as that
resulting from asthma (Thurston et al., 2017). More specifically, the recent ATS statement
expresses the view that the occurrence of "small lung function changes" in individuals with pre-
existing compromised function, such as asthma, "should be considered adverse ... even without
accompanying respiratory symptoms" (Thurston et al., 2017). In keeping with the intent of these
statements to avoid specific criteria, neither statement provides more specific descriptions of
such responses, such as with regard to magnitude, duration or frequency, for consideration of
such conclusions. The earlier ATS statement, in addition to emphasizing clinically relevant
effects, also emphasized both the need to consider changes in "the risk profile of the exposed
population," and effects on the portion of the population that may have a diminished reserve that
puts its members at potentially increased risk if affected by another agent (ATS, 2000). These
concepts, including the consideration of the magnitude of effects occurring in just a subset of
study subjects, continue to be recognized as important in the more recent ATS statement
(Thurston et al., 2017) and continue to be relevant to the evidence base for O3.
• Does the current evidence alter our understanding of populations that are
particularly at risk from O3 exposures? What are important uncertainties in that
evidence?
The newly available information has not altered our understanding of human populations
at particular risk of health effects from O3 exposures (ISA, section IS.4.4). For example, as
recognized in prior reviews, people with asthma are the key population at risk of Ch-related
effects. The respiratory effects evidence, extending decades into the past and augmented by new
studies in this review, supports this conclusion (ISA, sections IS.4.3.1). For example, numerous
epidemiological studies document associations with O3 with asthma-related health outcomes
(e.g., emergency department visits and hospital admissions). Such studies indicate the
associations to be strongest for populations of children which is consistent with their generally
greater time outdoors while at elevated exertion. Together, these considerations indicate people
with asthma, including particularly children with asthma, to be at relatively greater risk of O3-
related effects than other members of the general population (ISA, sections IS.4.3.1 and IS.4.4.2,
Appendix 3).38
With respect to people with asthma, the limited evidence from controlled human
exposure studies (which are primarily in adult subjects) indicates similar magnitude of FEVi
decrements as in people without asthma (ISA, Appendix 3, section 3.1.5.4.1). Across other
respiratory effects of O3 (e.g., increased respiratory symptoms, increased airway responsiveness
38 Populations or lifestages can be at increased risk of an air pollutant-related health effect due to one or more of a
number of factors. These factors can be intrinsic, such as physiological factors that may influence the internal
dose or toxicity of a pollutant, or extrinsic, such as sociodemographic, or behavioral factors.
3-32
-------
and increased lung inflammation), the evidence has also found the observed responses to
generally not differ due to the presence of asthma, although the evidence base is more limited
with regard to study subjects with asthma (ISA, Appendix 3, section 3.1.5.7). Further, the
features of asthma contribute to a risk of asthma-related responses, such as asthma exacerbation
in response to asthma triggers, which may increase the risk of more severe health outcomes
(ISA, section 3.1.5). For example, a particularly strong and consistent component of the
epidemiologic evidence is the appreciable number of epidemiologic studies that demonstrate
associations between ambient O3 concentrations and hospital admissions and emergency
department visits for asthma (ISA, section IS.4.4.3.1).39 We additionally recognize that in these
studies, the strongest associations (e.g., highest effect estimates) or associations more likely to be
statistically significant are those for childhood age groups, which are recognized in section 3.4 as
age groups most likely to spend time outdoors during afternoon periods (when O3 may be
highest) and at activity levels corresponding to those that have been associated with respiratory
effects in the human exposure studies (ISA, Appendix 3, sections 3.1.4.1 and 3.1.4.2).40 The
epidemiologic studies of hospital admissions and emergency department visits are augmented by
a large body of individual-level epidemiologic panel studies that demonstrated associations of
short-term ozone concentrations with respiratory symptoms in children with asthma. Additional
support comes from epidemiologic studies that observed ozone-associated increases in indicators
of airway inflammation and oxidative stress in children with asthma (ISA, section IS.4.3.1).
Together, this evidence continues to indicate the increased risk of population groups with asthma
(ISA, Appendix 3, section 3.1.5.7).
Children, and also outdoor adult workers, are at increased risk largely due to their
generally greater time spent outdoors while at elevated exertion rates (including in the summer
when O3 levels may be higher). This status makes them more likely to be exposed to O3 in
ambient air, under conditions contributing to increased dose due to greater air volumes taken into
the lungs (ISA, section IS.4.4.2; 2013 ISA, section 5.2.2.7). Thus, in light of the evidence
summarized in the prior paragraph, children and outdoor workers with asthma may be at
39 In addition to asthma exacerbation, the epidemiologic evidence also includes findings of positive associations of
increased O3 concentrations with hospital admissions or emergency department visits for COPD exacerbation and
other respiratory diseases (ISA, Appendix 3, sections 3.1.6.1.3 and 3.1.8).
40 There is limited data on activity patterns by health status. An analysis in the 2014 HREA indicated that asthma
status had little to no impact on the percent of people participating in outdoor activities during afternoon hours,
the amount of time spent, and whether they performed activities at elevated exertion levels (2014 HREA, section
5.4.1.5). Based on an updated evaluation of recent activity pattern data we found children, on average, spend
approximately 2xk hours of afternoon time outdoors, 80% of which is at a moderate or greater exertion level,
regardless of their asthma status (see Appendix 3D, section 3D.2.5.3). Adults spend approximately 2xk hours of
afternoon time outdoors regardless of their asthma status but the percent of afternoon time at moderate or greater
exertion levels for adults (about 55%) is lower than that observed for children.
3-33
-------
increased risk of more severe outcomes, such as asthma exacerbation. Further, with regard to
children, there is experimental evidence from early life exposures of nonhuman primates that
indicates the potential for effects in childhood (through adolescence) when human respiratory
systems are under development (ISA, sections IS.4.4.2 and IS.4.4.4.1). As noted in the ISA,
"these experimental studies indicate that early-life ozone exposure can cause structural and
functional changes that could potentially contribute to airway obstruction and increased airway
responsiveness" (ISA, p. IS-52). Overall, the evidence available in the current review continues
to indicate the increased susceptibility of these population groups.
The ISA in the last review additionally identified older adults as being at greater risk of
03-related health effects. That identification, however, was based largely on studies of short-term
O3 exposure and mortality, which are part of the larger evidence base that is now concluded to be
suggestive, but not sufficient to infer a causal relationship (ISA, sections IS.4.3.5 and IS.4.4.4.2,
Appendix 4, sections 4.1.16.1 and 4.1.17).41 Other evidence available in the current review adds
little to the evidence available at the time of the last review for consideration of susceptibility of
older adults (ISA, section IS.4.4.2).
The ISA in the last review concluded that the information available at the time for low
socioeconomic status (SES) as a factor associated with the risk of 03-related health effects,
provided suggestive evidence of potentially increased risk (2013 ISA, section 8.3.3 and p. 8-37).
The 2013 ISA concluded that" [o]verall, evidence is suggestive of SES as a factor affecting risk
of 03-related health outcomes based on collective evidence from epidemiologic studies of
respiratory hospital admissions but inconsistency among epidemiologic studies of mortality and
reproductive outcomes," additionally stating that "[further studies are needed to confirm this
relationship, especially in populations within the U.S." (2013 ISA, p. 8-28). The evidence
available in the current review adds little to the evidence available at the time of the last review
in this area (ISA, section IS.4.4.2 and Table IS-10). The ISA in the last review additionally
identified a role for dietary anti-oxidants such as vitamins C and E in influencing risk of 03-
related effects, such as inflammation, as well as a role for genetic factors to also confer either an
increased or decreased risk (2013 ISA, sections 8.1 and 8.4.1). No newly available evidence has
been evaluated that would inform or change these prior conclusions (ISA, section IS.4.4 and
Table IS-10).
41 As noted in the ISA, "[t]he majority of evidence for older adults being at increased risk of health effects related to
ozone exposure comes from studies of short-term ozone exposure and mortality evaluated in the 2013 Ozone
ISA" (ISA, p. IS-52).
3-34
-------
• What does the available information indicate with regard to the size of at-risk
populations and their distribution in the U.S.?
The magnitude and characterization of a public health impact is dependent upon the size
and characteristics of the populations affected, as well as the type or severity of the effects. As
summarized above, a key population most at risk of health effects associated with O3 in ambient
air is people with asthma. The National Center for Health Statistics data for 2017 indicate that
approximately 7.9% of the U.S. populations has asthma (Table 3-1; CDC, 2019). This is one of
the principal populations that the primary O3 NAAQS is designed to protect (80 FR 65294,
October 26, 2015). Table 3-1 below considers the currently available information that helps to
characterize key features of this population.
The age group for which the prevalence documented by these data is greatest is children
aged five to 19, with 9.7% of children aged five to 14 and 9.4% of children aged 15-19 having
asthma. In 2012 (the most recent year for which such an evaluation is available), asthma was the
leading chronic illness affecting children (Bloom et al., 2013). The prevalence is greater for boys
than girls (for those less than 18 years of age). Among populations of different races or
ethnicities, black non-Hispanic children aged five to 14 have the highest prevalence, at 16.1%.
Asthma prevalence is also increased among populations in poverty. For example, 11.7% of
people living in households below the poverty level have asthma compared to 7.3%, on average,
of those living above it. Populations groups with relatively greater asthma prevalence might be
expected to have a relatively greater potential for 03-related health impacts.42
42 As summarized in section 3.1 above, the current standard was set to protect at-risk populations, which include
people with asthma. Accordingly, populations with asthma living in areas not meeting the standard would be
expected to be at increased risk of effects.
3-35
-------
Table 3-1. National prevalence of asthma, 2017.
Number with Current
Percent with Current
characteristic"
Asthma (in thousands)B
Asthma
Total
25,191
7.9
Child (Age <18)
6,182
8.4
Adult (Aqe 18+)
19,009
7.7
All Aqe Groups
0-4 years
869
4.4
5-14 years
4,010
9.7
15-19 years
2,020
9.4
20-24 years
1,498
7.3
25-34 years
3,311
7.6
35-64 years
10,036
8.1
65+ years
3,447
7.0
Child Aqe Group
0-4 years
869
4.4
5-11 years
2,548
8.8
12-17 years
2,765
11.1
12-14 years
1,462
11.8
15-17 years
1,304
10.4
Sex
Males
10,035
6.4
Boys (Age <18)
3,569
9.5
Boys (Age 5-14)
2,165
10.3
Men (Age 18+)
6,466
5.4
Females
15,156
9.3
Girls (Age <18)
2,613
7.3
Girls (Age 5-14)
1,845
9.1
Women (Aqe 18+)
12,544
9.8
Race/Ethnicity
White NHC
15,718
8.1
Child (Age <18)
2,918
7.7
Child (Age 5-14)
1,841
8.8
Adult (Age 18+)
12,800
8.1
Black NH
3,910
10.1
Child (Age <18)
1,231
12.6
Child (Age 5-14)
895
16.1
Adult (Age 18+)
2,679
9.2
Other NH
1,871
6.7
Child (Age <18)
617
8.2
Child (Age 5-14)
403
9.4
Adult (Age 18+)
1,254
6.1
Hispanic, all
3,692
6.4
Child (Age <18)
1,416
7.7
Child (Age 5-14)
871
8.4
Adult (Aqe 18+)
2,276
5.8
3-36
-------
Characteristic A
Number with Current
Asthma (in thousands)B
Percent with Current
Asthma
Hispanic, Puerto Rican
Child (Age<18)
Child (Age 5-14)
Adult (Age 18+)
Hispanic, Mexican/Mexican-American
Child (Age<18)
Child (Age 5-14)
Adult (Aqe 18+)
695
195*
93*
499
1,822
753
481
1,069
12.8
11.3
10.1*
13.5
5.1
6.2
7.0
4.5
Federal Poverty Threshold
Below 100% of poverty level
5,020
11.7
100% to less than 250% of poverty
level
6,769
7.9
250% to less than 450% of poverty
level
6,061
7.3
450% of poverty level or higher
7,342
6.8
A Numbers within selected characteristics may not sum to total due to rounding
B Includes persons who answered "yes" to the questions "Have you EVER been told by a doctor or other health
professional that you had asthma" and "Do you still have asthma?"
c NH = non-Hispanic
* Relative standard error of the estimate is 30% - 50%; the estimate is unreliable.
Adapted from 2017 National Health Interview Survey, Tables 3-1 (https://www.cdc.gov/asthma/nhis/2017/table3-
1.htm) and 4-1 (https://www.cdc.gov/asthma/nhis/2017/table4-l.htmj.
Children under the age of 18 account for 16.7% of the total U.S. population, with 6.2% of
the total population being children under 5 years of age (U.S. Census Bureau, 2019). Based on a
prior analysis of data from the Consolidated Human Activity Database (CHAD)43 in the 2014
HREA, children ages 4-18 years old were found to more frequently spend time outdoors
compared to other age groups (e.g., adults aged 19-34) spending more than 2 hours outdoors,
particularly during the afternoon and early evening (e.g., 12:00 p.m. through 8:00 p.m.) (2014
HREA, section 5G-1.2). These results were confirmed by additional analyses of CHAD data
reported in the ISA, noting greater participation in afternoon outdoor events for children ages 6-
19 years old during the warm season months compared to other times of the day (ISA, Appendix
2, section 2.4.1, Table 2-1). The 2014 HREA also found that children ages 4-18 years old spent
79% of their outdoor time at moderate or greater exertion (2014 HREA, section 5G-1.4). Further
analyses performed for this review using the most recent version of CHAD generated similar
results, as described in section 3.4.5 below (Appendix 3D, section 3D.2.5.3 and Figure 3D-9).
Each of these analyses indicate children participate more frequently and spend more afternoon
43 The CHAD provides time series data on human activities through a database system of collected human diaries, or
daily time location activity logs.
3-37
-------
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.
Given that afternoon time outdoors and elevated exertion were determined most important in
understanding the fraction of the population that might experience O3 exposures of concern
(2014 HREA, section 5.4.2), they may be at greater risk of effects due to increased exposure to
O3 in ambient air.
About one third of workers were required to perform outdoor work in 2018 (Bureau of
Labor Statistics, 2019). Jobs in construction and extraction occupations and protective service
occupations required more than 90% of workers to spend at least part of their workday outdoors
(Bureau of Labor Statistics, 2017). Other employment sectors, including installation,
maintenance and repair occupations and building and grounds cleaning and maintenance
operations, also had a high percentage of employees who spent part of their workday outdoors
(Bureau of Labor Statistics, 2017). These occupations often include physically demanding tasks
and involve increased ventilation rates which when combined with exposure to O3, may increase
the risk of health effects.
3.3.3 Exposure Concentrations Associated with Effects
As at the time of the last review, the EPA's conclusions regarding exposure
concentrations of O3 associated with respiratory effects reflect the extensive longstanding
evidence base of controlled human exposure studies of short-term O3 exposures of people with
and without asthma.44 These studies have documented an array of respiratory effects, including
reduced lung function, respiratory symptoms, increased airway responsiveness, and
inflammation, in study subjects following 1- to 8-hour exposures, primarily while exercising.
The severity of observed responses, the percentage of individuals responding, and strength of
statistical significance at the study group level have been found to increase with increasing
exposure (ISA; 2013 ISA; 2006 AQCD). Factors influencing exposure include activity level or
ventilation rate, exposure concentration, and exposure duration (ISA; 2013 ISA; 2006 AQCD).
For example, evidence from studies with similar duration and exercise aspects (6.6-hour duration
with six 50-minute exercise periods) demonstrates an exposure-response relationship for O3-
44 As recognized elsewhere, the studies are largely conducted with adult subjects.
3-38
-------
induced reduction in lung function (Figure 3-2) 45 46 This specific evidence was integral to the
Administrator's judgments and decision in the last review (80 FR 65292, October 26, 2015).
A Adams (2006) ,
X Adams (2003) ^
~ Adams (2002)
~ Harstman et al (1990)
* Kim et al. (2011). McDonnell et al. (2012)
O McDonnell etal. (1991)
O Schelegle et al. (2009)
~ Folinsbee etal. (1988)
¦ Folinsbeeetal, (1994)
A Adams (2000)
• Adams and Ollison (1997)
ov
+ F
-Al
Ov
f
B
F.V
F
Kv
~
F = Face mask
A Av V=Varying
30 40 50 60 70 80 90 100 110 120 130
Ozone (ppb)
Figure 3-2. Group mean Ch-induced reduction in FEV1 from controlled human exposure
studies of healthy adults exposed for 6.6 hours with quasi-continuous exercise.
FEV1 values plotted reflect group mean Os-induced percent change in FEV1, based
on subtraction of the group mean filtered air percent change (post-pre exposure)
from the group mean O3 percent change in FEV1 (adapted from Appendix 3A;
ISA, Appendix 3, Figure 3 1). Concentrations are the time-weighted averages of
target concentrations across full 6.6-hour period in chamber studies (or across the
six exposures in face mask studies).
45 For a subset of the studies included in Figure 3-2 (those with face mask rather than chamber exposures), there is
no O3 exposure during some of the 6.6-hr experiment (e.g., during the lunch break). Thus, while the exposure
concentration during the exercise periods is the same for the two types of studies, the time-weighted average
(TWA) concentration across the full 6.6-hr period differs slightly. For example, in the facemask studies of 120
ppb, the TWA across the full 6.6-hour experiment is 109 ppb (Appendix 3A, Table 3A-2).
46 The relationship also exists for size of FEVi decrement with alternative exposure or dose metrics, including total
inhaled O3 and intake volume averaged concentration.
3-39
-------
• Does the current evidence alter our conclusions from the previous review regarding
the exposure duration and concentrations associated with health effects? Does the
currently available scientific evidence indicate health effects attributable to
exposures to O3 concentrations lower than previously reported?
The current evidence, including that newly available in this review, does not alter our
conclusions from the last review on exposure duration and concentrations associated with O3-
related health effects. These conclusions were largely based on the body of evidence from the
controlled human exposure studies. A limited number of controlled human exposure studies are
newly available in the current review, with none involving lower exposure concentrations than
those previously studied (e.g., Figure 3-2) or finding effects not previously reported (ISA,
Appendix 3, section 3.1.4).47
The extensive evidence base for O3 health effects, compiled over several decades,
continues to indicate respiratory responses to short-term exposures as the most sensitive effects
of O3. As summarized in section 3.3.1.1 above, an array of respiratory effects is well documented
in controlled human exposure studies of subjects exposed for 1 to 8 hours, primarily while
exercising. The risk of more severe health outcomes associated with such effects is increased in
people with asthma as illustrated by the epidemiological findings of positive associations
between O3 exposure and asthma-related ED visits and hospital admissions.
The magnitude of respiratory response (e.g., size of lung function reductions and
magnitude of symptom scores) documented in the controlled human exposure studies is
influenced by ventilation rate, exposure duration, and exposure concentration. When performing
physical activities requiring elevated exertion, ventilation rate is increased, leading to greater
potential for health effects due to an increased internal dose (2013 ISA, section 6.2.1.1, pp. 6-5 to
6-11). Accordingly, the exposure concentrations eliciting a given level of response after a given
exposure duration is lower for subjects exposed while at elevated ventilation, such as while
exercising (2013 ISA, pp. 6-5 to 6-6). For example, in studies of generally healthy young adults
exposed while at rest for 2 hours, 500 ppb is the lowest concentration eliciting a statistically
significant 03-induced group mean lung function decrement, while a 1- to 2-hour exposure to
120 ppb produces a statistically significant response in lung function when the ventilation rate of
the group of study subjects is sufficiently increased with exercise (2013 ISA, pp. 6-5 to 6-6).
The exposure conditions (e.g., duration and exercise) given primary focus in the past
several reviews are those of the 6.6-hour study design, which involves six 50-minute exercise
periods during which subjects maintain a moderate level of exertion to achieve a ventilation rate
47 No 6.6-hour studies are newly available (ISA, Appendix 3, section 3.1.4.1.1). The newly available studies are
generally for exposures of three hours or less, and in nearly all instances involve exposure (while at elevated
exertion) to concentrations above 100 ppb (ISA, Appendix 3, section 3.1.4).
3-40
-------
of approximately 20 L/min per m2 body surface area while exercising. The 6.6 hours of exposure
in these studies has generally occurred in an enclosed chamber and the study design includes
three hours in each of which is a 50-minute exercise period and a 10-minute rest period, followed
by a 35-minute lunch (rest) period, which is followed by three more hours of exercise and rest, as
before lunch.48 Most of these studies performed to date involve exposure maintained at a
constant (unchanging) concentration for the full duration, although a subset of studies have
concentrations that vary (generally in a stepwise manner) across the exposure period and are
selected so as to achieve a specific target concentration as the exposure average.49 No studies of
the 6.6-hour design are newly available in this review. The previously available studies of this
design document statistically significant Ch-induced reduction in lung function (FEVi) and
increased pulmonary inflammation in young healthy adults exposed to O3 concentrations as low
as 60 ppb. Statistically significant group mean changes in FEVi, also often accompanied by
statistically significant increases in respiratory symptoms, become more consistent across such
studies of exposures to higher O3 concentrations, such as 70 ppb and 80 ppb (Appendix 3A,
Table 3A-1). The lowest exposures concentration for which these studies document a statistically
significant increase in respiratoiy symptoms is somewhat above 70 ppb (Schelegle et al., 2009).50
In the 6.6-hour studies, the group means of Ch-induced51 FEVi reductions for exposure
concentrations below 80 ppb are at or below 6% (Figure 3-2, Table 3-2). For example, the group
means of Ch-induced FEVi decrements that have been found to be statistically significantly
different from the responses in filtered air are 6.1% for 70 ppb and 1.7% to 3.5% for 60 ppb
(Figure 3-2, Table 3-2). The group mean Ch-induced FEVi decrements generally increase with
increasing O3 exposures, reflecting increases in both the number of the individuals affected and
48 A few studies have involved exposures by facemask rather than in a chamber. To date, there is little research
differentiating between exposures conducted with a facemask and in a chamber since the pulmonary responses of
interest do not seem to be influenced by the exposure mechanism. However, similar responses have been seen in
studies using both exposure methods at higher O3 concentrations (Adams, 2002; Adams, 2003). In the facemask
designs, there is a short period of zero exposure, such that the total period of exposure is closer to 6 hours than 6.6
(Adams, 2000; Adams, 2002; Adams, 2003).
49 In these studies, the exposure concentration changes for each of the six hours in which there is exercise and the
concentration during the 35-minute lunch is the same as in the prior (third) hour with exercise. For example, in
the study by Adams (2006), the protocol for the 6.6-hour period is as follows: 60 minutes at 0.04 ppm, 60 minutes
at 0.07 ppm, 95 minutes at 0.09 ppm, 60 minutes at 0.07 ppm, 60 minutes at 0.05 ppm and 60 minutes at 0.04
ppm.
50 Measurements are reported in this study for each of the six 50-minute exercise periods, for which the mean is 72
ppb (Schelegle et al., 2009). Based on these data, the time-weighted average concentration across the full 6.6-
hour duration was 73 ppb (Schelegle et al., 2009). The study design includes a 35-minute lunch period following
the third exposure hour during which the exposure concentration remains the same as in the third hour.
51 Consistent with the ISA and 2013 ISA, the phrase "Ch-induced" decrement or reduction in lung function or FEVi
refers to the percent change from pre exposure measurement of the O3 exposure minus the percent change from
pre exposure measurement of the filtered air exposure (2013 ISA, p. 6-4).
3-41
-------
the magnitude of the FEVi reduction (Figure 3-2, Table 3-2). For example, following 6.6-hour
exposures to a lower concentration (40 ppb), for which decrements were not statistically
significant at the group mean level, none of 60 subjects across two separate studies experienced
an 03-induced FEVi reduction as large as 15% or more (Table 3-2; Appendix 3D, Table 3D-19).
Across the four experiments (with number of subjects ranging from 30 to 59 subjects) that have
reported results for 60 ppb target exposure, the number of subjects experiencing this magnitude
of FEVi reduction (at or above 15%) varied (zero of 30, one of 59, two of 31 and two of 30
exposed subjects). The response increases to three of 31 subjects for the study with a 70 ppb
target concentration (Appendix 3D, Table 3D-19; Schelegle et al., 2009). In addition to
illustrating the E-R relationship, these findings also illustrate the considerable variability in
magnitude of responses observed among study subjects (Table 3-2 and Figure 3-2; ISA,
Appendix 3, section 3.1.4.1.1; 2013 ISA, p. 6-13).
3-42
-------
Table 3-2. Summary of 6.6-hour controlled human exposure study-findings, healthy
adults.
Endpoint
O3 Target
Exposure
Concentration
Statistically
Significant
EffectB
03-lnduced Group
Mean ResponseB
Study
FEVi
Reduction
120 ppb
Yes
-10.3% to -15.9% c
See Appendix 3AD
100 ppb
Yes
-8.5% to -13.9% c
87 ppb
Yes
-12.2%
Scheleqle et al., 2009
80 ppb
Yes
-7.5%
Horstman et al., 1990
-7.7%
McDonnell et al., 1991
-6.5%
Adams, 2002
-6.2% to -5.5% c
Adams, 2003
-7.0% to -6.1% c
Adams, 2006b
-7.8%
Scheleqle et al., 2009
ND E
-3.5%
Kim et al., 2011 F
70 ppb
Yes
-6.1%
Scheleqle et al., 2009 1
60 ppb
Yes
G
-2.9%
-2.8%
Adams, 2006b; Brown et al., 2008
Yes
-1.7%
Kim et al., 2011
No
-3.5%
Scheleqle et al., 2009
40 ppb
No
-1.2%
Adams, 2002
No
-0.2%
Adams, 2006b
Increased
Respiratory
Symptoms
120 ppb
Yes
Increased symptom
scores
See Appendix 3A
100 ppb
Yes
87 ppb
Yes
80 ppb
Yes
70 ppb
Yes
60 ppb
No
40 ppb
No
Airway
Inflammation
80 ppb
Yes
Multiple indicatorsH
Devlin et al., 1991; Alexis et al., 2010
60 ppb
Yes
Increased neutrophils
Kim et al., 2011
Increased
Airway
Resistance and
Responsiveness
120 ppb
Yes
Increased
Horstman et al., 1990; Folinsbee et al.,
1994 (O3 induced sRaw not reported)
100 ppb
Yes
Horstman et al., 1990
80 ppb
Yes
Horstman et al., 1990
^ This is the average concentration across the six exercise periods as targeted
by authors. This differs from the time-weighted
average concentration for the full exposure periods (targeted or actual). For example, as shown in Appendix 3A, Table 3A-2, in
chamber studies implementing a varying concentration protocol with targets of 0.03, 0.07, 0.10, 0.15, 0.08 and 0.05 ppm, the
exercise period average concentration is 0.08 ppm while the time weighted average for the full exposure period (based on
targets) is 0.082 ppm due to the 0.6 hour lunchtime exposure between periods 3 and 4.
3 Statistical significance based on the O3 compared to filtered air response at the study group mean (rounded here to decimal).
3 Ranges reflect the minimum to maximum FEVi decrements across multiple exposure designs and studies. Study-specific
values and exposure details provided in Appendix 3A, Tables 3A-1 and 3A-2, respectively.
3 Citations for FEVi and respiratory symptoms findings for exposures above 80 ppb are in Appendix 3A and not repeated here.
= ND (not determined) indicates these data have not been subjected to statistical testing.
¦ The data for 30 subjects exposed to 80 ppb by Kim et al. (2011) are presented in Figure 5 of McDonnell et al. (2012).
3 Adams (2006b) reported FEVi data for 60 ppb exposure by both constant and varying concentration designs. Subsequent
analysis of the FEVi data from the former found the group mean O3 response to be statistically significant (p < 0.002) (Brown et
al., 2008; 2013 ISA, section 6.2.1.1). The varying-concentration design data were not analyzed by Brown et al., 2008.
H Increased numbers of bronchoalveolar neutrophils, permeability of respiratory tract epithelial lining, cell damage, production of
proinflammatory cytokines and prostaglandins (ISA, Appendix 3, section 3.1.4.4.1; 2013 ISA, section 6.2.3.1).
3-43
-------
For shorter exposure periods, ranging from one to two hours, higher exposure
concentrations, ranging from 80 ppb to 400 ppb, have been studied (Appendix 3A, Table 3A-3).
In these studies, some exposure protocols have included heavy intermittent or very heavy
continuous exercise, which results in 2-3 times greater ventilation rate than in the prolonged (6.6-
or 8-hour) exposure studies, which only incorporate moderate quasi-continuous exercise.52
Across these shorter-duration studies, the lowest exposure concentration for which statistically
significant respiratory effects were reported is 120 ppb, with the exposure combined with
continuous heavy exercise. As recognized above the increased ventilation rate associated with
increased exertion increases the amount of O3 entering the lung, where depending on dose and
the individual's susceptibility, it may cause respiratory effects (2013 ISA, section 6.2.1.1). Thus,
for exposures involving a lower exertion level, a comparable response would not be expected to
occur without a longer duration at this concentration (120 ppb), as is illustrated by the 6.6-hour
study results for this concentration (Appendix 3A, Table 3A-1).
We have also considered what may be indicated by the epidemiologic studies regarding
exposure concentrations associated with health effects, and particularly by such concentrations
that might occur in locations when the current standard is met. In so doing, however, we
recognize that these studies are generally focused on investigating the existence of a relationship
between O3 occurring in ambient air and specific health outcomes, and not on detailing the
specific exposure circumstances eliciting such effects. While the evidence base of epidemiologic
studies of associations between O3 and respiratory effects and health outcomes (e.g., asthma-
related hospital admission and emergency department visits), as a whole, provides strong support
for the conclusions of causality, as summarized in section 3.3.1 above,53 these studies generally
do not measure personal exposures of the study population or track individuals in the population
with a defined exposure to O3 alone. Notwithstanding this, we have considered the
epidemiologic studies identified in the ISA as to what they might indicate regarding O3 exposure
concentrations in this regard.
Among the epidemiologic studies finding a statistically significant positive relationship
of short- or long-term O3 concentrations with respiratory effects, there are no single-city studies
conducted in the U.S. in locations with ambient air O3 concentrations that would have met the
current standard for the entire duration of the study (ISA, Appendix 3, Tables 3-13, 3-14, 3-39,
3-41, 3-42 and Appendix 6, Tables 6-5 and 6-6; PA, Appendix 3B, Tables 3B-1). There are
52 A quasi-continuous exercise protocol is common to the prolonged exposure studies where study subjects complete
six 50-minute periods of exercise, each followed by 10-minute periods of rest (2013 ISA, section 6.2.1.1).
53 Combined with the coherent evidence from experimental studies, the epidemiologic studies "can support and
strengthen determinations of the causal nature of the relationship between health effects and exposure to ozone at
relevant ambient air concentrations" (ISA, p. ES-17).
3-44
-------
(among this large group of studies) two single city studies conducted in western Canada that
include locations for which the highest-monitor design values fell just below 70 ppb (PA,
Appendix 3B, Table 3B-1; Kousha and Rowe, 2014; Villeneuve et al., 2007). These studies did
not, however, include analysis of correlations with other co-occurring pollutants or of the
strength of the associations when accounting for effects of copollutants in copollutant models
(ISA, Tables 3-14 and 3-39). Thus, the studies pose significant limitations with regard to
informing conclusions regarding specific O3 exposure concentrations and elicitation of such
effects. There are also about a handful of multicity studies conducted in the U.S. or Canada54 in
which the O3 concentrations in a subset of the study locations and for a portion of the study
period appear to have met the current standard (Appendix 3B). Concentrations in other portions
of the study area or study period, however, do not meet the standard, or data were not available
in some cities for the earlier years of the study period when design values55 for other cities in the
study were well above 70 ppb. The extent to which reported associations with health outcomes in
the resident populations in these studies are influenced by the periods of higher concentrations
during times that did not meet the current standard is unknown. Additionally, with regard to
multicity studies, the reported associations were based on the combined dataset from all cities,
complicating interpretations regarding the contribution of concentrations in the small subset of
locations that would have met the current standard compared to that from the larger number of
locations that would have violated the standard (Appendix 3B, Table 3B-1 and Table 3B-2).56
Further, given that populations in such studies may have also experienced longer-term, variable
and uncharacterized exposure to O3 (as well as to other ambient air pollutants), "disentangling
the effects of short-term ozone exposure from those of long-term ozone exposure (and vice-
versa) is an inherent uncertainty in the evidence base" (ISA, p. IS-87 [section IS.6.1]). While
given the depth and breadth of the evidence base for O3 respiratory effects, such uncertainties do
not change our conclusions regarding the causal relationship between O3 and respiratory effects,
they affect the extent to which the two studies mentioned here (conducted in conditions that may
54 Consistent with the evaluation of the epidemiologic evidence of associations between short-term O3 exposure and
respiratory health effects in the ISA, we focus on those studies conducted in the U.S. and Canada, and most
particularly in the U.S., to provide a focus on study populations and air quality characteristics that are most
relevant to circumstances in the U.S. (ISA, Appendix 3, section 3.1.2).
55 As described in chapter 2, a design value is the metric used to describe air quality in a given area relative to the
level of the standard, taking the averaging time and form into account. For example, a design value of 70 ppb just
meets the current primary standard.
56 As recognized in the last review, "multicity studies do not provide a basis for considering the extent to which
reported O3 health effects associations are influenced by individual locations with ambient [air] O3 concentrations
low enough to meet the current O3 standard versus locations with O3 concentrations that violate this standard" (80
FR 64344, October 26, 2015).
3-45
-------
have met the current standard) can inform our conclusions regarding the potential for O3
concentrations allowed by the current standard to contribute to health effects.
We additionally considered the experimental animal evidence with regard to exposure
conditions associated with respiratory effects. As noted in section 3.3.1 above, however,
exposure concentrations in the animal studies were generally much greater than those examined
in the controlled human exposure studies (and accordingly higher than concentrations commonly
occurring in ambient air in areas of the U.S. where the current standard is met). This is also true
for the small number of early life studies in nonhuman primates recognized in section 3.3.1.1
above that reported O3 to contribute to allergic asthma-like effects in infant primates. The
exposures eliciting the effects in these studies included multiple 5-day periods with O3
concentrations of 500 ppb over 8-hours per day (ISA, Appendix 3, section 3.2.4.1.2).
With regard to short-term O3 and metabolic effects, the category of effects for which the
ISA concludes there to be a likely causal relationship with O3, the evidence base is comprised
primarily of experimental animal studies, as summarized in section 3.3.1.2 above (ISA,
Appendix 5, section 5.1). The exposure conditions from these studies, however, generally
involve much higher O3 concentrations than those examined in the controlled human exposure
studies for respiratory effects (and much higher than concentrations commonly occurring in
ambient air in areas of the U.S. where the current standard is met). For example, the animal
studies include 4-hour concentrations of 400 to 800 ppb (ISA, Appendix 5, Table 5-8).57 The two
epidemiologic studies reporting statistically significant positive associations of O3 with
metabolic effects (e.g., changes in glucose, insulin, metabolic clearance) are based in Taiwan and
South Korea, respectively.58 Given the potential for appreciable differences in air quality patterns
between Taiwan and South Korea and the U.S., as well as differences in other factors that might
affect exposure (e.g., activity patterns), those studies are of limited usefulness for informing our
understanding of exposure concentrations and conditions eliciting such effects in the U.S. (ISA,
Appendix 5, section 5.1).
57 The exposure concentration in the single controlled human exposure study of metabolic effects (e.g., 300 ppb) are
also well above those examined in the respiratory effect studies (ISA, Appendix 5, Table 5-7).
58 Of the five epidemiologic studies discussed in the ISA that investigate associations between short-term 03
exposure and metabolic effects, three are conducted in Asia or South America and two are conducted in the U.S.
The two U.S. studies report either a null or negative association of metabolic markers with O3 concentration (ISA,
Appendix 5, Tables 5-6 and 5-9). The South American study (focused on hospital admissions associated with
diabetes complications) reported positive associations with 24-hr average concentrations for some subgroups,
although no associations were statistically significant (ISA, Appendix 5, Table 5-9).
3-46
-------
3.3.4 Uncertainties in the Health Effects Evidence
• To what extent have previously identified uncertainties in the health effects evidence
been reduced or do important uncertainties remain?
We have not identified any new uncertainties in the evidence since the last review.
However, we continue to recognize important uncertainties that also existed in the last review.
This array of important areas of uncertainty related to the current health evidence, including that
newly available in this review, is summarized below.
Although the evidence clearly demonstrates that short-term O3 exposures cause
respiratory effects, as was the case in the last review, we continue to recognize uncertainties that
remain in several aspects of our understanding of these effects. Such uncertainties include those
associated with the severity and prevalence of responses to short (e.g., 6.6- to 8-hour) O3
exposures at and below 60 ppb and responses of some population groups not well represented in
the evidence base of controlled human exposure studies (e.g., children and people with asthma).
There are also uncertainties concerning the potential influence of exposure history and co-
exposure to other pollutants on the relationship between short-term O3 exposure and respiratory
effects. With regard to the full health effects evidence base, we also recognize as an important
uncertainty the extent to which O3 exposures are related to health effects other than respiratory
effects. The following discussion touches on each of these types of uncertainty.
The majority of the available studies have generally involved healthy young adult
subjects, although there are some studies involving subjects with asthma, and a limited number
of studies, generally of very short durations (i.e., less than four hours), involving adolescents and
adults older than 50. While there is evidence from short (6.6- to 8-hour) controlled exposure
studies of healthy adult subjects to concentrations as low as 40 ppb, the only controlled human
exposure study of such a duration (7.6 hours with quasi-continuous light exercise) conducted in
people with asthma was for an exposure concentration of 160 ppb (Appendix 3A, Table 3A-2).
Given a general lack of studies using subjects that have asthma, particularly those at exposure
concentrations likely to occur under conditions meeting the current standard, uncertainties
remain with regard to characterizing the response in people with asthma while at elevated
ventilation to lower exposure concentrations, e.g., below 80 ppb. The extent to which the
epidemiologic evidence, including that newly available, can inform this area of uncertainty also
may be limited.59 As discussed in section 3.3.2 above, given the effects of asthma on the
59 Associations of health effects with O3 that are reported in the epidemiologic analyses are based on air quality
concentration metrics used as surrogates for the actual pattern of O3 exposures experienced by study population
individuals over the period of a particular study. Therefore, the studies are limited in what they can convey
regarding the specific patterns of exposure circumstances (e.g., magnitude of concentrations over specific
duration and frequency) that might be eliciting reported health outcomes.
3-47
-------
respiratory system, exposures associated with significant respiratory responses in healthy people
may pose an increased risk of more severe responses, including asthma exacerbation, in people
with asthma. Such considerations remain areas of uncertainty in this review. Thus, uncertainty
remains with regard to the extent to which the controlled human exposure study evidence
describes the responses of the populations, such as children with asthma, that may be most at
risk of 03-related respiratory effects (e.g., through an increased likelihood of severe responses, or
greatest likelihood of response).
Other areas of uncertainty concerning the potential influence of O3 exposure history and
co-exposure to other pollutants on the relationship between short-term O3 exposures and
respiratory effects also remain from the last review. As in the epidemiologic evidence in the last
review, there is a limited number of studies that include copollutant analyses for a small set of
pollutants (e.g., PM or NO2). Recent studies with such analyses suggest that observed
associations between O3 concentrations and respiratory effects are independent of co-exposures
to correlated pollutants or aeroallergens (ISA, sections IS.4.3.1 and IS.6.1; Appendix 3, sections
3.1.10.1 and 3.1.10.2). Despite the increased prevalence of copollutant modeling in recent
epidemiologic studies, uncertainty still exists with regard to the independent effect of O3 given
the high correlations observed for some copollutants in some studies and the small fraction of all
atmospheric pollutants included in these analyses (ISA, section IS.4.3.1; Appendix 2, section
2.5). We also note that neither of the two epidemiologic studies of respiratory outcomes
conducting in Canadian areas that would have met the current standard included copollutant
modeling (as recognized in section 3.3.3 above).
Further, although there remains uncertainty in the evidence with regard to the potential
role of exposures to O3 in eliciting health effects other than respiratory effects, the evidence has
been strengthened since the last review with regard to metabolic effects. As noted in section
3.3.1.2 above, the ISA newly identifies metabolic effects as likely to be causally related to short-
term O3 exposures. The evidence supporting this relationship is limited and not without its own
uncertainties. For example, as noted in section 3.3.1.2 above, the conclusion is based primarily
on animal toxicological studies conducted at much higher O3 concentrations than those common
in ambient air in the U.S. A limited number of epidemiologic studies of short-term O3 exposure
and metabolic effects are available, many of which did not control for copollutants confounding;
just two studies, both in Asia, report significant positive associations with changes in markers of
glucose homeostasis (ISA, Appendix 5; sections 5.1.8 and 5.3). As noted in section 3.3.1.2
above, the ISA has also determined the evidence to be suggestive of, but not sufficient to infer, a
causal relationship between long-term O3 exposures and metabolic effects, and between O3
exposures and several other categories of health effects, including effects on the cardiovascular,
reproductive and nervous systems, and mortality (ISA, section IS.4.3). Additionally, the ISA
3-48
-------
finds the evidence to be inadequate to determine if a causal relationship exists with O3 and
cancer (ISA, section IS.4.3).
In summary, a variety of uncertainties from the last review remain, including those
related to the extent of effects at concentrations below those evaluated in controlled human
exposure studies, and the potential for more severe impacts in individuals with asthma, including
particularly children, and in other at-risk populations.
3.4 EXPOSURE AND RISK INFORMATION
Our consideration of the scientific evidence available in the current review, as at the time
of the last review (summarized in section 3.1 above), is informed by results from quantitative
analyses of estimated population exposure and consequent risk. Estimates from the exposure-
based analyses, particularly the comparison of daily maximum exposures to benchmark
concentrations, were most informative to the Administrator's decision in the last review (as
summarized in section 3.1.2 above). This largely reflected the EPA conclusion that "controlled
human exposure studies provide the most certain evidence indicating the occurrence of health
effects in humans following specific O3 exposures," and recognition that "effects reported in
controlled human exposure studies are due solely to O3 exposures, and interpretation of study
results is not complicated by the presence of co-occurring pollutants or pollutant mixtures (as is
the case in epidemiologic studies)" (80 FR 65343, October 26, 2015). In the last review, the
Administrator placed relatively less weight on the air quality epidemiologic-based risk estimates,
in recognition of an array of uncertainties, including, for example, those related to exposure
measurement error (80 FR 65346, October 26, 2015).60 Therefore, we have focused new
quantitative analyses in this review on exposure-based risk analyses. This reflects the emphasis
60 The 2015 decision notice recognized key uncertainties in utilizing the estimated air
concentrations and epidemiologic study relationships (often called epidemiologic-based risk
estimates) (80 FR 65316; 79 FR 75277-75279; 2014 HREA, sections 3.2.3.2 and 9.6). These
included the heterogeneity in effect estimates between locations, the potential for exposure
measurement errors, and uncertainty in the interpretation of the shape of concentration-response
functions at lower O3 concentrations, as well as uncertainties related to the public health
importance of increases in relatively low O3 concentrations following air quality adjustment.
Lower confidence was also placed in the results of the epidemiologic-based assessment of
respiratory mortality risks associated with long-term O3 exposures in consideration of several
factors, as noted in section 3.1 above. Importantly in this review, the causal determinations for
short-term O3 exposure with mortality in the current ISA differ from the 2013 ISA. The current
determinations for both short-term and long-term O3 exposure (as summarized in section 3.1
above) are that the evidence is "suggestive" but not sufficient to infer causal relationships for O3
with mortality (ISA, Table IS-1).
3-49
-------
given to these types of analyses and the characterization of their uncertainties in the last review,
along with the availability of new or updated information, models, and tools that address those
uncertainties (IRP, Appendix 5A).
In the sections below, we summarize the currently available exposure and risk
information for consideration in this review. In section 3.4.1, we summarize the conceptual
model for the assessment, as well as key aspects of the assessment design, including the study
areas, populations simulated, modeling tools, and exposure and risk metrics derived. Sections
3.4.2 and 3.4.3 summarize the assessment results. Key limitations and uncertainties associated
with the assessment estimates are identified in section 3.4.4. and potential public health
implications are discussed in section 3.4.5. An overarching consideration is whether the current
exposure and risk information alters overall conclusions reached in the previous review
regarding health risk associated with exposure to O3 in ambient air.
3.4.1 Conceptual Model and Assessment Approach
The long-standing evidence base for Ch-related health effects is comprised of a large
assemblage of controlled human exposure studies, laboratory animal research studies, and air
quality epidemiologic studies. Together, these health effect studies lead to the strongly supported
conclusion that O3 exposure causes respiratory effects (as summarized in section 3.3 above).
This conclusion is strongest with regard to short-term O3 exposures, for which the ISA and
science assessments in prior reviews have determined there to be a causal relationship. The ISA
additionally determines the relationship between long-term exposure and respiratory effects, as
well as between short-term exposures and metabolic effects to be likely causal, recognizing that
associated uncertainties remain in the evidence. Given the relatively greater strength of the
evidence and understanding of the relevant exposure conditions, as well as availability of
appropriate data and modeling tools, we focus the exposure and risk analysis in this review on
respiratory risks associated with short-term O3 exposures.
The controlled human exposure studies document the occurrence of an array of
respiratory effects in humans in a variety of short-term exposure circumstances. These studies, in
combination with the laboratory animal studies, inform our understanding of the mode of action
for 03-attributable effects, including those health outcomes associated with ambient air
concentrations in air quality epidemiologic studies (ISA, Appendix 3, section 3.1.3). Figure 3-3
below illustrates the conceptual model for Chin ambient air and respiratory effects, with a
particular focus on short-term exposures and including linkages with the risk metrics assessed in
the quantitative analyses performed for this review.
3-50
-------
o
p
o
UJ
u.
II.
Ul
*
09
S
x
2
a:
Emissions of 03 precursors to ambient air
O , in ambient air
Indoor Air *
03 originating
indoors
Children and adults (all and subgroups with asthma)
1
t
Inhalation while at moderate or greater exertion
1
f
Respiratory system
Lung function decrements (FEV., sRaw), inflammation, respiratory
symptoms, etc.
1
f
1
I
I
4
Exposures of Concern:
humbei & percent of
people expei lericing a
day with an exposure {at
elevated breathing rates 1
above benchmaiks
Lung Function Risk:
Numbei & pel cent of
people expenencmg a
day with an 0-induced
FEV, 1 eduction
C10% !'% 20-o)
Incidence of Respiratory
Health Outcomes:
{e.g., hospital emergency
department [ED] visits,
hospital admissions [HA],
mortality)
Figure 3-3. Conceptual model for exposure-based risk assessment. Solid lines indicate
processes explicitly modeled in the assessment. Dashed lines indicate relationships
that are not explicitly modeled.
Based on this conceptual model, as well as newly available information, an exposure-
based assessment was developed for this review. In this assessment, described in detail in
Appendix 3D, we have estimated O3 exposures and resulting risk for air quality conditions of
interest, most particularly air quality conditions that just meet the current primary O3 standard.
These analyses inform our understanding of the protection provided by the current primary
standard from effects that the health effects evidence indicates to be elicited in some portion of
exercising people exposed for several hours to elevated O3 concentrations.
The analysis approach employed is summarized in Figure 3-4 below and described in
detail in Appendices 3C and 3D. This approach incorporates the use of an array of models and
data to develop population exposure and risk estimates for a set of eight urban study areas.
Ambient air O3 concentrations were estimated in each study area using an approach that relies on
a combination of ambient air monitoring data, atmospheric photochemical modeling and
3-51
-------
statistical methods (described in detail in Appendix 3C). Population exposure and risk modeling
is employed to characterize exposures and related lung function risk associated with the ambient
air concentration estimates (described in detail in Appendix 3D). While the lung function risk
analysis focuses only on the specific Os effect of FEVi reduction, the comparison-to-benchmark
approach, with its use of multiple benchmark concentrations, provides for characterization of the
risk of other respiratory effects, the type and severity of which increase with increased exposure
concentration.
"ro
=3
a
Ambient Air Monitoring Data (hourly concentrations)
Hourly concentrations at monitoring sites
1 Adjustments 1
Photochemical
—
Air Quality
Modeling
CD
cr>
O
Q.
X
scenarios (just meeting the current standard and other design values)
Voronoi Neighbor Averaging (VNA) Interpolation I
Hourly concentrations at census tracts
Exposure Modeling (APEX)
(exposure concentrations and ventilation rate for each individual's exposure events)
cn
Population counts
of 7-hour daily
maximum 03
exposures at
elevated ventilation
Time series of 03
exposure events
(concentrations and
ventilation rates) for
each individual
Health-Based
Benchmark
Concentrations
Controlled Human
Exposure Data
(exposures involving
moderate or greater
exertion)
Population counts
of 7-hour daily
maximum 03
exposures at
elevated ventilation
MSS-FEV,
Lung Function
Risk Model
Exposure to Benchmark Comparison
Output: Number and percent of simulated at-risk
populations estimated to experience 1 or more days
with daily maximum exposures, at moderate or
greater exertion, at or above benchmark
concentrations (60 ppb, 70 ppb, 80 ppb)
Exposure-
Response
(E-R)
Function
1
Lung Function Risk
Output: Number and percent of simulated at-risk
populations estimated to experience 1 or more days
with specified 03-related lung function responses
(FEV, >10%, 15% and 20%)
Figure 3-4. Analysis approach for exposure-based risk analyses. Dashed lines and gray box
indicate the sole lung function risk approach used prior to 2014 HREA.
The analyses estimate exposure and risk for simulated populations in eight study areas in
Atlanta, Boston, Dallas, Detroit, Philadelphia, Phoenix, Sacramento and St. Louis. The eight
study areas represent a variety of circumstances with regard to population exposure to short-term
concentrations of Os in ambient air. The eight study areas range in total population size from
3-52
-------
approximately two to eight million and are distributed across the U.S. in seven different NOAA
climate regions: the Northeast, Southeast, Central, East North Central, South, Southwest and
West (Karl and Koss, 1984). Assessment of this set of study areas and the associated exposed
populations is intended to be informative to the EPA's consideration of potential exposures and
risks that may be associated with the air quality conditions that meet the current primary
standard.
This set of eight study areas represents a streamlined set as compared to the 15 study
areas in the last review but were chosen to ensure they reflect the full range of air quality and
exposure variation expected across major urban areas in the U.S. (2014 HREA, section 3.5).
Accordingly, while seven of the eight study areas were also included in the 2014 HREA, the
eighth study area is newly added in the current assessment to insure representation of a large city
in the southwest. Additionally, the years simulated reflect more recent emissions and
atmospheric conditions subsequent to data used in the 2014 HREA, and therefore represent O3
concentrations somewhat nearer the current standard than was the case for study areas included
in the HREA of the last review (Appendix 3C, Table 3C and 2014 HREA, Table 4-1). Thus, the
urban study areas (e.g., combined statistical areas that include urban and suburban populations)
for which the exposure and risk analyses have been conducted for this review reflect an array of
air quality, meteorological, and population exposure conditions.
Consistent with the health effects evidence in this review (summarized in section 3.3
above), the focus of the assessment is on short-term exposures of individuals in the population
during times when they are breathing at an elevated rate. Exposure and risk are characterized for
four population groups. Two are populations of school-aged children, aged 5 to 18 years:61 all
children and children with asthma. Two are populations of adults: all adults and adults with
asthma. Asthma prevalence estimates for the eight study areas ranges from 7.7 to 11.2%
(Appendix 3D, section 3D.3.1). For children, the study area asthma prevalence rates range from
9.2 to 12.3% (Appendix 3D, section 3D.3.1). Spatial variation within each study area related to
the population distribution of age, sex, and family income was also taken into account.62 For
children, this variation is greatest in the Detroit study area, with census tract level, age-specific
61 The child population group focuses on ages 5 to 18 in recognition of data limitations and uncertainties, including
those related to accurately simulating activities performed, estimating physiological attributes, as well as
challenges in asthma diagnoses for children younger than 5 years old.
62 As described in Appendix 3D, section 3D.2.2.2, asthma prevalence in each study area is estimated based on
combining regional national prevalence information from NHIS with U.S census tract level population data by
linking demographic information related to age, sex, and family income. Then, further adjustments were made
using state-level prevalence obtained from the U.S. Behavioral Risk Factor Surveillance System. See Appendix
3D, Attachment 1 for details.
3-53
-------
asthma prevalence estimates ranging from 6.4 to 13.2% for girls and from 7.7 to 25.5% for boys
(Appendix 3D, Table 3D-3).
Ambient air O3 concentrations were estimated in each study area for the air quality
conditions of interest by adjusting ambient air monitoring data using a photochemical model-
based approach and then applying a spatial interpolation technique to produce air quality
surfaces with high spatial and temporal resolution (Appendix 3C).63 The photochemical
modeling outputs included both modeled O3 concentrations and sensitivities of O3 concentrations
to changes in NOx emissions for each hour in a single year at all ambient air monitor locations
(Appendix 3C, sections 3C.4 and 3C.5). Linear regression was used with these single-year model
outputs to create relationships between the sensitivities and O3 concentrations for each hour of
each of the four seasons at each monitoring location. The relationships between hourly
sensitivities and hourly O3 for each season were then used with three years of ambient air
monitoring data at each location to predict hourly sensitivities for the complete 3-year record at
each monitoring location. From these, we calculated hourly O3 concentrations at each monitor
location based on iteratively increasing NOx reductions to determine the adjustments necessary
for the monitor location with the highest design value in each study area to just meet the target
value, e.g., 70 ppb for the current standard scenario (Appendix 3C, section 3C.5). Hourly O3
concentrations for all census tracts comprising each study area were then derived from the model
adjusted hourly concentrations at the ambient air monitor locations using the Voronoi Neighbor
Averaging (VNA) spatial interpolation technique (Appendix 3C, section 3C.6). The final product
was a dataset of ambient air O3 concentration estimates with high temporal and spatial resolution
(hourly concentrations in 500 to 1700 census tracts) for each of the eight study areas (Appendix
3C, section 3C.7).
The photochemical modeling approach involved use of the Comprehensive Air Quality
Model with Extensions (CAMx), version 6.5, instrumented with the higher order decoupled
direct method (HDDM) ,64 The CAMx-HDDM was run with emissions estimates and
meteorology data for calendar year 2016 to estimate the O3 sensitivities,65 and the linear
regressions of the modeled O3 concentrations to their respective sensitivities were applied to
hourly O3 concentrations reported at ambient air monitors for the 2015-2017 period to determine
the adjustments needed for each air quality scenario (Appendix 3C, sections 3C.4 and 3C.5). We
63 A similar approach was used to develop the air quality scenarios for the 2014 HREA.
64 Details on the models, methods and input data used to estimate ambient air concentrations for the eight study
areas are provided in Appendix 3C. The "higher order" aspect of the HDDM tool refers to the capability of
capturing nonlinear response curves (Appendix 3C, section 3C.5.1).
65 Sensitivities of O3 refer to predicted incremental changes in O3 concentrations in response to incremental changes
in precursor emissions (e.g., NOx emissions).
3-54
-------
maximized the spatial representation of the monitoring data by using all available monitors
within each study area (between 12 and 30) in addition to those within 50 km of the study area
boundaries (yielding between 5 and 31 additional monitors per area). Because we selected study
areas having design values close to the level of the current standard, the levels of NOx emissions
adjustments needed to meet the air quality scenarios of interest were generally lower than those
used in the 2014 HREA, thus reducing one of the important sources of uncertainty associated
with these air quality estimates.
Population exposures were estimated using the EPA's Air Pollutant Exposure model
(APEX) version 5, which probabilistically generates a large sample of hypothetical individuals
from 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 (Appendix 3D, section 3D.2).66 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 (Appendix 3D, section 3D.2). The APEX model generates
each simulated person or profile by probabilistically selecting values for a set of profile
variables, including demographic variables, health status and physical attributes (e.g., residence
with air conditioning, height, weight, body surface area) and ventilation rate (Appendix 3D,
section 3D.2).
The activity patterns of individuals are an important determinant of their exposure (2013
ISA, section 4.4.1). By incorporating individual activity patterns,67 the model estimates physical
exertion associated with each exposure event.68 This aspect of the exposure modeling is critical
66 The APEX model is a probabilistic model that estimates population exposure using a stochastic, event-based
microenvironmental approach. This model has a history of application, evaluation, and progressive model
development in estimating human exposure, dose, and risk for reviews of NAAQS for gaseous pollutants,
including the last review of the O3 NAAQS (U.S. EPA, 2008; U.S. EPA, 2009; U.S. EPA, 2010b; U.S. EPA,
2014; U.S. EPA, 2018).
67 To represent personal time-location-activity patterns of simulated individuals, the APEX model draws from the
CHAD developed and maintained by the EPA (McCurdy, 2000; U.S. EPA, 2019). The CHAD is comprised of
data from several surveys that collected activity pattern data at city, state, and national levels. Included are
personal attributes of survey participants (e.g., age, sex), the locations visited, and activities performed by survey
participants throughout a day, and the time-of-day activities occurred and their duration (Appendix 3D, section
3D.2.5.1).
68 An exposure event occurs when a simulated individual inhabits a microenvironment for a specified time, while
engaged at a constant exertion level and experiencing a particular pollutant concentration. If the
3-55
-------
in estimating exposure, ventilation rate, O3 intake (dose), and health risk resulting from ambient
air concentrations of O3.69 Because of variation in O3 concentrations among the different
microenvironments in which individuals are active, the amount of time spent in each location, as
well as the exertion level of the activity performed, will influence an individual's exposure to O3
from ambient air and potential for adverse health effects. Activity patterns vary both among and
within individuals, resulting in corresponding variations in exposure across a population and
over time (2013 ISA, section 4.4.1). For each exposure event, APEX tracks activity performed,
ventilation rate, exposure concentration, and duration. The time-series of exposure events serve
as the basis for calculating exposure and risk metrics of interest.
As in the last review, the quantitative analyses for this review uses the APEX model
estimates of population exposures for simulated individuals breathing at elevated rates70 to
characterize health risk based on information from the controlled human exposure studies on the
incidence of lung function decrements in study subjects who are exposed over multiple hours
while intermittently or quasi-continuously exercising (Appendix 3D, section 3D.2.8). In drawing
on this evidence base for this purpose, the assessment has given primary focus to the well-
documented controlled human exposure studies summarized in Appendix 3A, Table 3A-1 for
6.6-hour average exposure concentrations ranging from 40 ppb to 120 ppb (Figure 3-2; ISA,
Appendix 3, Figure 3-3). Health risk is characterized in two ways, producing two types of risk
metrics: one involving comparison of population exposures involving elevated exertion to
benchmark concentrations (that are specific to elevated exertion exposures), and the second
involving estimated population occurrences of ambient air 03-related lung function decrements
(Figure 3-2). The first risk metric is based on comparison of estimated daily maximum 7-hour
average exposure concentrations for individuals breathing at elevated rates to concentrations of
potential concern (benchmark concentrations). The second metric (lung function risk) uses E-R
information for O3 exposures and FEVi decrements to estimate the portion of the simulated at-
risk population expected to experience one or more days with an 03-related FEVi decrement of
at least 10%, 15% and 20%. Both of these metrics are used to characterize health risk associated
with O3 exposures among the simulated population during periods of elevated breathing rates.
Similar risk metrics were also derived in the HREA for the last review and the associated
microenvironmental concentration and/or activity/activity level changes, a new exposure event occurs (McCurdy
and Graham, 2003).
69 Indoor sources are generally minor in comparison to O3 from ambient air (ISA, Appendix 2, section 2.4.3) and are
not accounted for by the exposure modeling in this assessment.
70 Based on minute-by-minute activity levels, and physiological characteristics of the simulated person, APEX
estimates an equivalent ventilation rate (EVR), by normalizing the simulated individuals' activity-specific
ventilation rate to their body surface area (Appendix 3D, section 3D.2.2.3.3).
3-56
-------
estimates informed the Administrator's 2015 decision on the current standard (80 FR 65292,
October 26, 2015).
The general approach and methodology for the exposure-based assessment used in this
review is similar to that used in the last review. However, a number of updates and
improvements have been implemented in this review which result in differences from the
analyses in the prior review (Appendices 3C and 3D). These relate to the air quality, exposure,
and risk aspects of the assessment as summarized here.
• The ambient air monitoring data used is from a more recent period (e.g., 2015-2017)
during which O3 concentrations in the eight study areas are at or near the current standard
(Appendix 3C, Table 3C-1). This contrasts with the 2014 HREA use of 2006-2010 air
monitoring data, that for many study areas included design values (for unadjusted
concentrations) well above (e.g., by more than 10 ppb) the level of the then-existing
standard (2014 HREA, section 4.3.1.1, Table 4-1). The use of more recent ambient air
monitoring data in the current analysis allows for smaller adjustments to develop the air
quality conditions of interest, thus contributing to generally lesser uncertainty in the
concentrations estimated in each air quality scenario.
• The most recent CAMx model, with updates to the treatment of atmospheric chemistry
and physics within the model, is used to derive spatially and temporally varying
relationships between changes to emissions and modeled O3 concentrations, which are
then used in adjusting ambient air concentrations to just meet the air quality scenarios.
Model inputs represent recent year emissions, meteorology, and international transport
(e.g., 2016). The 2016-based inputs were derived using updated methods for calculating
emissions, as well as updated meteorological and hemispheric photochemical models
(described in more detail in Appendix 3C).
• Population exposure modeling inputs include the most recent U.S. Census demographics
and commuting data (i.e., 2010), meteorological data to reflect the assessment years
studied (e.g., 2015-2017), and updated estimates of asthma prevalence for all census
tracts in all study areas (e.g., 2013-2017). Regarding asthma prevalence, the more recent
information includes increased prevalence reported for adults and for children aged 10-17
years (Akinbami etal., 2016; CDC, 2016).71
• The APEX equations used to estimate of ventilation rate (Ve) and resting metabolic rate
have been updated such that the overall statistical model fit and predictability has been
improved (U.S. EPA, 2018, Appendix H).
• The approach for deriving population exposure estimates, both for comparison to
benchmark concentrations and for use in deriving lung function risk using the E-R
function, has been modified to provide for a better match of the simulated population
exposure estimates with the 6.6-hour duration of the controlled human exposure studies
and with the study subject ventilation rates (Appendix 3D, section 3D.2.8.1). The
modifications include deriving estimates for exposures of a duration and ventilation rate
71 For more information, see https://www.cdc.gov/nchs/products/databriefs/db239.htm.
3-57
-------
more closely corresponding to the duration and average ventilation rate across the 6.6-
hour duration in the controlled human exposure studies (Appendix 3D, section 3D.2.8.1).
72
• In addition to the E-R function, as updated in the 2014 HREA, an updated version of the
McDonnell Stewart Smith model (MSS-FEVt model, McDonnell et al., 2013) is used to
estimate individual-based lung function risk. Although the impact on risk estimates is
unclear, the updated MSS model has been described as better accounting for intra-subject
variability, yielding an improved model fit (McDonnell et al., 2013; Appendix 3D,
section 3D.2.8.2.2).
The exposure-to-benchmark comparison characterizes the extent to which individuals in
at-risk populations could experience O3 exposures, while engaging in their daily activities, with
the potential to elicit the effects reported in controlled human exposure studies for concentrations
at or above specific benchmark concentrations. Results are characterized using three benchmark
concentrations of O3: 60, 70, and 80 ppb. These are based on the three lowest concentrations
targeted in studies of 6- to 6.6-hour exposures, with quasi-continuous exercise, and that yielded
different occurrences and severity of respiratory effects (section 3.3.3 above; Appendix 3A,
section 3A.1; Appendix 3D, section 3D.2.8.1). The lowest benchmark, 60 ppb, represents the
lowest exposure concentration for which controlled human exposure studies have reported
statistically significant respiratory effects. At this concentration, there is evidence of a
statistically significant decrease in lung function and increase in airway inflammation (ISA,
Appendix 3, section 3.1.4.1.1; Brown et al., 2008; Adams, 2006b). Exposure to approximately 70
ppb73 averaged over a similar time resulted in a larger group mean lung function decrement, as
well as an increase in prevalence of respiratory symptoms over what was observed for 60 ppb
(Figure 3-3; ISA, Appendix 3, section 3.1.4.1.1; Schelegle et al., 2009). Studies of exposures to
approximately 80 ppb have reported larger lung function decrements at the study group mean
than following exposures to 60 or 70 ppb, in addition to an increase in airway inflammation,
increased respiratory symptoms, increased airway responsiveness, and decreased resistance to
other respiratory effects (Figure 3-3 and section 3.3.3, above; ISA, Appendix 3, sections 3.1.4.1-
72 Estimated exposures for a 7-hour duration are used in the comparison to benchmark concentrations (that are based
on the 6.6-hour exposure studies). The use of 7-hour exposure duration provides for a closer match of the duration
for the benchmark concentrations to the duration of population exposure concentration estimates than the 8-hour
exposure concentrations used in the last review. Additionally, an equivalent ventilation rate (EVR) of at least 17.3
L/min-m2 is used to more closely correspond to the average across the 6.6 hours of the controlled human
exposure studies (Appendix 3D, section 3D.2.8.1).
73 The design for the study on which the 70 ppb benchmark concentration is based, Schelegle et al. (2009), involved
varying concentrations across the full exposure period. The study reported the average O3 concentration measured
during each of the six exercise periods. The mean concentration across these six values is 72 ppb. The 6.6-hr time
weighted average based on the six reported measurements and the study design is 73 ppb (Schelegle et al., 2009).
Other 6.6-hr studies generally report an exposure concentration precision at or below 3 ppb (e.g., Adams, 2006b).
3-58
-------
3.1.4.4). The APEX-generated exposure concentrations for comparison to these benchmark
concentrations is the average of concentrations encountered by an individual while at an activity
level that elicits the specified elevated ventilation rate.74 The incidence of such exposures above
the benchmark concentrations are summarized for each simulated population, study area, and air
quality scenario as discussed in sections 3.4.2 and 3.4.3 below.
The lung function risk analysis provides estimates of the extent to which individuals in
the populations could experience decrements in lung function. Estimates were derived for risk of
experiencing a day with a lung function decrement at or above three different magnitudes, i.e.,
FEVi reductions of at least 10%, 15%, and 20%. Lung function decrement risk was estimated by
two different approaches, which utilize the evidence from the 6.6-hour controlled human
exposure studies in different ways.75 One, the population-based E-R function, uses quantitative
descriptions of the E-R relationships for study group incidence of the different lung function
decrements based on the individual study subject observations. The second, the individual-based
MSS model, uses quantitative estimations of biological processes identified as important in
eliciting the different sizes of decrements at the individual level, with a factor that also provides
a representation of intra- and inter-individual response variability (Appendix 3D, section
3D.2.8.2.2).
The E-R function used for estimating the risk of lung function decrements was developed
from the individual study subject measurements of 03-related FEVi decrements from the 6.6-
hour controlled human exposure studies targeting mean exposure concentrations from 120 ppb
down to 40 ppb (Appendix 3D, Table 3D-19; Appendix 3A, Figure 3A-1). The FEVi responses
reported in these studies have been summarized in terms of percent of study subjects
experiencing 03-related decrements equal to at least 10%, 15% or 20%. Across the exposure
range from 40 to 120 ppb, the percentage of exercising study subjects with asthma estimated to
have at least a 10% O3 related FEVi decrement increases from 0 to 7% (a statistically non-
74 The model averages the ventilation rate (Ve) for the exposed individual (based on the activities performed) over 7-
hour periods. This is done based on the APEX estimates of Ve and exposure concentration for every individual's
time-series of exposure events. For the exposure duration of interest (e.g., 7 hours), the model derives and outputs
the daily maximum average Ve (and hence an equivalent ventilation rate or EVR) and simultaneously occurring
exposure concentration for the specified duration for each simulated individual. To reasonably extrapolate the
ventilation rate of the controlled human study subjects (i.e., adults having a specified body size and related lung
capacity), who were engaging in quasi-continuous exercise during the study period, to individuals having varying
body sizes (e.g., children with smaller size and related lung capacity), an equivalent ventilation rate (EVR) was
calculated by normalizing the ventilation rate (L/min) by body surface area (m2). Seven-hour exposure
concentrations associated with 7-hour average EVR at or above the target of 17.3 + 1.2 L/min-m2 (i.e., the value
corresponding to average EVR across the 6.6-hour study duration in the controlled human exposure studies) are
compared to the benchmark concentrations (Appendix 3D, section 3D.2.8.1).
75 In so doing, the approaches also estimate responses associated with unstudied exposure circumstances and
population groups in different ways.
3-59
-------
significant response at exposures of 40 ppb) up to approximately 50 to 70% (at exposures of 120
ppb) (Appendix 3D, Section 3D.2.8.2.1, Table 3D-19). The E-R function relies on equations that
describe the fraction of the population experiencing a particular size decrement as a function of
the exposure concentration experienced while at the target ventilation rate. This type of risk
model, which has been used in risk assessments since the 1997 O3 NAAQS review, was updated
in the last review to include the more recently available study data (Appendix 3D, section
3D.2.8.2.1). In this review, the functions (fraction of the population having of a day or more per
simulation period with at least one decrement of one of the specified sizes) are applied to the
APEX estimates of 7-hour average exposure concentrations concomitant with the target
ventilation level estimated by APEX, with the results presented in terms of number of
individuals in the simulated populations (and percent of the population) estimated to experience
a day (or more) with a lung function decrement at or above 10%, 15% and 20%.
The MSS model, also used for estimating the risk of lung function decrements, was
developed using the extensive database from controlled human exposure studies that has been
compiled over the past several decades, and biological concepts based on that evidence
(McDonnell et al., 2012; McDonnell et al., 2013). The model mathematically estimates the
magnitude of FEV1 decrement as a function of inhaled O3 dose (based on concentration &
ventilation rate) over the time period of interest (Appendix 3D, section 3D.2.8.2.2). The
simulation of decrements is dynamic, based on a balance between predicted development of the
decrement in response to inhaled dose and predicted recovery (using a decay factor). Each
occurrence of decrements of interest (e.g., at or above 10%, 15% and 20%) is tallied. This model
was first applied in combination with the APEX model to generate lung function risk estimates
in the last review (80 FR 65314, October 26, 2015). As noted below, the MSS model used in the
current assessment has been updated since the previous review based on the most recent study by
its developers (McDonnell et al., 2013). In this review, the model is applied to the APEX
estimates of exposure concentration and ventilation for every exposure event experienced by a
simulated individual. The model then utilizes its mathematical descriptions of dose accumulation
and decay, and relationship of dose to response, to estimate the magnitude of O3 response
associated with the sequence of exposure events in each individual's day. We report the MSS
model risk results using the same metrics as for the E-R function, i.e., number of individuals in
the simulated populations (and percent of the population) estimated to experience a day (or
more) per simulation period with a lung function decrement at or above 10%, 15% and 20%.
The comparison-to-benchmark analysis (involving comparison of 7-hour average
exposure concentrations that coincide with a 7-hour average elevated ventilation rates) provides
perspective on the extent to which the air quality being assessed could be associated with
discrete exposures to O3 concentrations reported to result in respiratory effects. For example,
3-60
-------
estimates of such exposures can indicate the potential for Ch-related effects in the exposed
population, including effects for which we do not have E-R functions that could be used in
quantitative risk analyses (e.g., airway inflammation). The comparison-to-benchmark analysis
differs from the two lung function risk analyses which estimate the population incidence of one
or more days with specific lung function decrements of magnitudes of interest based on two
different uses of the health effects evidence.
3.4.2 Population Exposure and Risk Estimates for Air Quality Just Meeting the Current
Standard
In this section, we consider the exposure and risk estimates in the context of the
following questions.
• What are the nature and magnitude of O3 exposures and associated health risks for
air quality conditions just meeting the current standard? What portions of the
exposed populations are estimated to experience exposures of concern or lung
function decrements?
To address these questions, we consider the estimates provided by the exposure and risk
simulations for the eight urban study areas with air quality conditions adjusted to just meet the
current standard (Appendix 3D, sections 3D.3.2 through 3D.3.3). In considering these estimates
here and their associated limitations, uncertainties and implications in greater depth in sections
3.4.5 and 3.5 below, we particularly focus on the extent of protection provided by the standard
from O3 exposures of potential concern. As described in the prior section, the exposure and risk
analyses present two types of risk estimates for the 3-year simulation in each study area: (1) the
number and percent of simulated people experiencing exposures at or above the particular
benchmark concentrations of interest in a year, while breathing at elevated rates; and (2) the
number and percent of people estimated to experience at least one Ch-related lung function
decrement (specifically, FEVi reductions of a magnitude at or above 10%, 15% or 20%) in a
year and the number and percent of people estimated to experience multiple lung function
decrements associated with O3 exposures.
As an initial matter, we note that, as indicated by the use of an urban case study approach
(summarized in section 3.4.1 above), the exposure and risk analyses are not intended to provide a
comprehensive national assessment. Nor is the objective to present an exhaustive analysis of
exposure and risk in the areas that currently just meet the current standard and/or of exposure and
risk associated with air quality adjusted to just meet the current standard in areas that currently
do not meet the standard. Rather, the analyses are intended to provide assessments of an air
quality scenario just meeting the current standard for a diverse set of study areas and associated
exposed populations. The purpose is to assess, based on current tools and information, the
potential for exposures and risks beyond those indicated by the information available at the time
3-61
-------
the standard was established. Accordingly, capturing an appropriate diversity in study areas and
air quality conditions (that reflect the current standard scenario)76 is important to the role of the
exposure and risk analyses in informing the Administrator's conclusions on the public health
protection afforded by the current standard.
Of the two types of risk metrics derived in the exposure and risk analyses, we turn first to
the results for the benchmark-based risk metric with regard to the percent of the simulated
populations of all children and children with asthma estimated to experience at least one day per
year77 with a daily maximum 7-hour average exposure concentration at or above the different
benchmark concentrations while breathing at elevated rates under air quality conditions just
meeting the current standard (Table 3-3). Estimates for adults, in terms of percentages, are lower,
generally due to the lesser amount and frequency of time spent outdoors at elevated exertion
(Appendix 3D, section 3D.3.2). The exception to this is for outdoor workers, who due to the
requirements of their job spend more time outdoors. As information for this group, including
specific durations of time spent outdoors and activity data, is limited, the group was not
simulated in this assessment, although we note that a targeted analysis was performed in the
2014 HREA.78 Given the recognition of people with asthma as an at-risk population and the
relatively greater amount and frequency of time spent outdoors at elevated exertion of children,
we focus here on the estimates for children, including children with asthma.
76 A broad variety of spatial and temporal patterns of O3 concentrations can exist when ambient air concentrations
just meet the current standard. These patterns will vary due to many factors including the types, magnitude, and
timing of emissions in a study area, as well as local factors, such as meteorology and topography. We focused our
current assessment on specific study areas having ambient air concentrations close to conditions that reflect air
quality that just meets the current standard. Accordingly, assessment of these study areas is more informative to
evaluating the health protection provided by the current standard than would be an assessment that included areas
with much higher and much lower concentrations.
77 The three years of ambient air O3 concentrations analyzed in the exposure assessment analyses include
concentrations during the O3 seasons for that area. These seasons capture the times during the year when
concentrations are elevated (80 FR 65419-65420, October 26, 2015). While the duration of an O3 season for each
year may vary across the study areas, for the purposes of the exposure and risk analyses, the O3 season in each
study area is considered synonymous with a year.
78 Targeted analyses of outdoor workers in the 2014 HREA (single study area, single year) found an appreciably
greater portion of this population as compared to the full population of adults estimated to experience exposures
at or above benchmark concentration, and particularly to experience such exposures on multiple days (2014
HREA, section 5.4.3.2). The estimates for the outdoor worker population, for the single urban area and year
simulated, were also somewhat higher than those for the child population. For a number of reasons, including the
appreciable data limitations and associated uncertainties summarized in Table 3D-64 of Appendix 3D, outdoor
workers are not a population that has been explicitly simulated in the current analyses. It is expected that if an
approach similar to that used in the 2014 HREA were used for this assessment the distribution of exposures
(single day and multiday) would be similar to that estimated in the 2014 HREA (e.g., 2014 HREA, Figure 5 14),
although with slightly lower overall percentages (and based on the comparison of current estimates with estimates
from the 2014 HREA) (Appendix 3D, section 3D.3.2.4).
3-62
-------
Under air quality conditions just meeting the current standard, approximately 3% to
nearly 9% of each study area's simulated children with asthma, on average across the 3-year
period, are estimated to experience one or more days per year with a daily maximum 7-hour
average exposure at or above 60 ppb while breathing at elevated rates (Table 3-3). This range is
very similar for the populations of all children (Table 3-3). With regard to the 70 ppb benchmark,
the study areas' estimates for children with asthma are as high as 0.7 percent (0.6% for all
children), on average across the 3-year period, and range up to 1.0% in a single year (Table 3-3).
Less than 0.1% of any area's children with asthma, on average, were estimated to experience any
days per year with a daily maximum 7-hour average exposure at or above 80 ppb (Table 3-3).
Looking at estimates for multiple-day occurrences, we see that no children are estimated to
experience more than a single day with a daily maximum 7-hour average exposure at or above 80
ppb in any year simulated in any location (Table 3-3). For the 70 ppb benchmark, the estimate is
less than 0.1% of any area's children (on average across 3-year period), both those with asthma
and all children (Table 3-3, Figure 3-4). The estimates for the 60 ppb benchmark are slightly
higher, with up to 3% of children estimated to experience more than a single day with a daily
maximum 7-hour average exposure at or above 60 ppb, on average (and more than 4% in the
highest year across all eight study area locations) (Table 3-3).
These estimates are based on analyses that, while based on conceptually similar
approaches to those used in the 2014 HREA, reflect the updates and revisions to those
approaches that have been implemented since that time. Taking that into consideration, the
estimates for the 3-year period from the current assessment for air quality conditions simulated to
just meet the current standard are of a magnitude roughly similar, although slightly lower at the
upper end of the ranges, to the estimates for these same populations in the 2014 HREA. For
example, for air quality conditions just meeting the standard with a level of 70 ppb, the 2014
HREA estimated 0.1 to 1.2% of children to experience at least one day with exposure at or above
70 ppb, while at elevated ventilation (Section 3D.3.2.4, Table 3D-38). There are a number of
differences between the quantitative modeling and analyses performed in the current assessment
and the 2014 HREA that likely contribute to the small differences in estimates between the two
assessments (e.g., 2015-2017 vs. 2006-2010 distribution of ambient air concentrations, full
statistical distribution of ventilation rates vs. a 5th percentile point estimate, 7-hour vs. 8-hour
exposure durations).
3-63
-------
Table 3-3. Percent and number of simulated children and children with asthma
estimated to experience at least one or more days per year with a daily
maximum 7-hour average exposure at or above indicated concentration while
breathing at an elevated rate in areas just meeting the current standard.
Exposure
Concentration
(ppb)
One or more days
Two or more days
Four or more days
Average per
year
Highest in a
single year
Average
per year
Highest in a
single year
Average
per year
Highest in a
single year
Children with asthma - percent of simulated population A
>80
0B - <0.1 c
0.1
0
0
0
0
>70
0.2-0.7
1.0
<0.1
0.1
0
0
>60
3.3-8.8
11.2
0.6-3.2
4.9
<0.1 -0.8
1.3
-numl
b er of individuals A
>80
0-67
202
0
0
0
0
>70
93-1145
1616
3-39
118
0
0
>60
1517 - 8544
11776
282 - 2609
3977
23 - 637
1033
All children - percent of simulated population A
>80
0B - <0.1
0.1
0
0
0
0
>70
0.2-0.6
0.9
<0.1
0.1
0 - <0.1
<0.1
>60
3.2-8.2
10.6
0.6-2.9
4.3
<0.1 -0.7
1.1
-numl
b er of individuals A
>80
0-464
1211
0
0
0
0
>70
727 - 8305
11923
16-341
757
0-5
14
>60
14928 -
69794
96261
2601 -
24952
36643
158- 5997
9554
A Estimates for each study area were averaged across the 3-year assessment period. Ranges reflect the ranges of averages.
B A value of zero (0) means that there were no individuals estimated to have the selected exposure in any year.
c An entry of <0.1 is used to represent small, non-zero values that do not round upwards to 0.1 (i.e., <0.05).
In framing these same exposure estimates from the perspective of estimated protection
provided by the current standard, these results indicate that, in the single year with the highest
concentrations across the 3-year period, 99% of the population of children with asthma would
not be expected to experience such a day with an exposure at or above the 70 ppb benchmark;
99.9% would not be expected to experience such a day with exposure at or above the 80 ppb
benchmark. The estimates, on average across the 3-year period, indicate that over 99.9%, 99.3%
and 91.2% of the population of children with asthma would not be expected to experience a day
with a daily maximum 7-hour average exposure while at elevated ventilation that is at or above
80 ppb, 70 ppb and 60 ppb, respectively (Table 3-3 above). Further, with regard to multiple days,
more than approximately 97% of all children or children with asthma, on average across a 3-year
period, are estimated to be protected against multiple days of exposures at or above 60 ppb.
3-64
-------
These estimates are of a magnitude roughly consistent with the level of protection that was
described in establishing the now-current standard in 2015 (as summarized in section 3.1 above).
With regard to lung function risk, the estimates for all children and for children with
asthma are again roughly similar, with the higher end of the ranges for the eight study areas
being just slightly higher in some cases for the children with asthma (Table 3-4). The lung
function risk estimates from the MSS model are appreciably higher than those based on the E-R
function (full results in Appendix 3D, section 3D.3.3). This difference relates to the fact, noted in
section 3.4.1 above, that the two lung function risk approaches are based on different aspects of
the controlled human exposure study evidence and differ in how they extrapolate beyond the
exposure study conditions and observations. Accordingly, uncertainties associated with the two
modeling approaches also differ (as discussed in section 3.4.4 below). The E-R function risk
approach conforms more closely to the circumstances of the 6.6-hour controlled human exposure
studies, such that the 7-hour duration and moderate or greater exertion level are necessary for
nonzero risk. This approach does, however, use a continuous function which predicts responses
for exposure concentrations below those studied down to zero. As a result, exposures below
those studied in the controlled human exposures will result in a fraction of the population being
estimated by the E-R function to experience a lung function decrement (albeit to an increasingly
small degree with decreasing exposures). The MSS model, which has been developed based on a
conceptualization intended to reflect a broader set of controlled human exposure studies (e.g.,
including studies of exposures to higher concentrations for shorter durations), does not require a
7-hour duration for estimation of a response, and lung function decrements are estimated for
exertion below moderate or greater levels, as well as for exposure concentrations below those
studied (Appendix 3D, section 3D.3.4.2; 2014 HREA section 6.3.3). These differences in the
models, accordingly, result in differences in the extent to which they reflect the particular
conditions of the available controlled human exposure studies and the frequency and magnitude
of the measured responses.79
For example, the 6.6-hour controlled human exposure studies have reported
approximately 3% of subjects exposed to an average concentration of 60 ppb and 10% of
subjects exposed to 70 ppb to have at least a 15% FEVi decrement (Appendix 3D, Table 3D-20
and Figure 3D-11). Table 3-3 above shows that, at a maximum, approximately 11% and 1% of
children with asthma are estimated in a single year to have a day with daily maximum 7-hour
exposure at or above the 60 ppb and 70 ppb benchmarks, respectively, indicating that perhaps
10% (11% minus 1%) might be expected to have a day with an exposure at or above 60 ppb but
79 The two models, their bases in the evidence and associated limitations and uncertainties are discussed in detail in
Appendix 3D, sections 3D.2.8.2 and 3D.3.4.
3-65
-------
less than 70 ppb. If the simulated children had the same sensitivity as the controlled human
exposure study subjects, it might be expected that 0.3% (3% times 10%) of this group could have
a 15% (or larger) FEVi decrement resulting from concentrations at or above 60 ppb and less than
70 ppb and 0.1% (10% times 1%) of this group could have a 15% (or larger) decrement resulting
from concentrations at or above 70 ppb. Accordingly, this would yield an estimated lung
function risk for the simulated population of 0.4% for decrements of 15% or larger. This
contrasts with the estimates based on the E-R function, that are at most a 1% risk (Table 3-4),
and the MSS model estimates, that are at most an 8.7% risk (Table 3-4).
Table 3-4. Percent of simulated children and children with asthma estimated to
experience at least one or more days per year with a lung function decrement
at or above 10,15 or 20% while breathing at an elevated rate in areas just
meeting the current standard.
Lung Function
DecrementA
One or more days
Two or more days
Four or more days
Average
per year
Highest in a
single year
Average
per year
Highest in a
single year
Average per
year
Highest in a
single year
E-R Function
percent of simulatec
children with asthmaA
> 20%
0.2-0.3
0.4
0.1 -0.2
0.2
<0.1B - 0.1
0.1
>15%
O
O
I
LO
O
1.0
0.3-0.6
0.6
0.2-0.4
0.4
>10%
2.3-3.3
3.6
1.5-2.4
2.6
0.9-1.7
1.8
percent of all simulated childrenA
> 20%
0.2-0.3
0.4
0.1 -0.2
0.2
<0.1 -0.1
0.1
>15%
0.5-0.8
0.9
0.3-0.5
0.6
0.2-0.4
0.4
>10%
2.2-3.1
3.3
1.3-2.2
2.4
0.8-1.6
1.7
MSS Model
percent of simulatec
children with asthmaA
> 20%
1.8-3.5
3.9
0.8-2.1
2.5
0.3-1.1
1.3
>15%
4.5-8.2
8.7
2.2-4.9
5.3
1.1 -2.9
3.3
>10%
13.9-22
23.3
8.0-14.9
16
4.3-9.8
10.5
percent of all simulated childrenA
> 20%
1.7-3.1
3.6
0.8-1.7
2.0
0.3-0.9
1.1
>15%
4.1-7.1
7.8
2.1 -4.3
4.9
1.0-2.5
2.9
>10%
13.2-20.4
21.8
7.4-13.6
14.8
3.9-8.8
9.7
A Estimates for eac
across urban study
B An entry of <0.1 is
urban case stuc
area averages,
used to represe
y area were averaged across the 3-year assessment period. Ranges reflect the ranges
it small, non-zero values that do not round upwards to 0.1 (i.e., <0.05).
3.4.3 Population Exposure and Risk Estimates for Additional Air Quality Scenarios
In addition to estimating population exposure and risk for O3 concentrations simulated to
occur under air quality conditions when the current standard is just met, the exposure and risk
3-66
-------
analyses also estimated population exposure and risk in the eight study areas for two additional
air quality scenarios. In these scenarios, the air quality conditions were adjusted such that the
monitor location with the highest concentrations in each area had a design value just equal to
either 75 ppb or 65 ppb.
The results for the comparison-to-benchmarks analysis for these additional air quality
scenarios are summarized in Table 3-5 below for all three benchmark concentrations. The
estimates for these two additional scenarios differ markedly from the results for air quality just
meeting the current standard (summarized in Table 3-3 above). For simplicity, the summary of
the comparison discussed here focuses on the 70 ppb benchmark concentration, which falls just
below the time-weighted exposure concentration for which there was a statistically significant
lung function decrement and also a statistically significant increase in respiratory symptom score
in one of the controlled human exposure studies, as noted in section 3.3.3 (ISA, Appendix 3,
section 3.1.4.1.1; Schelegle et al., 2009). The pattern is similar for the other two benchmarks,
although in general, the differences from the results for the current standard (presented in section
3.4.2) are somewhat greater for the higher benchmark and slightly smaller for the lower
benchmark.
Under air quality conditions in the 75 ppb scenario, estimated percentages of children
with asthma expected to experience at least one day per year with exposures at or above the
benchmark concentrations are two or more times higher than the estimates discussed in section
3.4.2 above for air quality conditions just meeting the current standard. For example, the
minimum and maximum percentages, on average per year across the study areas, of children
with asthma estimated to experience one or more days with exposures at or above the 70 ppb
benchmark are five and three times, respectively, greater than the corresponding percentages for
conditions associated with the current standard (Table 3-3 and Table 3-5). The highest estimated
percentage in a single year for the 70 ppb benchmark is more than twice as high for the 75 ppb
scenario compared to conditions associated with the current standard. The corresponding
estimate for two or more days per year is even greater for the 75 ppb scenario versus the current
standard scenario (Table 3-3 and Table 3-5).
In contrast, under air quality conditions in the 65 ppb scenario, the estimated percentages
of children with asthma expected to experience at least one day per year with exposures above
the benchmark concentrations are at most one third the estimates discussed in section 3.4.2 above
for air quality conditions just meeting the current standard (Table 3-3 and Table 3-5). The
highest estimated percentage of children expected to experience two or more days a year at or
above the 70 ppb benchmark drops to zero for the 65 ppb scenario compared to <0.1% for air
quality conditions just meeting the current standard (Table 3-3 and Table 3-5).
3-67
-------
As with the estimates for air quality just meeting the current standard, and as expected
given the various exposure and risk analysis updates implemented, the estimates discussed here
for the additional air quality scenarios are also slightly different from the estimates for such
scenarios that were derived in the last review. However, the differences are not of such a
magnitude that the estimates for one air quality scenario in the current review are similar to
results for a different scenario in the last review. For example, while the current estimates for the
75 ppb air quality scenario are somewhat lower for some benchmarks than those for that scenario
in the last review, they are still higher than the estimates from the last review for the air quality
scenario just meeting the current standard.
3-68
-------
Table 3-5. Percent and number of simulated children and children with asthma
estimated to experience one or more days per year with a daily maximum 7-
hour average exposure at or above indicated concentration while breathing at
an elevated rate - additional air quality scenarios.
Exposure
One or more days
Two or more days
Four or more days
Concentration
(ppb)
Average per
year
Highest in a
single year
Average per
year
Highest in a
single year
Average per
year
Highest in a
single year
Air quality scenario for 75 ppb
Children with asthma
- percent of simulated population A
>80
<0.1B-0.3
0.6
0C - <0.1
<0.1
0
0
>70
1.1-2.1
3.9
0.1-0.4
0.8
0 - <0.1
0.1
>60
7.6-17.1
19.2
2.0-8.9
11.0
0.1 -3.3
4.4
- number of individuals A
>80
23-410
888
0-7
20
0
0
>70
502 - 2480
4544
36-316
637
0-33
99
>60
3538 - 14054
17673
1188 - 7232
8931
204 - 2708
3595
All child
ren
- percent of simulated population A
>80
<0.1 B - 0.3
0.6
0 c - <0.1
<0.1
0
0
>70
1.1-2.0
3.4
0.1-0.3
0.7
<0.1
<0.1
>60
6.6-15.7
17.9
1.7-8.0
9.9
0.1 -3.0
4.1
- number of individuals A
>80
129-3127
6658
0-54
121
0
0
>70
4915 -19794
34981
414 - 2750
5775
3-141
368
>60
34918 -133400
162894
11087 - 67747
83660
1813 - 25773
34902
Air quality scenario for 65 ppb
Children with asthma
- percent of simulated population A
>80
0 - <0.1
<0.1
0
0
0
0
>70
0-0.2
0.3
0
0
0
0
>60
0.5-2.5
4.3
<0.1-0.3
0.6
0 - <0.1
0.1
- number of individuals A
>80
0-23
68
0
0
0
0
>70
0-311
455
0
0
0
0
>60
212 - 3542
5165
13-386
709
0-14
42
All child
ren
- percent of simulated population A
>80
0 - <0.1
<0.1
0
0
0
0
>70
0-0.2
0.2
0 - <0.1
<0.1
0
0
>60
0.4-2.3
3.7
<0.1-0.3
0.5
0 - <0.1
<0.1
- number of individuals A
>80
0-38
114
0
0
0
0
>70
0 - 2495
3140
0-13
23
0
0
>60
1832 - 29486
39772
83 - 3681
7188
0-179
354
A Estimates for each study area were averaged across the 3-year assessment period. Ranges reflect the ranges of averages.
B An entry of <0.1 is used to represent small, non-zero values that do not round upwards to 0.1 (i.e., <0.05).
c A value of zero (0) means that there were no individuals estimated to have the selected exposure in any year.
3-69
-------
Lung function risk estimated for children and children with asthma in air quality
scenarios with design values just above and below the current standard are presented in detail in
Appendix 3D, section 3D.3.3. The patterns of the estimates are, as expected, higher for the 75
ppb air quality scenario and lower for the 65 ppb scenario. For each scenario, the differences in
risk estimates between the two models is similar to that which occurs with the risk estimates for
air quality just meeting the current standard (as discussed in section 3.4.2 above). These
estimates (for both lung function risk approaches) are less different from those for the current
standard air quality scenario than are differences noted above for the comparison-to-benchmarks
estimates. This is due to the greater influence on the risk results of exposures associated with the
low O3 concentrations that are less affected by air quality adjustments used to develop air
concentration surfaces for which the highest-concentration location has a design value just
meeting the different targets.
3.4.4 Key Uncertainties
In this section, we consider the uncertainties associated with the quantitative estimates of
exposure and risk, including those recognized by the characterization of uncertainty in Appendix
3D (section 3D.3.4). This characterization is based on an approach intended to identify and
compare the relative impact that important sources of uncertainty may have on the exposure and
risk estimates. The approach used has been applied in REAs for past NAAQS reviews for O3,
nitrogen oxides, carbon monoxide and SOx (U.S. EPA, 2008; U.S. EPA, 2010a; U.S. EPA, 2014;
U.S. EPA, 2018). The characterization of uncertainty for the current analyses utilized a largely
qualitative approach adapted from the World Health Organization (WHO) approach for
characterizing uncertainty in exposure assessment (WHO, 2008) accompanied by several
quantitative sensitivity analyses of key aspects of the assessment approach. This current
uncertainty characterization and the supporting quantitative sensitivity analyses described in
detail in Appendix 3D further build upon information generated from quantitative sensitivity
analyses analysis conducted for the 2014 HREA and a previously conducted quantitative
uncertainty analysis of the population-based exposure modeling performed for the O3 NAAQS
(Langstaff, 2007). The approach used varies from that of WHO (2008) in that the approach
placed a greater focus on evaluating the direction and the magnitude of the uncertainty (i.e.,
qualitatively rating how the source of uncertainty, in the presence of alternative improved
information or directly based on quantitative sensitivity analyses, may affect the estimated
exposures and health risk estimates).
The exposure and risk uncertainty characterization and quantitative sensitivity analyses,
presented in Appendix 3D, section 3D.3.4, involve consideration of the various types of inputs
and approaches that together result in the exposure and risk estimates for the eight study areas. In
3-70
-------
this way the limitations and uncertainties underlying these inputs and approaches and the extent
of their influence on the resultant exposure/risk estimates are considered. Consistent with the
WHO (2008) guidance, the overall impact of the uncertainty is scaled by considering the extent
or magnitude of the impact of the uncertainty as implied by the relationship between the source
of the uncertainty and the exposure and risk output. The characterization in Appendix 3D also
evaluated the direction of influence, indicating how the source of uncertainty was judged, or
found, to quantitatively affect the exposure and risk estimates, e.g., likely to over- or under-
estimate (Appendix 3D, section 3D.3.4.1).
• What are the important uncertainties associated with the exposure and risk
estimates?
Based on the uncertainty characterization and associated analyses in Appendix 3D and
consideration of associated policy implications, we recognize several areas of uncertainty as
particularly important in our consideration of the exposure and risk estimates, while also
recognizing several areas where new or updated information reduced uncertainties in the
exposure and risk estimates. In so doing, we note areas that pertain to estimates for both types of
risk metrics, as well as areas that pertain more to one type of estimate versus the other. We also
note differences in the uncertainties that pertain to each of the two approaches used for the lung
function risk metric.
An overarching and important area of uncertainty, which remains from the last review
and is important to our consideration of the exposure and risk analysis results, concerns the
extent to which the outcomes of the exposure and risk analysis represent important risks posed
by O3 present in ambient air under conditions meeting the current standard. As an initial matter,
we recognize that this analysis addresses an array of respiratory responses documented in the
controlled human exposure studies of 6.6-hour duration (e.g., lung function decrements,
respiratory symptoms, increased airway responsiveness and inflammation). The comparison-to-
benchmarks analysis is particularly focused on consideration of the potential for exposures that
pose a risk of experiencing such effects. We note, however, the lack of evidence from controlled
human exposure studies of 6.6-hour duration for people with asthma and the three benchmark
concentrations. As recognized in sections 3.3.1 and 3.3.4, the controlled human exposure study
evidence base for 6.6-hour studies of 60, 70 and 80 ppb does not include studies of people with
asthma or children. Although uncertainties remain, the limited evidence that informs our
understanding of risk to people with asthma indicates the potential for them to be at greater risk
relative to other population groups under similar exposure circumstances (e.g., of asthma
exacerbation), as summarized in section 3.3.4 above. Thus, the health effects documented in
controlled human exposure studies of healthy adults may contribute to more severe outcomes
when occurring in people with asthma. Such a conclusion is consistent with the epidemiological
3-71
-------
study findings of positive associations of O3 concentrations with asthma-related ED visits and
hospital admissions (and the higher effect estimates from these studies), as referenced in section
3.3.1 above and presented in detail in the ISA. Further, with regard to lung function decrements,
we note the lack of information on the factors contributing to increased susceptibility to O3-
induced lung function decrements among some people. Thus, there is uncertainty regarding the
interpretation of the exposure and risk estimates and the extent to which they represent the
populations at greatest risk of Ch-related respiratory effects.
Aspects of the analytical design that pertain to both exposure-based risk metrics include
the estimation of ambient air concentrations for the assessed scenarios, as well as the main
components of the exposure modeling. Key uncertainties identified include the ambient air
concentrations used in developing the ambient air quality data input to the exposure model, along
with the modeling approach used to adjust ambient air concentrations to meet the air quality
scenarios of interest and the method used to interpolate monitor concentrations to census tracts.
While we recognize the adjustment to conditions near, above, or just below the current standard
as an important area of uncertainty, the approach used has taken into account the currently
available information and selected study areas having design values near the level of the current
standard to minimize the size of the adjustment needed to meet a given air quality scenario,
along with the use of more recent data as inputs for the air quality modeling, such as more recent
O3 concentration data (2015-2017), meteorological data (2016) and emissions data (2016).
Further, we consider the number of ambient monitors sited in each of the eight study areas to
contribute to a reasonable representation of spatial and temporal variability in the eight study
areas for the air quality conditions simulated. Among other key areas, we additionally recognize
the uncertainty with regard to the simulation of study area populations (and at-risk populations)
and considering appropriate physical and personal attributes. As recognized in the 2014 HREA,
exposures could be underestimated for some population groups that are frequently and routinely
outdoors during the summer (e.g., outdoor workers, children). In addition, longitudinal activity
patterns do not exist for these and other important population groups (e.g., those having
respiratory conditions other than asthma), thus limiting the extent to which the exposure model
outputs reflect these groups that might routinely experience high exposure concentrations. We
recognize there are important uncertainties in the approach used to estimate energy expenditure
(i.e., metabolic equivalents of work or METs) which are ultimately used to estimate ventilation
rates. We consider the use of longer-term average MET distributions to derive short-term
estimates, along with extrapolating adult observations to children is reasonable, based on the
availability of relevant data and appropriate evaluations conducted to date. We note, however,
that the number of activities for which METs distributions are available has more than doubled
3-72
-------
since the last review and the added specificity and redevelopment of these distributions is
expected to more realistically estimate activity-specific energy expenditure.
With regard to the exposure and risk modeling aspects of the two risk metrics, we
recognize that there are some uncertainties that apply to the estimation of lung function risk (and
not related to the comparison-to-benchmarks analysis). Both of the lung function risk approaches
utilized in the risk analyses incorporate some degree of extrapolation beyond the exposure
circumstances that have been studied in the controlled human exposure studies. This is the case
in different ways and with differing impacts for the two approaches. One way in which both
approaches extrapolate beyond the exposure studies concerns estimates of lung function risk
derived for exposure concentrations below those represented in the evidence base. This is in
recognition of the potential for lung function decrements to be greater in unstudied at-risk
population groups than is evident from the available studies. In considering these risk estimates,
we recognize that the uncertainty in the response estimates likely increases with decreasing
exposure concentration below those evaluated in controlled exposure studies.
The two models differ in how they extrapolate beyond the exposure study conditions. In
recognition of the lack of data for some at risk groups and the potential for such groups, such as
children with asthma, to experience lung function decrements at lower exposures than healthy
adults, both models generate nonzero predictions for 7-hour concentrations below the 6.6-hour
concentrations investigated in the controlled human exposure studies. For example, the E-R
function risk approach generates nonzero predictions from the full range of potential nonzero
concentrations for 7-hour average durations for which the average exertion levels meets or
exceeds the target. The MSS model, which draws on evidence-based concepts of how human
physiological processes respond to O3, involves extrapolation beyond the controlled
experimental conditions, with regard to exposure concentration, as well as with regard to
exposure duration and ventilation rate (both magnitude and duration). The difference between
the two models in the extent of extrapolation beyond the studied exposure circumstances is
illustrated by differences in the percent of the risk estimates derived on days for which the
highest 7-hour average concentration is below the lowest 6.6-hour exposure concentration tested
(Table 3-6 and Table 3-7). For example, while 3 to 6% of the risk to children (based on single-
year estimates for three study areas) of experiencing at least one day with decrements greater
than 20% estimated by the E-R model is associated with exposure concentrations below 40 ppb
(the lowest exposure concentration studied, and at which no decrements of this severity occurred
in any study subjects), 25% to nearly 40% of MSS model estimates of decrements greater than
20% derive from exposures below 40 ppb (Table 3-6 and Table 3-7). Further, using ventilation
rates lower than those used for the E-R function risk approach (which are based on the controlled
human exposure study conditions) also contribute to relatively greater risks estimated by the
3-73
-------
MSS model. Limiting the MSS model results to estimates for individuals with at least the same
exertion level achieved by study subjects (>17.3 L/min-m2), reduces the risks of experiencing at
least one lung function decrement by an amount between 24 to 42% (Appendix 3D, Table 3D-
69).
The difference between the two models for risk contribution from low concentrations is
smaller for risk estimates for two or more days than the estimates for one or more days. This is
largely because the percent contribution to low-concentration risk for two or more decrement
days predicted by the E-R approach is, by design, greater than the corresponding contribution to
low-concentration risk for one or more days.80 This also occurs because the MSS model
estimates risk from a larger variety of exposure and ventilation conditions (Table 3-6 and Table
3-7). Further, many of the uncertainties previously identified as part of the 2014 HREA unique to
the MSS model remain as important uncertainties in the current assessment. For example, the
extrapolation of the MSS model age parameter down to age 5 (from the age range of 18- to 35-
year old study subjects to which the model was fit) is an important uncertainty given that
children are an at-risk population of particular interest in this assessment. Also, there is
uncertainty in estimating the frequency and magnitude of lung function decrements as a result of
the statistical form and parameters used for the MSS model inter- and intra-individual variability
terms. Each of these, among other newly identified MSS model uncertainties, are evaluated and
discussed in the current uncertainty characterization (Appendix 3D, section 3D.3.4). As a whole,
the differences between the two lung function risk approaches described above and the estimates
generated by these approaches indicate appreciably greater uncertainty associated with the MSS
model estimates than the E-R function estimates due to the significantly greater portion of
relatively low concentrations contributing to risk.
80 The E-R function approach uses the daily maximum exposure concentration for the simulated population. By
design, every individual would more than likely have a lower exposure on the second day than that experienced
on the first day, and so on for each progressive day throughout the simulation period. Therefore, if any risk is
estimated, the distribution of exposures would be shifted more so to lower concentrations for a greater proportion
of the population.
3-74
-------
Table 3-6. Percent of risk estimated for air quality just meeting the current standard in
three study areas using the E-R function approach on days where the daily
maximum 7-hour average concentration is below specified values.
Size of
Lung
Function
Decrement
Percent of child population at risk of decrement from specific 7-hour concentrations A
Percent of one-or-more-days risk
Percent of two-or-more-days risk
< 30ppb
< 40 ppb
< 50 ppb
< 60 ppb
< 30 ppb
< 40 ppb
< 50 ppb
< 60 ppb
> 20%
0.7-1%
3 - 6%
12-25%
39 - 70%
2 - 3%
7-12%
24 - 44%
67 - 93%
>15%
2 - 3%
6-11%
19-34%
48 - 78%
4 - 5%
12-18%
34 - 54%
75 - 95%
>10%
4 - 5%
11-16%
29 - 45%
61 - 86%
7 - 9%
18-25%
45 - 63%
83 - 97%
A The ranges presented are based on 1-year simulations in three study areas (Atlanta, Dallas, and St Louis); the values
presented here are rounded to whole numbers or at least one significant digit (full results are in Appendix 3D, section
3D.3.4.2, Table 3D-62).
Table 3-7. Percent of risk estimated for air quality just meeting the current standard in
three study areas using the MSS model approach on days where the daily
maximum 7-hour average concentration is below specified values.
Size of
Lung
Function
Decrement
Percent of child population at risk of decrement from specified 7-hour concentrations A
Percent of one-or-more-days risk
Percent of two-or-more-days risk
< 30 ppb
< 40 ppb
< 50 ppb
< 60 ppb
< 30 ppb
< 40 ppb
< 50 ppb
< 60 ppb
> 20%
5 - 9%
25 - 38%
63 - 78%
88 - 96%
5-10%
28 - 42%
66-81%
90 - 98%
>15%
11-18%
36-51%
72 - 84%
92 - 98%
11-19%
38 - 54%
74 - 87%
93 - 99%
>10%
25 - 32%
57 - 67%
84-91%
96 - 99%
26 - 33%
57 - 68%
84-91%
96 - 99%
A The ranges presented are based on 1-year simulations in three study areas (Atlanta, Dallas, and St Louis); the values
presented here are rounded to whole numbers or at least one significant digit (full results are in Appendix 3D, section
3D.3.4.2, Table 3D-63).
An additional area in which uncertainty has been reduced for the exposure estimates is
related to the approach to identifying when simulated individuals may be at moderate or greater
exertion. The approach used in the current review reduces the potential for overestimation of the
number of people achieving the associated ventilation rate, an important uncertainty identified in
the 2014 HREA. We also note that the exposure duration in the current review was a 7-hour
averaging time, which was selected to better represent the 6.6-hour exposures from the
controlled human exposure studies, compared to the 8-hour exposure durations used in the model
in the 2014 HREA and prior assessments.
In summary, among the multiple uncertainties and limitations in data and tools that affect
the quantitative estimates of exposure and risk and their interpretation in the context of
considering the current standard, we recognize several here as particularly important, noting that
some of these uncertainties are similar to those recognized in the last review. These include
3-75
-------
uncertainty related to estimation of the concentrations in ambient air for the current standard and
the additional air quality scenarios; lung function risk approaches that rely, to varying extents, on
extrapolating from controlled human exposure study conditions to lower exposure
concentrations, lower ventilation rates, and shorter durations; and, characterization of risk for
particular population groups that may be at greatest risk, particularly for people with asthma,
particularly children. We also recognize several areas in which uncertainty has been reduced by
new or updated information or methods, including more refined air quality modeling based on
selection of study areas with design values near the current standard and more recent model
inputs, as well as updates to several inputs to the exposure model including changes to the
exposure duration to better match those in the controlled human exposure studies and an
alternate approach to characterizing periods of activity while moderate or greater exertion for
simulated individuals.
3.4.5 Public Health Implications
In considering public health implications of the quantitative exposure and risk estimates
that may inform the Administrator's judgments in this area, this section discusses the information
pertaining to the following question.
• To what extent are the estimates of exposures and risks to at-risk populations
associated with air quality conditions just meeting the current standard reasonably
judged important from a public health perspective?
Several factors are important to the consideration of public health implications. These
include the magnitude or severity of the effects associated with the estimated exposures, as well
as their adversity at the individual and population scale. Other important considerations include
the size of the population estimated to experience such effects or to experience exposures
associated with such effects. Thus, the discussion here reflects consideration of the health
evidence, and exposure and risk estimates, as well as the consideration of potential public health
implications in previous NAAQS decisions and ATS policy statements (as also discussed in
section 3.3.2).
In considering the severity of responses associated with the exposure and risk estimates,
we take note of the health effects evidence for the different benchmark concentrations and
judgments made with regard to the severity of these effects in the last review. As in the last
review, we recognize the greater prevalence of more severe lung function decrements among
study subjects exposed to 80 ppb or higher concentrations compared to 60 or 70 ppb exposure
concentrations, as well as the prevalence of other effects such as respiratory symptoms; thus,
such exposures are appropriately considered to be associated with adverse respiratory effects
consistent with past and recent ATS position statements. At 70 ppb, statistically significant
3-76
-------
increases in lung function decrements (specifically reduced FEVi) and respiratory symptoms
have been reported, which has led to characterization of these exposure conditions as also being
associated with adverse responses, consistent with past ATS statements as summarized in section
3.1 above (e.g., 80 FR 65343, 65345, October 26, 2015). Studies of controlled human exposures
to the lowest benchmark concentration of 60 ppb have found small but statistically significant
03-related decrements in lung function and airway inflammation.
We additionally take note of the greater significance of estimates for multiple
occurrences of exposures at or above these benchmarks consistent with the evidence. This is
consistent with past O3 NAAQS reviews in which it was recognized, using the example of effects
such as inflammation, that while isolated occurrences can resolve entirely, repeated occurrences
from repeated exposure could potentially result in more severe effects (2013 ISA, section 6.2.3
and p. 6-76). The ascribing of greater significance to repeated occurrences of exposures of
potential concern is also consistent with public health judgments in NAAQS reviews for other
pollutants, such as SOx and carbon monoxide (84 FR 9900, March 18, 2019; 76 FR 54307,
August 31, 2011).
As in the last review, the exposure-based analyses include two types of metrics, one
involving comparison-to-benchmark concentrations corresponding to 6.6-hour exposure
concentrations to which exposures while at elevated ventilation have elicited lung function
decrements, and the second involving estimates of lung function risk with regard to such
decrements of magnitudes at or above 10%, 15% or 20%. Based on the currently available
evidence which is largely consistent with that available in the last review (as summarized in
section 3.3.1 above), the quantitative exposure and risk analyses results in which we have the
greatest confidence are estimates from the comparison-to-benchmarks analysis, as discussed in
section 3.4.4 above.
In light of the conclusions that people with asthma and children are at-risk populations
for 03-related health effects (summarized in section 3.3.2 above) and the exposure and risk
analysis findings of higher exposures and risks for children (in terms of percent of that
population), we have focused the discussion here on children, and specifically children with
asthma. We recognize that the exposure and risk estimates indicate that in some areas of the U.S.
where O3 concentrations just meet the current standard, on average across the 3-year period
simulated, less than 1%, and less than 0.1% of the simulated population of children with asthma
might be expected to experience a single day per year with a 7-hour exposure at or above 70 ppb
and 80 ppb, respectively, while breathing at an elevated rate. With regard to the lowest
benchmark considered (60 ppb), the corresponding percentage is less than approximately 9%,
with higher percentages in some individual years. The corresponding estimates for the air quality
scenario with higher O3 concentrations are notably higher (Table 3-5). For example, for the 75
3-77
-------
ppb air quality scenario, 1.1% to 2.1% of children with asthma, on average across the 3-year
design period, are estimated to experience at least one day with exposure concentrations at or
above 70 ppb, while at moderate or greater exertion, with as many as 3.9% in a single year
(Table 3-5). The estimates for the 65 ppb scenario are appreciably lower.
With regard to estimates of lung function decrements, we focus on the E-R model
estimates as having less associated uncertainty, as discussed in section 3.4.4 above. The exposure
and risk analysis estimates 0.2 to 0.3% of children with asthma, on average across the 3-year
design period to experience one or more days with a lung function decrement at or above 20%,
and 0.5 to 0.9 % to experience one or more days with a decrement at or above 15% (Table 3-4
above). In a single year, the highest estimate is 1.0% of this at-risk population expected to
experience one or more days with a decrement at or above 15%. The corresponding estimate for
two or more days is 0.6% (Table 3-4 above). As discussed in section 3.4.3 above, the estimates
for the 75 ppb air quality scenario are notably higher, while the estimates for the 65 ppb scenario
are notably lower (Table 3-5).
The size of the at-risk population (people with asthma, particularly children) in the U.S.
is substantial. As summarized in section 3.3.2, nearly 8% of the total U.S. population81 and 8.4%
of U.S. children have asthma. The asthma prevalence in U.S. child populations (younger than 18
years) of different races or ethnicities ranges from 6.2% for Hispanic, Mexican or Mexican-
American children to 12.6% for black non-Hispanic children (Table 3-1 above). This is well
reflected in the exposure and risk analysis study areas in which the asthma prevalence ranged
from 7.7% to 11.2% of the total populations and 9.2% to 12.3% of the children. In each study
area, the prevalence varies among census tracts, with the highest tract having a prevalence in
boys of 25.5% and a prevalence in girls of 17.1% (Appendix 3D, Table 3D-3).
The exposure and risk analyses inherently recognize that variability in human activity
patterns (where people go and what they do) is key to understanding the magnitude, duration,
pattern, and frequency of population exposures. For O3 in particular, the amount and frequency
of afternoon time outdoors at moderate or greater exertion is an important factor for
understanding the fraction of the population that might experience O3 exposures that have
elicited respiratory effects in controlled human exposure studies (2014 HREA, section 5.4.2). In
considering the available information regarding prevalence of behavior (time outdoors and
exertion levels) and daily temporal pattern of O3 concentrations, we take note of the findings of
evaluations of the data in the CHAD. Based on these evaluations of human activity pattern data,
81 The number of people in the US with asthma is estimated to be about 25 million. As shown in Table 3-1 the
estimated number of people with asthma was 25,191,000 in 2017. The updated estimate from the 2018 National
Health Interview Survey is 24,753,000 (CDC, 2020). For children (younger than 18 years), the 2017 estimate is
approximately 6,182,000 (Table 3-1), while the estimate for 2018 is slightly lower at 5,530,131.
3-78
-------
it appears that children and adults both, on average, spend about 2 hours of afternoon time
outdoors per day, but differ substantially in their participation in these events at elevated exertion
levels (rates of about 80% versus 60%, respectively) (2014 HREA, section 5.4.1.5), indicating
children are more likely to experience exposures that may be of concern. This is one basis for
their identification as an at-risk population for 03-related health effects. The human activity
pattern evaluations have also shown there is little to no difference in the amount or frequency of
afternoon time outdoors at moderate or greater exertion for people with asthma compared with
those who do not have asthma (2014 HREA, section 5.4.1.5). Further, recent CHAD analyses
indicate that while 46 - 73% of people do not spend any afternoon time outdoors at moderate or
greater exertion, a fraction of the population (i.e., between 5.5 - 6.8% of children) spend more
than 4 hours per day outdoors at moderate or greater exertion and may have greater potential to
experience exposure events of concern than adults (Appendix 3D, section 3D.2.5.3 and Figure
3D-9). It is this potential that contributes importance to consideration of the exposure and risk
estimates.
In considering the public health implications of the exposure and risk estimates across the
eight study areas, we note the purpose for the study areas is to illustrate exposure circumstances
that may occur in areas that just meet the current standard, and not to estimate exposure and risk
associated with conditions occurring in those specific locations today. To the extent that
concentrations in the specific areas simulated may differ from others across the U.S., the
exposure and risk estimates for these areas are informative to consideration of potential
exposures and risks in areas existing across the U.S. that have air quality and population
characteristics similar to the study areas assessed, and that have ambient concentrations of O3
that just meet the current standard today or that will be reduced to do so at some period in the
future. We note that numerous areas across the U.S. have air quality for O3 that is near or above
the existing standard.82 Thus, the air quality and exposure circumstances assessed in the eight
study areas are of particular importance in considering whether the currently available
information calls into question the adequacy of public health protection afforded by the current
standard.
The exposure and risk estimates for the study areas assessed for this review reflect
differences in exposure circumstances among those areas and illustrate the exposures and risks
that might be expected to occur in other areas with such circumstances under air quality
82 Based on the most recently available data from 2016-2018, 142 counties have 03 concentrations that exceed the
current standard. Population size in these counties ranges from approximately 20,000 to more than ten million,
with a total population of over 112 million living in counties that exceed the current standard. Air quality data are
from Table 4. Monitor Status in the Excel file labeled ozone_designvalues_20162018_final_06_28_19.xlsx
downloaded from https://www.epa.gov/air-trends/air-quality-design-values. Population sizes are based on 2017
estimates from the U.S. Census Bureau (https://www.census.gov/programs-surveys/popest.html).
3-79
-------
conditions that just meet the current standard (or the alternate conditions assessed). Thus, the
exposure and risk estimates indicate the magnitude of exposure and risk that might be expected
in many areas of the U.S. with O3 concentrations at or near the current standard. Although the
methodologies and data used to estimate population exposure and lung function risk in this
review differ in several ways from what was used in the last review, the findings and
considerations summarized here present a pattern of exposure and risk that is generally similar to
that considered in the last review (as described in section 3.4.2 above), and indicate a level of
protection generally consistent with that described in the 2015 decision.
In summary, the considerations raised here are important to conclusions regarding the
public health significance of the exposure and risk results. We recognize that such conclusions
also depend in part on public health policy judgments that will weigh in the Administrator's
decision in this review with regard to the adequacy of protection afforded by the current
standard. Such judgments that are common to NAAQS decisions include those related to public
health implications of effects of differing severity (75 FR 355260 and 35536, June 22, 2010; 76
FR 54308, August 31, 2011; 80 FR 65292, October 26, 2015). Such judgments also include those
concerning the public health significance of effects at exposures for which evidence is limited or
lacking, such as effects at the lower benchmark concentrations considered and lung function risk
estimates associated with exposure concentrations lower than those tested or for population
groups not included in the controlled exposure studies.
3.5 KEY CONSIDERATIONS REGARDING THE CURRENT PRIMARY
STANDARD
In considering what the currently available evidence and exposure/risk information
indicate with regard to the current primary O3 standard, the overarching question we consider is:
• Does the currently available scientific evidence- and exposure/risk-based information
support or call into question the adequacy of the protection afforded by the current
primary O3 standard?
To assist us in interpreting the currently available scientific evidence and the results of
recent quantitative exposure/risk analyses to address this question, we have focused on a series
of more specific questions, as detailed in sections 3.5.1 and 3.5.2 below. In considering the
scientific and technical information, we consider both the information available at the time of the
last review and information newly available in this review, which have both been critically
analyzed and characterized in the 2013 ISA for the last review and the ISA for the current
review, respectively. In so doing, a primary consideration is whether the information newly
available in this review alters our overall conclusions from the last review regarding health
effects associated with photochemical oxidants, including O3, in ambient air.
3-80
-------
3.5.1 Evidence-based Considerations
In considering the evidence with regard to the overarching question posed above
regarding the adequacy of the current standard, we address a series of more specific questions
that focus on policy-relevant aspects of the evidence. These questions begin with consideration
of the available evidence on health effects associated with exposure to photochemical oxidants,
and particularly O3.
• Is there newly available evidence that indicates the importance of photochemical
oxidants other than O3 with regard to abundance in ambient air, and potential for
human exposures and health effects?
No newly available evidence has been identified in this review regarding the importance
of photochemical oxidants other than O3 with regard to abundance in ambient air, and potential
for health effects.83 As summarized in section 2.1 above, O3 is one of a group of photochemical
oxidants formed by atmospheric photochemical reactions of hydrocarbons with nitrogen oxides
in the presence of sunlight, with O3 being the only photochemical oxidant other than nitrogen
dioxide that is routinely monitored in ambient air. Data for other photochemical oxidants are
generally derived from a few special field studies such that national scale data for these other
oxidants are scarce (ISA, Appendix 1, section 1.1; 2013 ISA, sections 3.1 and 3.6). Moreover,
few studies of the health impacts of other photochemical oxidants beyond O3 have been
identified by literature searches conducted for other recent O3 assessments (ISA, Appendix 1,
section 1.1). As stated in the ISA, "the primary literature evaluating the health.. .effects of
photochemical oxidants includes ozone almost exclusively as an indicator of photochemical
oxidants" (ISA, section IS. 1.1, p. IS-3). Thus, as was the case for previous reviews, the evidence
base for health effects of photochemical oxidants does not indicate an importance of any other
photochemical oxidants. For these reasons, discussion of photochemical oxidants in this
document focuses on O3.
• Does the currently available scientific evidence alter our conclusions from the last
review regarding the nature of health effects attributable to human exposure to O3
from ambient air?
The currently available evidence, including that newly available in this review, is largely
consistent with the conclusion reached in the last review regarding health effects causally related
to O3 exposures, specifically respiratory effects. Specifically, as in the last review, respiratory
effects are concluded to be causally related to short-term exposures to O3. Also, as in the last
83 Close agreement between past O3 measurements and the photochemical oxidant measurements upon which the
early photochemical oxidants NAAQS was based indicated the veiy minor contribution of other oxidant species
in comparison to O3 (DHEW, 1969).
3-81
-------
review, respiratory effects are concluded to be likely causally related to longer-term O3
exposures (ISA, section IS. 1.3.1, Appendix 3). Further, while a causal determination was not
made in the last review regarding metabolic effects, the ISA for this review finds there to be
sufficient evidence to concluded there to likely be a causal relationship of short-term O3
exposures and metabolic effects and finds the evidence to be suggestive of, but not sufficient to
infer, such a relationship between long-term O3 exposure and metabolic effects (ISA, section
IS. 1.3.1). This is based on evidence on these effects, largely from experimental animal studies,
that is newly available in this review (ISA, Appendix 5). Additionally, conclusions reached in
the current review differ with regard to cardiovascular effects and mortality, based on newly
available evidence in combination with uncertainties in the previously available evidence that
had been identified in the last review (ISA, Appendix 4, section 4.1.17 and Appendix 6, section
6.1.8). The current evidence base is concluded to be suggestive of, but not sufficient to infer,
causal relationships between O3 exposures (short- and long-term) and cardiovascular effects,
mortality, reproductive and developmental effects, and nervous system effects (ISA, section
IS. 1.3.1). As in the last review, the strongest evidence, including with regard to characterization
of relationships between O3 exposure and occurrence and magnitude of effects, is for respiratory
effects, and particularly for effects such as lung function decrements, respiratory symptoms,
airway responsiveness, and respiratory inflammation.
• Does the current evidence alter our understanding of populations that are
particularly at risk from O3 exposures?
The current evidence does not alter our understanding of populations at risk from health
effects of O3 exposures. As in the last review, people with asthma, and particularly children, are
the at-risk population groups for which the evidence is strongest. In addition to populations with
asthma, groups with relatively greater exposures, particularly those who spend more time
outdoors during times when ambient air concentrations of O3 are highest and while engaged in
activities that result in elevated ventilation, are recognized as at increased risk. Such groups
include outdoor workers and children. Other groups for which the recent evidence is less clear
include older adults, and recent evidence regarding individuals with reduced intake of certain
nutrients and individuals with certain genetic variants does not provide additional information for
these groups beyond the evidence available at the time of the last review (ISA, section IS.4.4).
• Does the current evidence alter our conclusions from the previous review regarding
the exposure duration and concentrations associated with health effects? To what
extent does the currently available scientific evidence indicate health effects
3-82
-------
attributable to exposures to O3 concentrations lower than previously reported and
what are important uncertainties in that evidence?
The currently available evidence regarding O3 exposures associated with health effects is
largely similar to that available at the time of the last review and does not indicate effects
attributable to exposures of shorter duration or lower concentrations than previously understood.
Respiratory effects continue to be the effects for which the experimental information regarding
exposure concentrations eliciting effects is well established, as summarized in section 3.3.3
above. Such information allows for characterization of potential population risk associated with
O3 in ambient air under conditions allowed by the current standard. The newly available
controlled human exposure studies, as discussed in section 3.3.3 above, are conducted over
shorter durations while at much higher concentrations than the key set of 6.6-hour studies that
have been the focus of the last several reviews. The respiratory effects evidence includes support
from a large number of epidemiologic studies. The positive associations of O3 with respiratory
health outcomes (e.g., asthma-related hospital admissions and emergency department visits)
reported in these studies are coherent with findings from the controlled human exposure and
experimental animal studies. All but a few of these studies, however, are conducted in areas
during periods when the current standard is not met, making them less useful with regard to
indication of effects of exposures allowed by the current standard.
Within the evidence base for the newly identified category of metabolic effects, the
evidence derives largely from experimental animal studies of exposures appreciably higher than
those for the 6.6-hour human exposure studies along with a small number of epidemiologic
studies. As discussed in section 3.3.3 above, these studies do not prove to be informative to our
consideration of exposure circumstances likely to elicit health effects.
Thus, the 6.6-hour controlled human exposure studies of respiratory effects remain the
focus for our consideration of exposure circumstances associated with O3 health effects. Based
on these studies, the exposure concentrations investigated range from as low as approximately 40
ppb to 120 ppb. This information on concentrations that have been found to elicit effects for 6.6-
hour exposures while exercising is unchanged from what was available in the last review. The
lowest concentration for which lung function decrements have been found to be statistically
significantly increased over responses to filtered air remains approximately 60 ppb, at which
group mean decrements on the order of 2% to 4% have been reported (Table 3-2 and Figure 3-2).
Respiratory symptoms were not increased with this exposure level.84 Exposure to concentrations
slightly above 70 ppb, with quasi-continuous exercise, has been reported to elicit statistically
84 A statistically significant increase in sputum neutrophils (a marker of increased airway inflammation) was
observed in one controlled human exposure study following 6.6-hour exposures to 60 ppb (Table 3-2 and Figure
3-2; Appendix 3A). An increase in respiratory symptoms has not been reported with this exposure level.
3-83
-------
significant increases in both lung function decrements and respiratory symptom scores, as
summarized in section 3.3.3 above. Still greater group mean and individual responses in lung
function decrements and respiratory symptom scores, as well as inflammatory response and
airway responsiveness, are reported for higher exposure concentrations.
• To what extent have previously identified uncertainties in the health effects evidence
reduced or do important uncertainties remain?
Uncertainties identified in the health effects evidence at the time of the last review
generally remain in the current evidence. These include uncertainties related to the susceptibility
of population groups not studied, the potential for effects to result from exposures to
concentrations below those included in controlled human exposure studies, and the potential for
increased susceptibility as a result of prior exposures. We additionally recognize uncertainties
associated with the epidemiologic studies (e.g., the potential for copollutant confounding and
exposure measurement error). In so doing, however, we note the appreciably greater strength in
the epidemiologic evidence in its support for determination of a causal relationship for
respiratory effects than that related to other categories, such as metabolic effects, newly
determined in this review to have a likely causal relationship with short-term O3 exposures (as
summarized in section 3.3.1 above).
3.5.2 Exposure/risk-based Considerations
Our consideration of the scientific evidence available in the current review, as at the time
of the last review, is informed by results from a quantitative analysis of estimated population
exposure and associated risk. The overarching consideration in this section is whether the current
exposure/risk information alters our overall conclusions from the previous review regarding
health risk associated with exposure to O3 in ambient air. As in our consideration of the evidence
in section 3.3.1 above, we have focused the discussion regarding the exposure/risk information
around key questions to assist us in considering the exposure/risk analyses of at-risk populations
living in a set of urban areas under air quality conditions simulated to just meet the existing
primary O3 standard. These questions are as follows.
• To what extent are the estimates of exposures and risks to at-risk populations
associated with air quality conditions just meeting the current standard reasonably
judged important from a public health perspective? What are the important
uncertainties associated with any exposure/risk estimates?
The exposure and risk analyses conducted for this review provide exposure and risk
estimates associated with air quality that might occur in an area under conditions that just meet
the current standard and, in so doing, they illustrate the differences likely to occur across various
locations with such air quality as a result of area-specific differences in emissions,
3-84
-------
meteorological and population characteristics. In understanding these results, we note that the
eight study areas provide a variety of circumstances with regard to population exposure to
concentrations of O3 in ambient air. These study areas reflect different combinations of different
types of sources of O3 precursor emissions, and also illustrate different patterns of exposure to O3
concentrations in a populated area in the U.S. (Appendix 3C, section 3C.2). In this way, the eight
areas provide a variety of examples of exposure patterns that can be informative to the EPA's
consideration of potential exposures and risks that may be associated with air quality conditions
occurring under the current O3 standard. While the same conceptual air quality scenario is
simulated in all eight study areas (i.e., conditions that just meet the existing standard), variability
in emissions patterns of O3 precursors, meteorological conditions, and population characteristics
in the study areas contribute to variability in the estimated magnitude of exposure and associated
risk across study areas.
In considering the exposure and risk results, we focus first on estimates for the eight
study areas from the comparison-to-benchmarks analysis, the results in which we have the
greatest confidence, as discussed in section 3.4.4 above. These results for urban areas with air
quality that just meets the current standard indicate that as many as 0.7% of children with
asthma, on average across the 3-year period, and up to 1.0% in a single year might be expected
to experience, while at elevated exertion, at least one day with a 7-hour average O3 exposure
concentration at or above 70 ppb (Table 3-3). As noted earlier, this benchmark concentration
reflects the finding of statistically significant 03-related decrements and increased respiratory
symptoms in a controlled human exposure study of individuals at elevated exertion. For the
benchmark concentration of 80 ppb (which reflects the potential for more severe effects), a much
lower percentage (0.1%) of children with asthma, on average across the 3-year period or in any
single year, might be expected to experience, while at elevated exertion, at least one day with
such a concentration (Table 3-3). By comparison, as many as 9% of children with asthma, on
average across the 3-year period, might be expected to experience one or more days with a 7-
hour average O3 exposure concentration at or above 60 ppb (the benchmark associated with less
severe effects), and just over 11% in a single year (Table 3-3). Regarding estimates for multiple
days, the percent of children with asthma estimated to experience two or more days with an
exposure at or above 70 ppb is less than 0.1%, on average across three years, and up to 0.1% in a
single year period (Table 3-3). There are no children with asthma estimated to experience more
than a single day per year with a 7-hour average O3 concentration at or above 80 ppb (Table 3-3).
With regard to the lowest benchmark concentration of 60 ppb, the percentages for more than a
single day occurrence are 3%, on average across the three years, and just below 5% in a single
year period (Table 3-3).
3-85
-------
The estimates for the additional air quality scenarios differ as would be expected. For the
75 ppb air quality scenario, the percent of children with asthma that might be expected to
experience at least one day with a 7-hour average O3 exposure concentration, while at elevated
exertion, at or above 70 ppb, is a factor of three or more higher than for the current standard
(Table 3-5). The corresponding estimates for multiple days are a factor of four or more higher
than those for air quality just meeting the current standard. By comparison, corresponding
estimates for the 65 ppb scenario are approximately a third those for the current standard
scenario, with a correspondingly smaller incremental difference in absolute number of children .
With regard to the 80 ppb benchmark, the difference of the 75 ppb scenario from the current
standard is a factor of three (for average across the 3-year period) to six (for the highest in a
single year). In contrast, the estimates for the 80 ppb benchmark (which is associated with the
more severe effects) in the 65 ppb air quality scenario are nearly identical to those for the current
standard.
With regard to the estimates of lung function risk, as an initial matter we note the
uncertainty associated with these estimates, as discussed in section 3.4.4 above. In so doing,
however, we recognize the lesser uncertainty associated with estimates derived using the E-R
function. Accordingly, we focus on those estimates here for air quality conditions just meeting
the current standard. The E-R lung function risk analysis for the eight study areas indicates that
the percent of children with asthma in an urban area that just meets the current standard that
might be expected to experience one or more days with a lung function decrement of at least
15% or 20% may be as high as 0.9% or 0.3%, respectively, on average across the three years,
and 1.0% or 0.4%, respectively, in a single-year period (Table 3-4). The estimates for a day with
a decrement of at least 10% may be as high as 3.3%, on average across the three years, and just
over 3.5% in a single-year period (Table 3-4). With regard to multiple day occurrences, the
percent of children with asthma that might be expected to experience two or more days with a
lung function decrement of at least 10% may be as high as 2.4%, on average across the three
years, and 2.6% in a single year (Table 3-4), with much smaller percentages for larger
decrements. For multiple days with a decrement of at least 15% or 20%, the percentages may be
as high as 0.6% or 0.2%, respectively, on average across the three years, and 0.6% or 0.2%,
respectively, in a single year period (Table 3-4).
We also consider the estimates from this assessment in light of the estimates from the
2014 HREA that were a focus of the decision on the standard in 2015. The estimates across all
study areas from this assessment are generally similar to those reported in the last review across
all study areas included in that HREA, particularly for the two or more occurrences and for the
3-86
-------
80 ppb benchmark (Table 3-8).85 In our consideration here, we focus on the full array of study
areas (e.g., rather than limiting to areas common to the two assessments) given the purpose of the
assessments in providing estimates across a range of study areas to inform decision making with
regard to the exposures and risks that may occur across the U.S. in areas that just meet the
current standard. In so doing, we note only slight differences observed for the lower benchmarks,
particularly in the estimates for the highest year. For example, consideration of the percentage of
children estimated to experience a day or more with an exposure at or above 70 ppb across the
three air quality conditions in the two assessments indicates that differences between air quality
scenarios in the current assessment remain appreciably larger than the slight differences in
estimates between the two assessments for a given scenario. The factors likely contributing to the
slight differences between the two assessments, such as for the lowest benchmark, include
greater variation in ambient air concentrations in some of the study areas in the 2014 HREA, as
well as the lesser air quality adjustments required in study areas for the current assessment due to
closer proximity of conditions to meeting the current standard (70 ppb).86 Other important
differences between the two assessments are the updates made to the ventilation rates used for
identifying when a simulated individual is at moderate or greater exertion and the use of 7 hours
for the exposure duration. Both of these changes were made to provide closer linkages to the
conditions of the controlled human exposure studies which are the basis for the benchmark
concentrations. Thus, we recognize there to be reduced uncertainty associated with the current
estimates. Overall, particularly in light of differences in the assessments, we conclude the current
estimates to be generally similar to those which were the focus in the 2015 decision on
establishing the current standard.
85 For consistency with the estimates highlighted in the 2015 review, Table 3-8 focuses on the simulated population
of all children (versus the simulated population for children with asthma that are focus in section 3.4).
86 The 2014 HREA air quality scenarios involved adjusting 2006-2010 ambient air concentrations, and some study
areas had design values in that time period that were well above the then-existing standard (and more so for the
current standard). Study areas included the current exposure analysis had 2015-2017 design values close to the
current standard, requiring less of an adjustment for the current standard (70 ppb) air quality scenario.
3-87
-------
Table 3-8. Comparison of current assessment and 2014 HREA (all study areas) for
percent of children estimated to experience at least one, or two, days with an
exposure at or above benchmarks while at moderate or greater exertion.
Air Quality
Scenario
(DV, ppb)
Estimated average % of simulated children
with at least one dav Der vear
at or above benchmark
(highest in single season)
Estimated average % of simulated children
with at least two davs Der vear
at or above benchmark
(highest in single season)
Current PA A
2014 HREA B
Current PA A
2014 HREA B
Benchmark Exposure Concentration of 80 ppb
75
<0.1 A -0.3 (0.6)
0-0.3(1.1)
0 - <0.1 (<0.1)
0(0.1)
70
0 - <0.1 (0.1)
0-0.1 (0.2)
0(0)
0(0)
65
0 - <0.1 (<0.1)
0(0)
0(0)
0(0)
Benchmark Exposure Concentration of 70 ppb
75
1.1-2.0 (3.4)
0.6-3.3 (8.1)
0.1-0.3 (0.7)
0.1-0.6 (2.2)
70
0.2-0.6 (0.9)
0.1 -1.2 (3.2)
<0.1 (0.1)
0-0.1 (0.4)
65
0-0.2(0.2)
0-0.2(0.5)
0 - <0.1 (<0.1)
0(0)
Benchmark Exposure Concentration of 60 ppb
75
6.6-15.7(17.9)
9.5-17.0 (25.8)
1.7-8.0 (9.9)
3.1 -7.6 (14.4)
70
3.2-8.2 (10.6)
3.3-10.2 (18.9)
0.6-2.9 (4.3)
0.5-3.5 (9.2)
65
0.4-2.3 (3.7)
0-4.2(9.5)
<0.1-0.3 (0.5)
0-0.8(2.8)
A For the current analysis, calculated percent is rounded to the nearest tenth decimal using conventional rounding. Values
equal to zero are designated by "0" (there are no individuals exposed at that level). Small, non-zero values that do not round
upwards to 0.1 (i.e., <0.05) are given a value of "<0.1"
B For the 2014 HREA. calculated percent was rounded to the nearest tenth decimal using conventional rounding. Values that
did not round upwards to 0.1 (i.e., <0.05) were given a value of "0".
3.5.3 CASAC Advice
In our consideration of the adequacy of the current primary O3 standard, in addition to the
evidence- and exposure and risk-based information discussed above, we have also considered the
advice and recommendations of the CASAC with regard to the adequacy of the current standard,
based on their review of the ISA and the earlier draft of this document, as well as comments
from the public on the earlier draft of this document.
A limited number of public comments have been received in this review to date,
including comments focused on the draft IRP or draft PA. Of the 11 commenters that addressed
adequacy of the current primary O3 standard, some expressed agreement with staff conclusions
in the draft PA, while others expressed the view that the standard should be more restrictive
based largely on advice from and considerations raised by the previous CASAC in the last
review regarding adequacy of the margin of safety.
3-88
-------
In the CASAC's review of the PA with regard to the primary standard, it agreed with the
draft PA that the evidence newly available in this review does not substantially differ from that
available in the 2015 review, stating that, " [t]he CASAC agrees that the evidence newly
available in this review that is relevant to setting the ozone standard does not substantially differ
from that of the 2015 Ozone NAAQS review" (Cox, 2020Cox, 2020, p. 12 of the Consensus
Responses). Further, some CASAC members "agree with the EPA that the available evidence
does not call into question the adequacy of protection provided by the current standard, and thus
support retaining the current primary standard" (Cox, 2020, p. 1 of letter). Other members
indicated their agreement with the previous CASAC's advice, based on review of the 2014 draft
PA, that a primary standard set at 70 ppb may not be protective of public health with an adequate
margin of safety (Cox, 2020, p. 1 of letter).87
The comments from the CASAC also included comments related to the quantitative
analyses summarized in section 3.4 above. These comments have been considered in completing
the analyses and associated characterizations in Appendices 3C and 3D. The CASAC
additionally took note of uncertainties that remain in this review of the primary standard.
Accordingly, it identified a number of additional areas for future research and data gathering that
would inform the next review of the primary O3 NAAQS (Cox, 2020, p. 14 of the Consensus
Responses). These are reflected in section 3.6 below.
3.5.4 Conclusions on the Primary Standard
This section describes our conclusions for the Administrator's consideration in this
review of the current primary O3 standard. These conclusions are based on considerations
described in the sections above, and in the discussion below regarding the currently available
scientific evidence (as summarized in the ISA, and the ISA and AQCDs from prior reviews), and
the risk and exposure information summarized above in chapter 3. Further, these conclusions
have taken into account advice from the CASAC and public comment on the draft PA.
Taking into consideration the discussions responding to specific questions above in this
and the prior chapter, this section addresses the following overarching policy question.
87 In the last review, the advice from the prior CASAC included a range of recommended levels for the standard,
with the CASAC concluding that "there is adequate scientific evidence to recommend a range of levels for a
revised primary ozone standard from 70 ppb to 60 ppb" (Frey, 2014, p. ii). In so doing, the prior CASAC noted
that" [i]n reaching its scientific judgment regarding a recommended range of levels for a revised ozone primary
standard, the CASAC focused on the scientific evidence that identifies the type and extent of adverse effects on
public health" and further acknowledged "that the choice of a level within the range recommended based on
scientific evidence is a policy judgment under the statutory mandate of the Clean Air Act" (Frey, 2014, p. ii). The
prior CASAC then described that its "policy advice [emphasis added] is to set the level of the standard lower than
70 ppb within a range down to 60 ppb, taking into account [the Administrator's] judgment regarding the desired
margin of safety to protect public health, and taking into account that lower levels will provide incrementally
greater margins of safety" (Frey, 2014, p. ii).
3-89
-------
• Does the currently available scientific evidence- and exposure/risk-based information
support or call into question the adequacy of the protection afforded by the current
primary O3 standard?
In considering this question, we recognize that, as is the case in NAAQS reviews in
general, the extent to which the current primary O3 standard is judged to be adequate will depend
on a variety of factors, including science policy judgments and public health policy judgments to
be made by the Administrator. These factors include public health policy judgments concerning
the appropriate benchmark concentrations on which to place weight, as well as judgments on the
public health significance of the effects that have been observed at the exposures evaluated in the
health effects evidence. The factors relevant to judging the adequacy of the standards also
include the interpretation of, and decisions as to the weight to place on, different aspects of the
results of the exposure assessment for the eight areas studied and the associated uncertainties.
Thus, we recognize that the Administrator's conclusions regarding the adequacy of the current
standard will depend in part on public health policy judgments, science policy judgments
regarding aspects of the evidence and exposure/risk estimates, and judgments about the degree of
protection that is requisite to protect public health with an adequate margin of safety.
Our response to the overarching question above takes into consideration the discussions
that address the specific policy-relevant questions in prior sections of this document (see section
3.2) and builds on the approach from the last review (summarized in section 3.1 above). We
focus first on consideration of the evidence, including that newly available in this review, and the
extent to which it alters key conclusions supporting the current standard. We then turn to
consideration of the quantitative exposure and risk estimates developed in this review, including
associated limitations and uncertainties, and the extent to which they indicate differing
conclusions regarding the magnitude of risk, as well as level of protection from adverse effects,
associated with the current standard. We additionally consider the key aspects of the evidence
and exposure/risk estimates emphasized in establishing the now-current standard, and the
associated public health policy judgments and judgments about the uncertainties inherent in the
scientific evidence and quantitative analyses that are integral to decisions on the adequacy of the
current primary O3 standard.
As an initial matter, we recognize the continued support in the current evidence for O3 as
the indicator for photochemical oxidants, as recognized in section 3.5.1 above. Of the
photochemical oxidants, O3 is the only one other than nitrogen dioxide (for which there are
separate NAAQS) that is routinely monitored in ambient air. Further, as stated in the ISA, "the
primary literature evaluating the health and ecological effects of photochemical oxidants
includes ozone almost exclusively as an indicator of photochemical oxidants" (ISA, section
IS. 1.1, p. IS-3). In summary, as was the case for previous reviews, the evidence base for health
3-90
-------
effects of photochemical oxidants does not indicate an importance of any other photochemical
oxidants as it includes O3 almost exclusively as an indicator of photochemical oxidants, thus
continuing to support the appropriateness of O3 as the indicator for photochemical oxidants.
In considering the extensive evidence base for health effects of O3, we give particular
attention to the longstanding evidence of respiratory effects causally related to O3 exposures.
This array of effects, and the underlying evidence base, was an integral basis for setting the
current standard. As summarized in section 3.3.1 above and addressed in detail in the ISA, the
evidence base in this review does not include new evidence of respiratory effects associated with
appreciably different exposure circumstances, including any that would be expected to occur
under air quality conditions associated with the current standard. Thus, in considering the
information available at this time, we continue to focus on exposure circumstances associated
with the current standard as those of importance in this review.
Further, while the evidence base has been augmented somewhat since the time of the last
review, we note that the newly available evidence does not lead to different conclusions
regarding the respiratory effects of O3 in ambient air or regarding exposure concentrations
associated with those effects; nor does it identify different populations at risk of 03-related
effects. In this way, the health effects evidence available in this review is consistent with
evidence available in the last review when the current standard was established. This strong
evidence base continues to demonstrate a causal relationship between short-term O3 exposures
and respiratory effects, including in people with asthma. This conclusion is primarily based on
evidence from controlled human exposure studies available at the time of the last review that
reported lung function decrements and respiratory symptoms in people exposed to O3 for 6.6
hours during which they engage in five hours of exercise. Support is also provided by the
experimental animal and epidemiologic evidence that is coherent with the controlled exposure
studies. The epidemiologic evidence, including that recently available, includes studies reporting
positive associations for asthma-related hospital admissions and emergency department visits,
which are strongest for children, with short-term O3 exposures. Based collectively on this
evidence, populations identified as at risk of such effects include people with asthma and
children.
As in the last review, the most certain evidence of health effects in humans elicited by
exposures to specific O3 exposure concentrations is provided by controlled human exposure
studies. This category of short-term studies includes an extensive evidence base of 1- to 3-hour
studies, conducted with continuous or intermittent exercise and generally involving relatively
3-91
-------
higher exposure concentrations (e.g., greater than 120 ppb).88 Given the lack of ambient air
concentrations of this magnitude in areas meeting the current standard (section 2.4.1 above), we
continue to focus primarily on a second group of somewhat longer-duration studies of much
lower exposure concentrations. These studies employ a 6.6-hour protocol that includes six 50-
minute periods of exercise at moderate or greater exertion. There are no new such studies
available in the current review. Thus, the evidence newly available in this review does not extend
our understanding of the range of exposure concentrations that elicit effects in such studies
beyond what was understood in the last review.
As in the last review, 60 ppb remains the lowest exposure concentration (target
concentration, as average across exercise periods) at which statistically significant lung function
decrements have been reported in the 6.6-hour exposure studies. Two studies have assessed
exposure concentrations at the lower concentration of 40 ppb, with no statistically significant
finding of Ch-related FEVi decrement for the group mean in either study (which is just above and
well below 1% in the two studies). At 60 ppb, the group mean Ch-related decrement in FEVi
ranges from approximately 2 to 4%, with associated individual study subject variability in
decrement size. In the single study assessing the next highest exposure concentration (just above
70 ppb),89 the group mean FEVi decrement (6%) was also statistically significant, as were
respiratory symptom scores. At higher exposure concentrations, the incidence of both respiratory
symptom scores and Ch-related lung function decrements in the study subjects is increased.
Other respiratory effects, such as inflammatory response and airway resistance are also increased
at higher exposures (ISA; 2013 ISA; 2006 AQCD).
In considering what may be indicated by the epidemiologic evidence with regard to
exposure concentrations eliciting effects, we recognize that very few of the numerous
epidemiologic studies of respiratory outcome associations with O3 in ambient air were conducted
in areas during times when the current standard was met. In fact, the vast majority of these
studies were conducted in locations and during time periods that would not have met the current
standard, thus making them less useful for considering the potential for O3 concentrations
allowed by the current standard to contribute to health effects. While there were about a handful
88 Table 3A-3 in Appendix 3 summarizes controlled human exposures to O3 for 1 to 2 hours during continuous or
intermittent exercise in contrast to similar exposure durations at rest. This table was adapted from Table 7-1 in the
1996 AQCD and Table AX6-1 in the 2006 CD, with additional studies from Table AX6-13 in the 2006 AQCD, as
well as more recent studies from the 2013 ISA and the ISA for this review.
89 As noted in sections 3.1.1 and 3.3.3 above, the 70 ppb target exposure comes from Schelegle et al. (2009). That
study reported, based on O3 measurements during the six 50-minute exercise periods, that the mean O3
concentration during the exercise portion of the study protocol was 72 ppb. Based on the measurements for the
six exercise periods, the time weighted average concentration across the full 6.6-hour exposure was 73 ppb
(Schelegle et al., 2009).
3-92
-------
of multi-city studies in which the O3 concentrations in a subset of the study locations and for a
portion of the study period appear to have met the current standard, data were not available in
some cities for the earlier years of the study period when design values for other cities were well
above 70 ppb (as discussed in section 3.3.3). We recognize that the study analyses and
associations reported were based on the combined dataset across the full time period (and, for
multicity studies, from all cities), and the extent to which risk associated with exposures derived
from the concentrations in the subset of years (and locations) that would have met the current
standard compared to that from the years (and locations) that would have violated the standard
influenced the study findings is not clear. There were no studies conducted in U.S. locations with
ambient air O3 concentrations that would meet the current standard for the entire duration of the
study (i.e., with design values90 at or below 70 ppb). Thus, the epidemiologic studies provide
limited insight regarding exposure concentrations associated with health outcomes that might be
expected under air quality conditions that meet the current standard (section 3.3.3 above).
In this review, as in the last review, we recognize some uncertainty, reflecting limitations
in the evidence base, with regard to the exposure levels eliciting effects in some population
groups not included in the available controlled human exposure studies, such as children and
individuals with asthma, as well as the severity of the effects. Further, we note uncertainty in the
extent or characterization of effects at exposure levels below those studied. In so doing, we
recognize that the controlled human exposure studies, primarily conducted in healthy adults, on
which the depth of our understanding of Ch-related health effects is based, provide limited, but
nonetheless important information with regard to responses in people with asthma or in children.
Additionally, some aspects of our understanding continue to be limited; among these aspects are
the potential for effects in some people exposed to concentrations below 60 ppb. Collectively,
these aspects of the evidence and associated uncertainties contribute to a recognition that for O3,
as for other pollutants, the available evidence base in a NAAQS review generally reflects a
continuum, consisting of exposure levels at which scientists generally agree that health effects
are likely to occur, through lower levels at which the likelihood and magnitude of the response
become increasingly uncertain.
As at the time of the last review, the exposure and risk estimates developed from
modeling exposures to O3 derived from precursors emitted into ambient air are critically
important to consideration of the potential for exposures and risks of concern under air quality
conditions of interest, and consequently are critically important to judgments on the adequacy of
public health protection provided by the current standard. In considering the public health
90 As described in chapter 2, a design value is the metric used to describe air quality in a given area relative to the
level of the standard, taking the averaging time and form into account. For example, a design value of 70 ppb just
meets the current primary standard.
3-93
-------
implications of estimated occurrences of exposures to the three benchmark concentrations, we
take note of guidance from the ATS, and judgments made by the EPA in considering similar
effects in previous NAAQS reviews (80 FR 65343, October 26, 2015). As recognized in section
3.3.2, an additional publication by the ATS that further addresses judgments on what constitutes
an adverse health effect of air pollution is newly available in this review (Thurston et al., 2017).
The more recent statement expands upon the 2000 statement, that was considered in the last O3
NAAQS review (e.g., giving increased weight to small lung function changes without
accompanying respiratory symptoms when they are documented to occur in individuals with
compromised lung function). We note that, in keeping with the intent of these statements to
avoid specific criteria, neither statement provides more specific descriptions of respiratory
responses, such as with regard to magnitude, duration or frequency, for consideration of such
conclusions. The earlier ATS statement, in addition to emphasizing clinically relevant effects,
also emphasized both the need to consider changes in "the risk profile of the exposed
population," and effects on the portion of the population that may have a diminished reserve that
puts its members at potentially increased risk if affected by another agent (ATS, 2000). These
concepts, including the consideration of the magnitude of effects occurring in just a subset of
study subjects, continue to be recognized as important in the more recent ATS statement
(Thurston et al., 2017) and continue to be relevant to the evidence base for O3.
As summarized in section 2.1 above, the decision in the last review considered the
breadth of the O3 respiratory effects evidence, recognizing the relatively greater significance of
effects reported for exposures at and above 80 ppb as well as the greater array of effects elicited.
The decision additionally emphasized consideration of the much less severe effects associated
with lower exposures, such as 60 ppb, in light of the need for a margin of safety in setting the
standard. The controlled human exposure study evidence as a whole provided context for
consideration of the 2014 HREA results for the exposures of concern (i.e., the comparison-to-
benchmarks analysis) (80 FR 65363, October 26, 2015).
In considering the exposure and risk analyses available in this review, we first note
several ways in which these analyses differ from and improve upon those available in the last
review. For example, we note the number of improvements to input data and modeling
approaches summarized in section 3.4.1 above. As in prior reviews, exposure and risk are
estimated from air quality scenarios designed to just meet an O3 standard in all its elements. That
is, the air quality scenarios are defined by the highest design value in the study area, which is the
location with the highest 3-year average of annual fourth highest daily maximum 8-hour O3
concentrations (e.g., equal to 70 ppb for the current standard scenario). The current risk and
exposure analyses include air quality simulations based on more recent ambient air quality data
that include O3 concentrations closer to the current standard. As a result, much smaller
3-94
-------
reductions in precursor emissions were needed in the photochemical modeling than was the case
with the 2014 HREA. Further, this modeling was updated to reflect the current state of the
science. Additionally, the approach for deriving population exposure estimates, both for
comparison to benchmark concentrations and for use in deriving lung function risk using the E-R
function approach, has been modified to provide for a better match of the simulated population
exposure estimates with the 6.6-hour duration of the controlled human exposure studies and with
the study subject ventilation rates. Together, these differences, as well as a variety of updates to
model inputs, are believed to reduce uncertainty associated with our interpretation of the analysis
results. As we consider the exposure and risk estimates, we also take note of the array of air
quality and exposure circumstances represented by the eight study areas. As summarized in
section 3.2.2 above, the areas fall into seven of the nine climate regions in the continental U.S.
The population sizes of the associated metropolitan areas range in size from approximately 2.4 to
8 million and vaiy in population demographic characteristics. While there are uncertainties and
limitations associated with the exposure and risk estimates, as noted in section 3.4.4 above, the
factors recognized here contribute to their usefulness in informing the current review.
As at the time of the last review, people of all ages with asthma, children, and outdoor
workers, are populations at increased risk of respiratory effects related to O3 in ambient air. The
size of the U.S. population with asthma is approximately 25 million. Children with asthma,
which number approximately six million in the U.S., may be particularly at risk (section 3.3.2
above). While there are more adults in the U.S. with asthma than children with asthma, the
exposure and risk analysis results in terms of percent of the simulated at-risk populations,
indicate higher frequency of exposures of potential concern and risks for children as compared to
adults. This finding relates to children's greater frequency and duration of outdoor activity, as
well as their greater activity level while outdoors (section 3.4.3 above). In light of these
conclusions and findings, we have focused our consideration of the exposure and risk analyses
here on children.
As can be seen by variation in exposure estimates across the study areas, the eight study
areas represent an array of exposure circumstances, including those contributing to relatively
higher and relatively lower exposures and associated risk. As recognized in Appendix 3D and in
section 3.4.3 above, the risk and exposure analyses are not intended to provide a comprehensive
national assessment. Rather, the analyses for this array of study areas and air quality patterns are
intended to indicate the magnitude of exposures and risks that may be expected in areas of the
U.S. that just meet the current standard but that may differ in ways affecting population
exposures of interest. In that way, the exposure and risk estimates are intended to be informative
to the EPA's consideration of potential exposures and risks associated with the current standard
and the Administrator's decision on the adequacy of protection provided by the current standard.
3-95
-------
While we note reduced uncertainty in several aspects of the exposure and risk analysis
approach (as summarized above), we continue to recognize the relatively greater uncertainty
associated with the lung function risk estimates compared to the results of the comparison-to-
benchmarks analysis. Further, with regard to the lung function risk estimates, we recognize
greater uncertainty with the estimates derived using the MSS model approach. Thus, we focus
primarily on the estimates of exposures at or above different benchmark concentrations that
represent different levels of significance of Ch-related effects, both with regard to the array of
effects and severity of individual effects.
Based on all of the above, and taking into consideration related information, limitations
and uncertainties, such as those recognized above, we address the extent to which the newly
available information in this review supports or calls into question the adequacy of protection
afforded by the current standard. In this context for the air quality scenario for the current
standard, we note that across all eight study areas, which provide an array of exposure situations,
less than 1% of children with asthma are estimated to experience, while breathing at an elevated
rate, a daily maximum 7-hour exposure per year at or above 70 ppb, on average across the 3-year
period, with a maximum of 1% for the study area with the highest estimates in the highest single
year (as summarized in section 3.4.2 above). Further, the percentage for at least one day with
such an exposure above 80 ppb is less than 0.1%, as an average across the 3-year period (and
0.1% or less in each of the three years simulated across the eight study areas). No simulated
individuals were estimated to experience more than a single such day with an exposure at or
above the 80 ppb benchmark. Although the exposure and risk analysis approaches have been
updated since the last review as summarized in section 3.4.1 above, these estimates are generally
similar to the comparable estimates for these benchmarks from the 2014 HREA considered at the
time the current standard was set,91 with only slight differences observed, e.g., for the lowest
benchmark. We take note, however, of the differences across air quality scenarios for both sets of
estimates which remain appreciably larger than the slight differences between the current and
2014 estimates. Thus, we conclude the current estimates of children and children with asthma
that might be expected to experience a day with an exposure while exercising at or above the
three benchmark concentrations to be generally similar to those that were a primary focus of the
decision in establishing the current standard in 2015.
91 For example, in the 2015 decision to set the standard level at 70 ppb, the Administrator took note of several
findings for the air quality scenarios for this level, noting that "a revised standard with a level of 70 ppb is
estimated to eliminate the occurrence of two or more exposures of concern to O3 concentrations at or above 80
ppb and to virtually eliminate the occurrence of two or more exposures of concern to O3 concentrations at or
above 70 ppb for all children and children with asthma, even in the worst-case year and location evaluated" (80
FR 65363, October 26, 2015). This statement remains true for the results of the current assessment (Table 3-8).
3-96
-------
We additionally consider the estimates of 7-hour exposures, at elevated ventilation, at or
above 60 ppb. In so doing, we recognize the role of this consideration in the 2015 decision to be
in the context of the Administrator's judgment regarding an adequate margin of safety for the
new standard. We additionally recognize the greater significance of risk for multiple occurrences
of days at or above this benchmark, given the associated greater potential for more lasting
effects. The exposure analysis estimates indicate fewer than 1% to just over 3% of children with
asthma, on average across the 3-year period to be expected to experience two or more days with
an exposure at or above 60 ppb, while at elevated ventilation. This finding of about 97% to more
than 99% of children protected from experiencing two or more days with exposures at or above
60 ppb while at elevated exertion is quite similar to the characterization of such estimates at the
time of the 2015 decision establishing the current standard (as summarized in section 3.1.2.4
above),92 and half that indicated by the comparable estimates for air quality just meeting the
slightly higher design value of 75 ppb. In addition to this level of protection at the lower
exposure level, the current information also indicates more than 99% of children with asthma, on
average per year, to be protected from a day or more with an exposure at or above 70 ppb. In
light of public health judgments by the EPA in prior NAAQS reviews, as well as ATS guidance,
we recognize a greater concern for 7-hour exposures generally at or above 70 and 80 ppb (while
at elevated exertion) than such exposures to O3 concentrations below 70 ppb, and a greater
concern for repeated (versus single) occurrences of such exposures at concentrations at or above
60 ppb up to 70 ppb. With this in mind, we find the exposure and risk estimates for the current
review indicate that the current standard is likely to provide a high level of protection from O3-
related health effects to at-risk populations of all children and children with asthma. We
additionally recognize such protection to be generally similar to what was estimated when the
standard was set in the last review.
As recognized above, the protection afforded by the current standard stems from its
elements collectively, including the level of 70 ppb, the averaging time of eight hours and the
form of the annual fourth-highest daily maximum concentration averaged across three years. The
current evidence as considered in the ISA, the current air quality information as analyzed in
chapter 2 of this document, and the current risk and exposure information presented in Appendix
92 For example, with regard to the 60 ppb benchmark, for which the 2015 decision placed relatively greater weight
on multiple (versus single) occurrences of exposures at or above it, the Administrator at that time noted the 2014
HREA estimates for the 70 ppb air quality scenario that estimated 0.5-3.5% of children to experience multiple
such occurrences on average across the study areas, stating that the now-current standard "is estimated to protect
the vast majority of children in urban study areas ... from experiencing two or more exposures of concern at or
above 60 ppb" (80 FR 65364, October 26, 2015). The corresponding estimates, on average across the 3-year
period in the current assessments, are remarkedly similar at 0.6 -2.9% (Table 3-8).
3-97
-------
3D and summarized here provide continued support to these elements, as well as to the current
indicator, as discussed earlier in this section.
In summarizing the information discussed thus far, we reflect on the key aspects of the
2015 decision that established the current standard. As an initial matter, effects associated with
6.6-hour exposures in the controlled human exposure studies, that included five hours of
exercise, to concentrations just above 70 ppb included both lung function decrements and
respiratory symptoms, which the EPA recognized to be adverse; this judgment was based on
consideration of the EPA decisions in prior NAAQS reviews and CASAC advice, as well as ATS
guidance (80 FR 65343, October 26, 2015). We note that the newly available information in this
review includes an additional statement from ATS on assessing adverse effects of air pollution
which is generally consistent with the earlier statement (available at the time of the 2015
decision), e.g., continuing to emphasize potentially at-risk groups including specific
consideration of effects in people with compromised lung function. While recognizing the
differences between the current and past exposure and risk analyses, as well as uncertainties
associated with such analyses, we note a rough consistency of the associated estimates when
considering the array of study areas in both reviews. Overall, the newly available quantitative
analyses appear to comport with the conclusions reached in the last review regarding control
expected to be exerted by the current standard on exposures of concern.
We additionally recognize that conclusions regarding the adequacy of the current
standard depend in part on public health policy judgments, such as those identified above, and
judgments about when a standard is requisite to protect the public health, including the health of
at-risk populations, allowing for an adequate margin of safety. In so doing, we take note of the
long-standing health effects evidence that documents the effects of 6.6-hour O3 exposures on
people exposed while breathing at elevated rates and recognize that these such effects have been
reported in a few individuals for the lowest concentration studied in exposure chambers (40 ppb).
Thus, in considering the exposure analysis estimates for 7-hour exposures at and above 60 ppb,
we also take note of the variability in the responses at low concentrations, including, for
example, the variation in average response to a 7-hour 60 ppb exposure with exercise (group
mean FEVi decrement of 1.7 to 3.5% change), as well as the lack of statistically significant
decrements in lung functions from such exposures at concentrations below 60 ppb. Consistent
with the EPA's judgments in the last review, we also recognize the greater potential for health
risk from repeated (versus isolated single) occurrences. In so doing, we note that the exposure
estimates indicate the current standard may be expected to protect more than 97% of populations
of children with asthma residing in areas just meeting the current standard from experiencing
more than a single day per year with an exposure at or above 60 ppb, on average over a 3-year
period. We additionally note the estimates that indicate protection of more than 99.9% of
3-98
-------
children with asthma living in such areas from experiencing any days with a 7-hour exposure
while at elevated exertion of 80 ppb or higher in a 3-year period, on average. In light of ATS
guidance, CASAC advice and EPA judgments and conclusions in past NAAQS reviews, these
results indicate a high level of protection of key at-risk populations from Ch-related health effects
that is a generally similar level of protection to what was articulated when the standard was set in
2015. Thus, the evidence and exposure/risk information, including that related to the lowest
exposures studied, lead us to conclude that the combined consideration of the body of evidence
and the quantitative exposure estimates including the associated uncertainties, do not call into
question the adequacy of protection provided by the current standard. Rather, this information
continues to provide support for the current standard, and thus supports consideration of
retaining the current standard, without revision.
In reaching this conclusion, we additionally note the CASAC conclusion that the newly
available evidence does not substantially differ from that available in the last review.
Accordingly, some of the CASAC concluded that the currently available evidence did not
support revision of the standard. Another segment of the CASAC indicated its agreement with
the prior CASAC comments on the 2014 draft PA, in which the prior CASAC opined that a
standard set at 70 ppb may not provide an adequate margin of safety (Cox, 2020, p. 1). In
considering this advice, we take note of the complete advice from the prior CASAC which based
on its focus on the scientific evidence recommended a range of levels from 70 ppb to 60 ppb and
recognized that the choice of a level within this range "is a policy judgment under the statutory
mandate of the Clean Air Act" (Frey, 2014, p. ii). We also note the judgments reached by the
Administrator in the last review, with consideration of this CASAC advice, in setting the
standard at a level of 70 ppb. (80 FR 65362, October 26, 2015).
In summary, the newly available health effects evidence, critically assessed in the ISA as
part of the full body of evidence, reaffirms conclusions on the respiratory effects recognized for
O3 in the last review. Further, we observe the general consistency of the current evidence with
the evidence that was available in the last review with regard to key aspects on which the current
standard is based. We additionally note the quantitative exposure and risk estimates for
conditions just meeting the current standard that indicate a generally similar level of protection
for at-risk populations from respiratory effects, as that described in the last review for the now-
current standard. We also recognize limitations and uncertainties associated with the available
information, similar to those at the time of the last review. Collectively, these considerations
(including those discussed above) provide the basis for the conclusion that consideration should
be given to retaining the current primary standard of 0.070 ppm O3, as the fourth-highest daily
maximum 8-hour concentration averaged across three years, without revision. Accordingly, and
3-99
-------
in light of this conclusion that it is appropriate to consider the current standard to be adequate,
we have not identified any potential alternative standards for consideration in this review.
3.6 KEY UNCERTAINTIES AND AREAS FOR FUTURE RESEARCH
In this section, we highlight key uncertainties associated with reviewing and establishing
the primary O3 standard, while additionally recognizing that research in these areas may be
informative to the development of more efficient and effective control strategies. The list in this
section includes key uncertainties and data gaps thus far highlighted in this review of the primary
standard. A critical aspect of our consideration of the evidence and the quantitative risk/exposure
estimates is our understanding of O3 effects below the lowest concentrations studied in
controlled human exposure studies, for longer exposures and for different population groups,
particularly including people with asthma. Additional information in several areas would reduce
uncertainty in our interpretation of the available information for purposes of risk characterization
and, accordingly, reduce uncertainty in characterization of Ch-related health effects. In this
section, we highlight areas for future health-related research, model development, and data
collection activities to address these uncertainties and limitations in the current scientific
evidence. These areas are similar to those highlighted in past reviews.
Exposure and Risk Assessment Data and Tools:
• An important aspect of risk assessment and characterization to inform decisions regarding
the primary standard is our understanding of the exposure-response relationship for O3
related health effects in at-risk populations. Additional research is needed to more
comprehensively assess risk of respiratory effects in at-risk individuals exposed to O3 in
the range of 40 to 80 ppb, and lower, for 6.6 hours while engaged in moderate exertion.
• Population- or cohort-based information on human exposure and associated health effects
for healthy adults and children and at-risk populations, including people with asthma, to
relevant levels and durations of O3 concentrations in ambient air, including exposure
information in various microenvironments and at varying activity levels, is needed to
better evaluate current and future O3 exposure and lung function risk models. Such
information across extended periods would facilitate evaluation of exposure models for
the O3 season.
• Collection of time-activity data over longer time periods, and particularly for children, is
needed to reduce uncertainty in the modeled exposure distributions that form an
important part of the basis for decisions regarding NAAQS for O3 and other air
pollutants. Research addressing energy expenditure and associated breathing rates in
various population groups, particularly healthy children and children with asthma, in
various locations, across the spectrum of physical activity, including sleep to vigorous
exertion, is needed.
3-100
-------
Health Effects Evidence Base:
• Epidemiologic studies assessing the influence of "long-term" or "short-term" O3
exposures is complicated by a lack of knowledge regarding the exposure history of study
populations. Further, existing studies generally focus on either long-term or short-term
exposure separately, thereby making it difficult to assess whether a single short-term
high-level exposure versus a repeated long-term low-level exposure, or a combination of
both short-term high-level and repeated long-term low-level exposures, influence health
outcomes of the study subjects. Epidemiologic studies that include exposure
measurements across a longer-term assessment period and can simultaneously assess the
impact of these various elements of exposure (i.e., magnitude, frequency, durations, and
pattern) are needed.
• The extent to which the broad mix of photochemical oxidants as well as other copollutants
in the ambient air (e.g., PM, NO2, SO2, etc.) may play a role in modifying or contributing
to the observed associations between ambient air O3 concentrations and reported health
outcomes continues to be an important research question. A better understanding of the
broader mixture of photochemical oxidants other than O3 in ambient air, the associated
human exposures, and of the extent to which effects of the mixture may differ from those
of O3, would be informative to future NAAQS reviews. Examine and improve analytical
approaches to better understand the role of copollutants, as well as temperature, in
contributing to potential confounding or effect modification in epidemiologic models.
• Most epidemiologic study designs remain subject to uncertainty due to use of fixed-site
ambient air monitors serving as a surrogate for exposure measurements. Measurements
made at stationary outdoor monitors have been used as independent variables for air
pollution, but the accuracy with which these measurements actually reflect subjects'
exposure is not yet fully understood. The degree to which discrepancies between
stationary monitor measurements and actual pollutant exposures introduces error into
statistical estimates of pollutant effects in epidemiologic studies needs to be investigated.
• For health endpoints reported in epidemiologic studies, such as respiratory hospital
admissions, ED visits, and premature mortality, a more comprehensive characterization
of the exposure circumstances (including ambient air concentrations, as well as duration
of exposure and activity levels of individuals) eliciting such effects is needed
• Further research investigating additional uncertainties and factors that modify
epidemiologic associations, particularly for different population groups would improve
our understanding in these areas.
• The current evidence base, expanded by evidence newly available in this review, indicates
a likely causal relationship between short-term O3 exposure and metabolic effects.
Further research characterizing perturbations of glucose and insulin homeostasis by O3 in
controlled human exposure studies at exertion and in animal toxicology studies at
concentrations closer to the current standard are needed inform decisions regarding the
primary standard. The collection of population-based information on clinical health
outcomes such as metabolic syndrome, diabetes, etc., as well as intermediate indicators
like insulin resistance is also needed. Such studies would provide an improved
understanding of relationships between O3 exposure and metabolic-related health
outcomes.
3-101
-------
Air Quality:
• To reduce uncertainties in photochemical modeling used in estimating O3 concentration
for different air quality scenarios, we need an improved understanding of the atmospheric
chemistry that produces elevated O3 concentrations and the extent to which key
influences and precursors vary seasonally and geographically across the U.S.
• An improved understanding of relationships between variations in the form of the current
standard would benefit consideration of the form in future reviews.
3-102
-------
REFERENCES
Adams, WC (2000). Ozone dose-response effects of varied equivalent minute ventilation rates. J
Expo Anal Environ Epidemiol 10(3): 217-226.
Adams, WC (2002). Comparison of chamber and face-mask 6.6-hour exposures to ozone on
pulmonary function and symptoms responses. Inhal Toxicol 14(7): 745-764.
Adams, WC (2003). Comparison of chamber and face mask 6.6-hour exposure to 0.08 ppm
ozone via square-wave and triangular profiles on pulmonary responses. Inhal Toxicol
15(3): 265-281.
Adams, WC (2006a). Human pulmonary responses with 30-minute time intervals of exercise and
rest when exposed for 8 hours to 0.12 ppm ozone via square-wave and acute triangular
profiles. Inhal Toxicol 18(6): 413-422.
Adams, WC (2006b). Comparison of chamber 6.6-h exposures to 0.04-0.08 PPM ozone via
square-wave and triangular profiles on pulmonary responses. Inhal Toxicol 18(2): 127-
136.
Akinbami, LJ, Simon, AE and Schoendorf, KC (2016). Trends in allergy prevalence among
children aged 0-17 years by asthma status, United States, 2001-2013. J Asthma 53(4):
356-362.
Alexis, NE, Lay, JC, Hazucha, M, Harris, B, Hernandez, ML, Bromberg, PA, Kehrl, H, Diaz-
Sanchez, D, Kim, C, Devlin, RB and Peden, DB (2010). Low-level ozone exposure
induces airways inflammation and modifies cell surface phenotypes in healthy humans.
Inhal Toxicol 22(7): 593-600.
Arjomandi, M, Balmes, JR, Frampton, MW, Bromberg, P, Rich, DQ, Stark, P, Alexis, NE,
Costantini, M, Hollenbeck-Pringle, D, Dagincourt, N and Hazucha, MJ (2018).
Respiratory Responses to Ozone Exposure. MOSES (The Multicenter Ozone Study in
Older Subjects). Am J Respir Crit Care Med 197(10): 1319-1327.
ATS (1985). Guidelines as to what constitutes an adverse respiratory health effect, with special
reference to epidemiologic studies of air pollution. Am Rev Respir Dis 131(4): 666-668.
ATS (2000). What constitutes an adverse health effect of air pollution? Am J Respir Crit Care
Med 161(2): 665-673.
Brown, JS, Bateson, TF and McDonnell, WF (2008). Effects of exposure to 0.06 ppm ozone on
FEV1 in humans: a secondary analysis of existing data. Environ Health Perspect 116(8):
1023-1026.
Bureau of Labor Statistics (2017). U.S. Department of Labor, The Economics Daily, Over 90
percent of protective service and construction and extraction jobs require work outdoors
on the Internet at https://www.bls.gov/opub/ted/2017/over-90-percent-of-protective-
3-103
-------
service-and-construction-and-extraction-jobs-require-work-outdoors.htm (visited August
27, 2019).
Bureau of Labor Statistics (2019). U.S. Department of Labor. Occupational Requirements
Survey - Accessed August 27, 2019. https://www.bls.gov/ors/home.htm.
CDC (2016). Current Asthma Prevalence by Weight Status Among Adults: United States, 2001—
2014. NCHS Data Brief No. 239, March 2016. Available at:
https://www.cdc.gov/nchs/products/databriefs/db239.htm.
CDC (2019). National Health Interview Survey, 2017. National Center for Health Statistics,
CDC. Washington, DC. Available at:
https://www.cdc.gov/asthma/most_recent_national_asthma_data.htm and
https://www.cdc.gov/asthma/nhis/2017/data.htm. Accessed August 27, 2019.
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 (External Review 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/264cbl227d55e02c85257402007446a4/4713
D217BC07103485258515006359BA/$File/EPA-CASAC-20-003.pdf.
Devlin, RB, McDonnell, WF, Mann, R, Becker, S, House, DE, Schreinemachers, D and Koren,
HS (1991). Exposure of humans to ambient levels of ozone for 6.6 hours causes cellular
and biochemical changes in the lung. Am J Respir Cell Mol Biol 4(1): 72-81.
DHEW (1969). Air Quality Criteria for Particulate Matter. National Air Pollution Control
Administration. Washington, D.C. U.S. Department of Health. January 1969.
Folinsbee, LJ, Horstman, DH, Kehrl, HR, Harder, S, Abdul-Salaam, S and Ives, PJ (1994).
Respiratory responses to repeated prolonged exposure to 0.12 ppm ozone. Am J Respir
Crit Care Med 149(1): 98-105.
Folinsbee, LJ, McDonnell, WF and Horstman, DH (1988). Pulmonary function and symptom
responses after 6.6-hour exposure to 0.12 ppm ozone with moderate exercise. JAPCA
38(1): 28-35.
Frey, HC. (2014). Letter from Dr. H. Christopher Frey, Chair, Clean Air Scientific Advisory
Committee to Honorable Gina McCarthy, Administrator, US EPA. Re: CASAC Review
of the EPA's Second Draft Policy Assessment for the Review of the Ozone National
Ambient Air Quality Standards. June 26, 2014. EPA-CASAC-14-004. Office of the
Administrator, Science Advisory Board Washington, DC. Available at:
https.Y/nepis. epa.gov/Exe/ZyPURL. cgi?Dockey=Pl 00JR6F. txt.
3-104
-------
Frey, HC and Samet, JM. (2012). Letter from Dr. H. Christopher Frey, Chair, Clean Air
Scientific Advisory Committee Ambient Air Monitoring & Methods Committee and
Jonathan Samet, Immediate Past Chair, Clean Air Scientific Advisory Committee, to
Administrator Lisa Jackson. Re: CASAC Review of the EPA's Policy Assessment for the
Review of the Ozone National Ambient Air Quality Standards (First External Review
Draft - August 2012). November 26, 2012. EPA-CASAC-13-003. Office of the
Administrator, Science Advisory Board Washington DC. Available at:
https://nepis. epa.gov/Exe/ZyPURL. cgi ?Dockey=Pl 00J 7PQ. txt.
Gong, H, Jr., Bradley, PW, Simmons, MS and Tashkin, DP (1986). Impaired exercise
performance and pulmonary function in elite cyclists during low-level ozone exposure in
a hot environment. Am J Respir Crit Care Med 134(4): 726-733.
Horstman, DH, Folinsbee, LJ, Ives, PJ, Abdul-Salaam, S and McDonnell, WF (1990). Ozone
concentration and pulmonary response relationships for 6.6-hour exposures with five
hours of moderate exercise to 0.08, 0.10, and 0.12 ppm. Am Rev Respir Dis 142(5):
1158-1163.
Karl, T and Koss, WJ (1984). Regional and national monthly, seasonal, and annual temperature
weighted by area, 1895-1983. 4-3. National Environmental Satellite and Data
Information Service (NESDIS). Asheville, NC.
Kim, CS, Alexis, NE, Rappold, AG, Kehrl, H, Hazucha, MJ, Lay, JC, Schmitt, MT, Case, M,
Devlin, RB, Peden, DB and Diaz-Sanchez, D (2011). Lung function and inflammatory
responses in healthy young adults exposed to 0.06 ppm ozone for 6.6 hours. Am J Respir
Crit Care Med 183(9): 1215-1221.
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.
Mar, TF and Koenig, JQ (2009). Relationship between visits to emergency departments for
asthma and ozone exposure in greater Seattle, Washington. Ann Allergy, Asthma
Immunol 103(6): 474-479.
McCurdy, T (2000). Conceptual basis for multi-route intake dose modeling using an energy
expenditure approach. J Expo Anal Environ Epidemiol 10(1): 86-97.
McCurdy, T and Graham, SE (2003). Using human activity data in exposure models: Analysis of
discriminating factors. J Expo Anal Environ Epidemiol 13(4): 294-317.
McDonnell, WF, Horstman, DH, Hazucha, MJ, Seal, E, Jr., Haak, ED, Salaam, SA and House,
DE (1983). Pulmonary effects of ozone exposure during exercise: Dose-response
characteristics. J Appl Physiol (1985) 54(5): 1345-1352.
McDonnell, WF, Kehrl, HR, Abdul-Salaam, S, Ives, PJ, Folinsbee, LJ, Devlin, RB, O'Neil, JJ
and Horstman, DH (1991). Respiratory response of humans exposed to low levels of
ozone for 6.6 hours. Arch Environ Health 46(3): 145-150.
3-105
-------
McDonnell, WF, Stewart, PW and Smith, MV (2013). Ozone exposure-response model for lung
function changes: an alternate variability structure. Inhal Toxicol 25(6): 348-353.
McDonnell, WF, Stewart, PW, Smith, MV, Kim, CS and Schelegle, ES (2012). Prediction of
lung function response for populations exposed to a wide range of ozone conditions.
Inhal Toxicol 24(10): 619-633.
Samet, JM. (2011). Letter from Jonathan Samet, Chair, Clean Air Scientific Advisory
Committee, to Administrator Lisa Jackson. Re: CASAC Response to Charge Questions
on the Reconsideration of the 2008 Ozone National Ambient Air Quality Standards..
March 30, 2011. EPA-CASAC-11-004. Office of the Administrator, Science Advisory
Board U.S. EPA HQ, Washington DC. Available at:
https://yosemite.epa.gov/sab/sabproduct.nsf/368203f97al5308a852574ba005bbd01/F08
BEB48C1139E2A8525785E006909AC/$File/EPA-CASAC-ll-004-unsigned+.pdf.
Schelegle, ES, Morales, CA, Walby, WF, Marion, S and Allen, RP (2009). 6.6-hour inhalation of
ozone concentrations from 60 to 87 parts per billion in healthy humans. Am J Respir Crit
Care Med 180(3): 265-272.
Silverman, RA and Ito, K (2010). Age-related association of fine particles and ozone with severe
acute asthma in New York City. J Allergy Clin Immunol 125(2): 367-373.
Stieb, DM, Szyszkowicz, M, Rowe, BH and Leech, JA (2009). Air pollution and emergency
department visits for cardiac and respiratory conditions: A multi-city time-series analysis.
Environ Health 8(25): 25.
Strickland, MJ, Darrow, LA, Klein, M, Flanders, WD, Sarnat, JA, Waller, LA, Sarnat, SE,
Mulholland, JA and Tolbert, PE (2010). Short-term associations between ambient air
pollutants and pediatric asthma emergency department visits. Am J Respir Crit Care Med
182(3): 307-316.
Thurston, GD, Kipen, H, Annesi-Maesano, I, Balmes, J, Brook, RD, Cromar, K, De Matteis, S,
Forastiere, F, Forsberg, B, Frampton, MW, Grigg, J, Heederik, D, Kelly, FJ, Kuenzli, N,
Laumbach, R, Peters, A, Rajagopalan, ST, Rich, D, Ritz, B, Samet, JM, Sandstrom, T,
Sigsgaard, T, Sunyer, J and Brunekreef, B (2017). A joint ERS/ATS policy statement:
what constitutes an adverse health effect of air pollution? An analytical framework. Eur
Respir J 49(1).
U.S. Census Bureau (2019). Current Population Survey, Annual Social and Economic
Supplement, 2018. Age and Sex Composition in the United States: 2018.
https://www. census.gov/data/tables/2018/demo/age-and-sex/2018-age-sex-
composition.html Accessed August 27, 2019.
U.S. EPA (1996). Review of national ambient air quality standards for ozone: Assessment of
scientific and technical information: OAQPS staff paper. Office of Air Quality Planning
and Standards. Research Triangle Park, NC. U.S. EPA. EPA-452/R-96-007. June 1996.
Available at: http.Y/nepis. epa.gov/Exe/ZyPDF.cgi?Dockey=2000DKJT.PDF.
3-106
-------
U.S. EPA (2008). Risk and Exposure Assessment to Support the Review of the NO2 Primary
National Ambient Air Quality Standard. EPA-452/R-08-008a. Office of Air Quality
Planning and Standards. Research Triangle Park, NC. Available at:
https://www3.epa.gov/ttn/naaqs/standards/nox/s_nox_cr_rea.html.
U.S. EPA (2009). Risk and Exposure Assessment to Support the Review of the SO2 Primary
National Ambient Air Quality Standard. Office of Air Quality Planning and Standards.
Research Triangle Park, NC. US EPA. EPA-452/R-09-007. Available at:
https://www3.epa.gOv/ttn/naaqs/standards/so2/data/200908S02REAFinalReport.pdf.
U.S. EPA (2010a). Quantitative Risk and Exposure Assessment for Carbon Monoxide -
Amended. Office of Air Quality Planning and Standards. Research Triangle Park, NC.
U.S. EPA. EPA-452/R-10-006. Available at: https://www.epa.gov/naaqs/carbon-
monoxide-co-standards-risk-and-exposure-assessments-current-review.
U.S. EPA (2010b). Quantitative Health Risk Assessment for Particulate Matter (Second External
Review Draft). Office of Air Quality Planning and Standards, Health and Environmental
Impacts Division. Research Triangle Park, NC. U.S. EPA. EPA-452/P-10-001 February
2010. Available at: https://nepis.epa.gov/Exe/ZyPURL.cgi?Dockey=P1006EST.txt.
U.S. EPA (2014). Health Risk and Exposure Assessment for Ozone. (Final Report). Office of Air
Quality Planning and Standards. Research Triangle Park, NC. U.S. EPA. EPA-452/R-14-
004a. August 2014. Available at:
https.Y/nepis. epa.gov/Exe/ZyPURL. cgi?Dockey=Pl OOKBUF. txt.
U.S. EPA (2018). Risk and Exposure Assessment for the Review of the Primary National
Ambient Air Quality Standard for Sulfur Oxides. Office of Air Quality Planning and
Standards. Research Triangle Park, NC. U.S. EPA. EPA-452/R-18-003. Available at:
https://www. epa.gov/sites/production/files/2018-05/documents/primary_so2_naaqs_ -
_final_rea_ -_may_2018.pdf.
U.S. EPA (2019). The Consolidated Human Activity Database (CHAD). Documentation and
User's Guide. Research Triangle Park, NC. US EPA. EPA-452/B-19-001. Available at:
https://www.epa.gov/healthresearch/consolidated-human-activity-database-chad-use-
human-exposure-and-health-studies-and.
Wells, BW, K.; Jenkins, S. (2012). Memorandum to Ozone NAAQS Review Docket (EPA-HQ-
OAR-2008-0699). Analysis of Recent U.S. Ozone Air Quality Data to Support the 03
NAAQS Review and Quadratic Rollback Simulations to Support the First Draft of the
Risk and Exposure Assessment. August 15, 2012.. Docket ID No. EPA-HQ-OAR-2008-
0699. Office of Air Quality Planning and Standards Research Triangle Park, NC.
Available at: https://www. regulations.gov/contentStreamer?documentId=EPA -HQ-OAR-
2008-0699-4253&contentType=pdf.
WHO (2008). Uncertainty and Data Quality in Exposure Assessment. The Internaltional
Programme on Chemical Safety. Geneva. WHO.
https://www.who.int/ipcs/publications/methods/harmonization/exposure_assessment.pdf.
3-107
-------
4 REVIEW OF THE SECONDARY STANDARD
This chapter presents and evaluates the policy implications of the currently available
scientific and technical information pertaining to this review of the O3 secondary standard. In so
doing, the chapter presents key aspects of the current evidence of the welfare effects of O3, as
documented in the ISA, with support from the prior ISA and AQCDs, and associated public
welfare implications, as well as key aspects of quantitative analyses of currently available air
quality and exposure-related information that is presented in appendices associated with this
chapter. Together this information provides the foundation for our evaluation of the current
scientific information regarding welfare effects of O3 in ambient air and the potential for welfare
effects to occur under air quality conditions associated with the current standard, as well as the
associated public welfare implications. Our evaluation is framed around key policy-relevant
questions derived from the questions included in the IRP (IRP, section 3.2.1) and also takes into
account conclusions reached in the last review. In so doing, we additionally take note of the
recent decision of the D.C. Circuit, summarized in section 4.1.2 below, remanding the secondary
standard established in the last review to the EPA for further justification or reconsideration
{Murray Energy Corp. v. EPA, 936 F.3d 597 [D.C. Cir. 2019]). In light of all of these
considerations, we will identify key policy-relevant considerations and summary conclusions
regarding the public welfare protection provided by the current standard for the Administrator's
consideration in this review of the secondary O3 standard.
Within this chapter, background information on the current standard, including
considerations in its establishment in the last review, is summarized in section 4.1. The general
approach for considering the currently available information in this review, including policy-
relevant questions identified to frame our policy evaluation, is summarized in section 4.2. Key
aspects of the currently available welfare effects evidence and associated public welfare
implications and uncertainties are addressed in section 4.3, and the current air quality and
exposure information, with associated uncertainties is addressed in section 4.4. Section 4.5
summarizes the key evidence- and air quality or exposure-based considerations identified in our
evaluation and presents associated summary conclusions of this analysis. Key remaining
uncertainties and areas for future research are identified in section 4.6.
4.1 BACKGROUND ON THE CURRENT STANDARD
The current standard was set in 2015 based on the scientific and technical information
available at that time, as well as the Administrator's judgments regarding the available welfare
effects evidence, the appropriate degree of public welfare protection for the revised standard, and
4-1
-------
available air quality information on seasonal cumulative exposures that may be allowed by such
a standard (80 FR 65292, October 26, 2015). With the 2015 decision, the Administrator revised
the level of the secondary standard for photochemical oxidants, including O3, to 0.070 ppm, in
conjunction with retaining the indicator (O3), averaging time (8 hours) and form (fourth-highest
annual daily maximum 8-hour average concentration, averaged across three years).
The welfare effects evidence base available in the 2015 review included more than fifty
years of extensive research on the phytotoxic effects of O3, conducted both in and outside of the
U.S. that documents the impacts of O3 on plants and their associated ecosystems (U.S. EPA,
1978, 1986, 1996, 2006, 2013). As was established in prior reviews, O3 can interfere with carbon
gain (photosynthesis) and allocation of carbon within the plant, making fewer carbohydrates
available for plant growth, reproduction, and/or yield. For seed-bearing plants, these
reproductive effects will culminate in reduced seed production or yield (U.S. EPA, 1996, pp. 5-
28 and 5-29). The strongest evidence for effects from O3 exposure on vegetation is from
controlled exposure studies, which "have clearly shown that exposure to O3 is causally linked to
visible foliar injury,1 decreased photosynthesis, changes in reproduction, and decreased growth"
in many species of vegetation (2013 ISA, p. 1-15). For example, "visible foliar injury occurs
only when sensitive plants are exposed to elevated O3 concentrations in a predisposing
environment," (2013 ISA, p. 9-39). Effects such as decreased photosynthesis, changes in
reproduction, and decreased growth at the plant scale can also be linked to an array of effects at
larger organizational (e.g., population, community, system) and spatial scales, with the evidence
available in the last review indicating that "O3 exposures can affect ecosystem productivity, crop
yield, water cycling, and ecosystem community composition" (2013 ISA, p. 1-15, Chapter 9,
section 9.4).
In light of this robust evidence base, the 2013 ISA concluded there to be causal
relationships between O3 and visible foliar injury, reduced vegetation growth, reduced
productivity in terrestrial ecosystems, reduced yield and quality of agricultural crops and
alteration of below-ground biogeochemical cycles. The 2013 ISA additionally concluded there
was likely to be a causal relationship between O3 and reduced carbon sequestration in terrestrial
ecosystems, alteration of terrestrial ecosystem water cycling and alteration of terrestrial
community composition (2013 ISA, p. lxviii and Table 9-19). Further, based on the then-
available evidence with regard to O3 effects on climate, the 2013 ISA also found there to be a
causal relationship between changes in tropospheric O3 concentrations and radiative forcing,
found there likely to be a causal relationship between tropospheric O3 concentrations and effects
1 This includes visible changes to leaves or needles such as the occurrence of small dots or bleaching (2013 ISA, p.
9-38).
4-2
-------
on climate as quantified through surface temperature response, and found the evidence to be
inadequate to determine if a causal relationship exists between tropospheric O3 concentrations
and health and welfare effects related to shielding of ultraviolet radiation at wavelengths of 280
to 320 nm (UV-B shielding) (2013 ISA, section 10.5).
The 2015 decision was a public welfare policy judgment made by the Administrator,
which drew upon the available scientific evidence for Ch-attributable welfare effects and on
quantitative analyses of exposures and public welfare risks based on impacts to vegetation,
ecosystems and their associated services, as well as judgments about the appropriate weight to
place on the range of uncertainties inherent in the evidence and analyses. The analyses utilized a
cumulative, concentration-weighted exposure index for O3, the W126 index, that gives greater
weight to elevated concentrations. Use of this metric was based on conclusions in the 2013 ISA
that exposure index that cumulate hourly O3 concentrations, giving greater weight to the higher
concentrations (such as the W126 index), perform well in describing exposure-response
relationships documented in crop and tree seedling studies (2013 ISA, section 9.5). Included in
this decision were judgments on the weight to place on the evidence of specific vegetation-
related effects estimated to result across a range of cumulative seasonal concentration-weighted
O3 exposures; on the weight to give associated uncertainties, including uncertainties of predicted
environmental responses (based on experimental study data); variability in occurrence of the
specific effects in areas of the U.S., especially in areas of particular public welfare significance;
and on the extent to which such effects in such areas may be considered adverse to public
welfare.
The decision was based on a thorough review in the 2013 ISA of the scientific
information on 03-induced environmental effects. The decision also took into account: (1)
assessments in the 2014 PA of the most policy-relevant information in the 2013 ISA regarding
evidence of adverse effects of O3 to vegetation and ecosystems, information on biologically-
relevant exposure metrics, 2014 welfare REA (WREA) analyses of air quality, exposure, and
ecological risks and associated ecosystem services, and staff analyses of relationships between
levels of a W126-based exposure index2 and potential alternative standard levels in combination
with the form and averaging time of the then-current standard; (2) additional air quality analyses
of the W126 index and design values based on the form and averaging time of the then-current
standard; (3) CASAC advice and recommendations; and (4) public comments received during
the development of these documents and on the proposal notice. In addition to reviewing the
2 The W126 index is a cumulative seasonal metric described as the sigmoidally weighted sum of all hourly O3
concentrations observed during a specified daily and seasonal time window, where each hourly O3 concentration
is given a weight that increases from zero to one with increasing concentration (80 FR 65373 74, October 26,
2015). Accordingly, W126 index values are in the units of ppm-hours (ppm-hrs).
4-3
-------
most recent scientific information as required by the CAA, the 2015 rulemaking also
incorporated the EPA's response to the judicial remand of the 2008 secondary O3 standard in
Mississippi v. EPA, 744 F.3d 1334 (D.C. Cir. 2013) and, in light of the court's decision in that
case, explained the Administrator's conclusions as to the level of air quality judged to provide
the requisite protection of public welfare from known or anticipated adverse effects.
Consistent with the general approach routinely employed in NAAQS reviews, the initial
consideration in the last review of the secondary standard was with regard to the adequacy of
protection provided by the then-existing standard. Key aspects of that consideration are
summarized in section 4.1.1 below. The subsequent selection of a standard concluded by the
Administrator to provide the requisite protection under the Act is summarized in section 4.1.2.
4.1.1 Considerations Regarding Adequacy of the Prior Standard
The Administrator's conclusion in the 2015 review regarding the adequacy of the
secondary standard that was set in 2008 (0.075 ppm, as annual fourth-highest daily maximum 8-
hour average concentration averaged over three consecutive years) gave primary consideration to
the evidence of growth effects in well-studied tree species and information on cumulative
seasonal O3 exposures occurring in Class I areas3 when the then-current standard was met (80 FR
65385-65386, October 26, 2015). In so doing, the exposure information for Class I areas was
evaluated in terms of the W126 cumulative seasonal exposure index, an index recognized by the
2013 ISA as a mathematical approach "for summarizing ambient air quality information in [a]
biologically meaningful form[] for O3 vegetation effects assessment purposes" (2013 ISA,
section 9.5.3, p. 9-99). The EPA focused on the W126 index for this purpose consistent with the
evidence in the 2013 ISA and advice from the CASAC (80 FR 65375, October 26, 2015).
In her decision making, the Administrator considered the effects of O3 on tree seedling
growth, as suggested by the CASAC, as a surrogate or proxy for the broader array of vegetation-
related effects of O3, ranging from effects on sensitive species to broader ecosystem-level effects
(80 FR 65369, 65406, October 26, 2015). The metric used for quantifying effects on tree
seedling growth in the review was relative biomass loss (RBL), with the evidence base providing
robust and established exposure-response (E-R) functions for seedlings of 11 tree species (80 FR
65391-92, October 26, 2015; 2014 PA, Appendix 5C).4 The Administrator used this proxy in
making her judgments on O3 effects to the public welfare.
3 Areas designated as Class I include all international parks, national wilderness areas which exceed 5,000 acres in
size, national memorial parks which exceed 5,000 acres in size, and national parks which exceed 6,000 acres in
size, provided the park or wilderness area was in existence on August 7, 1977. Other areas may also be Class I if
designated as Class I consistent with the CAA.
4 These functions for RBL estimate the reduction in a year's growth as a percentage of that expected in the absence
of O3 (2013 ISA, section 9.6.2; 2014 WREA, section 6.2).
4-4
-------
In considering the public welfare protection provided by the then-current standard, the
Administrator gave primary consideration to an analysis of cumulative seasonal exposures in or
near Class I areas during periods when the then-current standard was met and the associated
estimates of growth effects in tree seedlings, in terms of the O3 attributable reductions in RBL in
the median species for which E-R functions have been established (80 FR 65389-65390, October
26, 2015).5 The Administrator noted the occurrence of exposures for which the associated
median estimates of growth effects across the species with E-R functions extend above a
magnitude considered to be "unacceptably high" by the CASAC.6 This analysis estimated
cumulative exposures, in terms of 3-year average W126 index values, at and elevated above 19
ppm-hrs, that occurred under the then-current standard for nearly a dozen areas, distributed
across two NOAA climatic regions of the U.S (80 FR 65385-86, October 26, 2015) 7 The
Administrator gave particular weight to this analysis because of its focus on exposures in Class I
areas, which are lands that Congress set aside for specific uses intended to provide benefits to the
public welfare, including lands that are to be protected so as to conserve the scenic value and the
natural vegetation and wildlife within such areas, and to leave them unimpaired for the
enjoyment of future generations. This emphasis on lands afforded special government
protections, such as national parks and forests, wildlife refuges, and wilderness areas, some of
which are designated Class I areas under the CAA, was consistent with a similar emphasis in the
2008 review of the standard (73 FR 16485, March 27, 2008). The Administrator additionally
recognized that states, tribes and public interest groups also set aside areas that are intended to
provide similar benefits to the public welfare for residents on those lands, as well as for visitors
to those areas (80 FR 65390, October 26, 2015).
As noted across past reviews of O3 secondary standards, the Administrator's judgments
regarding effects that are adverse to public welfare consider the intended use of the ecological
receptors, resources and ecosystems affected (80 FR 65389, October 26, 2015; 73 FR 16496,
March 27, 2008). Thus, in the 2015 review, the Administrator utilized the median RBL estimate
5 In specifically evaluating exposure levels in terms of the W126 index as to potential for impacts on vegetation, the
Administrator focused on the median RBL estimate across the eleven tree species for which robust established E-
R functions were available. The presentation of these E-R functions for growth effects on tree seedlings (and
crops) included estimates ofRBL (and relative yield loss [RYL]) atarange of W126-based exposure levels (2014
PA, Tables 5C-1 and 5C-2). The median tree species RBL or crop RYL was presented for each W126 level (2014
PA, Table 5C 3; 80 FR 65391 [Table 4], October 26, 2015). The Administrator focused on RBL as a surrogate or
proxy for the broader array of vegetation-related effects of potential public welfare significance, which include
effects on growth of individual sensitive species and extend to ecosystem-level effects, such as community
composition in natural forests, particularly in protected public lands, as well as forest productivity (80 FR 65406,
October 26, 2015).
6 In the CASAC's consideration ofRBL estimates presented in the 2014 draft PA (for the 2015 review), it
characterized an estimate of 6% RBL in the median studied species as being "unacceptably high," (Frey, 2014).
7 The NOAA climatic regions are described in section 2.4.2 above and appendices 2B and 4D.
4-5
-------
for the studied species as a quantitative tool within a larger framework of considerations
pertaining to the public welfare significance of O3 effects. She recognized such considerations to
include effects that are associated with effects on growth and that the 2013 ISA determined to be
causally or likely causally related to O3 in ambient air, yet for which there are greater
uncertainties affecting estimates of impacts on public welfare. These other effects included
reduced productivity in terrestrial ecosystems, reduced carbon sequestration in terrestrial
ecosystems, alteration of terrestrial community composition, alteration of below-ground
biogeochemical cycles, and alteration of terrestrial ecosystem water cycles. Thus, in giving
attention to the CASAC's characterization of a 6% estimate for tree seedling RBL in the median
studied species as "unacceptably high", the Administrator, while mindful of uncertainties with
regard to the magnitude of growth impact that might be expected in the field and in mature trees,
was also mindful of related, broader, ecosystem-level effects for which the available tools for
quantitative estimates are more uncertain and those for which the policy foundation for
consideration of public welfare impacts is less well established. As a result, the Administrator
considered tree growth effects of O3, in terms of RBL "as a surrogate for the broader array of O3
effects at the plant and ecosystem levels" (80 FR 65389, October 26, 2015).
Based on all of these considerations, and taking into consideration CASAC advice, the
Administrator concluded that the protection afforded by the then-current standard was not
sufficient and that the standard needed to be revised to provide additional protection from known
and anticipated adverse effects to public welfare, related to effects on sensitive vegetation and
ecosystems, most particularly those occurring in Class I areas, and also in other areas set aside by
states, tribes and public interest groups to provide similar benefits to the public welfare for
residents on those lands, as well as for visitors to those areas. In so doing, she further noted that a
revised standard would provide increased protection for other growth-related effects, including
for relative yield loss (RYL) of crops, reduced carbon storage and for types of effects for which
it is more difficult to determine public welfare significance, as well as for other welfare effects of
O3, such as visible foliar injury (80 FR 65390, October 26, 2015).
4.1.2 Considerations for the Revised Standard
Consistent with the approach employed for considering the adequacy of the then-current
secondary standard, the approach for considering revisions that would result in a standard
providing the requisite protection under the Act also focused on growth-related effects of O3,
using RBL as a surrogate for the broad array of vegetation-related effects and included
judgments on the magnitude of such effects that would contribute to public welfare impacts of
concern. In considering the adequacy of potential alternative standards to provide protection
from such effects, the approach also focused on considering the cumulative seasonal O3
4-6
-------
exposures likely to occur with different alternative standards.
In light of the judicial remand of the 2008 secondary O3 standard referenced above, the
2015 decision on selection of a revised secondary standard first considered the available
evidence and quantitative analyses in the context of an approach for considering and identifying
public welfare objectives for such a standard (80 FR 65403-65408, October 26, 2015). The
robust and longstanding evidence of O3 effects on vegetation and associated terrestrial
ecosystems, including evidence newly available in the 2015 review, provided the foundation for
the Administrator's consideration of O3 effects, associated public welfare protection objectives,
and the revisions to the standard needed to achieve those objectives. In light of the extensive
evidence base in this regard, the Administrator focused on protection against adverse public
welfare effects of 03-related effects on vegetation. In so doing, she took note of effects that
compromise plant function and productivity, and associated effects on ecosystems. She had
particular concern about such effects in natural ecosystems, such as those in areas with
protection designated by Congress for current and future generations, as well as areas similarly
set aside by states, tribes and public interest groups with the intention of providing similar
benefits to the public welfare. The Administrator additionally recognized that providing
protection for this purpose will also provide a level of protection for other vegetation that is used
by the public and potentially affected by O3 including timber, produce grown for consumption
and horticultural plants used for landscaping (80 FR 65403, October 26, 2015).
As an initial matter, the Administrator considered the use of a cumulative seasonal
exposure index for purpose of assessing potential public welfare risks, and similarly, for
assessing potential protection achieved against such risks on a national scale. In consideration of
conclusions of the 2013 ISA and 2014 PA, as well as advice from the CAS AC and public
comments, the focus was on a W126 index described as a maximum 3-month, 12-hour index,
defined by the 3-consecutive-month period within the O3 season with the maximum sum of
W126-weighted hourly O3 concentrations during the period from 8:00 a.m. to 8:00 p.m. each day
(80 FR 65404, October 26, 2015). While recognizing that no one definition of an exposure
metric used for the assessment of protection for multiple effects at a national scale will be
exactly tailored to every species or each vegetation type, ecosystem and region of the country,
the Administrator judged that on balance, a W126 index derived in this way, and averaged over
three years would be appropriate for such purposes (80 FR 65403, October 26, 2015). Thus, in
considering revisions to the secondary standard that would specify a level of air quality to
provide the necessary public welfare protection, the Administrator focused on use of a
cumulative seasonal concentration-weighted exposure index (specifically the W126 index), for
assessing exposure, both for making judgments with regard to the potential harm to public
welfare posed by conditions allowed by various levels of air quality and for making the
4-7
-------
associated judgments regarding the appropriate degree of protection against such potential harm
(80 FR 65403, October 26, 2015).
Based on a number of considerations, the Administrator recognized greater confidence in
judgments related to public welfare impacts based on a 3-year average metric than a single-year
metric, and consequently concluded it to be appropriate to use an index averaged across three
years forjudging public welfare protection afforded by a revised secondary standard (80 FR
65404, October 26, 2015). For example, while recognizing that the scientific evidence
documents the effects on vegetation resulting from individual growing season exposures of
specific magnitude, including those that can affect the vegetation in subsequent years, the
Administrator was also mindful of both the strengths and limitations of the evidence and of the
information on which to base her judgments with regard to adversity of effects on the public
welfare. In this regard, she recognized uncertainties associated with interpretation of the public
welfare significance of effects resulting from a single-year exposure, and that the public welfare
significance of effects associated with multiple years of critical exposures are potentially greater
than those associated with a single year of such exposure. While recognizing the potential for
effects on vegetation associated with a single-year exposure, the Administrator concluded that
use of a 3-year average metric can address the potential for adverse effects to public welfare that
may relate to shorter exposure periods, including a single year (80 FR 65404, October 26, 2015).
While the Administrator recognized the scientific information and interpretations, as well
as CASAC advice, with regard to a single-year exposure index, she also took note of
uncertainties associated with judging the degree of vegetation impacts for single-year effects that
would be adverse to public welfare. It was noted that even in the case of annual crops, the
assessment of public welfare significance of such effects is unclear due to the role of crop
management and related agricultural practices. The Administrator was also mindful of the
variability in ambient air O3 concentrations from year to year, as well as year-to-year variability
in environmental factors, including rainfall and other meteorological factors, that influence the
occurrence and magnitude of 03-related effects in any year, and contribute uncertainties to
interpretation of the potential for harm to public welfare over the longer term (80 FR 65404,
October 26, 2015).
In reaching a conclusion on the amount of public welfare protection from the presence of
O3 in ambient air that is appropriate to be afforded by a revised secondary standard, the
Administrator gave particular consideration to the following: (1) the nature and degree of effects
of O3 on vegetation, including her judgments as to what constitutes an adverse effect to the
public welfare; (2) the strengths and limitations of the available and relevant information; (3)
comments from the public on the Administrator's proposed decision, including comments related
to identification of a target level of protection; and (4) the CASAC's views regarding the
4-8
-------
strength of the evidence and its adequacy to inform judgments on public welfare protection. The
Administrator recognized that such judgments include judgments about the interpretation of the
evidence and other information, such as the quantitative analyses of air quality monitoring,
exposure and risk. She also recognized that such judgments should neither overstate nor
understate the strengths and limitations of the evidence and information nor the appropriate
inferences to be drawn as to risks to public welfare. It was also noted that the CAA does not
require that a secondary standard be protective of all effects associated with a pollutant in the
ambient air but rather those known or anticipated effects judged "adverse to the public welfare"
(CAA section 109). She additionally recognized that the choice of the appropriate level of
protection is a public welfare policy judgment entrusted to the Administrator under the CAA
taking into account both the available evidence and the uncertainties (80 FR 65404-05, October
26, 2015).
With regard to the extensive evidence of welfare effects of O3, including the established
evidence base regarding O3 and visible foliar injury, in addition to the longstanding evidence
base on 03-attributable crop RYL, the information available for tree species was judged to be
more useful in informing judgments regarding the nature and severity of effects associated with
different air quality conditions and associated public welfare significance. Accordingly, the
Administrator gave particular attention to the effects related to native tree growth and
productivity, including forest and forest community composition, recognizing the relationship of
tree growth and productivity to a range of ecosystem services (80 FR 65405-06, October 26,
2015).
In so doing, the Administrator recognized that the robust evidence base documented a
broad array of 03-induced vegetation effects, among which were the occurrence of visible foliar
injury and growth and/or yield loss in 03-sensitive annual and perennial species, including crops
and other commercial species, such as timber, horticultural and landscaping plants, as well as
native species in unmanaged natural areas (80 FR 65405, October 26, 2015). In regard to visible
foliar injury, the Administrator recognized the potential for this effect to affect the public welfare
in the context of affecting value ascribed to natural forests, particularly those afforded special
government protection, with the significance of 03-induced visible foliar injury depending on the
extent and severity of the injury (80 FR 65407, October 26, 2015). In so doing, however, the
Administrator also took note of limitations in the available visible foliar injury information,
including the lack of established E-R functions that would allow prediction of visible foliar
injury severity and incidence under varying air quality and environmental conditions, a lack of
consistent quantitative relationships linking visible foliar injury with other 03-induced vegetation
effects, such as growth or related ecosystem effects, and a lack of established criteria or
objectives that might inform consideration of potential public welfare impacts related to this
4-9
-------
vegetation effect (80 FR 65407, October 26, 2015). Similarly, while 03-related growth effects on
agricultural and commodity crops had been extensively studied and robust E-R functions
developed for a number of species, the Administrator found this information less useful in
informing her judgments regarding an appropriate level of public welfare protection (80 FR
65405, October 26, 2015).8
Thus, and in light of the extensive evidence base in this regard, the Administrator focused
on trees and associated ecosystems in identifying the appropriate level of protection for the
secondary standard. Accordingly, the Administrator found the estimates of tree seedling growth
impacts (in terms of RBL) associated with a range of W126-based index values developed from
the E-R functions for 11 tree species (referenced in section 4.1.1 above) to be appropriate and
useful for considering the appropriate public welfare protection objective for a revised standard
(80 FR 65391-92, Table 4, October 26, 2015). The Administrator also incorporated into her
considerations the broader evidence base associated with forest tree seedling biomass loss,
including other less quantifiable effects of potentially greater public welfare significance. That is,
in drawing on these RBL estimates, the Administrator recognized she was not simply making
judgments about a specific magnitude of growth effect in seedlings that would be acceptable or
unacceptable in the natural environment. Rather, though mindful of associated uncertainties, the
Administrator used the RBL estimates as a surrogate or proxy for consideration of the broader
array of related vegetation and ecosystem effects of potential public welfare significance that
include effects on growth of individual sensitive species and extend to ecosystem-level effects,
such as community composition in natural forests, particularly in protected public lands, as well
as forest productivity (80 FR 65406, October 26, 2015). This broader array of vegetation-related
effects included those for which public welfare implications are more significant but for which
the tools for quantitative estimates were more uncertain.
In using the RBL estimates as a proxy, the Administrator recognized that the CASAC
gave weight to these relationships in formulating its advice and she took particular note of the
characterization by the CASAC of the 6% RBL level in the median studied species as
"unacceptably high," as this comment was provided in the context of the CASAC's consideration
of the significance of effects associated with a range of alternatives for the secondary standard
8 With respect to commercial production of commodities, the Administrator noted that judgments about the extent to
which 63-related effects on commercially managed vegetation are adverse from a public welfare perspective are
particularly difficult to reach, given that the extensive management of such vegetation (which, as the CASAC
noted, may reduce yield variability) may also to some degree mitigate potential 03-related effects. The
management practices used on such vegetation are highly variable and are designed to achieve optimal yields,
taking into consideration various environmental conditions. In addition, changes in yield of commercial crops and
commercial commodities, such as timber, may affect producers and consumers differently, further complicating
the question of assessing overall public welfare impacts (80 FR 65405, October 26, 2015).
4-10
-------
(Frey, 2014, pp. iii, 13, 14; 80 FR 65406, October 26, 2015). In consideration of CASAC advice;
strengths, limitations and uncertainties in the evidence; and the linkages of growth effects to
larger population, community and ecosystem impacts, the Administrator considered it
appropriate to focus on a standard that would generally limit cumulative exposures to those for
which the median RBL estimate for seedlings of the 11 species with robust and established E-R
functions would be somewhat below 6% (80 FR 65406-07, October 26, 2015).
In focusing on cumulative exposures associated with a median RBL estimate somewhat
below 6%, the Administrator considered the relationships between W126-based exposure and
RBL in the studied species (presented in the final PA and proposal notice), noting that the
median RBL estimate was 6% for a cumulative seasonal W126 exposure index of 19 ppm-hrs
(80 FR 65391-92, Table 4, October 26, 2015).9 Given the information on median RBL at
different W126 exposure levels, using a 3-year cumulative exposure index for assessing
vegetation effects, the potential for single-season effects of concern, and CASAC comments on
the appropriateness of a lower value for a 3-year average W126 index, the Administrator
concluded it was appropriate to identify a standard that would restrict cumulative seasonal
exposures to 17 ppm-hrs or lower, in terms of a 3-year W126 index, in nearly all instances (80
FR 65407, October 26, 2015). Based on such then-current information to inform consideration of
vegetation effects and their potential adversity to public welfare, the Administrator additionally
judged that the RBL estimates associated with marginally higher exposures in isolated, rare
instances are not indicative of effects that would be adverse to the public welfare, particularly in
light of variability in the array of environmental factors that can influence O3 effects in different
systems and uncertainties associated with estimates of effects associated with this magnitude of
cumulative exposure in the natural environment (80 FR 65407, October 26, 2015).
The Administrator's decisions regarding the revisions to the then-current standard that
would appropriately achieve these public welfare protection objectives were based on extensive
air quality analyses that extended from the then most recently available data (monitoring year
2013) back more than a decade (80 FR 65408, October 26, 2015; Wells, 2015). These analyses
evaluated the cumulative seasonal exposure levels in locations meeting different alternative
levels for a standard of the existing form and averaging time, indicating reductions in cumulative
exposures associated with air quality meeting lower levels of a standard of the existing form and
averaging time. Based on these analyses, the Administrator judged that the desired level of
public welfare protection could be achieved with a secondary standard having a revised level in
combination with the existing form and averaging time (80 FR 65408, October 26, 2015).
9 When stated to the first decimal place, the median RBL was 6.0% for a cumulative seasonal W126 exposure index
of 19 ppm-hrs. For 18 ppm-hrs, the median RBL estimate was 5.7%, which rounds to 6%, and for 17 ppm-hrs, the
median RBL estimate was 5.3%, which rounds to 5% (80 FR 65407, October 26, 2015).
4-11
-------
The air quality analyses described the occurrences of 3-year W126 index values of
various magnitudes at monitor locations where O3 concentrations met potential alternative
standards; the alternative standards were different levels for the current form and averaging time
(annual fourth-highest daily maximum 8-hour average concentration, averaged over three
consecutive years) (Wells, 2015). In the then-most recent period, 2011-2013, across the more
than 800 monitor locations meeting the then-current standard (with a level of 75 ppb), the 3-year
W126 index values were above 17 ppm-hrs in 25 sites distributed across different NOAA
climatic regions, and above 19 ppm-hrs at nearly half of these sites, with some well above. In
comparison, among sites meeting an alternative standard of 70 ppb, there were no occurrences of
a W126 value above 17 ppm-hrs and fewer than a handful of occurrences that equaled 17 ppm-
hrs.10 For the longer time period (extending back to 2001), among the nearly 4000 instances
where a monitoring site met a standard level of 70 ppb, the Administrator noted that there was
only "a handful of isolated occurrences" of 3-year W126 index values above 17 ppm-hrs, "all but
one of which were below 19 ppm-hrs" (80 FR 65409, October 26, 2015). The Administrator
concluded that that single higher value of 19.1 ppm-hrs, observed at a monitor for the 3-year
period of 2006-2008, was reasonably regarded as an extremely rare and isolated occurrence, and,
as such, it was unclear whether it would recur, particularly as areas across the U.S. took further
steps to reduce O3 to meet revised primary and secondary standards. Further, based on all of the
then available information, as noted above, the Administrator did not judge RBL estimates
associated with marginally higher exposures in isolated, rare instances to be indicative of adverse
effects to the public welfare. The Administrator concluded that a standard with a level of 70 ppb
and the current form and averaging time may be expected to limit cumulative exposures, in terms
of a 3-year average W126 exposure index, to values at or below 17 ppm-hrs, in nearly all
instances, and accordingly, to eliminate or virtually eliminate cumulative exposures associated
with a median RBL of 6% or greater (80 FR 65409, October 26, 2015). Thus, using RBL as a
proxy in judging effects to public welfare, the Administrator judged that a standard with a level
of 70 ppb would provide the requisite protection from adverse effects to public welfare by
limiting cumulative seasonal exposures to 17 ppm-hrs or lower, in terms of a 3-year W126 index,
in nearly all instances.
In summary, the Administrator judged that the revised standard would protect natural
forests in Class I and other similarly protected areas against an array of adverse vegetation
effects, most notably including those related to effects on growth and productivity in sensitive
tree species. The Administrator additionally judged that a revised standard set at a level of 70
10 The more than 500 monitors that would meet an alternative standard of 70 ppb during the 2011-2013 period were
distributed across all nine NOAA climatic regions and 46 of the 50 states (Wells, 2015 and associated dataset in
the docket [document identifier, EPA-HQ-OAR-2008-0699-4325]).
4-12
-------
ppb would be sufficient to protect public welfare from known or anticipated adverse effects. This
judgment by the Administrator appropriately recognized that the CAA does not require that
standards be set at a zero-risk level, but rather at a level that reduces risk sufficiently so as to
protect the public welfare from known or anticipated adverse effects. Thus, based on the
conclusions drawn from the air quality analyses which demonstrated a strong, positive
relationship between the 8-hour and W126 metrics and the findings that indicated the significant
amount of control provided by the fourth-high metric, the evidence base of O3 effects on
vegetation and her public welfare policy judgments, as well as public comments and CASAC
advice, the Administrator decided to retain the existing form and averaging time and revise the
level to 0.070 ppm, judging that such a standard would provide the requisite protection to the
public welfare from any known or anticipated adverse effects associated with the presence of O3
in ambient air (80 FR 65409-10, October 26, 2015).
As noted in Chapter 1, after publication of the final rule revising the standards, a number
of industry groups, environmental and public health organizations, and certain states sought
judicial review in the D.C. Circuit. On August 23, 2019, the court issued an opinion concluding,
in relevant part, that EPA had not provided a sufficient rationale for aspects of its decision on the
2015 secondary standard {Murray Energy Corp. v. EPA, 936 F.3d 597 [D.C. Cir. 2019]).
Accordingly, the court remanded the secondary standard to EPA for further justification or
reconsideration, particularly in relation to its decision to focus on a 3-year average for
consideration of the cumulative exposure, in terms of W126, identified as providing requisite
public welfare protection, and its decision to not identify a specific level of air quality related to
visible foliar injury.
4.2 GENERAL APPROACH AND KEY ISSUES IN THIS REVIEW
As is the case for all such reviews, this review of the secondary standard is most
fundamentally based on using the Agency's assessment of the current scientific evidence and
associated quantitative analyses to inform the Administrator's judgments regarding a secondary
standard that is requisite to protect the public welfare from known or anticipated adverse effects.
The approach planned for this review of the secondary O3 standard will build on the last review,
including the substantial assessments and evaluations performed over the course of that review,
and taking into account the more recent scientific information and air quality data now available
to inform understanding of the key policy-relevant issues in the current review. As noted above,
we are also considering the court's recent decision on the O3 secondary standard, recognizing
that issues raised by the court in its remand of the standard (recognized in section 4.1.2 above)
will be considered over the course of this review.
4-13
-------
The evaluations in the PA, of the scientific assessments in the ISA (building on prior such
assessments) augmented by quantitative air quality and exposure analyses, are intended to inform
the Administrator's public welfare policy judgments and conclusions, including his decisions as
to whether to retain or revise this standard. The PA considers the potential implications of
various aspects of the scientific evidence, the air quality, exposure or risk-based information, and
the associated uncertainties and limitations. In so doing, the approach for this PA involves
evaluating the available scientific and technical information to address a series of key policy-
relevant questions using both evidence- and exposure/risk-based considerations. Together,
consideration of the full set of evidence and information available in this review will inform the
answer to the following initial overarching question for the review:
Do the currently available scientific evidence and exposure/risk-based information
support or call into question the adequacy of the public welfare protection afforded by
the current secondary O3 standard?
In reflecting on this question in the remaining sections of this chapter, we consider the
available body of scientific evidence, assessed in the ISA, and considered as a basis for
developing or interpreting air quality and exposure analyses, including whether it supports or
calls into question the scientific conclusions reached in the last review regarding welfare effects
related to exposure to O3 in ambient air. Information available in this review that may be
informative to public policy judgments on the significance or adversity of key effects on the
public welfare is also considered. Additionally, the currently available exposure and risk
information, whether newly developed in this review or predominantly developed in the past and
interpreted in light of current information, is considered, including with regard to the extent to
which it may continue to support judgments made in the last review. Further, in considering this
question with regard to the secondary O3 standard, we give particular attention to exposures and
risks for effects with the greatest potential for public welfare significance.
The approach to reaching conclusions on the current secondary O3 standard and, as
appropriate, on potential alternative standards, including consideration of policy-relevant
questions that frame the current review, is illustrated in Figure 4-1.
4-14
-------
Does ciirrenly avaiabte evidence and. related
unceitsriiss stengtien orcal Ins tfiiesfart prior
conclusions?
¦ Evidence of welfare efecfc not previously ifeniled*?
¦ Evidence of eiees al tower leveis or for die-rani
exposure circumstances?
¦ Evidence tor vegetaton sleds from eusuSalve
exposures alowecf by tie current standard?
« Uncertairfles Berifecf. in fte lasr review reduced or
newuncertarfc:
> Mature, atagntate, and importance of t
exposures and risks associated vrith ueeing im
current standard?
>UnosrtaMes In t& e. p:sure and risk estates?
/ \
/ X
y
/
¦/
X
Does the
\
/ current tofomatiori
call into question
\ adequacy of
current standard?
NO j Consider retaining
/ \ current standard
YES
Consider Potential.Alternative Standards
Hndjrator, Averaging Tme, Form, Level
/"
Potential Alternative Standards for Consideration
Figure 4-1. Overview of general approach for review of the secondary O3 standard.
4-15
-------
The Agency's approach in its review of secondary standards is consistent with the
requirements of the provisions of the CAA related to the review of NAAQS and with how the
EPA and the courts have historically interpreted the CAA. As discussed in section 1.2 above,
these provisions require the Administrator to establish secondary standards that, in the
Administrator's judgment, are requisite (i.e., neither more nor less stringent than necessary) to
protect the public welfare from known or anticipated adverse effects associated with the presence
of the pollutant in the ambient air. In so doing, the Administrator considers advice from the
CASAC and public comment.
Consistent with the Agency's approach across all NAAQS reviews, the approach of this
PA to informing the Administrator's judgments is based on a recognition that the available
evidence generally reflects continuums that include ambient air exposures for which scientists
generally agree that effects are likely to occur through lower levels at which the likelihood and
magnitude of response become increasingly uncertain. The CAA does not require that standards
be set at a zero-risk level, but rather at a level that reduces risk sufficiently so as to protect the
public welfare from known or anticipated adverse effects. The Agency's decisions on the
adequacy of the current secondary standard and, as appropriate, on any potential alternative
standards considered in a review, are largely public welfare policy judgments made by the
Administrator. The four basic elements of the NAAQS (i.e., indicator, averaging time, form, and
level) are considered collectively in evaluating the protection afforded by the current standard, or
any alternative standards considered. Thus, the Administrator's final decisions in such reviews
draw upon the scientific information and analyses about welfare effects, environmental
exposures and risks, and associated public welfare significance, as well as judgments about how
to consider the range and magnitude of uncertainties that are inherent in the scientific evidence
and analyses.
4.3 WELFARE EFFECTS EVIDENCE
4.3.1 Nature of Effects
The welfare effects evidence base available in the current review includes more than fifty
years of extensive research on the phytotoxic effects of O3, conducted both in and outside of the
U.S., that documents the impacts of O3 on plants and their associated ecosystems (1978 AQCD,
1986 AQCD, 1996 AQCD, 2006 AQCD, 2013 ISA, 2020 ISA). As was established in prior
reviews, O3 can interfere with carbon gain (photosynthesis) and allocation of carbon within the
plant, making fewer carbohydrates available for plant growth, reproduction, and/or yield (1996
AQCD, pp. 5-28 and 5-29). For seed-bearing plants, reproductive effects can include reduced
seed or fruit production or yield. The strongest evidence for effects from O3 exposure on
vegetation was recognized at the time of the last review to be from controlled exposure studies,
4-16
-------
which "have clearly shown that exposure to O3 is causally linked to visible foliar injury,
decreased photosynthesis, changes in reproduction, and decreased growth" in many species of
vegetation (2013 ISA, p. 1-15). Such effects at the plant scale can also be linked to an array of
effects at larger spatial scales (and higher levels of biological organization), with the evidence
available in the last review indicating that "O3 exposures can affect ecosystem productivity, crop
yield, water cycling, and ecosystem community composition" (2013 ISA, p. 1-15, Chapter 9,
section 9.4). Beyond its effects on plants, the evidence in the last review also recognized O3 in
the troposphere as a major greenhouse gas (ranking behind carbon dioxide and methane in
importance), with associated radiative forcing and effects on climate, with accompanying "large
uncertainties in the magnitude of the radiative forcing estimate ... making the impact of
tropospheric O3 on climate more uncertain than the effect of the longer-lived greenhouse gases
(2013 ISA, sections 10.3.4 and 10.5.1 [p. 10-30]).
• Does the current evidence alter conclusions from the last review regarding the
nature of welfare effects attributable to O3 in ambient air? Is there new evidence on
welfare effects beyond those identified in the last review?
The evidence newly available in this review supports, sharpens and expands somewhat
on the conclusions reached in the last review (ISA, Appendices 8 and 9). Consistent with the
evidence in the last review, the currently available evidence describes an array of O3 effects on
vegetation and related ecosystem effects, as well as the role of tropospheric O3 in radiative
forcing and subsequent climate-related effects. Evidence newly available in this review augments
more limited previously available evidence related to insect interactions with vegetation,
contributing to conclusions regarding O3 effects on plant-insect signaling (ISA, Appendix 8,
section 8.7) and on insect herbivores (ISA, Appendix 8, section 8.6), as well as for ozone effects
on tree mortality (Appendix 8, section 8.4). Thus, conclusions reached in the last review are
supported by the current evidence base and conclusions are also reached in a few new areas
based on the now expanded evidence.
The current evidence base, including a wealth of longstanding evidence, supports the
conclusion of causal relationships between O3 and visible foliar injury, reduced vegetation
growth and reduced plant reproduction,11 as well as reduced yield and quality of agricultural
crops, reduced productivity in terrestrial ecosystems, alteration of terrestrial community
composition12, and alteration of belowground biogeochemical cycles (ISA, section IS.5). Based
on the current evidence base, the ISA also concluded there likely to be a causal relationship
11 The 2013 ISA did not include a separate causality determination for reduced plant reproduction. Rather, it was
included with the conclusion of a causal relationship of O3 with reduced vegetation growth (ISA, Table IS-12).
12 The 2013 ISA concluded alteration of terrestrial community composition to be likely causally related to O3 based
on the then available information (ISA, Table IS-12),
4-17
-------
between O3 and alteration of ecosystem water cycling, reduced carbon sequestration in terrestrial
ecosystems, and with increased tree mortality (ISA, section IS.5). Additional evidence newly
available in this review is concluded by the ISA to support conclusions on two additional plant-
related effects: the body of evidence is concluded to be sufficient to infer that there is likely to be
a causal relationship between O3 exposure and alteration of plant-insect signaling, and to infer
that there is likely to be a causal relationship between O3 exposure and altered insect herbivore
growth and reproduction (ISA, Table IS-12).
As in the last review, the strongest evidence and the associated findings of causal or
likely causal relationships with O3 in ambient air, and the quantitative characterizations of
relationships between O3 exposure and occurrence and magnitude of effects are for vegetation
effects. The scales of these effects range from the individual plant scale to the ecosystem scale,
with potential for impacts on the public welfare (as discussed in section 4.3.2 below). The
following summary addresses the identified vegetation-related effects of O3 across these scales.
The current evidence, consistent with the decades of previously available evidence,
documents and characterizes visible foliar injury in many tree, shrub, herbaceous, and crop
species as an effect of exposure to O3 (ISA, Appendix 8, section 8.2; 2013 ISA, section 9.4.2;
2006 AQCD, 1996 AQCD, 1986 AQCD, 1978 AQCD). As recognized in the last review with
regard to the then-available evidence, "[r]ecent experimental evidence continues to show a
consistent association between visible injury and ozone exposure" (ISA, Appendix 8, section 8.2,
p. 8-13; 2013 ISA, section 9.4.2, p. 9-41). Ozone-induced visible foliar injury symptoms on
certain tree and herbaceous species, such as black cherry, yellow-poplar and common milkweed,
have long been considered diagnostic of exposure to elevated O3 based on the consistent
association established with experimental evidence (ISA, Appendix 8, section 8.2; 2013 ISA, p.
1-10).13
The currently available evidence, consistent with that in past reviews, indicates that
"visible foliar injury usually occurs when sensitive plants are exposed to elevated ozone
concentrations in a predisposing environment," with a major factor for such an environment
being the amount of soil moisture available to the plant (ISA, Appendix 8, p. 8-23; 2013 ISA,
section 9.4.2). Further, the significance of O3 injury at the leaf and whole plant levels also
depends on an array of factors that include the amount of total leaf area affected, age of plant,
size, developmental stage, and degree of functional redundancy among the existing leaf area
13 As described in the ISA, "[tjypical types of visible injury to broadleaf plants include stippling, flecking, surface
bleaching, bifacial necrosis, pigmentation (e.g., bronzing), and chlorosis or premature senescence and [tjypical
visible injury symptoms for conifers include chlorotic banding, tip burn, flecking, chlorotic mottling, and
premature senescence of needles" (ISA, Appendix 8, p. 8 13).
4-18
-------
(ISA, Appendix 8, section 8.2; 2013 ISA, section 9.4.2). In this review, as in the past, such
modifying factors contribute to the difficulty in quantitatively relating visible foliar injury to
other vegetation effects (e.g., individual tree growth, or effects at population or ecosystem
levels), such that visible foliar injury "is not always a reliable indicator of other negative effects
on vegetation" (ISA, Appendix 8, section 8.2; 2013 ISA, p. 9-39).14
Consistent with conclusions in past reviews, the evidence, extending back several
decades, continues to document the detrimental effects of O3 on plant growth and reproduction
(ISA, Appendix 8, sections 8.3 and 8.4; 2013 ISA, p. 9-42). The available studies come from a
variety of different study types that cover an array of different species, effects endpoints, and
exposure methods and durations. In addition to studies on scores of plant species that have found
O3 to reduce plant growth, the evidence accumulated over the past several decades documents O3
alteration of allocation of biomass within the plant and plant reproduction (ISA, Appendix 8,
sections 8.3 and 8.4; 2013 ISA, p. 1-10). The biological mechanisms underlying the effect of O3
on plant reproduction include "both direct negative effects on reproductive tissues and indirect
negative effects that result from decreased photosynthesis and other whole plant physiological
changes" (ISA, section IS.5.1.2). A newly available meta-analysis of more than 100 studies
published between 1968 and 2010 summarizes effects of O3 on multiple measures of
reproduction (ISA, Appendix 8, section 8.4.1).
Studies involving experimental field sites have also reported effects on measures of plant
reproduction, such as effects on seeds (reduced weight, germination, and starch levels) that could
lead to a negative impact on species regeneration in subsequent years, and bud size that might
relate to a delay in spring leaf development (ISA, Appendix 8, section 8.4; 2013 ISA, section
9.4.3; Darbah et al., 2007, Darbah et al., 2008). A more recent laboratory study reported 6-hour
daily O3 exposures of flowering mustard plants to 100 ppb during different developmental stages
to have mixed effects on reproductive metrics. While flowers exposed early versus later in
development produced shorter fruits, the number of mature seeds per fruit was not significantly
affected by flower developmental stage of exposure (ISA, Appendix 8, section 8.4.1; Black et al.,
14 Similar to the 2013 ISA, the ISA for the current review states the following (ISA, pp. 8-23 to 8-24).
Although visible injury is a valuable indicator of the presence of phytotoxic concentrations of
ozone in ambient air, it is not always a reliable indicator of other negative effects on vegetation
[e.g., growth, reproduction; U.S. EPA (2013)]. The significance of ozone injury at the leaf and
whole-plant levels depends on how much of the total leaf area of the plant has been affected, as
well as the plant's age, size, developmental stage, and degree of functional redundancy among the
existing leaf area (U.S. EPA, 2013). Previous ozone AQCDs have noted the difficulty in relating
visible foliar injury symptoms to other vegetation effects, such as individual plant growth, stand
growth, or ecosystem characteristics (U.S. EPA, 2006, 1996). Thus, it is not presently possible to
determine, with consistency across species and environments, what degree of injury at the leaf
level has significance to the vigor of the whole plant.
4-19
-------
2012). Another study assessed seed viability for a flowering plant in laboratory and field
conditions, finding effects on seed viability of O3 exposures (90 and 120 ppb) under laboratory
conditions but less clear effects under more field-like conditions (ISA, Appendix 8, section 8.4.1;
Landesmann et al., 2013).
With regard to agricultural crops, the current evidence base, as in the last review, is
sufficient to infer a causal relationship between O3 exposure and reduced yield and quality (ISA,
section IS.5.1.2). The current evidence is augmented by new research in a number of areas,
including studies on soybean, wheat and other nonsoy legumes. The new information assessed in
the ISA remains consistent with the conclusions reached in the 2013 ISA (ISA, section IS.5.1.2).
The evidence base for trees includes a number of studies conducted at the Aspen free-air
carbon-dioxide and ozone enrichment (FACE) experiment site in Wisconsin (that operated from
1998 through 2011) and also available in the last review (ISA, IS.5.1 and Appendix 8, section
8.1.2.1; 2013 ISA, section 9.2.4). These studies, which occurred in a field setting (more similar
to natural forest stands than open-top-chamber studies), reported reduced tree growth when
grown in single or three species stands within 30-m diameter rings and exposed over one or more
years to elevated O3 concentrations (hourly concentrations 1.5 times concentrations in ambient
air at the site) compared to unadjusted ambient air concentrations (2013 ISA, section 9.4.3;
Kubiske et al., 2006, Kubiske et al., 2007).15
With regard to tree mortality, the 2013 ISA did not include a determination of causality
(ISA, Appendix 8, section 8.4). While the then-available evidence included studies identifying
ozone as a contributor to tree mortality, which contributed to the 2013 conclusion regarding O3
and alteration of community composition (2013 ISA, section 9.4.7.4), a separate causality
determination regarding O3 and tree mortality was not assessed (ISA, Appendix 8, section 8.4;
2013 ISA, Table 9-19). The evidence assessed in the 2013 ISA (and 2006 AQCD) was largely
observational, including studies that reported declines in conifer forests for which elevated O3
was identified as contributor but in which a variety of environmental factors may have also
played a role (2013 ISA, section 9.4.7.1; 2006 AQCD, sections AX9.6.2.1, AX9.6.2.2,
AX9.6.2.6, AX9.6.4.1 and AX9.6.4.2). Since the last review, three additional studies are now
available (ISA, Appendix 8, Table 8-9). Two of these are analyses of field observations, one of
15 Seasonal (90-day) W126 index values for unadjusted O3 concentrations over six years of the Aspen FACE
experiments ranged from 2 to 3 ppm-hrs, while the elevated exposure concentrations (reflecting addition of O3 to
ambient air concentrations) ranged from somewhat above 20 to somewhat above 35 ppm-hrs (ISA, Appendix 8,
Figure 8-17).
4-20
-------
which is set in the Spanish Pyrenees.16 A second study is a large-scale empirical statistical
analysis of factors potentially contributing to tree mortality in eastern and central U.S. forests
during the 1971-2005 period, which reported O3 (county-level 11-year [1996-2006] average 8
hour metric)17 to be ninth among the 13 potential factors assessed18 and to have a significant
positive correlation with tree mortality (ISA, section IS.5.2, Appendix 8, section 8.4.3; Dietze
and Moorcroft, 2011). A newly available experimental study also reported increased mortality in
two of five aspen genotypes grown in mixed stands under elevated O3 concentrations (ISA,
section IS.5.1.2; Moran and Kubiske, 2013). Coupled with the plant-level evidence of
phytotoxicity discussed above, as well as consideration of community composition effects, this
evidence was concluded to indicate the potential for elevated O3 concentrations to contribute to
tree mortality (ISA, section IS.5.1.2 and Appendix 8, sections 8.4.3 and 8.4.4). Based on the
current evidence, the ISA concludes there is likely to be a causal relationship between O3 and
increased tree mortality (ISA, Table IS-2, Appendix 8, section 8.4.4).
A variety of factors in natural environments can either mitigate or exacerbate predicted
03-plant interactions and are recognized sources of uncertainty and variability. Such factors at
the plant level include multiple genetically influenced determinants of O3 sensitivity, changing
sensitivity to O3 across vegetative growth stages, co-occurring stressors and/or modifying
environmental factors (ISA, Appendix 8, section 8.12).
Ozone-induced effects at the scale of the whole plant have the potential to translate to
effects at the ecosystem scale, such as reduced productivity and carbon storage, and altered
terrestrial community composition, as well as impacts on ecosystem functions, such as
belowground biogeochemical cycles and ecosystem water cycling. For example, under the
relevant exposure conditions, 03-related reduced tree growth and reproduction, as well as
increased mortality, could lead to reduced ecosystem productivity. Recent studies from the
Aspen FACE experiment and modeling simulations indicate that 03-related negative effects on
ecosystem productivity may be temporary or may be limited in some systems (ISA, Appendix 8,
16 The concentration gradient with altitude in the Spanish study, includes - at the highest site - annual average April-
to-September O3 concentrations for the 2004 to 2007 period that range up to 74 ppb (Diaz-de-Quijano et al.,
2016), indicating O3 concentrations likely to exceed the current U.S. secondary standard.
17 As indicated in Figures 2 11 and 2-12, annual fourth highest daily maximum 8-hour O3 concentrations in these
regions were above 80 ppb in the early 2000s and the median design values at national trend sites was nearly 85
ppb.
18 This statistical analysis, which utilized datasets from within the 1971-2005 period, included an examination of the
sensitivity of predicted mortality rate to 13 different covariates. On average across the predictions for 10 groups
of trees (based on functional type and major representative species), the order of mortality rate sensitivity to the
covariates, from highest to lowest, was: sulfate deposition, tree diameter, nitrate deposition, summer temperature,
tree age, elevation, winter temperature, precipitation, O3 concentration, tree basal area, topographic moisture
index, slope and topographic radiation index (Dietze and Moorcroft, 2011).
4-21
-------
section 8.8.1). Previously available studies had reported impacts on productivity in some forest
types and locations, such as ponderosa pine in southern California and other forest types in the
mid-Atlantic region (2013 ISA, section 9.4.3.4). Through reductions in sensitive species growth,
and related ecosystem productivity, O3 could lead to reduced ecosystem carbon storage (ISA,
IS.5.1.4; 2013 ISA, section 9.4.3). With regard to forest community composition, available
studies have reported changes in tree communities composed of species with relatively greater
and relatively lesser sensitivity to O3, such as birch and aspen, respectively (ISA, section
IS.5.1.8.1, Appendix 8, section 8.10; 2013 ISA, section 9.4.3; Kubiske et al., 2007). As the ISA
concludes, " [t]he extent to which ozone affects terrestrial productivity will depend on more than
just community composition, but other factors, which both directly influence [net primary
productivity] (i.e., availability of N and water) and modify the effect of ozone on plant growth"
(ISA, Appendix 8, section 8.8.1). Thus, the magnitude of O3 impact on ecosystem productivity,
as on forest composition, can vary among plant communities based on several factors, including
the type of stand or community in which the sensitive species occurs (e.g., single species versus
mixed canopy), the role or position of the species in the stand (e.g., dominant, sub-dominant,
canopy, understoiy), and the sensitivity of co-occurring species and environmental factors (e.g.,
drought and other factors).
The effects of O3 on plants and plant populations have implications for ecosystem
functions. Two such functions, effects with which O3 is concluded to be likely causally or
causally related, are ecosystem water cycling and belowground biogeochemical cycles,
respectively (ISA, Appendix 8, sections 8.11 and 8.9). With regard to the former, the effects of
O3 on plants (e.g., via stomatal control, as well as leaf and root growth and changes in wood
anatomy associated with water transport) can affect ecosystem water cycling through impacts on
root uptake of soil moisture and groundwater as well as transpiration through leaf stomata to the
atmosphere (ISA, Appendix 8, section 8.11.1). These "impacts may in turn affect the amount of
water moving through the soil, running over land or through groundwater and flowing through
streams" (ISA, Appendix 8, section 8.11, p. 8-161). Evidence newly available in this review is
supportive of previously available evidence in this regard (ISA, Appendix 8, section 8.11.6). The
current evidence, including that newly available, indicates the extent to which the effects of O3
on plant leaves and roots (e.g., through effects on chemical composition and biomass) can impact
belowground biogeochemical cycles involving root growth, soil food web structure, soil
decomposer activities, soil microbial respiration, soil carbon turnover, soil water cycling and soil
nutrient cycling (ISA, Appendix 8, section 8.9).
Additional vegetation-related effects with implications beyond individual plants include
the effects of O3 on insect herbivore growth and reproduction and plant-insect signaling (ISA,
Table IS-12, Appendix 8, sections 8.6 and 8.7). With regard to insect herbivore growth and
4-22
-------
reproduction, the evidence includes multiple effects in an array of insect species, although
without a consistent pattern of response for most endpoints (ISA, Appendix 8, Table 8-11). As
was also the case with the studies available at the time of the last review,19 in the newly available
studies the individual-level responses are highly context- and species-specific and not all species
tested showed a response (ISA, p. IS-64, Table IS-12, section IS.5.1.3 and Appendix 8, section
8.6). Evidence on plant-insect signaling that is newly available in this review comes from
laboratory, greenhouse, open top chambers (OTC) and FACE experiments (ISA, section IS.5.1.3
and Appendix 8, section 8.7). The available evidence indicates a role for elevated O3 in altering
and degrading emissions of chemical signals from plants and reducing detection of volatile plant
signaling compounds (VPSCs) by insects, including pollinators. Elevated O3 concentrations
degrade some VPSCs released by plants, potentially affecting ecological processes including
pollination and plant defenses against herbivoiy. Further, the available studies report elevated O3
conditions to be associated with plant VPSC emissions that may make a plant either more
attractive or more repellant to herbivorous insects, and to predators and parasitoids that target
phytophagous (plant-eating) insects (ISA, section IS.5.1.3 and Appendix 8, section 8.7).
Ozone welfare effects also extend beyond effects on vegetation and associated biota due
to it being a major greenhouse gas and radiative forcing agent.20 As in the last review, the current
evidence, augmented since the 2013 ISA, continues to support a causal relationship between the
global abundance of O3 in the troposphere and radiative forcing, and a likely causal relationship
between the global abundance of O3 in the troposphere and effects on temperature, precipitation,
and related climate variables21 (ISA, section IS.5.2 and Appendix 9; Myhre et al., 2013). As was
also true at the time of the last review, tropospheric O3 has been ranked third in importance for
global radiative forcing, after carbon dioxide and methane, with the radiative forcing of O3 since
pre-industrial times estimated to be about 25 to 40% of the total warming effects of
anthropogenic carbon dioxide and about 75% of the effects of anthropogenic methane (ISA,
Appendix 9, section 9.1.3.3). Uncertainty in the magnitude of radiative forcing estimated to be
attributed to tropospheric O3 is a contributor to the relatively greater uncertainty associated with
19 During the last review, the 2013 ISA stated with regard to O3 effects on insects and other wildlife that "there is no
consensus on how these organisms respond to elevated O3 (2013 ISA, section 9.4.9.4, p. 9-98).
20 Radiative forcing is a metric used to quantify the change in balance between radiation coming into and going out
of the atmosphere caused by the presence of a particular substance. The ISA describes it more specifically as "a
perturbation in net radiative flux at the tropopause (or top of the atmosphere) caused by a change in radiatively
active forcing agent (s) after stratospheric temperatures have readjusted to radiative equilibrium (stratospherically
adjusted RF)" (ISA, Appendix 9, section 9.1.3.3).
21 Effects on temperature, precipitation, and related climate variables were referred to as "climate change" or
"effects on climate" in the 2013 ISA (ISA, p. IS 82; 2013 ISA, pp. 114, 10-31).
4-23
-------
climate effects of tropospheric O3 compared to such effects of the well mixed greenhouse gases,
such as carbon dioxide and methane (ISA, section IS.6.2.2).
Lastly, the evidence regarding tropospheric O3 and UV-B shielding was evaluated in the
2013 ISA and determined to be inadequate to draw a causal conclusion (2013 ISA, section
10.5.2). The current ISA concludes there to be no new evidence since the 2013 ISA relevant to
the question of UV-B shielding by tropospheric O3 (ISA, IS. 1.2.1 and Appendix 9, section
9.1.3.4).
4.3.2 Public Welfare Implications
The public welfare implications of the evidence regarding O3 welfare effects are
dependent on the type and severity of the effects, as well as the extent of the effect at a particular
biological or ecological level of organization. We discuss such factors here in light of judgments
and conclusions made in prior reviews regarding effects on the public welfare.
As provided in section 109(b) (2) of the CAA, the secondary standard is to "specify a
level of air quality the attainment and maintenance of which in the judgment of the
Administrator... is requisite to protect the public welfare from any known or anticipated adverse
effects associated with the presence of such air pollutant in the ambient air." The secondary
standard is not meant to protect against all known or anticipated 03-related welfare effects, but
rather those that are judged to be adverse to the public welfare, and a bright-line determination of
adversity is not required in judging what is requisite (78 FR 3212, January 15, 2013; 80 FR
65376, October 26, 2015; see also 73 FR 16496, March 27, 2008). Thus, the level of protection
from known or anticipated adverse effects to public welfare that is requisite for the secondary
standard is a public welfare policy judgment to be made by the Administrator. In each review,
the Administrator's judgment regarding the currently available information and adequacy of
protection provided by the current standard is generally informed by considerations in prior
reviews and associated conclusions.
• Is there information newly available in this review relevant to consideration of the
public welfare implications of 03-related welfare effects?
The categories of effects identified in the CAA to be included among welfare effects are
quite diverse,22 and among these categories, any single category includes many different types of
effects that are of broadly varying specificity and level of resolution. For example, effects on
vegetation, is a category identified in CAA section 302(h), and the ISA recognizes numerous
22 Section 302(h) of the CAA states that language referring to "effects on welfare" in the CAA "includes, but is not
limited to, effects on soils, water, crops, vegetation, manmade materials, animals, wildlife, weather, visibility, and
climate, damage to and deterioration of property, and hazards to transportation, as well as effects on economic
values and on personal comfort and well-being" (CAA section 302(h)).
4-24
-------
vegetation-related effects of O3 at the organism, population, community and ecosystem level, as
summarized in section 4.3.1 above (ISA, Appendix 8). The significance of each type of
vegetation-related effect with regard to potential effects on the public welfare depends on the
type and severity of effects, as well as the extent of such effects on the affected environmental
entity, and on the societal use of the affected entity and the entity's significance to the public
welfare. For example, a key consideration with regard to public welfare implications in prior
reviews of the O3 secondary standard was the intended use of the affected or sensitive vegetation
and the significance of the vegetation to the public welfare (73 FR 16496, March 27, 2008; 80
FR 65292, October 26, 2015).
More specifically, judgments regarding public welfare significance in the last two O3
NAAQS decisions gave particular attention to O3 effects in areas with special federal protections,
and lands set aside by states, tribes and public interest groups to provide similar benefits to the
public welfare (73 FR 16496, March 27, 2008; 80 FR 65292, October 26, 2015). For example, in
the decision to revise the secondary standard in the 2008 review, the Administrator took note of
"a number of actions taken by Congress to establish public lands that are set aside for specific
uses that are intended to provide benefits to the public welfare, including lands that are to be
protected so as to conserve the scenic value and the natural vegetation and wildlife within such
areas, and to leave them unimpaired for the enjoyment of future generations" (73 FR 16496,
March 27, 2008). As further recognized in the 2008 notice, "[s]uch public lands that are
protected areas of national interest include national parks and forests, wildlife refuges, and
wilderness areas" (73 FR 16496, March 27, 2008).23,24 Such areas include Class I areas25 which
are federally mandated to preserve certain air quality related values. Additionally, as the
Administrator recognized, "States, Tribes and public interest groups also set aside areas that are
intended to provide similar benefits to the public welfare, for residents on State and Tribal lands,
as well as for visitors to those areas" (73 FR 16496, March 27, 2008). The Administrator took
note of the "clear public interest in and value of maintaining these areas in a condition that does
23 For example, the fundamental purpose of parks in the National Park System "is to conserve the scenery, natural
and historic objects, and wild life in the System units and to provide for the enjoyment of the scenery, natural and
historic objects, and wild life in such manner and by such means as will leave them unimpaired for the enjoyment
of future generations" (54 U.S.C. § 100101).
24 As a second example, the Wilderness Act of 1964 defines designated "wilderness areas" in part as areas
"protected and managed so as to preserve [their] natural conditions" and requires that these areas "shall be
administered for the use and enjoyment of the American people in such manner as will leave them unimpaired for
future use and enjoyment as wilderness, and so as to provide for the protection of these areas, [and] the
preservation of their wilderness character..." (16 U.S.C. § 1131 (a) and (c)).
25 Areas designated as Class I include all international parks, national wilderness areas which exceed 5,000 acres in
size, national memorial parks which exceed 5,000 acres in size, and national parks which exceed six thousand
acres in size, provided the park or wilderness area was in existence on August 7, 1977. Other areas may also be
Class I if designated as Class I consistent with the CAA.
4-25
-------
not impair their intended use and the fact that many of these lands contain 03-sensitive species"
(73 FR 16496, March 27, 2008). Similarly, in the 2015 review, the Administrator indicated
particular concern for Ch-related effects on plant function and productivity and associated
ecosystem effects in natural ecosystems "such as those in areas with protection designated by
Congress for current and future generations, as well as areas similarly set aside by states, tribes
and public interest groups with the intention of providing similar benefits to the public welfare"
(80 FR 65403, October 26, 2015).
The 2008 and 2015 decision notices recognized that the degree to which effects on
vegetation in specially protected areas, such as those identified above, may be judged adverse
involves considerations from the species level to the ecosystem level, such that judgments can
depend on the intended use for, or service (and value) of, the affected vegetation, ecological
receptors, ecosystems and resources and the significance of that use to the public welfare (73 FR
16496, March 27, 2008; 80 FR 65377, October 26, 2015). Uses or services provided by areas
that have been afforded special protection can flow in part or entirely from the vegetation that
grows there. For example, ecosystem services are the "benefits that people derive from
functioning ecosystems" (Costanza et al., 2017; ISA, section IS.5.1).26 Ecosystem services range
from those directly related to the natural functioning of the ecosystem to ecosystem uses for
human recreation or profit, such as through the production of lumber or fuel (Costanza et al.,
2017). Aesthetic value and outdoor recreation depend, at least in part, on the perceived scenic
beauty of the environment. Further, there have been analyses that report the American public
values - in monetary as well as nonmonetary ways - the protection of forests from air pollution
damage (Haefele et al., 1991). In fact, public surveys have indicated that Americans rank as very
important the existence of resources, the option or availability of the resource and the ability to
bequest or pass it on to future generations (Cordell et al., 2008). The spatial, temporal and social
dimensions of public welfare impacts are also influenced by the type of service affected. For
example, a national park can provide direct recreational services to the thousands of visitors that
come each year, but also provide an indirect value to the millions who may not visit but receive
satisfaction from knowing it exists and is preserved for the future (80 FR 65377, October 26,
2015).
The different types of effects on vegetation discussed in section 4.3.1 above differ with
regard to aspects important to judging their public welfare significance. In the case of crop yield
loss, such judgments depend on considerations related to the heavy management of agriculture in
26 Ecosystem services analyses were one of the tools used in the last review of the secondary standards for oxides of
nitrogen and sulfur to inform the decisions made with regard to adequacy of protection provided by the standards
and as such, were used in conjunction with other considerations in the discussion of adversity to public welfare
(77 FR 20241, April 3, 2012).
4-26
-------
the U.S., while judgments for other categories of effects may generally relate to considerations
regarding forested areas, including specifically those not managed for harvest. For example,
effects on tree growth and reproduction, and also visible foliar injury, have the potential to be
significant to the public welfare through impacts in Class I and other areas given special
protection in their natural/existing state, although they differ in how they might be significant.
As described in section 4.3.1 above, O3 effects on tree growth and reproduction could,
depending on severity, extent and other factors, lead to effects on a larger scale including
reduced productivity, altered forest and forest community (plant, insect and microbe)
composition, reduced carbon storage and altered ecosystem water cycling (ISA, section
IS.5.1.8.1; 2013 ISA, Figure 9-1, sections 9.4.1.1 and 9.4.1.2). For example, forest or forest
community composition can be affected through O3 effects on growth and reproductive success
of sensitive species in the community, with the extent of compositional changes dependent on
factors such as competitive interactions (ISA, section IS.5.1.8.1; 2013 ISA, sections 9.4.3 and
9.4.3.1). Impacts on some of these characteristics (e.g., forest or forest community composition)
may be considered of greater public welfare significance when occurring in Class I or other
protected areas, due to value for particular services that the public places on such areas.
Depending on the type and location of the affected ecosystem, however, a broader array
of services benefitting the public can be affected in a broader array of areas well. For example,
other services valued by people that can be affected by reduced tree growth, productivity and
associated forest effects include aesthetic value, food, fiber, timber, other forest products, habitat,
recreational opportunities, climate and water regulation, erosion control, air pollution removal,
and desired fire regimes, as summarized in Figure 4-2 (ISA, section IS.5.1; 2013 ISA, sections
9.4.1.1 and 9.4.1.2). In the decisions to revise the secondary standard in the last two reviews, the
Administrator recognized that by providing protection based on consideration of effects in
natural ecosystems in areas afforded special protection, the revised secondary standard would
also "provide a level of protection for other vegetation that is used by the public and potentially
affected by O3 including timber, produce grown for consumption and horticultural plants used
for landscaping" (80 FR 65403, October 26, 2015). For example, locations potentially vulnerable
to 03-related impacts but not necessarily identified for special protection might be forested lands,
both public and private, where trees are grown for timber production. Forests in urbanized areas
also provide a number of services that are important to the public in those areas, such as air
pollution removal, cooling, and beautification. There are also many other tree species, such as
various ornamental and agricultural species (e.g., Christmas trees, fruit and nut trees), that
provide ecosystem services that may be judged important to the public welfare.
Depending on its severity and spatial extent, visible foliar injury, which affects the
physical appearance of the plant, also has the potential to be significant to the public welfare
4-27
-------
through impacts in Class I and other similarly protected areas. In cases of widespread and severe
injury during the growing season (particularly when sustained across multiple years, and
accompanied by obvious impacts on the plant canopy), Ch-induced visible foliar injury might be
expected to have the potential to impact the public welfare in scenic and/or recreational areas,
particularly in areas with special protection, such as Class I areas.27 The ecosystem services most
likely to be affected by 03-induced visible foliar injury (some of which are also recognized
above for tree growth-related effects) are cultural services, including aesthetic value and outdoor
recreation.
The geographic extent of protected areas that may be vulnerable to public welfare effects
of O3, such as impacts to outdoor recreation, is potentially appreciable. For example,
biomonitoring surveys that were routinely administered by the U.S. Forest Service (USFS) as far
back as 1994 in the eastern U.S. and 1998 in the western U.S. include many field sites at which
there are plants sensitive to Ch-related visible foliar injury; there are 450 field sites across 24
states in the North East and North Central regions (Smith, 2012).28 Since visible foliar injury is a
visible indication of O3 exposure in species sensitive to this effect, a number of such species
have been established as bioindicator species, and such surveys have been used by federal land
managers as tools in assessing potential air quality impacts in Class I areas (U.S. Forest Service,
2010). Additionally, the USFS has developed categories for the scoring system they use for
purposes of describing and comparing injury severity at biomonitoring sites. The sites are termed
biosites and the scoring system involves deriving biosite index scores that may be described with
regard to one of several categories. For example, biosite index scores of zero to five are
described as "little or no foliar injury," scores above five to 15 as "low" or "light to moderate"
foliar injury, scores from 15 to 25 as "moderate foliar injury" and scores above 25 as "severe
injury" (Campbell et al., 2007; Smith et al., 2007; Smith, 2012).29 As noted in section 4.3.1
above, there is not an established quantitative relationship between visible foliar injury and other
27 For example, although analyses specific to visible foliar injury are of limited availability, there have been analyses
developing estimates of recreation value damages of severe impacts related to other types of forest effects, such
as tree mortality due to bark beetle outbreaks (e.g., Rosenberger et al., 2013). Such analyses estimate reductions
in recreational use when the damage is severe (e.g., reductions in the density of live, robust trees). Such damage
would reasonably be expected to also reflect damage indicative of injury with which a relationship with other
plant effects (e.g., growth and reproduction) would be also expected. Similarly, a couple of studies from the
1970s and 1980s indicated likelihood for reduced recreational use in areas with stands of pine in which moderate
to severe injury was apparent from 30 or 40 feet.
28 This aspect of the USFS biomonitoring surveys has apparently been suspended, with the most recent surveys
conducted in 2011 (USFS, 2013, USFS, 2017).
29 Early in the USFS biomonitoring program data, there were suggestions of what might be considered "assumptions
of risk" related to scores in these categories, e.g., none, low, moderate and high for BI scores of zero to five, five
to 15, 15 to 25 and above 25, respectively (e.g., Smith et al., 2003; Smith et al., 2012).
4-28
-------
effects, such as reduced growth and productivity as visible foliar injury "is not always a reliable
indicator of other negative effects" (ISA, Appendix 8, section 8.2).
Public welfare implications associated with visible foliar injury might further be
considered to relate largely to effects on scenic and aesthetic values. The available information
does not yet address or describe the relationships expected to exist between some level of injury
severity (e.g., little, low/light, moderate or severe) and/or spatial extent affected and scenic or
aesthetic values. This gap impedes consideration of the public welfare implications of different
injury severities, and accordingly judgments on the potential for public welfare significance.
That notwithstanding, some level of severity and widespread occurrence of visible foliar injury,
particularly if occurring in specially protected areas, such as Class I areas, where the public can
be expected to place value (e.g., for recreational uses), might reasonably be concluded to impact
the public welfare. Thus, key considerations for public welfare significance of this endpoint in
past reviews have related to qualitative consideration of the potential for such effects to affect the
aesthetic value of plants in protected areas, such as Class I areas (73 FR 16490, March 27, 2008).
While, as noted above, public welfare benefits of forested lands can be particular to the
type of area in which the forest occurs, some of the potential public welfare benefits associated
with forest ecosystems are not location dependent. A potentially extremely valuable ecosystem
service provided by forested lands is carbon sequestration or storage (ISA, section IS.5.1.4 and
Appendix 8, section 8.8.3; 2013 ISA, section 2.6.2.1 and p. 9-37).30 As noted above, the EPA has
concluded that effects on this ecosystem service are likely causally related to O3 in ambient air
(ISA, Table IS-12). The importance of carbon sequestration to the public welfare relates to its
role in counteracting the impact of greenhouse gases on radiative forcing and related climate
effects. As summarized in section 4.3.1 above, O3 is also a greenhouse gas and O3 abundance in
the troposphere is causally related to radiative forcing and likely causally related to subsequent
effects on temperature, precipitation and related climate variables (ISA, section IS.6.2.2).
Accordingly, such effects also have important public welfare implications, although their
quantitative evaluation in response to O3 concentrations in the U.S. is complicated by "[c]urrent
limitations in climate modeling tools, variation across models, and the need for more
comprehensive observational data on these effects" (ISA, section IS.6.2.2). The service of carbon
storage is of paramount importance to the public welfare no matter in what location the trees are
growing or what their intended current or future use (e.g., 2013 ISA, section 9.4.1.2). In other
words, the benefit exists as long as the trees are growing, regardless of what additional functions
and services it provides.
30 While carbon sequestration or storage also occurs for vegetated ecosystems other than forests, it is relatively
larger in forests given the relatively greater biomass for trees compared to other plants.
4-29
-------
With regard to agriculture-related effects, the EPA has recognized other complexities
related to areas and plant species that are heavily managed to obtain a particular output (such as
commodity crops or commercial timber production). For example, the EPA has recognized that
the degree to which O3 impacts on vegetation that could occur in such areas and on such species
would impair the intended use at a level that might be judged adverse to the public welfare has
been less clear (80 FR 65379, October 26, 2015; 73 FR 16497, March 27, 2008). While having
sufficient crop yields is of high public welfare value, important commodity crops are typically
heavily managed to produce optimum yields. Moreover, based on the economic theory of supply
and demand, increases in crop yields would be expected to result in lower prices for affected
crops and their associated goods, which would primarily benefit consumers. These competing
impacts on producers and consumers complicate consideration of these effects in terms of
potential adversity to the public welfare (2014 WREA, sections 5.3.2 and 5.7). When agricultural
impacts or vegetation effects in other areas are contrasted with the emphasis on forest ecosystem
effects in Class I and similarly protected areas, it can be seen that the Administrator has in past
reviews judged the significance to the public welfare of 03-induced effects on sensitive
vegetation growing within the U.S. to differ depending on the nature of the effect, the intended
use of the sensitive plants or ecosystems, and the types of environments in which the sensitive
vegetation and ecosystems are located, with greater significance ascribed to areas identified for
specific uses and benefits to the public welfare, such as Class I areas, than to areas for which
such uses have not been established (80 FR 65292, October 26, 2015; FR 73 16496-16497,
March 27, 2008).
Categories of effects newly identified as likely causally related to O3 in ambient air, such
as alteration of plant-insect signaling and insect herbivore growth and reproduction, also have
potential public welfare implications. For example, given the role of plant-insect signaling in
such important ecological processes as insect herbivore growth and reproduction. The potential
to contribute to adverse effects to the public welfare, e.g., given the role of the plant-insect
signaling process in pollination and seed dispersal, as well as natural plant defenses against
predation and parasitism, particular effects on particular signaling processes can be seen to have
the potential for adverse effects on the public welfare (ISA, section IS.5.1.3).However,
uncertainties and limitations in the current evidence (e.g., as summarized in sections 4.3.3.3 and
4.3.4 below) preclude an assessment of the extent and magnitude of O3 effects on these
endpoints, which thus also precludes an evaluation of the potential for associated public welfare
implications, particularly under exposure conditions expected to occur in areas meeting the
current standard.
In summary, several considerations are recognized as important to judgments on the
public welfare significance of the array of effects of different O3 exposure conditions on
4-30
-------
vegetation. While there are uncertainties and limitations associated with the consideration of the
magnitude of key vegetation effects that might be concluded to be adverse to ecosystems and
associated services, there are numerous locations where the presence of Ch-sensitive tree species
may contribute to a vulnerability to impacts from O3 on tree growth, productivity and carbon
storage and their associated ecosystems and services. Exposures that may elicit effects and the
significance of the effects in specific situations can vary due to differences in exposed species
sensitivity, the severity and associated significance of the observed or predicted Ch-induced
effect, the role that the species plays in the ecosystem, the intended use of the affected species
and its associated ecosystem and services, the presence of other co-occurring predisposing or
mitigating factors, and associated uncertainties and limitations.
4-31
-------
Ill
(ft Ui
o5„
ur u. u)
&¦*¦* H ij
all
F m ~
O
a.
e Reductions in plait reproduction' and growth
. rref easeJ free , and related effects fin
,,4- Crop yield and quality
" - Terrestrial ecosystem productivity
- Ofw seci/esMicx: >r< terrestrial ecos /stms
; - Terrealrial community composition
| - Bel owg round biogeoc hemic atcycle*
; - Ecos,stew f,9ter o^dlrg
ir
Visible
foliar
injury
*
Effects on federally protected Class I and other areas, such as wilderness areas and lands provided similar protections by state,
tribal and othergroups.
• Impaimientofscenicvalue.naturalvegetationandwildlife
• Reduced aesthetic and outdoor recreation value
Aiierea ecosystem community composition
Changes to timber, fruit vegetable and fiber (for fabrics} production, including reductions in some spectficproduds/materiab
Reduced efccfweness of we species of trees/plants for cooling and beautleati® in residential are®
Erosion in populated areas
Reducedf ametish an-cJ recreational fisheries and other effects related to reduced late-season stream low
Climate-related impacts, such as net changes in agricultural procfuSiitf, property damage, and changes in beating/raoing costs
Non; Welsrt «flK& causaiy km » 0» exposure aw ifi ted text. tad My cans# tfeft ait teksed, Dashed wifere efects texts oomn ertjf Ik#/ causal etects. Doted
toes conned ftose etecs c aher sleds » whefcisy could ptS&ftialy cwtttome.
Figure 4-2. Potential effects of O3 on the public welfare.
4-32
-------
4.3.3 Exposures Associated with Effects
The types of effects identified in section 4.3.1 above vary widely with regard to the
extent and level of detail of the available information that describes the O3 exposure
circumstances that may elicit them. The discussion in this section is organized in recognition of
this. We focus first on growth and yield effects, a category of effects for which the information
on exposure metric and E-R relationships is most advanced (section 4.3.3.1). Section 4.3.3.2
discusses the current information regarding exposure metrics and relationships between exposure
and the occurrence and severity of visible foliar injury. The availability of such information for
other categories of effects is addressed in section 4.3.3.3.
4.3.3.1 Growth-related Effects
4.3.3.1.1 Exposure Metric
The longstanding body of vegetation effects evidence includes a wealth of information on
aspects of O3 exposure that are important in influencing effects on plant growth and yield (1996
AQCD; 2006 AQCD; 2013 ISA; 2020 ISA). A variety of factors have been investigated,
including "concentration, time of day, respite time, frequency of peak occurrence, plant
phenology, predisposition, etc." (2013 ISA, section 9.5.2), and the importance of the duration of
the exposure as well as the relatively greater importance of higher concentrations over lower
concentrations have been consistently well documented (2013 ISA, section 9.5.3). Based on the
associated improved understanding of the biological basis for plant response to O3 exposure, a
number of mathematical approaches have been developed for summarizing O3 exposure for the
purpose of assessing effects on vegetation, including those that cumulate exposures over some
specified period while weighting higher concentrations more than lower (2013 ISA, sections
9.5.2 and 9.5.3; ISA, Appendix 8, section 8.2.2.2).
In the last several reviews, based on the then-available evidence, as well as advice from
the CASAC, the EPA focused on the use of a cumulative, seasonal31 concentration-weighted
index for considering the growth-related effects evidence and in quantitative exposure analyses
for purposes of reaching conclusions on the secondary standard. More specifically, the Agency
used the W126-based cumulative, seasonal metric (80 FR 65404, October 26, 2015; ISA, section
IS.3.2, Appendix 8, section 8.13). This metric, commonly called the W126 index, is a non-
threshold approach described as the sigmoidally weighted sum of all hourly O3 concentrations
observed during a specified daily and seasonal time window, where each hourly O3 concentration
31 In describing the form as "seasonal," the EPA is referring generally to an index focused on a time period of a
duration that may relate to that of a growing season for Ch-sensitive vegetation, not to the seasons of the year
(spring, summer, fall, winter).
4-33
-------
is given a weight that increases from zero to one with increasing concentration (2013 ISA, p. 9-
101).
Across the last several reviews of the O3 NAAQS, several different exposure metrics
have been evaluated, primarily for their ability to summarize ambient air concentrations in a way
that best correlates with effects on vegetation, particularly growth-related effects. Based on
extensive review of the published literature on different types of E-R metrics, including
comparisons between metrics, the EPA has generally focused on cumulative, concentration-
weighted indices of exposure, recognizing them as the most appropriate biologically based
metrics to consider in this context (1996 AQCD; 2006 AQCD; 2013 ISA).32 Quantifying
exposure in this way has been found to improve the explanatory power of E-R models for growth
and yield over using indices based only on mean and peak exposure values (2013 ISA, section
2.6.6.1, p. 2-44). The most well-studied datasets in this regard are those for 11 tree species
seedlings and 10 crops referenced above and described further in section 4.3.3.2 below (e.g., Lee
and Hogsett, 1996, Hogsett et al., 1997). The most detailed and well analyzed information in this
regard are two datasets established two decades ago (and described in section 4.3.3.1.2 below),
one for growth effects on seedlings of a set of tree species and the second for quality and yield
effects for a set of crops. These datasets, which include growth and yield response information
across a range of multiple seasonal cumulative exposures, were used to develop robust
quantitative E-R functions for reduced growth (termed relative biomass loss or RBL) in
seedlings of the tree species and E-R functions for RYL for a set of common crops (ISA,
Appendix 8, section 8.13.2; 2013 ISA, section 9.6.2). The EPA's conclusions regarding exposure
levels of O3 associated with vegetation-related effects at the time of the last review were based
primarily on these established E-R functions.
Along with the continuous weighted, W126 index, two other cumulative indices that have
received greatest attention across the past several O3 NAAQS reviews have been the threshold
32 The Agency has focused its analyses in the last several reviews on metrics that characterize cumulative exposures
over a season or seasons: SUM06 in the 1997 review (61 FR 65716, December 13, 1996; 62 FR 38856, July 18,
1997) and W126 in both the 2008 and 2015 reviews (72 FR 37818, July 11, 2007; 73 FR 16436, March 27, 2008;
80 FR 65373-65374, October 26, 2015). This approach to characterizing O3 exposure concentrations with regard
to potential vegetation effects, particularly growth, has received strong support from CASAC in the past two
reviews (Henderson, 2006; Samet, 2010; Frey, 2014).
4-34
-------
weighted indices, AOT6033 and SUM06.34 Accordingly, some studies of O3 vegetation effects
have reported exposures using these metrics.
Alternative methods for characterizing O3 exposure to predict various plant responses
have, in recent years, included flux models (models that are based on the amount of O3 that
enters the leaf). However, as was the case in the last review, there remain a variety of
complications, limitations and uncertainties associated with this approach. For example, "[w]hile
some efforts have been made in the U.S. to calculate ozone flux into leaves and canopies, little
information has been published relating these fluxes to effects on vegetation" (ISA, section
IS.3.2). Further, as flux of O3 into the plant under different conditions of O3 in ambient air is
affected by several factors including temperature, vapor pressure deficit, light, soil moisture, and
plant growth stage, use of this approach to quantify the vegetation impact of O3 would require
information on these various types of factors (ISA, section IS.3.2). In addition to these data
requirements, each species has different amounts of internal detoxification potential that may
protect species to differing degrees. The lack of detailed species- and site-specific data required
for flux modeling in the U.S. and the lack of understanding of detoxification processes continues
to make this technique less viable for use in risk assessments in the U.S. (ISA, section IS.3.2).
Among the studies newly available since the last review, no new exposure indices for
assessing effects on vegetation growth or other physiological process parameters have been
identified. In the literature available since the 2013 ISA, the SUM06, AOTx (e.g., AOT60) and
W126 exposure metrics remain the metrics that are most commonly discussed (ISA, Appendix 8,
section 8.13.1). The ISA notes that "[c]umulative indices of exposure that differentially weight
hourly concentrations [which would include the W126 index] have been found to be best suited
to characterize vegetation exposure to ozone with regard to reductions in vegetation growth and
yield" (ISA, section ES.3). Accordingly, in this review, as in the last two reviews, we use the
seasonal W126-based cumulative, concentration-weighted metric for consideration of the effects
evidence and quantitative exposure analyses, particularly related to growth effects (as
summarized in sections 4.3.3.2 and 4.4 below).
33 The AOT60 index is the seasonal sum of the difference between an hourly concentration above 60 ppb, minus 60
ppb (2006 AQCD, p. AX9-161). More recently, some studies have also reported O3 exposures in terms of
AOT40, which is conceptually similar but with 40 substituted for 60 in its derivation (ISA, Appendix 8, section
8.13.1).
34 The SUM06 index is the seasonal sum of hourly concentrations at or above 0.06 ppm during a specified daily time
window (2006 AQCD, p. AX9-161; 2013 ISA, section 9.5.2). This may sometimes be referred to as SUM60, e.g.,
when concentrations are in terms of ppb. There are also variations on this metric that utilize alternative reference
points above which hourly concentrations are summed. For example, SUM08 is the seasonal sum of hourly
concentrations at or above 0.08 ppm and SUMO is the seasonal sum of all hourly concentrations.
4-35
-------
The first step in calculating the seasonal W126 index for a specific year, as described and
considered in this review, is to sum the weighted hourly O3 concentrations in ambient air during
daylight hours (defined as 8:00 a.m. to 8:00 p.m. local standard time) within each calendar
month, resulting in monthly index values. The monthly W126 index values are calculated from
hourly O3 concentrations as follows.35
Monthly W126 = Zdt=a Zh=8: Cdh
' l+4403*exp (-126*Cdh)
where,
Nis the number of days in the month
dis the day of the month (d = 1, 2, ..., N)
his the hour of the day (h = 0, 1, ..., 23)
Cdh is the hourly O3 concentration observed on day d, hour h, in parts per million
The W126 index value for a specific year is the maximum sum of the monthly index values for
three consecutive months within a calendar year (i.e., January to March, February to April, ...
October to December). Three-year average W126 index values are calculated by taking the
average of seasonal W126 index values for three consecutive years (e.g., as described in
Appendix 4D, section 4D.2.2).
4.3.3.1.2 Relationships Between Exposure Levels and Effects
Across the array of 03-related welfare effects, consistent and systematically evaluated
information on E-R relationships across multiple exposure levels is limited. Most prominent is
the information on E-R relationships for growth effects on tree seedlings and crops,36 which has
been available for the past several reviews. The information on which these functions are based
comes primarily from the U.S. EPA's National Crop Loss Assessment Network (NCLAN)37
project for crops and the NHEERL-WED project for tree seedlings, projects implemented
primarily to define E-R relationships for major agricultural crops and tree species, thus
advancing understanding of responses to O3 exposures (ISA, Appendix 8, section 8.13.2). These
projects included a series of experiments that used OTCs to investigate tree seedling growth
response and crop yield over a growing season under a variety of O3 exposures and growing
conditions (2013 ISA, section 9.6.2; Lee and Hogsett, 1996). These experiments have produced
35 In situations where data are missing, an adjustment is factored into the monthly index (as described in Appendix
4D, section 4D.2.2).
36 The E-R functions estimate 03-related reduction in a year's tree seedling growth or crop yield as a percentage of
that expected in the absence of O3 (Appendix 4A; ISA, Appendix 8, section 8.13.2).
37 The NCLAN program, which was undertaken in the early to mid-1980s, assessed multiple U.S. crops, locations,
and O3 exposure levels, using consistent methods, to provide the largest, most uniform database on the effects of
O3 on agricultural crop yields (1996 AQCD, 2006 AQCD, 2013 ISA, sections 9.2, 9.4, and 9.6; ISA, Appendix 8,
section 8.13.2).
4-36
-------
multiple studies that document O3 effects on tree seedling growth and crop yield across multiple
levels of exposure. Importantly, the information on exposure includes hourly concentrations
across the season (or longer) exposure period which can then be summarized in terms of the
various seasonal metrics.38 In the initial analyses of these data, exposure was characterized in
terms of several metrics, including seasonal SUM06 and W126 indices (Lee and Hogsett, 1996;
1997 Staff Paper, sections IV.D.2 and IV.D.3; 2007 Staff Paper, section 7.6;), while use of these
functions in the last review focused on their implementation in terms of seasonal W126 index
(2013 ISA, section 9.6; 80 FR 65391-92, October 26, 2015). This information for seedlings of
the 11 tree species, in combination with air quality analyses, was a key consideration in the 2015
decision on the level for the revised secondary standard (80 FR 65292, October 26, 2015).
The 11 species for which robust and well-established E-R functions for RBL are
available are black cherry, Douglas fir, loblolly pine, ponderosa pine, quaking aspen, red alder,
red maple, sugar maple, tulip poplar, Virginia pine, and white pine (Figure 4-3; Appendix 4A;
2013 ISA, section 9.6).39 While these 11 species represent only a small fraction of the total
number of native tree species in the contiguous U.S., this small subset includes eastern and
western species, deciduous and coniferous species, and species that grow in a variety of
ecosystems and represent a range of tolerance to O3 (Appendix 4B; 2013 ISA, section 9.6.2). The
established E-R functions for most of the 11 species were derived using data from multiple
studies or experiments involving a wide range of exposure and/or growing conditions. From the
available data, separate E-R functions were developed for each combination of species and
experiment (2013 ISA, section 9.6.1; Lee and Hogsett, 1996). From these separate species-
experiment-specific E-R functions, species-specific composite E-R functions were developed
(Appendix 4A). Biomass growth loss predictions using these functions were evaluated in the ISA
for the last review based on a recent study for aspen (2013 ISA, section 9.6.2; ISA, Appendix 8,
section 8.13.2).
38 This underlying database for the exposure is a key characteristic that sets this set of studies (and their associated
E-R analyses) apart from other available studies.
39 A quantitative analysis of E-R information for an additional species was considered in the 2014 WREA. But the
underlying study, rather than being an OTC controlled exposure study, involves exposure to ambient air along an
existing gradient of O3 concentrations in the New York City metropolitan area, such that O3 and climate
conditions were not controlled (2013 ISA, section 9.6.3.3). Based on comments from the CASAC on the WREA
cautioning against placing too much emphasis on these data (e.g., saying that the eastern cottonwood response
data from a single study "receive too much emphasis," explaining that these "results are from a gradient study
that did not control for ozone and climatic conditions and show extreme sensitivity to ozone compared to other
studies" and that" [although they are important results, they are not as strong as those from other experiments
that developed E-R functions based on controlled ozone exposure") (Frey, 2014, p. 10), the EPA did not include
the E-R function for eastern cottonwood among the set of tree seedling E-R functions given focus in the WREA,
or relied on in decision-making for the last review (80 FR 65292, October 26, 2015.)
4-37
-------
The 11 species-specific composite median functions are described in Appendix 4A (see
section 4 A. 1.1). For some of these species, the E-R function is based on a single study (e.g., red
maple), while for other species there were as many as 11 studies available (e.g., ponderosa pine).
In total, the 11 species-specific E-R functions are based on 51 tree seedling studies or
experiments. A stochastic analysis performed for the 2014 WREA provides a sense of the
variability and uncertainty associated with the estimated E-R relationships among and within
species (Appendix 4A, section 4 A. 1.1, Figure 4A-13). Based on the species-specific E-R
functions, the studied tree species appear to vary widely in sensitivity to reduced growth at the
seedling stage (Figure 4-3). Since the initial set of studies were completed, several additional
studies, focused on aspen, have been published based on the Aspen FACE experiment in a
planted forest in Wisconsin; the findings were consistent with many of the QTC studies (ISA,
Appendix 8, section 8.13.2).
With regard to crops, established E-R functions are available for 10 crops: barley, field
corn, cotton, kidney bean, lettuce, peanut, potato, grain sorghum, soybean and winter wheat
(Figure 4-4; Appendix 4A; ISA, Appendix 8, section 8.13.2). Studies available in the last review
increased our confidence in the predictability of the crop E-R functions. Since then, new
evidence is available for seven soybean cultivars that confirms the reliability of the soybean E-R
functions developed from NCLAN data and indicates that they extend in applicability to recent
cultivars (ISA, Appendix 8, section 8.13.2). In the last review, these E-R functions were used to
characterize the estimated growth reduction across the studied species for a range of seasonal
W126 index exposures (Appendix 4A, section 4A.1).
o
Red Maple
• Sugar Maple
• Red Alder
Tulip Poplar
Ponderosa Pine
• White Pine
• Loblolly Pine
Virginia Pine
• Aspen
• Black Cherry
• Douglas Fir
ID
o
_i
m
DC
o
CM
d
o
o
0
10
20
30
40
50
W126 (ppm-hrs)
Figure 4-3. Established RBL functions for seedlings of 11 tree species.
4-38
-------
o
00
o
Barley
• Field Corn
• Cotton
Kidney Bean
Lettuce
>-
O
o
CD
o
o
c\j
o
• Peanut
• Potato
Grain Sorghum
• Soybean
• Winter Wheat
0
10 20 30 40 50
W126 (ppm-hrs)
Figure 4-4. Established RYL functions for 10 crops.
Newly available studies that investigated growth effects of O3 exposures are also
consistent with the existing evidence base, and generally involved particular aspects of the effect
rather than expanding the conditions under which plant species, particularly trees, have been
assessed (ISA, section IS.5.1.2). The ISA notes the recent availability of a compilation of
previously available studies on plant biomass response to O3 (in terms of AOT40); the
compilation reports linear regressions conducted on the associated varying datasets (ISA,
Appendix 8, section 8.13.2). Little is presented in this compilation with regard to evaluation of
consistent and similar O3 exposure and biomass response measurements, and the exposure
durations, which are not reported for each study, are reported to vary, with the shortest being 21
days (van Goethem et al., 2013).40 These aspects of the publication limit its usefulness with
regard to describing E-R relationships that might provide for estimation of specific impacts
associated with air quality conditions meeting the current standard. As was noted in the 2013
ISA, "[i]n order to support quantitative modeling of exposure-response relationships, data should
preferably include more than three levels of exposure, and some control of potential confounding
40 The set of studies included in this compilation were described as meeting a set of criteria, such as: including O3
only exposures in conditions described as "close to field" exposures (which were expressed as AOT40);
including at least 21 days exposure above 40 ppb O3; and, having a maximum hourly concentration that was no
higher than 100 ppb (van Goethem et al., 2013).
4-39
-------
or interacting factors should be present in order to model the relationship with sufficient
accuracy" (2013 ISA, p. 9-118). The 2013 ISA further discussed the differences across available
studies, recognizing that the majority of studies contrast only two (or sometimes three with the
addition of a carbon filtration) O3 exposure levels. While such studies can be important for
verifying more extensive studies, they "do not provide exposure-response information that is
highly relevant to reviewing air quality standards" (2013 ISA, p. 9-118).
4.3.3.2 Visible Foliar Injury
With regard to visible foliar injury, as with the evidence available in the last review, the
evidence newly available in this review "continues to show a consistent association between
visible injury and ozone exposure," while also recognizing the role of modifying factors such as
soil moisture and time of day (ISA, section IS.5.1.1). The ISA, in concluding that the newly
available information is consistent with conclusions of the 2013 ISA, also summarizes several
recently available studies that continue to document that O3 elicits visible foliar injury in many
plant species, including a synthesis of previously published studies that categorizes studied
species (and their associated taxonomic classifications) as to whether or not Ch-related foliar
injury has been reported. Although this recent publication identifies many species in which
visible foliar injury has been documented to occur in the presence of elevated O3,41 it does not
provide quantitative information regarding specific exposure conditions or analyses of E-R
relationships (ISA, Appendix 8, section 8.3). Additionally, one recent study is identified as
reporting visible foliar injury in a non-native, yet established, and invasive tree species in a
location with O3 concentrations corresponding to a seasonal W126 index of 11.6 ppm-hrs (ISA,
Appendix 8, sections 8.2 and 8.2.1). The annual fourth highest 8-hour daily maximum
concentration for the study year and location of this study (monitoring site 42-027-9991) is 76
ppb. The design value for the 3-year design period encompassing the year and location of this
study exceeds 70 ppb (monitoring site 42-027-9991 for 2011-2013 design period), indicating that
the air quality associated with the exposure would not have met the current secondary standard.42
The evidence in the current review, as was the case in the last review, while documenting
that elevated O3 conditions in ambient air generally results in visible foliar injury in sensitive
41 The publication identifies 245 species across 28 plant genera, many native to the U.S., in which 03-related visible
foliar injury has been reported (ISA, Appendix 8, section 8.3).
42 Ozone design values for this period are available at: https://www.epa.gov/air-trends/air-quality-design-values. The
year 2011 is the first year for which data are available and adequate for use in deriving a design value at this
monitoring site.
4-40
-------
species (when in a predisposing environment)43, does not include a quantitative description of
the relationship of incidence or severity of visible foliar injury in sensitive species in natural
locations in the U.S. with specific metrics of O3 exposure. Several studies of the extensive USFS
field-based dataset of visible foliar injury incidence in forests across the U.S.44 illustrate the
limitations of current understanding of this relationship. For example, a study that was available
in the last review presents a trend analysis of these data for sites located in 24 states of the
northeast and north central U.S. for the 16-year period from 1994 through 2009 that provides
some insight into the influence of changes in air quality and soil moisture on visible foliar injury
and the difficulty inherent in predicting foliar injury response under different air quality and soil
moisture scenarios (Smith, 2012, Smith et al., 2012; U.S. EPA, 2018; ISA, Appendix 8, section
8.2). This study, like prior analyses of such data, shows the dependence of foliar injury incidence
and severity on local site conditions for soil moisture availability and O3 exposure. For example,
while the authors characterize the ambient air O3 concentrations to be the "driving force" behind
incidence of injury and its severity, they state that "site moisture conditions are also a very strong
influence on the biomonitoring data" (Smith et al., 2003). In general, the USFS data analyses
have found foliar injury prevalence and severity to be higher during seasons and sites that have
experienced the highest O3 than during other periods (e.g., Campbell et al., 2007; Smith, 2012).
Studies of the incidence of visible foliar injury in national forests, wildlife refuges, and
similar areas have often used cumulative indices such as SUM06 to investigate variations in
incidence of foliar injury (e.g., Hildebrand et al., 1996). For example, a study of six years of
USFS biosite data for three western states found that the biosites with the highest O3 exposure
(SUM06 at or above 25 ppm-hrs) had the highest percentage of biosites with injury and the
highest mean biosite index, with little discernable difference among the lower exposure
categories; this study also identified "better linkage between air levels and visible injury" as an
O3 research need (Campbell et al., 2007). More recent studies of the complete 16 years of data in
24 northeast and north central states have suggested that a cumulative exposure index alone may
not completely describe the 03-related risk of this effect (Smith et al., 2012; Smith, 2012). For
example, Smith (2012) observed there to be a declining trend in the 16-year dataset, "especially
43 As noted in the 2013 ISA and the ISA for the current review, visible foliar injury usually occurs when sensitive
plants are exposed to elevated ozone concentrations in a predisposing environment, with a major modifying factor
being the amount of soil moisture available to a plant. Accordingly, diy periods are concluded to decrease the
incidence and severity of ozone-induced visible foliar injury, such that the incidence of visible foliar injury is not
always higher in years and areas with higher ozone, especially with co-occurring drought (ISA, Appendix 8, p. 8-
23; Smith, 2012; Smith et al., 2003).
44 These data were collected as part of the U.S. Forest Service Forest Health Monitoring/Forest Inventory and
Analysis (USFS FHM/FIA) biomonitoring network program (2013 ISA, section 9.4.2.1; Campbell et al., 2007,
Smith et al., 2012).
4-41
-------
after 2002 when peak ozone concentrations declined across the entire region" thus suggesting a
role for peak concentrations.
Some studies of visible foliar injury incidence data have investigated the role of peak
concentrations quantified by an O3 exposure index that is a count of hourly concentrations (e.g.,
in a growing season) above a threshold 1-hour concentration of 100 ppb, N100 (e.g., Smith,
2012; Smith et al., 2012). For example, the study by Smith (2012) discussed injury patterns at
biosites in 24 states in the Northeast and North Central regions in the context of the SUM06
index and N100 metrics (although not in statistical combination).45 That study of 16 years of
biomonitoring data from these sites suggested that there may be a threshold exposure needed for
injury to occur, and the number of hours of elevated O3 concentrations during the growing
season (such as what is captured by a metric like N100) may be more important than cumulative
exposure in determining the occurrence of foliar injury (Smith, 2012).46 The study's authors
noted this finding to be consistent with findings reported by a study of statistical analyses of
seven years of visible foliar injury data from a wildlife refuge in the mid-Atlantic (Davis and
Orendovici, 2006, Smith et al., 2012). The latter study investigated the fit of multiple models that
included various metrics of cumulative O3 (e.g., SUM06, SUMO, SUM08), alone and in
combination with some other variables (Davis and Orendovici, 2006). Among the statistical
models investigated, the model with the best fit to the visible foliar injury incidence data was
found to be one that included N100 and W126 indices, as well as drought index (Davis and
Orendovici, 2006).47
The 2013 ISA and 2006 AQCD noted the established significant role of higher or peak
O3 concentrations, as well as pattern of their occurrence, in plant responses. In identifying
support with regard to foliar injury as the response, these assessments both cite studies that
support the "important role that peak concentrations, as well as the pattern of occurrence, plays
in plant response to O3" (2013 ISA, p. 9-105; 2006 AQCD, p. AX9-169). For example, a study
of European white birch saplings reported that peak concentrations and the duration of the
exposure event were important determinants of foliar injury (2013 ISA, section 9.5.3.1; Oksanen
and Holopainen, 2001). This study also evaluated tree growth, which was found to be more
45 The current ISA, 2013 ISA and prior AQCDs have not described extensive evaluation of specific peak-
concentration metrics such as the N100 that might assist in identifying the one best suited for such purposes.
46 In summarizing this study in the last review, the ISA observed that "[o]verall, there was a declining trend in the
incidence of foliar injury as peak 03 concentrations declined" (2013 ISA, p. 9-40).
47 The models evaluated included several with cumulative exposure indices alone. These included SUM60, SUMO,
and SUM80. They did not include a model with W126 that did not also include N100. Across all of these models,
the model with the best fit to the data was found to be the one that included N100 and W126, along with the
drought index (Davis and Orendovici, 2006).
4-42
-------
related to cumulative exposure (2013 ISA, p. 9-105).48 A second study that was cited by both
assessments that focused on aspen, reported that "the variable peak exposures were important in
causing injury, and that the different exposure treatments, although having the same SUM06,
resulted in very different patterns of foliar injury (2013 ISA, p. 9-105; 2006 AQCD, p. AX9-169;
Yun and Laurence, 1999). As noted in the 2006 AQCD, the cumulative exposure indices (e.g.,
SUM06, W126) were "originally developed and tested using only growth/yield data, not foliar
injury" and " [t]his distinction is critical in comparing the efficacy of one index to another" (2006
AQCD, p. AX9-173). It is also recognized that where cumulative indices are highly correlated
with the frequency or occurrence of higher hourly average concentrations, they could be good
predictors of such effects (2006 AQCD, section AX9.4.4.3).
In a more recent study that is cited in the current ISA a statistical modeling analysis was
performed on a subset of the years of data that were described in Smith (2012). This analysis,
which involved 5,940 data records from 1997 through 2007 from the 24 northeast and north
central states, tested a number of models for their ability to predict the presence of visible foliar
injury (a nonzero biosite score), regardless of severity, and generally found that the type of O3
exposure metric (e.g., SUM06 versus N100) made only a small difference, although the models
that included both a cumulative index (SUM06) and N100 had a just slightly better fit (Wang et
al., 2012). Based on their investigation of 15 different models, using differing combination of
several types of potential predictors, the study authors concluded that they were not able to
identify environmental conditions under which they "could reliably expect plants to be damaged"
(Wang et al., 2012). This is indicative of the current state of knowledge, in which there remains a
lack of established quantitative functions describing E-R relationships that would allow
prediction of visible foliar injury severity and incidence under varying air quality and
environmental conditions.
The available information related to O3 exposures associated with visible foliar injury of
varying severity also includes the dataset developed by the EPA in the last review from USFS
biosite index (BI) scores, collected during the years 2006 through 2010 at locations in 37 states,
that were combined with estimates of soil moisture49 and estimates of seasonal cumulative O3
48 The study authors concluded that "high peak concentrations were important for visible injuries and stomatal
conductance, but less important for determining growth responses" (Oksanen and Holopainen, 2001).
49 Soil moisture categories (dry, wet or normal) were assigned to each biosite record based on the NOAA Palmer Z
drought index values obtained from the NCDC website for the April-through-August periods, averaged for the
relevant year; details are provided in Appendix AC, section 4C.2. There are inherent uncertainties in this
assignment, including the substantial spatial variation in soil moisture and large size of NOAA climate divisions
(hundreds of miles). Uncertainties and limitations in the dataset are summarized in Appendix AC, section 4C.5).
4-43
-------
exposure in terms of W126 index50 (Smith and Murphy, 2015; Appendix 4C). This dataset
includes more than 5,000 records of which more than 80 percent have a BI score of zero
(indicating a lack of visible foliar injury).51 While the estimated W126 index assigned to records
in this dataset (described in Appendix 4C) ranges from zero to somewhat above 50 ppm-hrs, only
8% of the records have W126 index values above 15 ppm-hrs. Beyond an analysis included from
the last review, the presentations in Appendix 4C are primarily descriptive (as compared to
statistical analysis); this is in recognition of the limitations and uncertainties of the dataset as
summarized in Appendix 4C, section 4C.5. The presentation in Appendix 4C describes the BI
scores for the records in the dataset in relation to the W126 index estimate for each record, using
bins of increasing W126 index values. The presentation indicates the occurrence of injury (and in
severity) across the W126 index bins to be variable, and also finds the greatest incidence of
records with BI scores above zero, five, or higher to occur for records with the highest W126
index values (i.e., the bin for W126 index estimates greater than 25 ppm-hrs), as seen in Figure
4-5 for records in the normal soil moisture category52 (see also Appendix 4C, Table 4C-6).
The average BI score per W126 index bin is also variable, although for records
categorized as normal soil moisture, the average BI score in the highest W126 bin is noticeably
greater than for lower W126 bin scores (Figure 4-5). For example, the average BI score for the
normal soil moisture category is 7.9 among records with W126 index estimates greater than 25
ppm-hrs, compared to 1.6 among records for W126 index estimates between 19 and 25 ppm-hrs.
For records categorized as wet soil moisture, the sample size for the W126 bins above 13 ppm-
hrs is quite small (including only 18 of the 1,189 records in that soil moisture category),
precluding meaningful interpretation.53
50 The W126 index values assigned to the biosite locations are estimates developed for 12 kilometer (km) by 12 km
cells in a national-scale spatial grid for each year. The grid cell estimates were derived from applying a spatial
interpolation technique to annual W126 values derived from O3 measurements at ambient air monitoring locations
for the years corresponding to the biosite surveys (details in Appendix AC, sections 4.C.2 and 4C.5).
51 In the scheme used by the USFS to categorize severity of biosite scores the lowest category encompasses BI
scores from zero to just below 5; scores of this magnitude are described as "little or no foliar injury" (Smith et al.,
2012). The next highest category encompasses scores from five to just below 15 and is described as "light to
moderate foliar injury," BI scores of 15 up to 25 are described as "moderate" and above 25 is described as
"severe" (Smith, 2012; Smith et al., 2012)..
52 The number of records per W126 bin in Figure 4-5 ranges from a low of 15 in the ">19-25" bin to 158 in the "<7"
bin (Appendix 4C, Table 4C-4).
53 The full database includes only 18 records at sites in the wet soil moisture category and a W126 index value above
13 ppm-hrs, with 9 or fewer (less than 1%) in each of the W126 bins above 13 ppm-hrs (Appendix AC, Table 4C-
3). Across the W126 bins in which at least 1% of the wet soil moisture records are represented, differences of
incidence or average score of lower bins from the highest bin is less than a factor of two (Appendix AC, section
4C.4.2).
4-44
-------
While for BI scores above zero, the data may indicate a suggestion of increased incidence
among records in the W126 bins just below the highest (e.g., for the dry or normal soil moisture
categories), for BI scores above 5, there is little or no difference across the W126 bins except for
the highest bin, which is for W126 above 25 ppm-hrs (Appendix 4C, Table 4C-6). For example,
among records in the normal soil category, the proportion of records with BI above five
fluctuates between 5% and 13% across all but the highest W126 bin (>25 ppm-hrs) for which the
proportion is 41% (Appendix 4C, Table 4C-6). The same pattern is observed for BI scores above
15 at sites with normal and dry soil moisture conditions, albeit with lower incidences. For
example, the incidence of normal soil moisture records with BI score above 15 in the bin for
W126 index values above 25 ppm-hrs was 20% but fluctuates between 1% and 4% in the bines
with W126 index values at or below 25 ppm-hrs (Appendix 4C, Table 4C-6).
£
o
o
>7-9 >9-11 >11-13 >13-15 >15-17 >17-19 >19-25 >25
<7
Key: The boxes denote the 25th, 50th and 75th percentiles, the x's the arithmetic mean, and the whiskers
denote the value equal to the 75th percentile plus 1.5 times the interquartile range (75th minus 25th percentile).
Circles show scores higher than that.
Figure 4-5. Distribution of nonzero BI scores at USFS biosites (normal soil moisture)
grouped by assigned W126 index estimates.
Overall, the dataset described in Appendix 4C generally indicates the risk of injury, and
particularly light, moderate or greater injury, to be higher at the highest W126 index values, with
appreciable variability in the data for the lower bins. A number of factors may contribute to the
observed variability in BI scores and lack of a clear pattern with W126 index bin; among others,
these may include uncertainties in assignment of W126 estimates and soil moisture categories to
biosite locations, variability in biological response among the sensitive species monitored, and
the potential role of other aspects of O3 air quality not captured by the W126 index. This appears
4-45
-------
to be consistent with the conclusions of the studies of detailed quantitative analyses, summarized
above, that the pattern is stronger at higher O3 concentrations while uncertainty remains
regarding the tools for and the appropriate metric for quantifying influence of O3 exposures, as
well as perhaps soil moisture conditions (Davis and Orendovici, 2006, Smith et al., 2012; Wang
et al., 2012). Thus, the limitations recognized in the last review remain in our ability to
quantitatively estimate incidence and severity of visible foliar injury likely to occur in areas
across the U.S. under different air quality conditions over a year, or over a multi-year period
(Appendix 4C, section 4C.5).
Dose modeling or flux models, discussed in section 4.3.3.1.1 above, have also been
considered for quantifying O3 dose that may be related to plant injury. Among the newly
available evidence is a study examining relationships between short-term flux and leaf injury on
cotton plants that described a sensitivity parameter that might characterize the influence on the
flux-injury relationship of diel and seasonal variability in plant defenses (among other factors)
and suggested additional research might provide for such a sensitivity parameter to "function
well in combination with a sigmoidal weighting of flux, analogous to the W126 weighting of
concentration", and perhaps an additional parameter (Grantz et al., 2013, p. 1710; ISA, Appendix
8, section 8.13.1). However, the ISA recognizes there is "much unknown" with regard to the
relationship between O3 uptake and leaf injury, and relationships with detoxification processes
(ISA, Appendix 8, section 8.13.1 and p. 8-184). These uncertainties have made this technique
less viable for assessments in the U.S., precluding use of a flux-based approach at this time (ISA,
Appendix 8, section 8.13.1 and p. 8-184).
4.3.3.3 Other Effects
With regard to radiative forcing and subsequent climate effects associated with the global
tropospheric abundance of O3, the newly available evidence in this review does not provide more
detailed quantitative information regarding O3 concentrations at the national scale. Although
tropospheric O3, at the global scale, continues to be recognized as having a causal relationship
with radiative forcing and a likely causal relationship with subsequent effects on temperature,
precipitation and related climate variables, the non-uniform distribution of O3 (spatially and
temporally) makes the development of quantitative relationships between the magnitude of such
effects and differing O3 concentrations in the U.S. challenging (ISA, Appendix 9). Additionally,
"the heterogeneous distribution of ozone in the troposphere complicates the direct attribution of
spatial patterns of temperature change to ozone induced [radiative forcing]" and there are "ozone
climate feedbacks that further alter the relationship between ozone [radiative forcing] and
temperature (and other climate variables) in complex ways" (ISA, Appendix 9, section 9.3.1, p.
9-19). Thus, the ISA recognizes that "[c]urrent limitations in climate modeling tools, variation
4-46
-------
across models, and the need for more comprehensive observational data on these effects
represent sources of uncertainty in quantifying the precise magnitude of climate responses to
ozone changes, particularly at regional scales" (ISA, section IS.6.2.2, p. 9-22). While these
complexities impede our ability to consider specific O3 concentrations in the U.S. with regard to
specific magnitudes of impact on radiative forcing and subsequent climate effects, we note that
our ability to estimate growth-related impacts of trees can also inform our consideration of the
sequestration of carbon in terrestrial ecosystems, a process that can reduce tropospheric
abundance of CO2, the pollutant ranked first in importance as a greenhouse gas and radiative
forcing agent.
With regard to the two newly identified categories of effects, there are multiple
limitations and uncertainties regarding characterization of exposure conditions that might elicit
effects and the comprehensive characterization of the effects. For example, with regard to
alteration of herbivore growth and reproduction, although "there are multiple studies
demonstrating ozone effects on fecundity and growth in insects that feed on ozone-exposed
vegetation", "no consistent directionality of response is observed across studies and
uncertainties remain in regard to different plant consumption methods across species and the
exposure conditions associated with particular severities of effects " (ISA, pp. ES-18, IS-64, IS-
91 and Appendix 8, section 8.6.3). Such limitations and uncertainties in the evidence base for
this category of effects preclude broader characterization, as well as quantitative analysis related
to air quality conditions meeting the O3 standard. As characterized in the ISA, uncertainties
remain in the evidence; these relate to the different plant consumption methods across species
and the exposure conditions associated with particular responses, as well as variation in study
designs and endpoints used to assess O3 response (ISA, IS.6.2.1 and Appendix 8, section 8.6).
Thus, while the evidence describes changes in nutrient content and leaf chemistry following O3
exposure (ISA, p. IS-73), the effect of these changes on herbivores consuming the leaves is not
well characterized or clear.
The evidence for a second newly identified category of effects, alteration of plant-insect
signaling, draws on new research that has provided clear evidence of O3 modification of VPSCs
and behavioral responses of insects to these modified chemical signals. Most of these studies,
however, have been carried out in laboratory conditions rather than in natural environments, and
involve a relatively small number of plant species and plant-insect associations. While the
evidence documents effects on plant production of signaling chemicals and on the atmospheric
persistence of signaling chemicals, as well as on the behaviors of signal-responsive insects, it is
limited with regard to characterization of mechanisms and the consequences of any modification
of VPSCs by O3 (ISA, section IS.6.2.1) Further, the available studies vary with regard to the
experimental exposure circumstances in which the different types of effects have been reported
4-47
-------
(most of the studies have been carried out in laboratory conditions rather than in natural
environments), and many of the studies involve quite short controlled exposures (hours to days)
to elevated concentrations, posing limitations for our purposes of considering the potential for
impacts associated with the studied effects to be elicited by air quality conditions that meet the
current standard (ISA, section IS.6.2.1 and Appendix 8, section 8.7).
With regard to previously recognized categories of vegetation-related effects, other than
growth and visible foliar injury, such as reduced plant reproduction, reduced productivity in
terrestrial ecosystems, alteration of terrestrial community composition and alteration of below-
ground biogeochemical cycles, the newly available evidence includes a variety of studies, as
identified in the ISA (ISA, Appendix 8, sections 8.4, 8.8 and 8.10). Across the studies, a variety
of metrics (including AOT40, 4- to 12-hour mean concentrations, and others) are used to
quantify exposure over varying durations and various countries. The ISA additionally describes
publications that summarize previously published studies in several ways. For example, a meta-
analysis of reproduction studies categorized the reported O3 exposures into bins of differing
magnitude, grouping differing concentration metrics and exposure durations together, and
performed statistical analyses to reach conclusions regarding the presence of an Ch-related effect
(ISA, Appendix 8, section 8.4.1). While such studies continue to support conclusions of the
ecological hazards of O3, they do not improve capabilities for characterizing the likelihood of
such effects under varying patterns of environmental O3 conditions that occur under the current
standard.
As at the time of the last review, growth impacts, most specifically as evaluated by RBL
for tree seedlings and RYL for crops, remain the type of vegetation-related effects for which we
have the best understanding of exposure conditions likely to elicit them. Thus, as was the case in
the decision for the last review, the quantitative analyses of exposures occurring under air quality
that meets the current standard (summarized in section 4.4 below) is focused primarily on the
W126 index, given its established relationship with growth effects.
4.3.4 Key Uncertainties
The type of uncertainties for each category of effects tends to vary in relation to the
maturity of the associated evidence base from those associated with overarching
characterizations of the effects to those associated with quantification of the cause and effect
relationships. For example, given the longstanding nature of the evidence for many of the
vegetation effects identified in the ISA as causally or likely causally related to O3 in ambient air,
the key uncertainties and limitations in our understanding of these effects relate largely to the
implications or specific aspects of the evidence, as well as to current understanding of the
quantitative relationships between O3 concentrations in the environment and the occurrence and
4-48
-------
severity (or relative magnitude) of such effects or understanding of key influences on these
relationships. For more newly identified categories of effects, the evidence may be less
extensive, thus precluding consideration of such details.
• What are important uncertainties in the evidence? To what extent have important
uncertainties in the evidence identified in the last review been reduced and/or have
new uncertainties been recognized?
Among the categories of effects identified in past reviews, key uncertainties remain in the
current evidence. The category of O3 welfare effects for which current understanding of
quantitative relationships is strongest is reduced plant growth. As a result, this category was the
focus of the Administrator's decision-making in the last review, with RBL in tree seedlings
playing the role of surrogate (or proxy) for the broader array of vegetation-related effects that
range from the individual plant level to ecosystem services. Limitations in the evidence base and
associated uncertainties recognized in the last review remain and include a number of
uncertainties that affect characterization of the magnitude of cumulative exposure conditions
eliciting growth reductions in U.S. forests.
As recognized in the last review there are uncertainties in the extent to which the 11 tree
species for which there are established E-R functions encompass the range of O3 sensitive
species in the U.S., and also the extent to which they represent U.S. vegetation as a whole. These
11 species include both deciduous and coniferous trees with a wide range of sensitivities and
species native to every NOAA climate region across the U.S. and in most cases are resident
across multiple states and regions. In considering this issue in the last review, the CASAC stated
that there is "considerable uncertainty in extrapolating from the [studied] forest tree species to all
forest tree species in the U.S.," and additionally expressed the view that it should be anticipated
that there are highly sensitive vegetation species for which we do not have E-R functions and
others that are insensitive (Frey, 2014, p. 15). The CASAC also expressed the view, in the last
review, that it "should not be assumed that species of unknown sensitivity are tolerant to ozone"
and "[i]t is more appropriate to assume that the sensitivity of species without E-R functions
might be similar to the range of sensitivity for those species with E-R functions" (Frey, 2014, p.
11).
We additionally recognize important uncertainties in the extent to which the E-R
functions for reduced growth in tree seedlings are also descriptive of such relationships during
later lifestages, for which there is a paucity of established E-R relationships. Although such
information is limited with regard to mature trees, analyses in the 2013 ISA indicated that
reported growth response of young aspen over six years was similar to the reported growth
response of seedlings (ISA, Appendix 8, section 8.13.2; 2013 ISA, section 9.6.3.2). Additionally,
there are uncertainties with regard to the extent to which various factors in natural environments
4-49
-------
can either mitigate or exacerbate predicted Ch-plant interactions and contribute variability in
vegetation-related effects, including reduced growth. Such factors include multiple genetically
influenced determinants of O3 sensitivity, changing sensitivity to O3 across vegetative growth
stages, co-occurring stressors and/or modifying environmental factors.
Another area of uncertainty affects interpretation of the potential for harm to public
welfare over multi-year periods of air quality that meet the current standard. For example, there
is variability in ambient air O3 concentrations from year to year, as well as year-to-year
variability in environmental factors, including rainfall and other meteorological factors that
affect plant growth and reproduction, such as through changes in soil moisture. Accordingly,
these variabilities contribute uncertainties to estimates of the occurrence and magnitude of O3-
related effects in any year, and to such estimates over multi-year periods. Accordingly,
limitations in our ability to estimate growth effects over tree lifetimes of year-to-year variation in
O3 concentrations, particularly those associated with conditions meeting the current standard,
contribute uncertainty to estimates of cumulative growth (biomass) effects over multi-year
periods in the life of individual trees and associated populations, as well as related effects in
associated communities and ecosystems.
These uncertainties in estimates stem from limitations and imprecision in our tools, and
variation in aspects of the underlying data. For example, the studies on which the established E-
R functions for the 11 tree species are based vary in exposure duration (e.g., from periods of 82
to 140 days over a single year to periods of 180 to 555 days across two years) and in whether
measurements were made immediately following exposure period or in subsequent spring. The
E-R functions were derived based on the exposure duration of the experiment and, adjusted or
normalized to 3-month periods based on assumptions regarding relationships between duration,
cumulative exposure in terms of W126 index and plant growth response (see Lee and Hogsett,
1996, section 1.3). For example, while the functions are defined as describing a seasonal
response, some were derived by distributing responses observed at the end of two seasons of
varying exposures equally across the two seasons (essentially applying the average to both
seasons). The evidence for seasonal growth effects on trees is also somewhat limited with regard
to multi-year studies (particularly longer than two years) that have reported detailed O3
concentration data throughout the exposure. This contributes uncertainty, and accordingly a lack
of precision, to an understanding of the quantitative impacts of seasonal O3 exposure, including
its year-to-year variability, on tree growth and annual biomass accumulation. This uncertainty
limits our understanding of the extent to which tree biomass would be expected to appreciably
differ at the end of multi-year exposures for which the overall average exposure is the same, yet
for which the individual year exposures varies in different ways (e.g., as analyzed in Appendix
4D). For example, the extent of any differences in tree biomass for two multi-year scenarios with
4-50
-------
the same 3-year average W126 index but differing single-year indices is not clear, including for
exposures associated with O3 concentrations that would meet the current standard.54
A study of multi-year growth effects is available for aspen (King et al., 2005). This study
was assessed in the 2013 ISA and summarized in the ISA for the current review with regard the
extent to which it confirmed Ch-related biomass impacts estimated using the established E-R
functions for aspen (2013 ISA, section 9.6.3.2; current ISA, Appendix 8, section 8.13.2). The
2013 assessment applied the E-R functions to O3 exposure (quantified as cumulative average
seasonal W126 index) at each of six consecutive years and compared the estimated aboveground
biomass to estimates based on data reported for each year by the study (2013 ISA, section
9.6.3.2). The conclusions reached were that the experimental observations are "very close" to
estimates based on the established E-R function for aspen, and that "the function based on one
year of growth was shown to be applicable to subsequent years" (2013 ISA, p. 9-135; ISA,
Appendix 8, p. 8-186).
Further, while the tree seedling E-R relationships for 11 species are long-established, we
recognize the large variation among the species regarding how much experimental evidence is
available. For example, the E-R function for aspen was derived from 13 experimental studies,
while the E-R functions for the red maple and Virginia pine were each derived from a single
study (Appendix 4A, section 4A.2, Table 4A-6; 1996 AQCD, Table 5-28; Lee and Hogsett,
1996). Additionally, while the evidence is longstanding and robust for growth effects of O3, there
is variation across the 11 species for which we have established E-R functions with regard to the
extent to which the studies include O3 treatment levels reflecting cumulative O3 exposures, in
terms of W126 index, lower than 20 ppm-hrs. Studies for five of the eleven species include
cumulative exposures likely to correspond to W126 index values below 20 ppm-hrs (Appendix
4A, Table 4A-5).55 Further, among studies for these five species, the findings for at least one
study reported statistical significance only for biomass effects observed for higher O3 exposures
(e.g., Appendix 4A, Table 4A-6, black cherry). All of the factors identified here contribute to
imprecision or inexactitude in estimates based on the E-R functions.
54 Variation in annual W126 index values is described in Appendix 4D, indicating for the period, 2016-2018, that the
amount by which annual W126 index values at a site differ from the 3-year average varies, but generally falls
below 10 ppm-hrs across all sites and generally below 5 ppm-hrs at sites with design values at or below 70 ppb
(Appendix 4D, Figure 4D-7)..
55 For five of the species in Table 4A-5 in Appendix 4A, SUM06 index values below 25 ppm-hrs range from 12 to
21.7. In considering these values, we note that an approach used in the 2007 Staff Paper on specific temporal
patterns of O3 concentrations concluded that a SUM06 index value of 25 ppm-hrs would be estimated to
correspond to a W126 index value of approximately 21 ppm-hrs (U.S. EPA, 2007, Appendix 7B, p. 7B-2). This
would imply that a SUM06 value of 21 ppm-hrs would be expected to correspond to a W126 index value below
20 ppm-hrs.
4-51
-------
Our consideration of the magnitude of tree growth effects that might cause or contribute
to adverse effects for trees, forests, forested ecosystems or the public welfare is also complicated
by various uncertainties or limitations in the evidence base, including those associated with
relating magnitude of tree seedling growth reduction to larger-scale forest ecosystem impacts.
Further, other factors can influence the degree to which Ch-induced growth effects in a sensitive
species affect forest and forest community composition and other ecosystem service flows (e.g.,
productivity, belowground biogeochemical cycles and terrestrial ecosystem water cycling) from
forested ecosystems. These include (1) the type of stand or community in which the sensitive
species is found (i.e., single species versus mixed canopy); (2) the role or position the species has
in the stand (i.e., dominant, sub-dominant, canopy, understoiy); (3) the O3 sensitivity of the other
co-occurring species (O3 sensitive or tolerant); and (4) environmental factors, such as soil
moisture and others. The lack of such established relationships with O3 complicates
consideration of the extent to which different estimates of impacts on tree seedling growth would
indicate significance to the public welfare. Further, efforts to estimate O3 effects on carbon
sequestration are handicapped by the large uncertainties involved in attempting to quantify the
additional carbon uptake by plants as a result of avoided 03-related growth reductions. Such
analyses require complex modeling of biological and ecological processes with their associated
sources of uncertainty.
With regard to crop yield effects, as at the time of the last review, we recognize the
potential for greater uncertainty in estimating the impacts of O3 exposure on agricultural crop
production than that associated with O3 impacts on vegetation in natural forests. This relates to
uncertainty in the extent to which agricultural management methods influence potential for 03-
related effects and accordingly, the applicability of the established E-R functions for RYL in
current agricultural areas.
With regard to visible foliar injury, for which longstanding evidence documents a causal
role for O3, important uncertainties and limitations fall into two categories. The first category
relates to our understanding of the key aspects of O3 concentrations - and other key variables
(e.g., soil moisture) - that have a direct bearing on the severity and incidence of vegetation
injury, while the second concerns the impacts on aesthetic and recreational values of various
severities and incidences of injury. With regard to the former, there is a lack of detailed
understanding of specific patterns of O3 concentrations over a growing season and the key
aspects of those patterns (e.g., incidence of concentrations of particular magnitude) that
contribute to an increased incidence and severity of injury occurrence in the U.S. For example,
"the incidence of visible foliar injury is not always higher in years and areas with higher ozone,
especially with co-occurring drought" (ISA, Appendix 8, p. 8-24). Accordingly, there are no
established, quantitative E-R functions that document visible foliar injury severity and incidence
4-52
-------
under varying air quality and environmental conditions (e.g., soil moisture). As discussed in
section 4.3.3.2 above, the available studies that have investigated the role of different variables,
including different metrics for characterizing O3 concentrations over a growing season, do not
provide a basis for characterizing the potential for different patterns of O3 concentrations to
contribute to different incidences and severity of foliar injury in U.S. forests. The second
category of uncertainties and limitations concerns the information that would support associated
judgments on the public welfare significance of different patterns of and severity of foliar injury,
such as the extent to which such effects in areas valued by the public for different uses may be
considered adverse to public welfare. In considering this issue, we note that some level of
severity to a tree stand would be obvious to the casual observer (e.g., when viewing a stand
covering a hillside from a distance), and some level of severity (e.g., leaf and crown damage that
appreciably affects overall plant physiology) would also be expected to affect plant growth and
reproduction. The extent to which recreational values are affected by lesser levels of injury
severity and incidence is not clear from the available information. Thus, limitations and
uncertainties in the available information, such as those described above, complicate our ability
to comprehensively estimate the potential for visible foliar injury, its severity or extent of
occurrence for specific air quality conditions, and associated public welfare implications, thus
affecting a precise identification of air quality conditions that might be expected to provide a
specific level of protection for this effect.
During the last review, the 2013 ISA did not assess the evidence of O3 exposure and tree
mortality with regard to its support for inference of a causal relationship. Evidence available in
the last several reviews included field studies of pollution gradients that concluded O3 damage to
be an important contributor to tree mortality although several confounding factors such as
drought, insect outbreak and forest management were identified as potential contributors (2013
ISA, section 9.4.7.1). Since the last review, three additional studies have been identified, as
summarized in section 4.3.1 above, contributing to the ISA conclusion of sufficient evidence to
infer a likely causal relationship for O3 with tree mortality (ISA, Appendix 8, section 8.4). As
noted in the ISA, there is only limited evidence from experimental studies that isolate the effect
of O3 on tree mortality, with the recently available Aspen FACE study of aspen survival
involving cumulative seasonal exposures above 30 ppm-hrs during the first half of the 11-year
study period (ISA, Appendix 8, Tables 8-8 and 8-9). Evidence is lacking regarding exposure
conditions closer to those occurring under the current standard and any contribution to tree
mortality.
In the case of the two newly identified categories of effects, the key uncertainties relate to
comprehensive characterization of the effects. For example, with regard to alteration of herbivore
growth and reproduction, although "there are multiple studies demonstrating ozone effects on
4-53
-------
fecundity and growth in insects that feed on ozone-exposed vegetation", "no consistent
directionality of response is observed across studies and uncertainties remain in regard to
different plant consumption methods across species and the exposure conditions associated with
particular severities of effects " (ISA, pp. ES-18, IS-64, IS91 and Appendix 8, section 8.6.3).
Such limitations and uncertainties in the evidence base for this category of effects preclude
broader characterization, as well as quantitative analysis related to air quality conditions meeting
the O3 standard. As characterized in the ISA, uncertainties remain in the evidence; these relate to
the different plant consumption methods across species and the exposure conditions associated
with particular responses, as well as variation in study designs and endpoints used to assess O3
response (ISA, IS.6.2.1 and Appendix 8, section 8.6). Thus, while the evidence describes
changes in nutrient content and leaf chemistry following O3 exposure, the effect of these changes
on herbivores consuming the leaves is not well characterized or clear (ISA, p. IS-73).
The evidence for a second newly identified category of effects, alteration of plant-insect
signaling, draws on new research that has provided clear evidence of O3 modification of VPSCs
and behavioral responses of insects to these modified chemical signals. Most of these studies,
however, have been carried out in laboratory conditions rather than in natural environments, and
involve a relatively small number of plant species and plant-insect associations. While the
evidence documents effects on plant production of signaling chemicals and on the atmospheric
persistence of signaling chemicals, as well as on the behaviors of signal-responsive insects, it is
limited with regard to characterization of mechanisms and the consequences of any modification
of VPSCs by O3 (ISA, section IS.6.2.1) Further, the available studies vary with regard to the
experimental exposure circumstances in which the different types of effects have been reported
(most of the studies have been carried out in laboratory conditions rather than in natural
environments), and many of the studies involve quite short controlled exposures (hours to days)
to elevated concentrations, posing limitations for our purposes of considering the potential for
impacts associated with the studied effects to be elicited by air quality conditions that meet the
current standard (ISA, section IS.6.2.1 and Appendix 8, section 8.7).
With regard to radiative forcing and climate effects, "uncertainty in the magnitude of
radiative forcing estimated to be attributed to tropospheric ozone is a contributor to the relatively
greater uncertainty associated with climate effects of tropospheric ozone compared to such
effects of the well mixed greenhouse gases (e.g., carbon dioxide and methane)" (ISA, section
IS.6.2.2). With regard to O3 effects on temperature, "the heterogeneous distribution of ozone in
the troposphere complicates the direct attribution of spatial patterns of temperature change to
ozone induced RF" and the existence of O3 climate feedbacks "further alter the relationship
between ozone RF and temperature (and other climate variables) in complex ways" (ISA,
Appendix 9, section 9.3.1). Thus, various uncertainties "render the precise magnitude of the
4-54
-------
overall effect of tropospheric ozone on climate more uncertain than that of the well-mixed
GHGs" (ISA, Appendix 9, section 9.3.3). Further, "[c]urrent limitations in climate modeling
tools, variation across models, and the need for more comprehensive observational data on these
effects represent sources of uncertainty in quantifying the precise magnitude of climate responses
to ozone changes, particularly at regional scales" (ISA, Appendix 9, section 9.3.3).
4.4 EXPOSURE AND AIR QUALITY INFORMATION
Several different exposure and risk analyses were conducted in the last review of the
secondary O3 standard, as summarized in the IRP for this review. Uncertainties associated with
the results for some analyses limited their use in the Administrator's decision-making, while
uncertainties regarding public welfare significance of the findings for other analyses also limited
such use of those analyses. In general, decision-making in the last review placed greatest weight
on estimates of cumulative exposures to vegetation based on ambient air monitoring data and
consideration of those estimates in light of E-R functions for 03-related reduction in tree seedling
growth (summarized in section 4.3.3 above). These analyses supported the consideration of the
potential for O3 effects on tree growth and productivity, as well as its associated impacts on a
range of ecosystem services, including forest ecosystem productivity and community
composition (80 FR 65292, October 26, 2015).
The air quality and exposure analyses considered in the last review were of two types: (1)
W126-based cumulative exposure estimates in Class I areas during 3-year periods that met the
then-current standard (80 FR 65485-86, Table 3, October 26, 2015); and, (2) analyses for all U.S.
monitoring locations and time periods that met the then-current and several potential alternative
standards (Wells, 2015; 80 FR 65292, October 26, 2015). In these analyses, W126 index values56
occurring in locations with air quality meeting the then-current standard (or potential
alternatives) were considered in the context of the magnitude of W126 exposure index associated
with an estimate of 6% RBL in tree seedlings for the median tree species among the 11 species
for which there are established E-R relationships (80 FR 65391-92, Table 4, October 26, 2015).
That magnitude of W126 index is 19 ppm-hrs (80 FR 65391-65392). The second set of analyses
also included an evaluation of relationships between W126 index values and design values57
based on the form and averaging time of the then-current secondary standard (Wells, 2015). As
summarized in the IRP, we identified these analyses to be updated in this review in recognition
56 Based on judgments in the last review, the W126 metric analyzed and considered in the 2015 decision was the 3-
year average of consecutive year seasonal W126 index values (derived as described in section 4.3.3.1 above).
57 As described in earlier chapters, a design value is a statistic that describes the air quality status of a given area
relative to the level of the standard, taking the averaging time and form into account. For example, a design value
of 75 would have indicated O3 concentrations that just met the prior standard in a specific 3-yr period.
4-55
-------
of the relatively reduced uncertainty associated with the use of these types of analyses (compared
to the national or regional-scale modeling performed in the last review) to inform a
characterization of cumulative O3 exposure (in terms of the W126 index) associated with air
quality just meeting the current standard (IRP, section 5.2.2). This lesser uncertainty of these air
quality monitoring-based analyses contributed to their being more informative in the last review.
The sections below present findings of the updated analyses that have been performed in the
current review using the now available information.
Analyses in the current review are based on the expanded set of air monitoring data now
available,58 which includes 1,557 monitoring sites with sufficient data for derivation of design
values (Appendix 4D, section 4D.2.2). The current analyses are described in detail in Appendix
4D. As in the last review, we have analyzed the data both for the most recent periods, and also
across the full historical period back to 2000, which is now expanded from that available in the
last review59. We have performed analyses for all sites in the U.S., as well as for the subset of
sites in or near Class I areas. The most recent data analyzed are those for the design value period
from 2016 to 2018. For all monitoring sites with valid design values for this period, Figure 4-6
presents the average seasonal W126 index for the recent 3-year period (2016-2018) and also
denotes whether each site meets the current standard.
58 In addition to being expanded with regard to data for more recent time periods than were available during the last
review, the current dataset also includes a small amount of newly available older data for some monitoring sites
that are now available in the AQS.
59 In the last review, the dataset analyzed included data from 2000 through 2013 (Wells, 2015).
4-56
-------
• 1-7 ppm-hrs (663 sites) G 14-15 ppm-hrs (25 sites) • 18-62 ppm-hrs (92 sites)
• 8-13 ppm-hrs (312 sites) Q 16-17 ppm-hrs (30 sites) A 4th Max Value > 70 ppb
Figure 4-6. W126 index at monitoring sites with valid design values (2016-2018 average).
4-57
-------
4.4.1 Influence of Form and Averaging Time of Current Standard on W126 Index
In revising the standard in 2015 to the now-current standard, the Administrator concluded
that, with revision of the standard level, the existing form and averaging time provided the
control needed to achieve the cumulative seasonal exposure circumstances identified for the
secondary standard (80 FR 65408, October 26, 2015). The focus on cumulative seasonal
exposure primarily reflects the evidence on E-R relationships for plant growth. The 2015
conclusion was based on the air quality data analyzed at that time (80 FR 65408, October 26,
2015). Analyses in the current review of the now expanded set of air monitoring data, which
includes 1,557 monitoring sites with sufficient data for derivation of design values (Appendix
4D, section 4D.2.2), document similar findings as from the analysis of data from 2000-2013
described in the last review. The current analyses, which span 19 years and 17 3-year periods,
are described in detail in Appendix 4D.
One aspect of these analyses documents the positive nonlinear relationship that is
observed between cumulative seasonal exposure, quantified using the W126 index, and design
values, based on the form and averaging time of the current standard. This is shown for both the
average W126 index across the 3-year design value period (Figure 4-7, left) and for annual index
values within the period (Figure 4-7, right). From both of these presentations, it is clear that
cumulative seasonal exposures, assessed in terms of W126 index, are lower at monitoring sites
with lower design values. This is seen both for design values above the level of the current
standard (70 ppb), where the slope is steeper (due to the sigmoidal weighting of higher
concentrations by the W126 index function), as well as for lower design values that meet the
current standard (Figure 4-7; Appendix 4D, Figure 4D-4).
These presentations also indicate some regional differences. For example, as shown in
Figure 4-6 for the 2016-2018 period, sites meeting the current standard in the regions outside of
the West and Southwest regions, all W126 index values are at or below 13 ppm-hrs (for either
W126 metric). Ozone concentrations, and W126 index values, are generally higher in the West
and Southwest regions (Figure 4-6). However, the positive relationship between the W126 index
and the design value is evident in all regions (Figure 4-7).
4-58
-------
« oat ;;;t
40 50 60 70 80 90 100 110
3-year Average 4th Max Value (ppb)
Figure 4-7. Relationship between the W126 index and design values for the current standard (2016-2018). The W126
analyzed in terms of averages across the 3-year design value period (left) and annual values (right).
index is
• Central
o EastNorthCentral
o NorthEast
o Northwest
o South
o SouthEast
o Southwest
• West
• WestNorthCentral
® Central A 2016
O EastNorthCentral ¦ 2017
® NorthEast • 2018
0 Northwest
o South
o SouthEast
O Southwest
• West
• WestNorthCentral
60 70 80 90
3-year Average 4th Max Value (ppb)
4-59
-------
An additional analysis assesses the relationship between long-term changes in design
value and long-term changes in the W126 index. This type of analysis, which was also
performed in the last review with the then-available data, is presented in detail in Appendix 4D
(section 4D.3.2.2). The current analysis focuses on the relationship between changes (at each
monitoring site) in the 3-year design value (termed "4th max" in Appendix 4D and Figures 4-7
and 4-8) across the 16 design value periods from 2000-2002 to 2016-2018 and changes in the
W126 index over the same period.60 This analysis, performed using either the 3-year average
W126 index or annual values, shows there to be a positive, linear relationship between the
changes in the W126 index and the changes in the design value at monitoring sites across the
U.S. (Figure 4-8). This means that a change in the design value at a monitoring site was
generally accompanied by a similar change in the W126 index. Nationally, the W126 index (in
terms of 3-year average) decreased by approximately 0.62 ppm-hrs per ppb decrease in design
value over the full period from 2000 to 2018. This relationship varies across the NOAA climate
regions, with the greatest change in the W126 index per unit change in design value observed in
the Southwest and West regions. Thus, the regions which had the highest W126 index values at
site meeting the current standard (Figure 4-6) also showed the greatest improvement in the W126
index per unit decrease in their design values over the past 19 years (Appendix 4D, Table 4D-
1 land Figure 4D-12). This indicates that going forward as design values are reduced in areas
that are presently not meeting the current standard, the W126 index in those areas would also be
expected to decline (Appendix 4D, section 4D.3.2.3 and 4D.4).
The overall trend showing reductions in the W126 concurrent with the design value
metric for the current standard is positive whether the W126 index is expressed in terms of the
average across the 3-year design value period or the annual value (Appendix 4D, section
4D.3.2.3). This similarity is consistent with the relationship between the W126 index and the
design value metric for the current standard summarized above, which shows a strong positive
relationship between those metrics (Figure 4-7, Appendix 4D, section 4D.3.1.2).
60 At each site, the trend in values of a metric (W126 or 4th max), in terms of a per-year change in metric value, is
calculated using the Theil-Sen estimator, a type of linear regression method that chooses the median slope among
all lines through pairs of sample points. For example, if applying this method to a dataset with metric values for
four consecutive years (e.g., W126i, W1262, WI263, WI264), the trend would be the median of the different per-
year changes observed in the six possible pairs of values ([Wl 264- W126s]/l, [W1263- W126z]/l, [W1262
W126i]/1, [W1264-W1262]/2, [W1263-W126i]/2, [W1264- W126i]/3).
4-60
-------
o SouthEast
o SouthWest
• Central Q SouthEast
o EastNorthCentral ® SouthWest
© NorthEast • West
o Northwest • WestNorthCentral
o South
-2 -1 0
Trend in 3-year Average 4th Max Value (ppb/yr)
® Central o SouthEast
o EastNorthCentral o SouthWest
© NorthEast • West
® Northwest • WestNorthCentral
o South
1-3 -2 -1 0 1
Trend in 3-year Average 4th Max Value (ppb/yr)
Figure 4 8. Relationship between trends in the W126 index and trends in design values across a 19-year period (2000-2018) at
U.S. monitoring sites. W126 is analyzed in terms of averages across 3-year design value periods (left) and annual
values (right).
4-61
-------
In considering quantitative analyses concerning the control of the current form and
averaging time on vegetation exposures of potential concern, we additionally take note of the
evidence discussed in section 4.3.3.2 above regarding the potential for days with particularly
high O3 concentrations to play a contributing role in visible foliar injury. While the occurrence
and severity of visible foliar injury indicates some relationship with cumulative concentration-
weighted indices such as SUM06 and W126, the evidence also indicates a contributing role for
occurrences of peak concentrations. We note that the current standard's form and averaging time,
by their very definition, limit such occurrences. For example, the peak 8-hour average
concentrations are lower at sites with lower design values, as illustrated by the declining trends
in annual fourth highest MDA8 concentrations that accompany the declining trend in design
values described in chapter 2 (e.g., Figure 2-11). Additionally, with regard to hourly
concentrations, analyses summarized in Appendix 2A document decreasing frequency of
elevated 1-hour concentrations (e.g., concentrations at or above 100 ppb) with decreasing design
values (Appendix 2A, Tables 2A-2 through 2A-4). For example, in the most recent design value
period (2016-2018) across all sites with adequate data to derive design values, the mean number
of observations per site at or above 100 ppb was well below one (0.19) for sites that meet the
current standard, compared to well above one (8.09) for sites not meeting the current (Appendix
2A, Table 2A-2).
In summary, monitoring sites with lower O3 concentrations as measured by the design
value metric (based on the current form and averaging time of the secondary standard) have
lower cumulative seasonal exposures, as quantified by the W126 index (as well as lower short-
term peak concentrations). As the form and averaging time of the secondary standard have not
changed since 1997, the analyses performed have been able to assess the amount of control
exerted by these aspects of the standard, in combination with reductions in the level (i.e., from 80
ppb in 1997 to 75 ppb in 2008 to 70 ppb in 2015) on cumulative seasonal exposures in terms of
W126 index (and on the magnitude of short-term peak concentrations). The analyses have found
that the reductions in design value, presumably associated with implementation of the revised
standards, have been accompanied by reductions in cumulative seasonal exposures in terms of
W126 index, as well as reductions in short-term peak concentrations.
4.4.2 Environmental Exposures in Terms of W126 Index
To inform the Administrator's exposure/risk-based considerations in the current review,
we have developed updates to the air quality analyses of O3 concentrations and W126 index
values that were developed in the last review. Given the evidence indicating the W126 index to
be strongly related to growth effects and its use in the E-R functions for tree seedling RBL,
exposure is quantified using the W126 metric (Figure 4-9). In light of the importance placed on
4-62
-------
Class I areas in past secondary standard reviews and the greater public welfare significance of O3
related impacts in such areas, as discussed in section 4.3.2 above, a separate evaluation is
conducted on cumulative O3 exposure at monitoring sites in or near Class I areas61, in addition to
that at all monitoring sites nationwide. The potential for impacts of interest is assessed through
considering the magnitude of estimated exposure in light of current information and in
comparison to levels given particular focus in the 2015 decision on the current standard (80 FR
65292; October 26, 2015).62
Consideration of occurrences of W126 exposures
of different magnitudes
All monitoring locations
Class I Areas
Cumulative seasonal concentration-weighted exposure index
Hourly concentrations
0, in Ambient Air
Figure 4 9. Analytical approach for characterizing vegetation exposure.
The updated analyses discussed here and described in greater detail in Appendix 4D
include assessment of all monitoring sites nationally and also a focused evaluation in Class I
areas for which such monitoring data are available. The analyses include air quality monitoring
data for the most recent 3-year period (2016 to 2018) for which data were available when the
analyses were performed, and also all 3-year periods going back as far as the 2000-2002 period.
Design values (3-year average annual fourth-highest 8-hour daily maximum concentration, also
termed "4th max metric" in this analysis) and VV126 index values (in terms of the 3 year average)
were calculated at each site where sufficient data were available.63 Across the seventeen 3-year
periods from 2000-2002 to 2016-2018, the number of monitoring sites with sufficient data for
calculation of valid design values and W126 index values ranged from a low of 992 in 2000-
2002 to a high of 1119 in 2015-2017. As specific monitoring sites differed somewhat across the
61 Included are monitors sited within Class I areas or the closest monitoring site within 15 km of the area boundary.
62 The W126 index values were rounded to the nearest unit ppm-hr for these comparisons to a specific whole-
number W126 level (Appendix 4D, section 4D.2).
63 Data adequacy requirements and methods for these calculations are described in Appendix 4D, section 4D.2.
4-63
-------
19 years, there were 1,557 sites with sufficient data for calculation of valid design values and
W126 index values for at least one 3-year period between 2000 and 2018, and 543 sites had such
data for all seventeen 3-year periods. The sections below discuss key aspects of these analyses
and what they indicate with regard to protection from vegetation-related effects of potential
public welfare significance.
The analyses of cumulative seasonal exposures included a focus on the W126 index in
terms of the average seasonal index across the 3-year design value period, with additional
analyses also characterizing the annual W126 index. Among the analyses performed is an
evaluation of the variability of annual W126 index values across the 3-year period (Appendix
4D, section 4D.3.1.2). This evaluation was performed for all monitoring sites in the most recent
3-year period, 2016 to 2018. This analysis indicates the extent to which single-year values within
the 3-year period deviate from the average for the period. Across the full set of sites, regardless
of W126 index magnitude (or whether or not the current standard is met), single-year W126
index values differ no more than 12 or 13 ppm-hrs from the average for the 3-year period
(Appendix 4D, Figure 4D-6). Focusing on the approximately 850 sites (Appendix 4D, Table 4D-
1) meeting the current standard (design value at or below 70 ppb), over 99% of single-year
W126 values in this subset differ from the 3-year average by no more than 5 ppm-hrs, and 87%
by no more than 2 ppm-hrs (Appendix 4D, Figure 4D-7).
The following discussion is framed by a key policy-relevant question based on those
identified in the IRP. The question considers all areas nationally, with particular focus on air
quality data for Class I areas.
• What are the nature and magnitude of vegetation exposures associated with
conditions meeting the current standard at sites across the U.S., particularly in
specially protected areas, such as Class I areas, and what do they indicate regarding
the potential for 03-related vegetation impacts?
To further address this question, we considered both recent air quality (2016-2018) and
air quality since 2000. These air quality analyses of cumulative seasonal exposures associated
with conditions meeting the current standard nationally provide conclusions generally similar to
those based on the data available at the time of the last review when the current standard was set,
when the most recent data were available for 2011 to 2013 (Wells, 2015). Cumulative exposures
vary across the U.S, with the highest W126 index values for sites that met the current standard
being located exclusively in the Southwest and West climate regions (Figure 4-6, Appendix 4D,
Table 4D-1). In all other NOAA climate regions, W126 index values at sites meeting the current
standard are generally at or below 13 ppm-hrs (Figure 4-6 and Appendix 4D, Figure 4D-2). In
the Southwest and West, W126 index values at all sites meeting the current standard are at or
below 17 ppm-hrs on virtually all occasions in the most recent 3-year period and across all of the
4-64
-------
seventeen 3-year periods in the full dataset evaluated.64 That is, among sites meeting the current
standard in the most recent period of 2016 to 2018, there are none with a W126 index, based on
the 3-year average, above 19 ppm-hrs, and just one with such a value above 17 ppm-hrs (Table
4-1). Additionally, the historical dataset includes no occurrences of a 3-year average W126 index
above 19 ppm-hrs at sites meeting the current standard, and just a small number of occurrences
(limited to eight, all but one from a period prior to 2011) of a W126 index above 17 ppm-hrs,
with the highest just equaling 19 ppm-hrs (Table 4-1; Appendix 4D, section 4D.3.2.1).
Given the recognition of more significant public welfare implications of effects in
protected areas, such as Class I areas (as discussed in section 4.3.2 above), we give particular
attention to Class I areas (Appendix 4D, section 4D.3.2.3). In so doing, we consider the updated
air quality analysis presented in Appendix 4D for 65 Class I areas. The findings for these sites,
which are distributed across all nine NOAA climate regions in the contiguous U.S., as well as
Alaska and Hawaii, mirror all U.S. sites. Among the Class I area sites meeting the current
standard (i.e., having a design value at or below 70 ppb) in the most recent period of 2016 to
2018, there are none with a W126 index (as average over design value period) above 17 ppm-hrs
(Table 4-1). The historical dataset includes just seven occurrences (all dating from the 2000-2010
period) of a Class I area site meeting the current standard and having a 3-year average W126
index above 17 ppm-hrs, and no such occurrences above 19 ppm-hrs (Table 4-1).
The W126 exposures at sites with design values above 70 ppb range up to approximately
60 ppm-hrs (Table 4-1, Appendix 4D, Table 4D-17). Among all sites across the U.S. that do not
meet the current standard in the 2016 to 2018 period, more than a quarter have average W126
index values above 19 ppm-hrs and a third exceed 17 ppm-hrs (Table 4-1).65 A similar situation
exists for Class I area sites (Table 4-1). Thus, as was the case in the last review, the currently
available quantitative information continues to indicate appreciable control of seasonal W126
index-based cumulative exposure at all sites with air quality meeting the current standard.
64 On 99.9 percent of occasions across all sites with valid design values at or below 70 ppb during the 2000 to 2018
period, the W126 metric (seasonal W126, averaged over three years) was at or below 17 ppm-hrs (Table 4-1). All
but one of the eight occasions when it was above 17 ppm-hrs (the highest was 19 ppm-hrs) occurred in the
Southwest region during a period before 2011. The eighth occasion occurred at a site in the West region when the
3-year average W126 index value was 18 ppm-hrs. On more than 97 percent of occasions in the full dataset with
valid design values at or below 70 ppb, the 3-year average W126 index was at or below 13 ppm-hrs (Appendix
4D, section 4D.3.2).
65 As described above and in detail in Appendix 4D, W126 index values were rounded to the nearest unit ppm-hr for
comparisons to a specific whole-number W126 level.
4-65
-------
Table 4-1. Distribution of 3-yr average seasonal W126 index for sites in Class I areas and
across U.S. that meet the current standard and for those that do not.
3-year periods
Number of Occurrences or Site-DVsA
In Class I Areas
Across All Monitoring Sites (urban and rural)
Total
W126 (ppm-hrs)
>19 | >17 | <17
Total
W126 (ppm-hrs)
>19
>17
<17
At sites that meet the current standard (design value at or below 70 ppb)
2016-2018
47
0
0
47
849
0
1
848
All from 2000 to 2018
498
0
7
491
8,292
0
8
8,284
At sites that exceed t
he current standard (design value a
bove 70 ppb
2016-2018
11
8
9
2
273
78
91
182
All from 2000 to 2018
362
159
197
165
10,695
2,317
3,174
7,521
A The counts presented here are drawn from Appendix D, Tables 4D-2, 4D-4, 4D-5, 4D-6, 4D-9, 4D-10 and 4D-14 through 17.
As discussed in section 4.3.3 above, the evidence currently available leads us to similar
conclusions regarding exposure levels associated with effects as in the last review. Based largely
on this evidence in combination with the use of RBL as a surrogate or proxy for all vegetation-
related effects, the value of 17 ppm-hrs was the average W126 index (over three years) was
generally identified as a target level for protection in the 2015 decision (80 FR 65393; October
26, 2015). As summarized above, the information available in this review continues to indicate
that average cumulative seasonal exposure levels at virtually all sites and 3-year periods with air
quality meeting the current standard fall at or below the level of 17 ppm-hrs that, as summarized
in section 4.1 above, was identified when the current standard was established (80 FR 65393;
October 26, 2015). Additionally, the full dataset indicates that at sites meeting the current
standard, annual W126 index values were less than or equal to 19 ppm-hrs well over 99% of the
time (Appendix 4D, section 4D.3.2.1). Additionally, the average W126 index in Class I areas
that meet the current standard for the most recent 3-year period is below 17 ppm-hrs in all areas
for which there is a monitor in or nearby (Appendix 4D, Table 4D-16). Further, with the
exception of seven values that occurred prior to 2011, cumulative seasonal exposures in all Class
I areas during periods that met the current standard were no higher than 17 ppm-hrs. This
contrasts with the occurrence of much higher seasonal W126 index values in sites when the
current standard was not met. For example, out of the 11 Class I area sites with design values
above 70 ppb during the most recent period, eight had had a W126 index (based on 3-year
average) above 19 ppm-hrs (ranging up to 47 ppm-hrs) and nine sites had a W126 index above
17 ppm-hrs (Table 4-1; Appendix 4D, Table 4D-17).
4-66
-------
4.4.3 Limitations and Uncertainties
• What are the important uncertainties associated with any exposure estimates and
associated characterization of potential for public welfare effects?
The analyses described above in sections 4.1 and 4.2 are based primarily on the hourly air
monitoring dataset that is available at O3 monitoring sites nationwide. While there are inherent
limitations in any air monitoring network, the monitors for O3 are distributed across the U.S.,
covering all NOAA regions and all states (e.g., Figure 4-6).
There is uncertainty about whether areas that are not monitored would show the same
patterns of exposure as areas with monitors. There are limitations in the distributions of the
monitors, and some geographical areas are more densely covered than others, which may have
sparse or no data. For example, only about 40% of all Federal Class I Areas have or have had O3
monitors within 15 km with valid design values, thus allowing inclusion in the Class I area
analysis. Even so, the dataset includes sites in 27 states distributed across all nine NOAA
climatic regions across the contiguous U.S, as well as Hawaii and Alaska. Some NOAA regions
have far fewer numbers of Class I areas with monitors than others. For instance, the Central,
North East, East North Central, and South regions all have three or fewer Class I areas in the
dataset. However, these areas also have appreciably fewer Class 1 areas in general when
compared to the Southwest, Southeast, West, and West North Central regions, which are more
well represented in the dataset. The West and Southwest regions are identified as having the
largest number of Class I areas, and they have approximately a third of those areas represented
with monitors, which include locations where W126 index values are generally higher, thus
playing a prominent role in the analysis.
We also recognize a limitation that accompanies any analysis, i.e., that it is based on
currently available information. Thus, it may or may not reflect conditions far out into the future
as air quality and patterns of O3 concentrations in ambient air continue to change in response to
changing circumstances, such as changes in precursor emissions to meet the current standard
across the U.S. That said, we note that findings from these analyses in the current review are
largely consistent with those from analyses of the data available in the last review. Further, we
note the findings of the analysis in Appendix 4D of how changes in O3 patterns in the past have
affected the relationship between W126 index and the averaging time and form of the current
standard, as represented by design values (Appendix 4D, section 4D.3.2.3). This analysis finds a
positive, linear relationship between trends in design values and trends in the W126 index (both
in terms of single-year W126 index and averages over 3-year design value period), as was also
the case for similar analyses conducted for the data available at the time of the last review
(Wells, 2015). While this relationship varies across NOAA regions, the regions showing the
greatest potential for exceeding W126 index values of interest (e.g., with 3-year average values
4-67
-------
above 17 and/or 19 ppm-hrs) also showed the greatest improvement in the W126 index per unit
decrease in design value over the historical period assessed (Appendix 4D, section 4D.3.2.3).
Thus, the available data and this analysis appear to indicate that as design values are reduced to
meet the current standard in areas that presently do not, W126 values in those areas would also
be expected to decline (Appendix 4D, section 4D.4).
4.5 KEY CONSIDERATIONS REGARDING THE CURRENT
SECONDARY STANDARD
In considering what the currently available evidence and exposure/risk information
indicate with regard to the current secondary O3 standard, the overarching question we address
is:
• Does the currently available scientific evidence and air quality and exposure analyses
support or call into question the adequacy of the protection afforded by the current
secondary O3 standard?
To assist us in interpreting the currently available scientific evidence and the results of
recent quantitative analyses to address this question, we have focused on a series of more
specific questions. In considering the scientific and technical information, we consider both the
information available at the time of the last review and information newly available since then
which has been critically analyzed and characterized in the current ISA, the 2013 ISA and prior
AQCDs. In so doing, an important consideration is whether the information newly available in
this review alters the EPA's overall conclusions from the last review regarding welfare effects
associated with photochemical oxidants, including O3, in ambient air. We also consider the
currently available quantitative information regarding environmental exposures, characterized by
the pertinent metric, likely to occur in areas of the U.S. where the standard is met. Additionally,
we consider the significance of these exposures with regard to the potential for 03-related
vegetation effects, their potential severity and any associated public welfare implications.
4.5.1 Evidence and Exposure/Risk-based Considerations
In considering first the currently available evidence with regard to the overarching
question posed above regarding the protection provided by the current standard from welfare
effects, we address a series of more specific questions that focus on policy-relevant aspects of the
evidence. These questions relate to three main areas of consideration: (1) the available evidence
on welfare effects associated with exposure to photochemical oxidants, and particularly O3
(section 4.5.1.2); (2) the risk management framework or approach for reaching conclusions on
the adequacy of protection provided by the secondary standard (section 4.5.1.2); and, (3)
4-68
-------
findings from the air quality and exposure analyses pertaining to public welfare protection under
the current standard (section 4.5.1.3).
4.5.1.1 Welfare Effects Evidence
• Is there newly available evidence that indicates the importance of photochemical
oxidants other than O3 with regard to abundance in ambient air, and potential for
welfare effects?
No newly available evidence has been identified in this review regarding the importance
of photochemical oxidants other than O3 with regard to abundance in ambient air, and potential
for welfare effects.66 As summarized in section 2.1 above, O3 is one of a group of photochemical
oxidants formed by atmospheric photochemical reactions of hydrocarbons with nitrogen oxides
in the presence of sunlight, with O3 being the only photochemical oxidant other than nitrogen
dioxide that is routinely monitored in ambient air (ISA, Appendix 1, section 1.1).67 Data for
other photochemical oxidants are generally derived from a few special field studies; such that
national scale data for these other oxidants are scarce (ISA, Appendix 1, section 1.1; 2013 ISA,
sections 3.1 and 3.6). Moreover, few studies of the welfare effects of other photochemical
oxidants beyond O3 have been identified by literature searches conducted for the 2013 ISA and
prior AQCDs (ISA; Appendix 1, section 1.1). As stated in the current ISA, "the primary
literature evaluating the health and ecological effects of photochemical oxidants includes ozone
almost exclusively as an indicator of photochemical oxidants" (ISA, section IS. 1.1). Thus, as was
the case for previous reviews, the evidence base for welfare effects of photochemical oxidants
does not indicate an importance of any other photochemical oxidants. For these reasons,
discussion of photochemical oxidants in this document focuses on O3.
• Does the current evidence alter conclusions from the last review regarding the
nature of welfare effects attributable to O3 in ambient air?
The current evidence, including that newly available in this review, supports, sharpens
and expands somewhat on the conclusions reached in the last review (ISA, sections IS. 1.3.2 and
IS.5 and Appendices 8 and 9). A wealth of scientific evidence, spanning more than six decades,
demonstrates effects on vegetation and ecosystems of O3 in ambient air (ISA, section IS.6.2.1;
2013 ISA, 2006 AQCD, 1997 AQCD, 1986 AQCD; U.S. DHEW, 1970). Accordingly, consistent
with the evidence in the last review, the currently available evidence describes an array of O3
effects on vegetation and related ecosystem effects. The evidence also describes climate effects
66 Close agreement between past ozone measurements and the photochemical oxidant measurements upon which the
early NAAQS (for photochemical oxidants including O3) was based indicated the very minor contribution of
other oxidant species in comparison to O3 (U.S. DHEW, 1970).
67 Consideration of welfare effects associated with nitrogen oxides in ambient air is addressed in the review of the
secondary NAAQS for oxides of nitrogen and oxides of sulfur (U.S. EPA, 2018).
4-69
-------
of tropospheric O3, through a role in radiative forcing and subsequent effects on temperature,
precipitation and related climate variables. Evidence newly available in this review strengthens
previous conclusions, provides further mechanistic insights and augments current understanding
of varying effects of O3 among species, communities and ecosystems (ISA, section IS.6.2.1).
The current evidence, including a wealth of longstanding evidence, supports conclusions reached
in the last review of causal relationships between O3 and visible foliar injury, reduced yield and
quality of agricultural crops, reduced vegetation growth and plant reproduction,68 reduced
productivity in terrestrial ecosystems, and alteration of belowground biogeochemical cycles. The
current evidence, including a wealth of longstanding evidence, also supports conclusions reached
in the last review of likely causal relationships between O3 and reduced carbon sequestration in
terrestrial systems, and alteration of terrestrial ecosystem water cycling (ISA, section IS.1.3.2).
Additionally, as in the last review, the current ISA determines there to be a causal relationship
between tropospheric O3 and radiative forcing and a likely causal relationship between
tropospheric O3 and temperature, precipitation and related climate variables (ISA, section
IS. 1.3.3). Further, the current evidence has led to an updated conclusion on the relationship of O3
with alteration of terrestrial community composition to causal (ISA, sections IS.I.3.2). Lastly, the
current ISA concludes the current evidence sufficient to infer likely causal relationships of O3
with three additional categories of effects (ISA, sections IS.I.3.2). While previous recognition of
O3 as a contributor to tree mortality in a number of field studies was a factor in the 2013
conclusion regarding composition composition, it has been separately assessed in this review.
Additionally, evidence newly available in this review on two additional plant-related effects
augments more limited previously available evidence related to insect interactions with
vegetation, contributing to additional conclusions that the body of evidence is sufficient to infer
likely causal relationships between O3 and alterations of plant-insect signaling and insect
herbivore growth and reproduction (ISA, Appendix 8, sections 8.6 and 8.7).69
As in the last review, the strongest evidence and the associated findings of causal or
likely causal relationships with O3 in ambient air, and quantitative characterizations of
relationships between O3 exposure and occurrence and magnitude of effects are for vegetation-
related effects, and particularly those identified in the last review. The evidence base for the
newly identified category of increased tree mortality includes previously available evidence
68 As noted in section 4.3.1 above, the ISA in this review includes a causality determination specific to reduced plant
reproduction, while this category of effects was considered in combination with reduced plant growth in the last
review (ISA, Table IS. 13).
69 As in the last review, the ISA again concludes that the evidence is inadequate to determine if a causal relationship
exists between changes in tropospheric ozone concentrations and UV-B effects (ISA, Appendix 9, section 9.1.3.4;
2013 ISA, section 10.5.2).
4-70
-------
largely comprised of field observations from locations and periods of O3 concentrations higher
than are common today and three more recently available publications assessing O3 exposures
not expected under conditions meeting the current standard. Among the three more recent
publications, one assessed survival of aspen clones across an 11-year period under O3 exposures
that included single-year seasonal W126 index values ranging above 30 ppm-hrs during the first
four years, and the other two were analyses based on field observations during periods when O3
concentrations were such that they would not be expected to meet the current standard, as
summarized in section 4.3.1 above (ISA, Appendix 8, section 8.4.3).
The information available regarding the newly identified categories of plant-insect
signaling and insect herbivore growth and reproduction does not provide for a clear
understanding of the specific environmental effects that may occur in the natural environment
under specific exposure conditions (as discussed in sections 4.3.1, 4.3.3.2 and 4.3.4 above). For
example, while the evidence base for effects on herbivore growth and reproduction is expanded
in this review, "there is no clear trend in the directionality of response for most metrics," such
that some show an increased effect and some show reductions (ISA, p. IS-64; section IS.5.1.3
and section 8.6). More specifically "no consistent directionality of response is observed across
the literature, and uncertainties remain in regard to different plant consumption methods across
species and the exposure conditions associated with particular severities of effects" (ISA, p. IS-
91). Additionally, while the current evidence base documents effects of O3 on some plant VPSCs
(e.g., changing the floral scent composition and reducing dispersion), and indicates reduced
pollinator attraction, decreased plant host detection and altered plant-host preference in some
insect species in the presence of elevated O3 concentrations, characterization of such effects is
still "an emerging area of research with information available on a relatively small number of
insect species and plant-insect associations," and with gaps remaining in the consequences of
modification of signaling compounds by O3 in natural environments (ISA, p. IS-91 and section
IS.6.2.1). Accordingly, the focus in this review is on other vegetation effects described above,
rather than these two newly identified categories.
With regard to tropospheric O3 and effects on climate, we recognize the strength of the
conclusion that tropospheric O3 is a greenhouse gas at the global scale, with associated effects on
climate (ISA, section 9.1.3.3). Accordingly, as indicated by the ISA causal determinations, O3
abundance in the troposphere contributes to radiative forcing and likely also to subsequent
climate effects. There is appreciable uncertainty, however, associated with understanding
quantitative relationships involving regional O3 concentrations near the earth's surface and
climate effects of tropospheric O3 on a global scale As recognized in the ISA (and summarized in
sections 4.3.3.3 and 4.3.4 above), there are limitations in our modeling tools and associated
uncertainties in interpretations related to capabilities for quantitatively estimating effects of
4-71
-------
regional-scale lower tropospheric O3 concentrations on climate. Thus, while additional
characterizations of tropospheric O3 and climate have been completed since the last review,
uncertainties and limitations in the evidence that were also recognized in the last review remain.
As summarized in sections 4.3.3.3 and 4.3.4 above, these affect our ability to make a quantitative
characterization of the potential magnitude of climate response to changes in O3 concentrations
in ambient air, particularly at regional (vs global) scales, and thus our ability to assess the impact
of changes in ambient air O3 concentrations in regions of the U.S. on global radiative forcing or
temperature, precipitation and related climate variables. Consequently, the current evidence in
this area is not informative to our consideration of the adequacy of public welfare protection of
the current standard.
• To what extent does the available evidence provide E-R information (e.g.,
quantitative E-R relationships) for 03-related effects that can inform judgments on
the likelihood of occurrence of such effects in areas with air quality that meets the
current standard? Does the currently available evidence provide new or altered such
information since the last review?
In considering what the currently available information indicates with regard to
exposures associated with welfare effects and particularly in the context of what is indicated for
exposures associated with air quality conditions that meet the current standard, we focus
particularly on the availability of quantitatively characterized E-R relationships for key effects.
While the ISA describes additional studies of welfare effects associated with O3 exposures since
the last review, the established E-R functions for tree seedling growth and crop yield that have
been available in the last several reviews continue to be the most robust descriptions of E-R
relationships for welfare effects. These well-established E-R functions for seedling growth
reduction in 11 tree species and yield loss in 10 crop species are based on response information
across multiple levels of cumulative seasonal exposure (estimated from extensive records of
hourly O3 concentrations across the exposure periods). Studies of some of the same species,
conducted since the E-R function derivation, provide supporting information for these functions
(ISA, Appendix 8, section 8.13.2; 2013 ISA, sections 9.6.3.1 and 9.6.3.2). The E-R functions
provide for estimation of growth-related effects for a range of cumulative seasonal exposures.
The evidence newly available in this review does not include new studies that assessed
reductions in tree growth or crop yield responses across multiple O3 exposures and for which
sufficient data are available for analyses of the shape of the E-R relationship across the range of
cumulative exposure levels (e.g., in terms of W126 index) relevant to conditions associated with
the current standard. For example, among the newly available studies are several that summarize
4-72
-------
previously available studies or draw from them, such as for linear regression analyses.70
However, as discussed in sections 4.3.3.2 above, these do not provide robust E-R functions or
cumulative seasonal exposure levels associated with important vegetation effects that define the
associated exposure circumstances in a consistent manner, limiting their usefulness for our
purposes here with regard to considering the potential for occurrence of welfare effects in air
quality conditions that meet the current standard. Thus, robust E-R functions are not available for
growth or yield effects on any additional tree species or crops in this review.
Based on these established E-R functions for tree seedling growth reductions in 11
species, the tree seedling RBL for the median tree species is 5.3% for a W126 index of 17 ppm-
hrs, rising to 5.7% for 18 ppm-hrs, 6.0% for 19 ppm-hrs and 6.4% for 20 ppm-hrs. Below 17
ppm-hrs, the median estimates include 4.9% for 16 ppm-hrs, 4.5% for 15 ppm-hrs, 4.2% for 14
ppm-hrs and 3.8% for 13 ppm-hrs (Appendix4A, Table 4A-5). These RBL estimates are
unchanged from what was indicated by the evidence in the last review. As summarized in section
4.1 above, the RBL estimates were used in the 2015 decision as a surrogate or proxy for the
broader array of vegetation-related effects.
With regard to visible foliar injury, as in the last review, we lack established E-R
relationships that would quantitatively describe relationships between visible foliar injury
(occurrence and incidence, as well as, injury severity) and O3 exposure, as well as factors
influential in those relationships, such as soil moisture conditions. As discussed in section
4.3.3.2 above, the currently available evidence continues to include both experimental studies
that document foliar injury in specific plants in response to O3 exposures, and quantitative
analyses of the relationship between environmental O3 exposures and occurrence of foliar injury.
The analyses involving environmental conditions, while often using cumulative exposure metrics
to quantify O3 exposures additionally have reported there to also be a role for a metric that
quantifies the incidence of "high" O3 days (2013 ISA, p. 9-10; Smith, 2012; Wang et al., 2012).
However, such analyses have not established specific air quality metrics and associated
quantitative functions for describing the influence of ambient air O3 on incidence and severity of
visible foliar injury.
70 For example, among the newly available publications cited in the ISA is a publication on tree and grassland
species that compiles EC10 values (estimated concentration at which 10% lower biomass [compared to zero O3] is
predicted) derived using linear regression of previously published data on plant growth response and O3
concentration quantified as AOT40. The data were from studies of various experimental designs, that involved
various durations ranging up from 21 days, and involving various concentrations no higher than 100 ppb as a
daily maximum hourly concentration. More detailed analyses of consistent, comparable E-R information across a
relevant range of seasonal exposure levels, accompanied by detailed records of O3 concentrations, that would
support derivation of robust E-R functions for purposes discussed here are not available (ISA, Appendix 8,
section 8.10.1.2).
4-73
-------
Multiple studies have involved quantitative analysis of data collected as part of the USFS
biosite biomonitoring program (e.g., Smith, 2012). These analyses continue to indicate the
limitations in capabilities for predicting the exposure circumstances under which visible foliar
injury would be expected to occur, as well as the circumstances contributing to increased injury
severity. As noted in section 4.3.3.2 above, expanded summaries of the dataset compiled in the
2015 review from several years of USFS biosite records does not clearly and consistently
describe the shape of a relationship between incidence of foliar injury or severity (based on
individual site scores) and W126 index estimates. Overall, however, the dataset indicates that the
proportion of records having different levels of severity score is generally highest in the group of
records for sites with the highest W126 index (e.g., greater than 25 ppm-hrs for the normal and
dry soil moisture categories). Thus, the currently available evidence indicates increased
occurrence and severity at the highest category of exposures in the dataset (above 25 ppm-hrs in
terms of a W126 index), but does not provide for identification of air quality conditions, in terms
of O3 concentrations associated with the relatively lower environmental exposures most common
in the USFS dataset that would correspond to a specific magnitude of injury incidence or severity
scores across locations.
Thus, based on considering the available information for the array of O3 welfare effects,
we again recognize the E-R relationships available in the last review for purposes of considering
O3 exposure levels associated with growth-related impacts to be the most robust E-R information
available. The currently available evidence for growth-related effects, including that newly
available in this review, does not indicate the occurrence of growth-related responses attributable
to cumulative O3 exposures lower than was established at the time of the last review. With regard
to visible foliar injury, the available information continues to be limited with regard to estimating
occurrence and severity across a range of air quality conditions, providing for only limited and
somewhat qualitative conclusions related to potential occurrence and/or severity under different
air quality conditions. The quantitative information for other effects is still more limited, as
recognized in sections 4.3.3 and 4.3.4 above. Thus, the newly available evidence does not
appreciably address key limitations or uncertainties needed to expand capabilities for estimating
welfare impacts that might be expected as a result of differing patterns of O3 concentrations in
the U.S.
• Does the current evidence continue to support a cumulative, seasonal exposure
index, such as the W126 function, as a biologically relevant and appropriate metric
for assessment of vegetation-related effects of O3 in ambient air?
As in the last review, the currently available evidence continues to support a cumulative,
seasonal exposure index as a biologically relevant and appropriate metric for assessment of the
evidence of exposure/risk information for vegetation, most particularly for growth-related
4-74
-------
effects. The most commonly used such metrics are the SUM06, AOT40 (or AOT60) and W126
indices (ISA, section IS.3.2).71 The evidence for growth-related effects continues to support
important roles for cumulative exposure and for weighting higher concentrations over lower
concentrations. Thus, among the various such indices considered in the literature, the cumulative,
concentration-weighted metric, defined by the W126 function, continues to be best supported for
purposes of relating O3 air quality to growth-related effects. Accordingly, in our consideration of
the potential for vegetation-related effects to occur under air quality conditions associated with
the current standard, we continue to focus on the W126 index as the appropriate metric. In so
doing, we also recognize, as recognized in the past, that this metric may not well describe the key
circumstances of O3 exposure for occurrences of other effects, particularly, visible foliar injury.
As discussed in section 4.3.3.2 above, the evidence indicates an important role for peak
concentrations (e.g., N100) in influencing the occurrence and severity of visible foliar injury.
4.5.1.2 General Approach for Considering Public Welfare Protection
The general approach and risk management framework applied in 2015 for making
judgements and reaching conclusions regarding the adequacy of public welfare protection
provided by the newly established secondary standard is summarized in section 4.1 above. In
light of the current evidence and air quality information, we discuss here key considerations in
judging public welfare protection provided by the O3 secondary standard. In so doing, we
address a series of questions.
• Does the newly available information continue to support the use of tree seedling
RBL as a proxy for the broad array of vegetation-related effects?
As summarized in section 4.3 above, the currently available evidence is largely consistent
with that available in the last review and does not call into question conceptual relationships
between plant growth impacts and the broader array of vegetation effects. Rather, the ISA for the
current review describes (or relies on) such relationships in considering causality determinations
for ecosystem-scale effects such as altered terrestrial community composition and reduced
productivity, as well as reduced carbon sequestration, in terrestrial ecosystems (ISA, Appendix 8,
sections 8.8 and 8.10). Thus, the current evidence continues to support the use of tree seedling
RBL as a proxy for the broad array of vegetation-related effects, most particularly those
conceptually related to growth.
71 While the evidence includes some studies reporting 03-reduced soybean yield and perennial plant biomass loss
using AOT40 (as well as W126) as the exposure metric, no newly available analyses are available that compare
AOT40 to W126 in terms of the strength of association with such responses. Nor are studies available that
provide analyses of E-R relationships for AOT with reduced growth or RBL with such extensiveness as the
analyses supporting the established E-R functions for W126 with RBL and RYL.
4-75
-------
In the last review, the Administrator recognized the view that an appropriate
consideration of RBL was as a surrogate for an array of adverse welfare effects. Based on
consideration of ecosystem services and potential for impacts to the public, as well as conceptual
relationships between vegetation growth effects and ecosystem-scale effects, biomass loss could
be appropriately described as a scientifically valid surrogate of a variety of adverse effects to
public welfare (Frey, 2014, pp. iii, 9-10).72 In light of this, and in consideration of the broader
evidence base and public welfare implications, including associated strengths, limitations and
uncertainties, the Administrator focused on RBL, not simply in making judgments specific to a
magnitude of growth effect in seedlings that would be acceptable or unacceptable in the natural
environment, but as a surrogate or proxy for consideration of the broader array of vegetation-
related effects of potential public welfare significance, that included effects on growth of
individual sensitive species and extended to ecosystem-level effects, such as community
composition in natural forests, particularly in protected public lands (80 FR 65406, October 26,
2015). The information available in this review does not call into question this approach,
indicating there to be continued support for the use of tree seedling RBL as a proxy for the broad
array of vegetation-related effects, most particularly those conceptually related to growth.
Beyond tree seedling growth, on which RBL is specifically based, two other vegetation
effect categories with extensive evidence bases are crop yield and visible foliar injury. In setting
the current standard in 2015, as summarized in section 4.1 above, the Administrator considered
what those evidence bases indicated regarding the need for additional protection specifically for
those effects, judging that the available information, with associated limitations and
uncertainties, did not provide support for such a conclusion. With regard to crop yield, she
recognized the significant role of agricultural management practices in agricultural productivity,
as well as market variability, concluding that, in describing her public welfare protection
objectives, additional attention to this endpoint was not necessary. The rough similarities in
estimated W126 levels of median crops and tree species are also noteworthy. In the 2015
decision, the Administrator concluded that a standard set based on public welfare protection
72 The CASAC letter on the second draft PA in that review stated the following (Frey, 2014, p. 9-10):
For example, CASAC concurs that trees are important from a public welfare perspective because
they provide valued services to humans, including aesthetic value, food, fiber, timber, other forest
products, habitat, recreational opportunities, climate regulation, erosion control, air pollution
removal, and hydrologic and fire regime stabilization. Damage effects to trees that are adverse to
public welfare occur in such locations as national parks, national refuges, and other protected
areas, as well as to timber for commercial use. The CASAC concurs that biomass loss in trees is a
relevant surrogate for damage to tree growth that affects ecosystem services such as habitat
provision for wildlife, carbon storage, provision of food and fiber, and pollution removal. Biomass
loss may also have indirect process-related effects such as on nutrient and hydrologic cycles.
Therefore, biomass loss is a scientifically valid surrogate of a variety of adverse effects to public
welfare.
4-76
-------
objectives described in terms of cumulative exposures and relationships with tree seedling RBL
was an appropriate means to, and would, provide appropriate protection for the array of
vegetation-related effects. In summary, the information available in the current review does not
call into question such conclusions and continues to be supportive of the use of tree seedling
RBL as a proxy for the broad array of vegetation-related effects. Such effects and key
considerations with regard to protection afforded by the current standard are addressed in section
4.5.1.3 below.
• To what extent does the available information alter our understanding of an
appropriate magnitude of RBL, in its role as a surrogate or proxy, reasonably
expected to be of public welfare significance?
In considering the RBL estimate on which to focus in its role as a surrogate or proxy for
the full array of vegetation effects in the last review, the Administrator endeavored to identify a
secondary standard that would limit 3-year average O3 exposures somewhat below W126 index
values associated with a 6% RBL median estimate from the established species-specific E-R
functions.73 This led to identification of a seasonal W126 index value of 17 ppm-hrs that the
Administrator concluded appropriate as a target at or below which the new standard would
generally restrict cumulative seasonal exposures (80 FR 65407, October 26, 2015). In identifying
this exposure level as a target, the Administrator, recognizing limitations and uncertainties in the
evidence and variability in biota and ecosystems in the natural environment, additionally judged
that RBL estimates associated with isolated rare instances of marginally higher cumulative
exposures (in terms of a 3-year average W126 index), e.g., those that round to 19 ppm-hrs
(which corresponds to 6% RBL as median of 11 established E-R functions), were not indicative
of adverse effects to the public welfare (80 FR 65409, October 26, 2015).
The information newly available in this review does not differ from that available in the
last review with regard to a magnitude of RBL in the median species appropriately considered a
reference for judgments concerning potential vegetation-related impacts to the public welfare.
The currently available evidence continues to indicate conceptual relationships between reduced
growth and the broader array of vegetation-related effects. Quantitative representations of such
relationships have been used to study potential impacts of tree growth effects on such larger-
scale effects as community composition and productivity with the results indicating the array of
complexities involved (e.g., ISA, Appendix 8, section 8.8.4). Given their purpose in exploring
complex ecological relationships and their responses to environmental variables, as well as
limitations of the information available for such work, these analyses commonly utilize
73 The CASAC in the last review stated that 6% RBL in seedlings for the median tree species (a metric it described
as a valid surrogate for consideration of broader public welfare impacts) was "unacceptably high" in the context
of protecting against "current and anticipated welfare effects of ozone" (Frey, 2014, p. iii).
4-77
-------
somewhat general representations. This work indicates how established the existence of such
relationships is, while also identifying complexities inherent in quantitative aspects of such
relationships and interpretation of estimated responses. Thus, the currently available evidence is
little changed from the last review with regard to informing identification of an RBL reference
point reflecting ecosystem-scale effects with public welfare impacts elicited through such
linkages.
• What does the information available in the current review indicate with regard to
support for use of a 3-year average seasonal W126 index as the cumulative exposure
metric (associated with a target value of 17 ppm-hrs) for describing the public
welfare protection objectives for the secondary standard?
In setting the current standard, as described in section 4.1 above, the Administrator
focused on control of seasonal cumulative exposures in terms of a 3-year average W126 index.
The evaluations in the PA for that review recognized there to be limited information to discern
differences in the level of protection afforded for cumulative growth-related effects by a standard
focused on a single-year W126 index as compared to a 3-year W126 index (80 FR 65390).
Accordingly, the identification of the 3-year average for considering the seasonal W126 index
recognized that there was year-to-year variability not just in O3 concentrations, but also in
environmental factors, including rainfall and other meteorological factors, that influence the
occurrence and magnitude of Os-related effects in any year (e.g., through changes in soil
moisture), contributing uncertainties to projections of the potential for harm to public welfare (80
FR 65404 October 26, 2015). Based on this recognition, as well as other considerations, the
Administrator expressed greater confidence in judgments related to projections of public welfare
impacts based on seasonal W126 index estimated by a 3-year average and accordingly, relied on
that metric.
Among the factors referenced in the Administrator's 2015 decision to focus on a 3-year
average W126 in assessing potential impacts on vegetation and protection from Os in ambient
air, were consideration of the strengths and limitations of the evidence, and of the information on
which to base her judgments with regard to adversity of effects on the public welfare. With
regard to the current evidence we first consider the evidence and information underlying the E-R
functions and the extent to which it is specific to a single seasonal exposure, e.g., as compared to
an average across multiple seasons. In so doing, we also take note of aspects of the evidence that
reflect variability in organism response under different experimental conditions and the extent to
which this variability is represented in the available data, which might indicate an
appropriateness of assessing environmental conditions using a mean across seasons in
recognition of the existence of such year-to-year variability in conditions and responses. An
additional aspect of the information underlying the E-R functions that may be relevant to
4-78
-------
consider is the extent to which the exposure conditions represented include those associated with
0'! concentrations that meet the current standard, and the extent to which tree seedling growth
responses to such conditions may have been found to not be significantly different from
responses to the control (e.g., zero Os) conditions. The extent to which E-R predictions are
extrapolated beyond the tested exposure conditions also contributes to uncertainty which may
argue for a less precise interpretation, such as an average across multiple seasons.
As an initial matter, we note that while the tree seedling E-R functions for the 11 species
have been derived in terms of a seasonal (single-year) W126 index, the experiments from which
they were derived vary in duration from periods of 82 to 140 days over a single year to periods
of 180 to 555 days across two years (Appendix 4A, Table 4A-5; Lee and Hogsett, 1996). In order
to produce E-R functions for 3-month periods from the experiments of variable durations
adjustments were made based on assumptions regarding relationships between duration,
cumulative exposure in terms of W126 index, and plant growth response (Lee and Hogsett,
1996). Specifically, the E-R functions were derived based on the exposure duration of the
experiment and adjusted or normalized to 3-month periods based on assumptions regarding
relationships between duration, cumulative exposure in terms of the W126 index and plant
growth response (see Lee and Hogsett, 1996, section 1.3). For example, while the functions are
defined as describing a seasonal response, some were derived by distributing responses observed
at the end of two seasons of varying exposures equally across the two seasons (essentially
applying the average to both seasons). Thus, the growth response data for some of the 11 species
with established E-R functions were collected over time periods longer than a single year's
growing season and those functions are reflecting the average E-R relationship across the longer
period. Consequently, they cannot provide precise estimates of response from a single year's
exposure (e.g., vs averages over a longer period which may span multiple growing seasons).
Additionally, the number of experiments available for each species for which E-R
functions have been established also varies. For example, there are 14 experimental studies for
aspen (seven for which the E-R function for wild aspen has been derived and seven supporting a
function for aspen clones) and 11 for ponderosa pine (Appendix 4A, Table 4A-5). The number of
studies available for the other species, however, is much lower, e.g., only two or three for the
three species generally exhibiting greater sensitivity than aspen and ponderosa pine (based on the
available data) (Appendix 4A, section 4A-2, Table 4A-5; 1996 AQCD, Table 5-28; Lee and
Hogsett, 1996). The E-R functions for both the aspen and ponderosa pine experiments illustrate
appreciable variability in response across experiments (Appendix 4A, Figure 4A-10). Reasons
for this variability may relate to a number of factors, including variability in seasonal response
related to variability in non-03 related environmental influences on growth, such as rainfall,
temperature and other meteorological variables, as well as biological variability across individual
4-79
-------
seedlings, in addition to potentially variability in the pattern of O3 concentrations contributing to
similar cumulative exposures.
Regarding the extent or strength of the database underlying the E-R functions for
cumulative exposure levels of interest in the current review, we note that the data appear to be
more extensive for relatively higher (e.g., at/above a SUM06 of 30 ppm-hrs) vs lower seasonal
exposures (Appendix 4A, Table 4A-6). Additionally, there are differences across the
experimental studies in the extent to which they include O3 exposure levels, in terms of W126
index, that commonly occur under air quality conditions that meet the current standard. For
example, the studies appear to be somewhat limited for W126 index values below 20 ppm-hrs.
As recognized in section 4.3.4 above, studies for five of the 11 species appear to have included
exposure treatments likely to correspond to W126 index values at or below 20 ppm-hrs
(Appendix 4A, Table 4A-5).74 We additionally note that for at least one of these species, black
cherry, the growth impacts for the lower exposure were not statistically significantly different
from those for the next higher cumulative exposure (Appendix 4A, Table 4A-6).
There is also limited evidence that allows for specific evaluation of the predictability of
growth impacts from single-year versus multiple-year average exposure estimates. Such
evidence includes multi-year studies reporting results for each year of the study, which are the
most informative to the question of plant annual and cumulative responses to individual years
(high and low) over multiple-year periods. As summarized in section 4.3.4 above, the evidence is
quite limited with regard to studies of O3 effects that report seasonal observations across multi-
year periods and that also include detailed hourly O3 concentration records (to allow for
derivation of exposure index values). One such study, which tracked exposures across six years,
is available for aspen (King et al., 2005; 2013 ISA, section 9.6.3.2; ISA, Appendix 8, section
8.13.2). This study was used in a presentation of the 2013 ISA that compared the observed
growth response to that predicted from the E-R function for aspen. Specifically, the observed
aboveground biomass (and RBL) after each of the six growing seasons was compared to
estimates derived from the aspen E-R function based on the cumulative multiple-year average
seasonal W126 index values for each year75 (2013 ISA, section 9.6.3.2). The conclusions reached
74 For five of the species in Table 4A-5 in Appendix 4A, SUM06 index values below 25 ppm-hrs range from 12 to
21.7. In considering these values, we note that an approach used in the 2007 Staff Paper on specific temporal
patterns of O3 concentrations concluded that a SUM06 index value of 25 ppm-hrs would be estimated to
correspond to a W126 index value of approximately 21 ppm-hrs (U.S. EPA, 2007, Appendix 7B, p. 7B-2).
Accordingly, we conclude that a SUM06 value of 21 ppm-hrs would be expected to correspond to a W126 index
value below 20 ppm-hrs.
75 Although not emphasized or explained in detail in the 2013 ISA, the W126 estimates used to generate the
predicted growth response were cumulative average. To clarify, the cumulative average W126 for year 1 is
simply the W126 index for that year (e.g., based on highest 3 months). For year 2, it is the average of the year 1
4-80
-------
were that the agreement between the set of predictions and the Aspen FACE observations were
"very close" (2013 ISA, p. 9-135). The results indicate that when considering O3 impacts across
multiple years, a multi-year average index yields predictions close to observed measurements
(2013 ISA, section 9.6.3.2 and Figure 9-20; Appendix 4A, section 4.A.3). Further analyses using
observations from the multi-year study analyzed in the ISA (King et al., 2005) are presented in
Appendix 4A. These analyses suggest that estimation of aboveground aspen biomass over a
multi-year 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 may yield similar total biomass estimates after
multiple years (Appendix 4A, section 4A.3).
Thus, while the E-R functions are based on strong evidence of seasonal and cumulative
seasonal O3 exposure reducing tree growth, and while they provide for quantitative
characterization of the extent of such effects across O3 exposure levels of appreciable magnitude,
there is uncertainty associated with the resulting RBL predictions. Further, the current evidence
does not indicate single-year seasonal exposure in combination with the established E-R
functions to be a better predictor of RBL than a seasonal exposure based on a multi-year average,
or vice versa (Appendix 4A, section 4A.3.1). Rather, there is uncertainty, implying an
imprecision or inexactitude, in the resulting predictions. In light of this, the current evidence
does not support concluding there to be an appreciable difference in the effect of three years of
exposure held at 17 ppm-hrs compared to a 3-year exposure that averaged 17 ppm-hrs yet varies
by 5 to 10 ppm (e.g., 7 ppm-hrs) from that in any of the three years. All of the factors identified
here, the currently available evidence and recognized limitations, variability and uncertainties,
contribute uncertainty and resulting imprecision or inexactitude to RBL estimates of single-year
seasonal W126 index values. Thus, the available information indicates no lesser support for use
of an average seasonal W126 index derived from multiple years (with their representation of
variability in environmental factors), such as for a 3-year period, for estimating median RBL
using the established E-R functions, than for a single-year index.
• What does the currently available information indicate for considering potential
public welfare protection from 03-related visible foliar injury?
In establishing the current secondary standard in 2015 and its underlying public welfare
protection objectives, as summarized in section 4.1, above, the Administrator focused primarily
on RBL in tree seedlings as a proxy or surrogate for the full array of vegetation related effects of
O3 in ambient air, from sensitive species to broader ecosystem-level effects. In so doing, she also
seasonal Wl 26 and year 2 seasonal W126, and so on. So that for year 6 it is the average of each of the six year's
seasonal Wl 26 index values.
4-81
-------
concluded there to be support for establishing a strengthened standard provided by the then-
available information regarding visible foliar injury, taking note that the available analyses of
USFS biosite data, which indicated declines in BI scores with reductions in cumulative W126
from well above 20 ppm-hrs to lower levels (80 FR 65407-65408, October 26, 2015). She also
concluded, however, that, due to associated uncertainties and complexities, the evidence was not
conducive to use for identifying a quantitative public welfare protection objective focused
specifically visible foliar injury. In reaching this conclusion, she specifically recognized
significant challenges in judging the specific extent and severity at which such effects should be
considered adverse to public welfare, in light of the variability in the occurrence of visible foliar
injury and the lack of clear quantitative relationships (including robust exposure-response
functions) that would allow prediction of visible foliar injury severity and incidence under
varying air quality and environmental conditions, as well as the lack of established criteria or
objectives that might inform consideration of potential public welfare impacts related to this
vegetation effect (80 FR 65407, October 26, 2015).
As an initial matter, we note that, as recognized in the last review, some level of visible
foliar injury can impact public welfare and thus might reasonably be judged adverse to public
welfare.76 As summarized in section 4.3.2 above, depending on its spatial extent and severity,
there are many locations in which visible foliar injury can adversely affect the public welfare.
For example, significant, readily perceivable and widespread injury in national parks and
wilderness areas can adversely impact the perceived scenic beauty of these areas, impacting the
aesthetic experience for both outdoor enthusiasts and the occasional park visitor. Such
considerations have also been recognized by the Agency in past reviews, in which decisions to
revise the O3 secondary standard emphasized protection of Class I areas, which are areas such as
national wilderness areas and national parks given special protections by the Congress (e.g., 73
FR 16496, March 27, 2008, "the Administrator concludes it is appropriate to revise the
76 As stated in the 2015 decision notice: "both tree growth-related effects and visible foliar injury have the potential
to be significant to the public welfare" (80 FR 65377, October 26, 2015); "Ch-induced visible foliar injury also
has the potential to be significant to the public welfare through impacts in Class I and other similarly protected
areas" (80 FR 65378, October 26, 2015); "[depending on the extent and severity, 03-induced visible foliar injury
might be expected to have the potential to impact the public welfare in scenic and/or recreational areas during the
growing season, particularly in areas with special protection, such as Class I areas. (80 FR 65379, October 26,
2015); "[t]he Administrator also recognizes the potential for this effect to affect the public welfare in the context
of affecting values pertaining to natural forests, particularly those afforded special government protection (80 FR
65407, October 26, 2015). The CAS AC in the last review also stated that visible foliar injury "can impact public
welfare" (Frey, 2014, p. 10).
4-82
-------
secondary standard, in part, to provide increased protection against Ch-caused impairment to
such protected vegetation and ecosystems").77
Further, as discussed in section 4.3.3.2 above, a study identified in the 2013 ISA that
focused on visible foliar injury in west coast states observed that both percentage of USFS
biosites with injury and average biosite index were higher for sites with average cumulative O3
concentrations above 25 ppm-hrs in terms of SUM06 as compared to groups of sites with lower
average cumulative exposure concentrations, with little difference apparent between the two
lower exposure groups (80 FR 65395, October 26, 2015; Smith and Murphy, 2015; Campbell et
al., 2007, Figures 27 and 28 and p. 30).78 Similarly, a county-scale analysis of USFS biosite data
in the 2007 Staff Paper (from earlier years than those analyzed in the 2015 review) indicated a
somewhat smaller incidence of nonzero BI biosites in counties with O3 exposures below a
SUM06 metric of 15 ppm-hrs (or a fourth-high metric of 74 ppb) as compared to larger groups
that also included sites with SUM06 values up to 25 ppm-hrs (or fourth-high metric up to 84
ppb) (U.S. EPA 2007, pp. 7-63 to 7-64; 80 FR 65395, October 26, 2015). This indication that the
averaging time and form of the current standard, which emphasize peak concentrations through a
short (8-hour) averaging time and a rare-occurrence form (annual fourth highest daily
maximum), exert some control on the incidence of sites with visible foliar injury has a
conceptual similarity to the finding of the most recent and extensive USFS data analysis that
reductions in peak 1-hour concentrations have influenced the declining trend in visible foliar
injury since 2002 (Smith, 2012).
Additional characterization of USFS biosite data in the current review is based on a more
complete and further description of a dataset developed in the last review by combining USFS
biosite records with W126 index estimates and a categorization of soil moisture (Appendix 4C).
While recognizing limitations in the dataset79 and considering the records for the normal or dry
77 In the discussion of the need for revision of the 1997 secondary standard, the 2008 decision noted that "[i]n
considering what constitutes a vegetation effect that is adverse from a public welfare perspective, ... the
Administrator has taken note of a number of actions taken by Congress to establish public lands that are set aside
for specific uses that are intended to provide benefits to the public welfare, including lands that are to be protected
so as to conserve the scenic value and the natural vegetation and wildlife within such areas, and to leave them
unimpaired for the enjoyment of future generations" (73 FR 16496, March 27, 2008). This passage of the 2008
decision notice clarified that "[s]uch public lands that are protected areas of national interest include national
parks and forests, wildlife refuges, and wilderness areas" (73 FR 16496, March 27, 2008).
78 In considering their findings, the authors expressed the view that" [although the number of sites or species with
injury is informative, the average biosite injury index (which takes into account both severity and amount of
injury on multiple species at a site) provides a more meaningful measure of injury" for their assessment at a
statewide scale (Campbell et al., 2007).
79 For example, the majority of these data are records with W126 index estimates at or below 9 ppm-hrs, and fewer
than 10% of the records have W126 estimates above 15 ppm-hrs. Additionally, the BI scores are quite variable
4-83
-------
soil moisture categories, for which there is somewhat better representation of W126 index levels
above 13 ppm-hrs,80 we note that the percentage of USFS records recording visible foliar injury
(of any severity level) presents no clear trend across W126 index estimates below 17 ppm-hrs.
Among records in the normal soil category, BI scores are noticeably increased in the highest
W126 index bin (above 25 ppm-hrs) compared to the others. The percentages of records in the
greater than 25 ppm-hrs bin that have BI scores above 15 ("moderate" and "severe" injury) and
above 5 ("light," "moderate" and "severe" injury) are more than three times greater than
percentages for these score levels in any of the lower W126 bins. Additionally, the average BI of
7.9 in the greater-then-25-ppm-hrs bin is more than three times the next highest bin average. The
average BI in the next two lower W126 bins (which vary inversely with W126 index) are just
slightly higher than average Bis for the rest of the bins, and the average BI for all bins at or
below 25 ppm-hrs are well below 5. Among records in the dry soil moisture category, the two
highest W126 bins (which together include the W126 index estimates above 19 ppm-hrs) exhibit
the highest percentages of records with BI above 15 or above 5. The average score for all dry soil
moisture records in each W126 index bin is highest for W126 above 25 ppm and somewhat
higher for W126 above 17 ppm-hrs. Yet all three of those averages are below 5, indicating little
or no injury.
Thus, the strongest conclusions that can be reached from the USFS dataset described in
Appendix 4C are that the incidence/prevalence of sites with more severe injury (e.g., BI score
above 15 or 5) is also lower at sites with W126 index values below 25 ppm-hrs than at sites with
higher index values and that clear trends in such incidence/prevalence related to increasing
W126 index levels are not evident across the lower W126 estimates. When scores characterized
as "little injury" by the USFS classification scheme are included (i.e., when considering all
scores above zero for the normal soil moisture category), there is a slight suggestion of increased
incidence of records for the W126 index bins above 19 or 17 ppm-hrs, although for of the bins at
or below 25 ppm-hrs, the incidence is less than 5%. As discussed in section 4.3.3.2 above,
variability in the data across sites, and uncertainty with regard to the role of peak O3
concentrations as an influence on occurrence of visible foliar injury separate from cumulative
W126 index, lead to the conclusion that the currently available information does not support
precise conclusions as to the severity and extent of such injury associated with the lower values
of W126 index most common at USFS sites during the time of the dataset (2006-2010).
Notwithstanding this, records categorized as normal soil moisture indicate there to be an
across the full dataset, with even the bin for the lowest W126 index estimates (below 7 ppm-hrs) including BI
scores well above 15 (Appendix AC, section 4C.4.2).
80 In the case of records in the wet soil moisture category, nearly 90% of the records are for W126 estimates at or
below 9 ppm-hrs, limiting interpretations for higher W126 bins (Appendix AC, Table 4C.4 and section 4C.6).
4-84
-------
appreciable difference in severity of injury between records with W126 index estimates above 25
ppm-hrs and those with estimates at or below 25 ppm-hrs (e.g., Appendix 4C, Figures 4C-5 and
4C-6 and Table 4C-5). The records categorized as dry soil moisture do not indicate such a clear
pattern. The records categorized as wet soil moisture are much too limited (and variability) for
W126 index estimates above 13 ppm-hrs to support a conclusion (Appendix 4C). Thus, we
conclude, based primarily on the records categorized as having normal soil moisture, that under
conditions that maintain W126 index values below 25 ppm-hrs a reduced severity (average BI
score below 5) and incidence of visible foliar injury, as quantified by biosite index scores, would
be expected.
As discussed in section 4.3.3.2 above, consistent relationships of injury extent and
severity with vegetation exposure circumstances have not been developed. The current evidence
indicates a role for cumulative seasonal concentration-weighted metrics such as SUM06 and
W126 indices, while also indicating an importance of the occurrence of particularly high
concentrations (e.g., hours above a concentration such as 100 ppb). Thus, in making judgements
regarding air quality conditions of concern with regard to impacts associated with incidence and
severity of visible foliar injury, it is appropriate to consider both cumulative concentration-
weighted seasonal exposures and the occurrence of peak concentrations. In this context, it is
appropriate to recognize the control of peak concentrations exerted by the form and averaging
time of the current standard. For example, as noted in chapter 2, daily maximum 1-hour, as well
as 8-hour average O3 concentrations have declined over the past 15 years, a period in which there
have been two revisions of the level of the secondary standard, each providing greater
stringency, while retaining the same averaging time and form as the current standard (e.g.,
Figures 2-8, 2-9 and 2-14).
Further, we note that judgments related to the extent of public welfare impacts of visible
foliar injury depend on the severity and extent of the injury, as well as the location where the
effects occur and the associated intended use. As noted in section 4.3.2 above, aesthetic value
and outdoor recreation depend, at least in part, on the perceived scenic beauty of the
environment. Accordingly, depending on its spatial extent and severity, visible foliar injury in
national parks and wilderness areas can adversely impact the aesthetic experience for both
outdoor enthusiasts and the occasional park visitor. Beyond the limitations associated with the
evidence for descriptive quantitative relationships for O3 concentrations and visible foliar injury,
we further face a paucity of information clearly relating differing severity and prevalence of
injury to conditions in natural areas that would reasonably be concluded to impact public use and
enjoyment in a way that might suggest adversity to the public welfare. The available information
does not yet address or describe the relationships expected to exist between some level of
severity and/or extent of location affected and scenic or aesthetic values (e.g., reflective of visitor
4-85
-------
enjoyment and likelihood of frequenting such areas). However, it might reasonably be expected
that in cases of widespread and relatively more severe injury during the growing season
(particularly when sustained across multiple years, and accompanied by obvious impacts on the
plant canopy), Ch-induced visible foliar injury could adversely impact the public welfare in
scenic and/or recreational areas, particularly in parks and other areas with special protection,
such as Class I areas. In summary, the available evidence does not include characterization of
USFS biosite scores with regard to public perception and potential impacts on public enjoyment.
Nor does it address this in combination with information on whether air quality conditions in
sites with scores of a particular severity level do or do not meet the current standard.
In consideration of all of the above, we recognize the appreciable limitations of the
current information touched on above with regard to providing a foundation for judgments on
public welfare protection objectives specific to visible foliar injury. In light of such limitations
and in light of the above discussion, we then consider what the available information indicates
with regard to potential for adverse effects to the public welfare related to visible foliar injury
under air quality conditions allowed by the current standard, which was established with a focus
on protecting against RBL as a surrogate/proxy for the broad array of vegetation-related effects.
As recognized in section 4.3.1 above, while the evidence continues to show a consistent
association between the occurrence of visible injury and ozone, "visible foliar injury is not
always a reliable indicator of other negative effects on vegetation" (ISA, Appendix 8, section
8.2). Based on the USFS biosite data, the conditions associated with visible foliar injury in
locations with sensitive species appear to relate to peak concentration as well as sustained
exposure to higher concentrations over the growing season, such that cumulative exposure
metrics may not well or completely describe or predict the occurrence and severity of injury.
• What does the currently available information indicate for considering potential
public welfare protection from 03-related climate effects?
In considering the currently available information for the effects of the global abundance
of O3 in the troposphere on radiative forcing, and temperature, precipitation and related climate
variables, we note as an initial matter that, as summarized in section 4.3.3 above, there are
limitations and uncertainties in the associated evidence bases with regard to assessing potential
for occurrence of climate-related effects as a result of varying O3 concentrations in ambient air of
locations in the U.S. The current evidence is limited with regard to support for such quantitative
analyses that might inform considerations related to the current standard. For example, as stated
in the ISA, "[c]urrent limitations in climate modeling tools, variation across models, and the
need for more comprehensive observational data on these effects represent sources of uncertainty
in quantifying the precise magnitude of climate responses to ozone changes, particularly at
regional scales" (ISA, section 9.3.1). These are "in addition to the key sources of uncertainty in
4-86
-------
quantifying ozone RF changes, such as emissions over the time period of interest and baseline
ozone concentrations during preindustrial times" (ISA, section IS.9.3.1). Together such
uncertainties limit development of quantitative estimates of climate-related effects in response to
earth surface O3 concentrations at the regional scale, such as in the U.S. While these
complexities inhibit our ability to consider tropospheric O3 effects, such as radiative forcing, we
note that our consideration of O3 growth-related impacts on trees inherently encompasses
consideration of the potential for O3 to reduce carbon sequestration in terrestrial ecosystems
(e.g., through reduced tree biomass as a result of reduced growth). That is, limiting the extent of
03-related effects on growth would be expected to also limit reductions in carbon sequestration,
a process that can reduce the tropospheric abundance of CO2, the greenhouse gas ranked highest
in importance (section 4.3.3.3 above; ISA, section 9.1.1).
4.5.1.3 Public Welfare Implications of Air Quality under the Current Standard
Our consideration of the scientific evidence available in the current review, as at the time
of the last review, is informed by results from a quantitative analysis of estimated exposure and
associated risk. An overarching consideration is whether the current exposure/risk and air quality
information calls into question the adequacy of protection provided by the now-current standard.
As in our consideration of the evidence above, we have organized the discussion regarding the
information related to exposures and potential risks around a key question to assist us in
considering the quantitative analyses of air quality at U.S. locations nationwide, particularly
including those in Class I areas. In so doing, we consider first analyses particular to cumulative
O3 exposures, in terms of the W126 index, given its established relationship with growth-related
effects and specifically RBL as the identified proxy or surrogate for the full array of such effects.
To understand the cumulative O3 exposures likely occurring under the current standard,
we consider the air quality analyses summarized in section 4.4 above and presented in detail in
Appendix 4D. These air quality analyses of monitoring data at sites across the U.S., including
sites in Class I areas, document seasonal cumulative concentration-weighted exposures occurring
when the current standard is met. In so doing, they indicate that, as described in section 4.4.2
above, with very few exceptions (one in the most recent 3-year period), the seasonal W126 index
at sites nationwide (including those in Class I areas), as assessed by the 3-year average, are at or
below 17 ppm-hrs when the current standard is met. Further, such exposures are generally well
below 17 ppm-hrs across most of the U.S. The overall pattern for single-year seasonal W126
index values at monitors meeting the current standard in the recent period is generally similar,
with about a dozen of the 849 sites nationwide having a single-year W126 index above 19 ppm-
4-87
-------
hrs (and under two dozen above 17 ppm-hrs).81 The frequency of higher single-year W126 index
values during periods when the current standard is met is much lower for the Class I area
monitors. During the most recent three years, when the average seasonal W126 index is at or
below 17 ppm-hrs in all Class I areas meeting the current standard, there were just three single-
year W126 index values above 17 ppm-hrs and none above 19 ppm-hrs (Appendix 4D, Table
4D-16).82
Combining this information regarding likely W126-based exposure levels with the
established E-R functions for 11 tree seedling species indicates that based on monitoring data for
locations meeting the current standard during the most recent design period, the median species
RBL for tree seedlings is at or below 5.3% based on the 3-year average W126, with very few
exceptions, with the highest estimates occurring in areas not near or within Class I areas.
Looking at the data over a longer time period (2000-2018) confirms this general pattern for the
bulk of the data, with some infrequent higher occurrences, and virtually all RBL estimates below
6%.83 Further, given the variability and uncertainty associated with the data underlying the E-R
functions (as discussed in section 4.5.1.2 above), the few higher single-year occurrences are
reasonably considered to be of less significance than 3-year average values.
With regard to visible foliar injury, as discussed in section 4.3.3.2 above, the evidence is
somewhat limited and unclear with regard to the metric and quantitative approach that well
describes a relationship between incidence or severity of injury in U.S. forests across a broad
range of air quality conditions. As indicated in the Appendix 4C presentation of the dataset
developed from USFS biosite records, W126 index estimates and categorizations of soil
moisture, while most of the records (more than 95%) in the dataset are for W126 exposure index
estimates below 17 ppm-hrs - and there is appreciable variability in incidence of records with
nonzero BI scores, and more importantly with records above 5 or 15 (scores associated with
injury considered greater than "a little" by the USFS scheme), the increased incidence of such
scores appears most consistently with higher W126 estimates. The incidence is greatest in the bin
for the highest estimated exposures, i.e., W126 index above 25 ppm-hrs, which are not seen to
81 These highest W126 index values occur in the South West and West regions in which there are nearly 150
monitor locations meeting the current standard (Figure 4-6; Appendix 4D, Figure 4D-5, Table 4D-1).
82 Across the full 19-year dataset for Class I area monitors meeting the current standard (58 monitors with at least
one such period), there are 15 design value periods with single-year W126 index values above 19 ppm-hrs, all of
which are prior to the 2013-2015 period (Appendix 4D, section 4D.3.2.4).
83 Although potential for effects on crop yield was not given particular emphasis in the last review (for reasons
similar to those summarized earlier), we additionally note that combining the exposure levels summarized for
areas across the U.S. where the current standard is met with the E-R functions established for 10 crop species
indicates a median RYL across crops to be at or below 5.1%, on average, with very few exceptions. Further,
estimates based on W126 index at the great majority of the areas are below 5%.
4-88
-------
occur in Class I area monitoring sites that meet the current standard (Appendix 4C, section
4C.3). Further, as discussed in section 4.3.3.2 above, the evidence indicates a role for
occurrences of higher concentrations, such as above 100 ppb, the frequency of which has
declined in U.S. monitoring sites over the past 15 years. The analyses of hourly concentrations in
section 2 A. 2 of Appendix 2A demonstrates the substantial control of peak 1-hour concentrations
exerted by the current standard. For example, in three different datasets of monitoring data since
2000, the average number of observations at or above 100 ppb per site-design value period is
well below one for when sites were meeting the current standard, and well above at sites when
not meeting the current standard (Appendix 2A, Tables 2A-2, 2A-3 and 2A-4). Thus, although
the current information does not establish a metric or combination of metrics that well describes
the relationship between occurrence and severity of visible foliar injury across a broad range of
O3 concentration patterns from those more common in the past to those in areas recently meeting
today's standard, the current evidence and currently available air quality information indicates
that the exposure conditions occurring at sites with air quality meeting the current standard are
not those that might reasonably be concluded to elicit the occurrence of significant foliar injury
(with regard to severity and extent).
• Are such exposures (in terms of W126 index) that occur in areas that meet the
current standard indicative of welfare effects reasonably judged important from a
public welfare perspective? What are important associated uncertainties?
Given the findings summarized in section 4.4 above regarding W126 index values in
areas where the current standard is met, we reflect on the potential public welfare significance of
vegetation-related effects that may be associated with such exposures. This consideration is
important to informing the Administrator's judgment on the secondary standard, which is not
meant to protect against all known or anticipated 03-related welfare effects, but rather those that
are judged to be adverse to the public welfare (as noted in section 4.3.2 above). Accordingly, for
the purposes of informing that judgment, we consider here the exposures indicated to occur
under conditions that meet the current standard, the associated potential for effects and the
potential public welfare implications.
As an initial matter, we recognize the increased significance to the public welfare of
effects in areas that have been accorded special protection, such as Class I areas, while noting
some general similarities of the exposure estimates in Class I areas for periods when the current
standard was met to such estimates at monitoring sites in other areas, as documented in the larger
air quality data analysis. Across both datasets, and extending back 19 years, the cumulative
exposure estimates, averaged over the design value period, for these air quality conditions were
virtually all at or below 17 ppm-hrs, with most of the W126 index values below 13 ppm-hrs
(Appendix 4D, Table 4D-9), corresponding to median RBL estimates of 3.8% or less (based on
4-89
-------
the established tree seedling E-R relationships detailed in Appendix 4A). We additionally note
that single-year W126 index values in Class I areas over the 19-year dataset evaluated were
generally at or below 19 ppm-hrs, particularly in the more recent years (Appendix 4D, section
4D.3.2.3).
Regarding the effects associated with the exposures commonly occurring, we consider
first the categories of effects for which the quantitative information related to exposure and
associated effects is most well developed. As in the last review, these are effects on plant growth.
Based on the median of RBL estimates derived from the established E-R functions for 11 tree
species seedlings, W126 index values at or below 17 ppm-hrs correspond to median species tree
seedling RBL estimates at or below 5.3% (Appendix 4A, Table 4A-5). Judgments in the last
review (in the context of the framework considered in section 4.5.1.2 above) concluded isolated
rare occurrences of exposures for which median RBL estimates might be at or just above 6% to
not be indicative of conditions adverse to the public welfare, particularly considering the
variability in the array of environmental factors that can influence O3 effects in different systems,
and the uncertainties associated with estimates of effects in the natural environment.
In the last review, the Administrator focused on cumulative exposure estimates derived as
the average W126 index over the 3-year design value period, concluding variations of single-
year W126 index from the average to be of little significance. This focus generally reflected the
judgment that estimates based on the average adequately, and appropriately reflected the
precision of current understanding of Ch-related growth reductions, given the various limitations
and uncertainties in such predictions. Additional analyses have been explored in the current
review to further examine this issue, as summarized in section 4.5.1.2 above. The current air
quality data indicates single-year W126 index values generally to vary by less than 5 ppm-hrs
from the 3-year average when the 3-year average is below 20 ppm-hrs (which is the case for
locations meeting the current standard). With such variation, year-to-year differences in tree
growth responding to each year's seasonal exposure from estimated response based on the 3-year
average of those seasonal exposures would, given the offsetting impacts of seasonal exposures
above and below the average, reasonably be expected to generally be small over tree lifetimes.
Additionally, we have also further considered the experimental data underlying the E-R
functions for estimating RBL, particularly those pertaining to cumulative exposures on the order
of 17 ppm-hrs and informing estimates of multiyear impacts. We note limitations in the evidence
base in these regards, as discussed further in section 4.5.1.2 above, that contribute to imprecision
or inexactitude to estimates of growth impacts associated with multi-year exposures in this range.
Further, the information newly available in the current review does not appreciably address these
limitations and uncertainties to improve the certainty or precision in RBL estimates for such
exposures.
4-90
-------
With regard to visible foliar injury, as noted in section 4.3.3.2 above, the dataset based on
USFS biomonitoring data that was developed in the last review (see Appendix 4C) does not
provide for a clear predictive relationship between O3, in terms of W126 index, and incidence of
injury or magnitude of injury score. It additionally indicates variability in the incidence of
nonzero injury scores (or of scores of relatively greater severity), particularly for records with
generally lower W126 index estimates (e.g., below 19 ppm-hrs). Further, the sample size for
sites with wet soil moisture conditions is quite limited for sites with W126 index above 13 ppm-
hrs, limiting conclusions for those situations. Further, as discussed in sections 4.3.3.2 and 4.5.1.2
above, a quantitative description of the relationship between O3 concentrations and visible foliar
injury extent or incidence, as well as severity, that would support estimation of injury under
varying air quality and environmental conditions (e.g., moisture), most particularly for locations
that meet the current standard is not yet established in the evidence.
In light of the potential role of peak O3 concentrations (e.g., hourly concentrations at or
above 100 ppb) as an influence on visible foliar injury occurrence and severity, it is of interest to
take note of analyses in Chapter 2 and Appendix 2A. These indicate that the magnitude of daily
maximum 1-hour concentrations has declined appreciably since 2000. For example, the median
annual 2nd highest MDA1 concentration across U.S. monitoring sites has declined from 100 to 80
ppb (Figure 2-17 above). The analysis in Appendix 2A of three recent design value periods
(covering 2014 through 2018) and three periods more than ten years prior (covering 2000
through 2004) show that for all sites with DVs that would meet the current standard during these
periods, there is less than one observation per site, on average, at or above 100 ppb. There are
roughly 40 times more such observations per site, on average, for sites with DVs that would not
meet the current standard (Appendix 2A, section 2A.2). These data indicate that the current
standard provides appreciable control of peak 1-hour concentrations, and thus, to the extent that
such peak concentrations play a role in the occurrence and severity of visible foliar injury, the
current standard also provides appreciable control of these.
Additionally, as discussed in section 4.3.2 above, the public welfare implications
associated with visible foliar injury (when considered as an effect separate from effects on plant
physiology) relate largely to effects on scenic and aesthetic values. The available information
does not yet address or describe the relationships expected to exist for some level of visible foliar
injury severity (below that at which broader physiological effects on plant growth and survival
might also be expected) and/or extent of location or site injury (e.g., BI) scores with values held
by the public and associated impacts on public uses of the locations.84 As discussed in section
84 Information with some broadly conceptual similarity to this has been used forjudging public welfare implications
of visibility effects of PM in setting the PM secondary standard (78 FR 3086, January 15, 2012).
4-91
-------
4.3.2 above, this gap limits our ability to identify air quality conditions that might be expected to
provide a specific level of protection from public welfare effects of this endpoint (e.g., separate
from effects that might relate to plant growth and reproduction under conditions where foliar
injury may also be severe). Additionally, as recognized in the sections above, no criteria have
been established regarding a level or prevalence of visible foliar injury considered to be adverse
to the affected vegetation as the current evidence does not provide for determination of a degree
of leaf injury that would have significance to the vigor of the whole plant (ISA, Appendix 8, p.
8-24). Thus, key considerations of this endpoint in past reviews have related to qualitative
consideration of potential impacts related to the plant's aesthetic value in protected forested areas
and the somewhat general, nonspecific judgment that a more restrictive standard is likely to
provide increased protection. Nevertheless, while minor spotting on a few leaves of a plant may
easily be concluded to be of little public welfare significance, it is reasonable to conclude that
cases of widespread and relatively severe injury during the growing season (particularly when
sustained across multiple years, and accompanied by obvious impacts on the plant canopy)
would likely impact the public welfare in scenic and/or recreational areas, particularly in areas
with special protection, such as Class I areas.
The currently available evidence, as discussed above, as well as in sections 4.3.3.2 and
4.5.1.2 (with consideration of presentations in Appendix 4C and air quality analyses in
Appendices 2A and 4D) do not indicate that a situation of widespread and relatively severe
visible foliar injury is likely associated with air quality that meets the current standard. While the
USFS biosite dataset includes appreciable variability in biosite index scores, including the
occurrence of scores above 15 in records with W126 index estimates in the lowest bin, it does
not demonstrate a clear trend in biosite index score across the lower W126 index bins. That
notwithstanding, as noted in section 4.5.1.2 above, records with W126 index estimates below 25
ppm-hrs appear to have BI scores below 5, on average, with fewer than 10% of records having a
BI score above 15 or 5. The latter compares to some 20 to 40% of records with higher W126
index estimates having scores above 5 (in the USFS dataset presented in Appendix 4C, see Table
4C-6). In this context, we note that the current air quality analyses indicate that virtually all 3-
year average, and single-year, W126 index values at locations meeting the current standard are at
or below 25 ppm-hr. Further, the average number of observations of 1-hour concentrations at or
above 100 ppb per site and design value period are well below one during periods when the
current standard is met. Thus, while the current evidence is limited for the purposes of
identifying public welfare protection objectives related to visible foliar injury in terms of specific
air quality metrics, the current information indicates that the occurrence of injury categorized as
more severe than "little" by the USFS categorization (i.e., a BI scores above 5 or above 15)
would be expected to be infrequent in areas that meet the current standard. Based on the USFS
4-92
-------
dataset presentations as well as the air quality analyses of W126 index values and frequency of
1-hour observations at or above 100 ppb, the prevalence of injury scores categorized as severe,
which, depending on spatial extent, might reasonably be concluded to have potential to be
adverse to the public welfare would be expected to be appreciably more likely for air quality
conditions unlikely to meet the current standard.
With regard to other vegetation-related effects, including those at the ecosystem scale,
such as alteration in community composition or reduced productivity in terrestrial ecosystems, as
recognized in section 4.5.1.1, the available evidence is not clear with regard to the risk of such
impacts (and their magnitude or severity) associated with the environmental O3 exposures
estimated to occur under air quality conditions meeting the current standard (e.g., W126 index at
or below 17 ppm-hrs). In considering effects on crop yield, the air quality analyses at monitoring
locations that meet the current standard indicate estimates of RYL for such conditions to be at
and below 5.1%, based on the median estimate derived from the established E-R functions for 10
crops (Appendix 4A, Table 4A-5). We additionally recognize there to be complexities involved
in interpreting the significance of such small estimates in light of the factors also recognized in
the last review. These included the extensive management of crops in agricultural areas that may
to some degree mitigate potential Ch-related effects, as well as the use of variable management
practices to achieve optimal yields, while taking into consideration various environmental
conditions. We also recognize, as was recognized in the last review, that changes in yield of
commercial crops and commercial commodities may affect producers and consumers differently,
further complicating the question of assessing overall public welfare impacts for such RYL
estimates (80 FR 65405, October 26, 2015).
4.5.2 CASAC advice
In our consideration of the adequacy of the current secondary O3 standard, in addition to
the evidence- and air quality/exposure/risk-based information discussed above, we have also
considered the advice and recommendations of the CASAC, based on their review of the ISA
and the earlier draft of this PA, as well as comments from the public on the earlier draft of this
PA. A limited number of public comments have been received in this review to date, including
comments focused on the draft IRP or draft PA. Of the commenters that addressed adequacy of
the current secondary O3 standard, most expressed agreement with staff conclusions in the draft
PA, while some expressed the view that the standard should be revised to a W126-based form
based on advice from the previous CASAC in the last review. The CASAC provided its advice
regarding the current secondary standard in the context of its review of the draft PA (Cox, 2020).
In so doing, the CASAC concurred with the PA conclusions, stating that it "finds, in agreement
with the EPA, that the available evidence does not reasonably call into question the adequacy of
4-93
-------
the current secondary ozone standard and concurs that it should be retained" (Cox, 2020, p. 1).
The CAS AC additionally stated that it "commends the EPA for the thorough discussion and
rationale for the secondary standard" while also providing comments particular to the
consideration of climate and growth-related effects (Cox, 2020, pp. 2, 22).
With regard to O3 effects on climate, the CASAC recommended quantitative uncertainty
and variability analyses, with associated discussion (Cox, 2020, pp. 2 , 22) ,85 With regard to
growth-related effects and consideration of the evidence in quantitative exposure analyses, it
stated that the W126 index, "appears reasonable and scientifically sound" "particularly related to
growth effects" (Cox, 2020, p 16). Additionally, with regard to the Administrator in the last
review expressing greater confidence in judgements related to public welfare impacts based on
seasonal W126 Index estimated by a three-year average and accordingly relying on that metric
the CASAC expressed the view that it "appears of reasonable thought and scientifically sound"
(Cox, 2020, p. 19). Further, the CASAC stated that "RBL appears to be appropriately considered
as a surrogate for an array of adverse welfare effects and based on consideration of ecosystem
services and potential for impact to the public as well as conceptual relationships between
vegetation growth effects and ecosystem scale effects" and that it agrees "that biomass loss, as
reported in RBL, is a scientifically-sound surrogate of a variety of adverse effects that could be
exerted to public welfare," concurring that this approach is not called into question by the current
evidence which continues to support "the use of tree seedling RBL as a proxy for the broader
array of vegetation related effects, most particularly those related to growth that could be
impacted by ozone" (Cox, 2020, p 21). The CASAC additionally concurred that the strategy of a
secondary standard that limits 3-yr average W126 index values somewhat below those associated
with a 6% RBL in the median species is "still scientifically reasonable" and that, accordingly, a
W126 index target value of 17 ppm-hrs for generally restricting cumulative exposures "is still
effective in particularly protecting the public welfare in light of vegetation impacts from ozone"
(Cox, 2020, p 21).
With regard to the court's remand of the 2015 secondary standard to the EPA for further
justification or reconsideration ("particularly in relation to its decision to focus on a 3-year
average for consideration of the cumulative exposure, in terms of W126, identified as providing
requisite public welfare protection, and its decision to not identify a specific level of air quality
related to visible foliar injury"), while the CASAC stated that it was not clear whether the draft
85 As recognized in the ISA, [c] and "[c]urrent limitations in climate modeling tools, variation across models, and
the need for more comprehensive observational data on these effects represent sources of uncertainty in
quantifying the precise magnitude of climate responses to ozone changes, particularly at regional scales" (ISA,
section IS.6.2.2, Appendix 9, section 9.3.3, p. 9-22). As noted in section 4.3.3.3 above, these complexities impede
our ability to consider specific 03 concentrations in the U.S. with regard to specific magnitudes of impact on
radiative forcing and subsequent climate effects.
4-94
-------
PA had fully addressed this concern (Cox, 2020, p. 21), it described there to be a solid scientific
foundation for the current secondary standard and also commented on areas related to the
remand.. With regard to the focus on the 3-year index, in addition to the comments summarized
above, the CASAC concluded that the EPA Administrator's focus on 3-year average and her
judgments in doing so "appears of reasonable thought and scientifically sound" (Cox, 2020, p
19). Further, while recognizing the existence of established E-R functions that relate cumulative
seasonal exposure of varying magnitudes to various incremental reductions in expected tree
seedling growth (in terms of RBL) and in expected crop yield, the CASAC letter also noted that
while decades of research also recognizes visible foliar injury as an effect of O3, "uncertainties
continue to hamper efforts to quantitatively characterize the relationship of its occurrence and
relative severity with ozone exposures" (Cox, 2020, p 20). In summary, the CASAC stated that
the approach described in the draft PA to considering the evidence for welfare effects "is laid out
very clearly, thoroughly discussed and documented, and provided a solid scientific underpinning
for the EPA conclusion leaving the current secondary standard in place" (Cox, 2020, p. 22).
4.5.3 Conclusions
This section describes conclusions for the Administrator's consideration in this review of
the current secondary O3 standard. These conclusions are based on consideration of the
assessment and integrative synthesis of the evidence (as summarized in the ISA, and the 2013
ISA and AQCDs from prior reviews), and the information on quantitative exposure and air
quality analyses summarized above, as well as CASAC advice and public comment on the draft
PA. Taking into consideration the discussions above in this chapter, this section addresses the
following overarching policy question.
• Does the currently available scientific evidence and air quality and exposure analyses
support or call into question the adequacy of the protection afforded by the current
secondary O3 standard?
In considering this question, we first recognize what the CAA specifies with regard to
protection to be provided by the secondary standard. Under section 109(b)(2) of the CAA, a
secondary standard must "specify a level of air quality the attainment and maintenance of which,
in the judgment of the Administrator, based on such criteria, is requisite to protect the public
welfare from any known or anticipated adverse effects associated with the presence of [the]
pollutant in the ambient air." Accordingly, as noted in section 4.3.2 above, the secondary
standard is meant to protect against Ch-related welfare effects that are judged to be adverse to the
public welfare (78 FR 8312, January 15, 2013; see also 73 FR 16496, March 27, 2008). Thus,
our consideration of the currently available information regarding welfare effects of O3 is in this
context, while recognizing that the level of protection from known or anticipated adverse effects
4-95
-------
to public welfare that is requisite for the secondary standard is a public welfare policy judgment
to be made by the Administrator.
As is the case in NAAQS reviews in general, the extent to which the protection provided
by the current secondary O3 standard is judged to be adequate will depend on a variety of factors,
including science policy judgments and public welfare policy judgments. These factors include
public welfare policy judgments concerning the appropriate benchmarks on which to place
weight, as well as judgments on the public welfare significance of the effects that have been
observed at the exposures evaluated in the welfare effects evidence. The factors relevant to
judging the adequacy of the standard also include the interpretation of, and decisions as to the
weight to place on, different aspects of the quantitative analyses of air quality and cumulative O3
exposure and any associated uncertainties. Thus, we recognize that the Administrator's
conclusions regarding the adequacy of the current standard will depend in part on public welfare
policy judgments, science policy judgments regarding aspects of the evidence and exposure/risk
estimates, as well as judgments about the level of public welfare protection that is requisite under
the Clean Air Act.
Our response to the overarching question above takes into consideration the discussions
that address the specific policy-relevant questions in prior sections of this document and the
approach described in section 4.2 that builds on the approach from the last review, with some
further attention to the issues highlighted in the court's remand. We focus first on consideration
of the evidence, including that newly available in this review, and the extent to which it alters
key conclusions supporting the current standard. We then turn to consideration of the
quantitative analyses, including associated limitations and uncertainties, and the extent to which
they indicate differing conclusions regarding level of protection indicated to be provided by the
current standard from adverse effects. We additionally consider the key aspects of the evidence
and air quality/exposure information emphasized in establishing the now-current standard, and
the associated public welfare policy judgments and judgments about inherent uncertainties that
are integral to decisions on the adequacy of the current secondary O3 standard.
In considering the currently available evidence, we recognize the longstanding evidence
base of the vegetation-related effects of O3, augmented in some aspects since the last review.
Consistent with the evidence in the last review, the currently available evidence describes an
array of O3 effects on vegetation and related ecosystem effects, as well as the role of
tropospheric O3 in radiative forcing and subsequent effects on temperature, precipitation and
related climate variables. The current evidence base, including the wealth of longstanding
evidence, supports the conclusion of causal relationships between O3 and visible foliar injury,
reduced plant growth and reproduction, as well as reduced yield and quality of agricultural crops,
reduced productivity in terrestrial ecosystems, alteration of terrestrial community composition,
4-96
-------
and alteration of belowground biogeochemical cycles (ISA, section IS.5). This current evidence
base also supports likely causal relationships for O3 with alteration of terrestrial ecosystem water
cycling and reduced carbon sequestration in terrestrial ecosystems, and also with increased tree
mortality (ISA, section IS.5). Evidence available in this review also supports Agency
conclusions on two additional plant-related effects: the body of evidence is determined to be
sufficient to infer a likely causal relationship between O3 exposure and alteration of plant-insect
signaling, and to infer a likely causal relationship between O3 exposure and altered insect
herbivore growth and reproduction (ISA, section IS.5).
We additionally recognize that uncertainties in categories of effects newly identified in
this review, such as alteration of plant-insect signaling and insect herbivore growth and
reproduction, limit our consideration of the protection that might be provided by the current
standard against these effects. Depending on a number of factors, such effects may have a
potential for adverse effects to the public welfare, e.g., given the role of plant-insect signaling in
such important ecological processes as pollination and seed dispersal, as well as, natural plant
defenses against predation and parasitism (as discussed in section 4.3.2 above). Uncertainties in
the current evidence, however, preclude a full understanding of such effects, the air quality
conditions that might elicit them, the potential for impacts in a natural ecosystem and,
consequently, the potential for such impacts under air quality conditions associated with meeting
the current standard. As one example of such uncertainties, although there are multiple
statistically significant measures of O3 effects on insect herbivore growth and reproductive
endpoints, there is no clear trend in the directionality of response for most endpoints studied.
Additionally, the characterization of effects on plant VPSCs in natural ecosystems is still an
emerging area of research that includes knowledge gaps with regard to the role of O3, including
an understanding of the air quality conditions and O3 concentrations that would be expected to
cause effects in the natural environment, and the magnitude or severity of such effects.
As was the case in the last review, a category of effects for which the evidence supports
quantitative description of relationships between air quality conditions and response is plant
growth or yield. The evidence base continues to indicate growth-related effects as sensitive
welfare effects, with the potential for ecosystem-scale ramifications. For this category of effects,
there are established E-R functions that relate cumulative seasonal exposure of varying
magnitudes to various incremental reductions in expected tree seedling growth (in terms of RBL)
and in expected crop yield (in terms of RYL). Many decades of research also recognize visible
foliar injury as an effect of O3, although uncertainties continue to hamper efforts to quantitatively
characterize the relationship of its occurrence and relative severity with O3 exposures. The
evidence for these categories of O3 effects is discussed further below.
4-97
-------
Before focusing further on the key vegetation-related effects identified above, we first
recognize the strong evidence documenting tropospheric O3 as a greenhouse gas causally related
to radiative forcing, and likely causally related to subsequent effects on variables such as
temperature and precipitation. In so doing, however, we take note of the limitations and
uncertainties in the evidence base that affect characterization of the extent of any relationships
between O3 concentrations in ambient air in the U.S. and climate-related effects. Accordingly,
we recognize, as was recognized at the time of the last review, the lack of important quantitative
tools with which to consider such effects in this context (as summarized in sections 4.3.3.3 and
4.3.4 above).86 Notwithstanding consideration of these effects, a focus in this review, as in the
last, on the protection offered by the standard against vegetation-related effects is expected to
also have positive implications for climate change protection through the protection of terrestrial
ecosystem carbon storage.
Turning next to consideration of visible foliar injury, the available information has been
examined and analyzed as to what it indicates and supports with regard to adequacy of protection
provided by the current standard (e.g., as discussed in section 4.5.1 above). Visible foliar injury
is an effect for which an association with O3 in ambient air is well documented, and the public
welfare significance of visible foliar injury of vegetation in areas not closely managed for
harvest, particularly specially protected natural areas, has generally been considered in the
context of potential effects on aesthetic and recreational values, such as the aesthetic value of
scenic vistas in protected natural areas such as national parks and wilderness areas (e.g., 73 FR
16496, March 27, 2008). Accordingly, depending on its severity and spatial extent, as well as the
location(s) and the associated intended use, its effects on the physical appearance of the plant
have the potential to be significant to the public welfare. For example, cases of widespread and
relatively severe injury during the growing season (particularly when sustained across multiple
years and accompanied by obvious impacts on the plant canopy) might reasonably be expected to
have the potential to adversely impact the public welfare in scenic and/or recreational areas,
particularly in areas with special protection, such as Class I areas. Thus, we consider the
currently available information with regard to the potential for such an occurrence with air
quality conditions that meet the current standard. In so doing, we recognize that important
uncertainties remain in the understanding of the O3 exposure conditions that will elicit visible
86 With regard to radiative forcing and effects on temperature, precipitation, and related climate variables, while
additional characterizations have been completed since the last review, uncertainties and limitations in the
evidence that were also recognized in the last review remain. As summarized in sections 4.3.3.3 and 4.3.4 above,
these affect our ability to make a quantitative characterization of the magnitude of climate response to changes in
O3 concentrations in ambient air at regional (vs global) scales.
4-98
-------
foliar injury (and its severity), and particularly in light of the other environmental variables that
influence its occurrence. For example, as discussed in section 4.5.1.2 above, while analyses of
USFS data for foliar injury often consider O3 concentrations in terms of a cumulative exposure
metric, multiple studies also have indicated a role for an additional metric related to the
occurrence of days with relatively high concentrations (e.g., number of days with a 1-hour
concentration at or above 100 ppb), although, there has not yet been extensive work done to
confirm the specific peak concentration that would be appropriate for such a metric. With regard
to an implied importance of peak or elevated short-term (e.g., 1-hour) concentrations, the dataset
of BI scores at USFS biosites (sites with 03-sensitive vegetation assessed for visible foliar
injury) analyzed in Appendix 4C indicates variability in incidence of BI scores indicative of
moderate or greater severity injury across the bins for W126 index values most common in areas
where the standard is met (Appendix 4C, section 4.C.3; Appendix 4D, Figures 4D-3 and 4D-4,
and section 4D.3.2.3). The incidence of nonzero scores, and of relatively higher scores87 appears
to markedly increase only with W126 index values above 25 ppm-hrs, a magnitude not seen to
occur at monitoring locations (including in or near Class I areas) where the current standard is
met (Appendix 4C, section 4C.3; Appendix 4D, section 4D.3.2.3).
Publications related to the evidence base for the USFS biosite monitoring program
document reductions in the incidence of the higher BI scores over the 16-year period of the
program (1994 through 2010), especially after 2002, leading to researcher conclusions of a
"declining risk of probable impact" on the monitored forests over this period (e.g., Smith, 2012).
These reductions parallel the O3 concentration trend information nationwide that show clear
reductions in cumulative seasonal exposures, as well as in peak O3 concentrations such as the
annual fourth highest daily maximum 8-hour concentration, from 2000 through 2018 (Appendix
4D, Figure 4D-9 and Figure 2-11 above). The foliar injury reductions also parallel reductions in
the occurrence of 1-hour concentrations above 100 ppb (Appendix 2A, Tables 2A-2 to 2A-4).
Thus, the extensive air quality evidence of trends across the past nearly 20 years indicate
reductions in peak concentrations that some studies have suggested to be influential in the
severity of visible foliar injury, as discussed in section 4.5.1 above.
Further, we note the paucity of established approaches for interpreting specific levels of
severity and extent of foliar injury in protected forests with regard to impacts on public welfare
effects, e.g., related to recreational services.88 As discussed in sections 4.3.2 and 4.5.1 above,
injury to whole stands of trees of a severity apparent to the casual observer (e.g., when viewed as
87 In the USFS categorization, scores from zero to just below 5 are described as "little or no foliar injury."
88 This contrasts with another welfare effect, visibility, for which there is evidence relating to levels of visibility
found to be acceptable by the public that was considered in judging the public welfare protection provided by the
particulate matter secondary standard (78 FR 3226-3228, January 15, 2013).
4-99
-------
a whole from a distance) would reasonably be expected to affect recreational values. Current
information, however, particularly in locations meeting the current standard or with W126 index
estimates likely to occur under the current standard does not indicate a significant extent and
degree of injury (e.g., based on BI scores analyzed in Appendix 4C) or specific impacts on
recreational or related services for areas, such as wilderness areas or national parks. Thus, the
evidence does not appear to suggest public welfare significance for BI scores reported at sites
likely to meet the current standard.
We additionally take note of the recognition by the CASAC that "uncertainties continue
to hamper efforts to quantitatively characterize the relationship of [visible foliar injury]
occurrence and relative severity with ozone exposures" (Cox, 2020, p. 20 of the Response to
Charge), and of the CASAC advice, which concurred with the draft PA preliminary conclusion
"that the available evidence does not reasonably call into question the adequacy of the current
secondary ozone standard" (Cox, 2020, p. 1). Based on all of the above considerations it appears
reasonable to conclude that the current evidence and quantitative exposure information for
visible foliar injury does not call into question the adequacy of protection provided by the current
standard.
Uncertainties additionally affect our understanding of the extent to which RYL estimates
on the order of 5% (or less), based on the set of 10 established E-R functions, would be expected
to be of public welfare significance, given the extensive management of such crops, and other
factors summarized in sections 4.3.2 and 4.5.1.3 above. Further, we recognize uncertainties in
the details and quantitative aspects of relationships between plant-level effects such as growth
and reproduction, and ecosystem impacts, the occurrence of which are influenced by many other
ecosystem characteristics and processes. These examples illustrate the role of public welfare
policy judgments, both with regard to the extent of protection that is requisite and concerning the
weighing of uncertainties and limitations of the underlying evidence base and associated
quantitative analyses. Such judgments will inform the Administrator's decision in the current
review, as they did in the setting of the current standard in 2015, as summarized in section 4.1.2
above.
We recognize that public welfare policy judgments play an important role in each review
of a secondary standard, just as public health policy judgments have important roles in primary
standard reviews. One type of public welfare policy judgment focuses on how to consider the
nature and magnitude of the array of uncertainties that are inherent in the scientific evidence and
analyses. These judgments are traditionally made with a recognition that current understanding
of the relationships between the presence of a pollutant in ambient air and associated welfare
effects is based on a broad body of information encompassing not only more established aspects
of the evidence but also aspects in which there may be substantial uncertainty. This may be true
4-100
-------
even of the most robust aspect of the evidence base. In the case of the secondary O3 standard
review, as an example, we recognize increased uncertainty, and associated imprecision, at lower
cumulative exposures in application of the established and well-founded E-R functions, and in
the current understanding of aspects of relationships of such estimated effects with larger-scale
impacts, such as those on populations, communities and ecosystems, as summarized in sections
4.5.1.3 and 4.3.4 above.
The category of effects for which the evidence is most certain with regard to quantitative
functions describing relationships between O3 in ambient air and response continues to be
reduced plant growth or yield. The evidence base includes established E-R functions for
seedlings of 11 tree species that relate cumulative seasonal exposure of varying magnitudes to
various incremental reductions in expected tree seedling growth (in terms of RBL) and in
expected crop yield. These functions are well established and have been recognized across
multiple O3 NAAQS reviews. Uncertainties related to use of the RBL estimates include the
limited information regarding the extent to which they reflect growth impacts in mature trees,
and the fact that the 11 species represent a very small portion of the tree species across the U.S.
While recognizing these and other uncertainties, RBL estimates based on the median of the 11
species were used as a surrogate in the last review for comparable information on other species
and lifestages, as well as a proxy or surrogate for other vegetation-related effects, including
larger-scale effects. Use of this approach continues to appear to be a reasonable judgment in this
review. More specifically, the currently available information continues to support (and does not
call into question) the use of RBL as a useful and evidence-based approach for consideration of
the extent of protection from the broad array of vegetation-related effects associated with O3 in
ambient air. The currently available evidence, while somewhat expanded since the last review
does not indicate an alternative metric for such a use; nor is an alternative approach evident. As
noted in section 4.5.2 above the CAS AC concurred that this approach is not called into question
by the current evidence which continues to support "the use of tree seedling RBL as a proxy for
the broader array of vegetation related effects" (Cox, 2020x, p. 21).
In considering the use of RBL and recognizing the role of the established E-R functions
for 11 species of tree seedlings, and the median across the 11 species, we note that assessment of
cumulative exposure in the 2015 review focused on use of the 3-year average seasonal W126
index with these functions. Thus, the discussion in the sections above (e.g., section 4.5.1.2) of
the information available in the current review summarized and expanded on the technical issues
that had informed the EPA's focus on the average over the 3-year design period in the last
review.
Summarizing key points of that discussion here, we take note of the uncertainties
associated with RBL estimates derived from the E-R functions. For example, as discussed in
4-101
-------
section 4.3.4 above, while the E-R functions for the 11 species have been derived in terms of a
seasonal (single-year) W126 index, the experiments from which they were derived vary in
duration from periods of 82 to 140 over a single year days to periods of 180 to 555 days across
two years. Thus, the "adjustment" to a single season carries uncertainties and contributes some
imprecision to the resulting function and estimates derived using it. Additionally, we note that
the exposure levels represented in the data underlying the E-R functions are somewhat limited at
the lower cumulative exposure levels, such as those most commonly associated with the current
standard (as characterized in Appendix 4D). Further, we recognize the variability that is
associated with tree growth in the natural environment (e.g., related to variability in plant, soil,
meteorological and other factors). In similar manner, we note the variability associated with
plant responses to O3 exposures in the natural environment. For example, we note the
comparisons performed in the 2013 ISA and current ISA of RBL estimates based on either
cumulative average multi-year W126 index or single-year W126 with estimates derived from
measurements from a multi-year O3 exposure study (as summarized in Appendix 4A, section
4.A.3.1). These presentations illustrate the variability inherent in the magnitude of growth
impacts of O3 and in the quantitative relationship of O3 exposure and RBL, while also providing
general agreement of predictions (based on either metric) with observations. Further, an
illustrative example in Appendix 4A also provides quantitative estimates of potential differences
in growth impacts of O3 exposure controlled in terms of a 3-year average, such that the single-
year values may vary while meeting the value specified for the average, compared to exposure
controlled to such a value annually. While simplistic in nature, this example illustrates that based
on the magnitude of variation documented for annual W126 index values occurring under the
current standard, the magnitude of any differences in tree biomass between single-year and
multi-year average approaches to controlling cumulative exposure would be expected to be quite
small (Appendix 4A, section 4A.3). All of the factors identified here, the currently available
evidence and recognized limitations, variability and uncertainties, contribute uncertainty and
resulting imprecision or inexactitude to RBL estimates of single-year seasonal W126 index
values, thus providing support for use of an average seasonal W126 index derived from multiple
years (with their representation of variability in environmental factors), such as for a 3-year
period, for estimating median RBL using the established E-R functions. Additionally, we take
note of the CASAC advice in this review which affirmed the EPA's focus on a 3-year average
W126, concluding that the EPA Administrator's focus on a 3-year average and her greater
confidence in judgments related to public welfare impacts in doing so "appears of reasonable
thought and scientifically sound" (Cox, 2020x, p. 19).
In considering tree growth effects, we also take note of the public welfare policy
judgments inherent in the Administrator's decision in establishing the current standard in 2015.
4-102
-------
Those judgments include her adoption of the median tree seedling RBL estimate for the studied
species as a surrogate for the broad array of vegetation related effects that extend to the
ecosystem scale, and her identification of cumulative seasonal exposures (in terms of the average
W126 index across the 3-year design period for the standard) associated with a median RBL
somewhat below 6% as an appropriate focus for considering target levels of protection for the
2015 standard. The newly available information in this review does not appear to call into
question such judgements, indicating them to continue to appear reasonable in this review, and
the current CASAC has agreed that this strategy "is still scientifically reasonable" (Cox 2020, p.
21).
Reviews of secondary NAAQS also require judgments on the extent to which particular
welfare effects (e.g., with regard to type, magnitude/severity or extent) are important from a
public welfare perspective. In the case of O3, such a judgment includes consideration of the
public welfare significance of small magnitude estimates of RBL and associated unquantified
potential for larger-scale related effects. In establishing the current standard in 2015 with a focus
on RBL as a proxy or surrogate for the broad array of vegetation effects, the Administrator took
note of the 2014 CASAC characterization of 6% RBL (in seedlings of median tree species). As
described in section 4.1 above, the rationale provided by the CASAC with this characterization
was primarily conceptual and qualitative, rather than quantitative. The conceptual
characterization recognized linkages between effects at the plant scale and broader ecosystem
impacts, with the CASAC recommending that the Administrator consider RBL as a surrogate or
proxy for the broader impacts that could be elicited by O3. In the 2015 decision, the
Administrator took note of this CASAC advice regarding use of RBL as a proxy and set the
standard with an "underlying objective of a revised secondary standard that would limit
cumulative exposures in nearly all instances to those for which the median RBL estimate would
be somewhat lower than 6%" (80 FR 65407, October 26, 2015). While noting the CASAC view
regarding 6% RBL in describing this objective, the Administrator did not additionally find that a
cumulative seasonal exposure, for which such a magnitude of median species RBL was
estimated, represented conditions that were adverse to the public welfare. Rather the 2015
decision noted that "the Administrator does not judge RBL estimates associated with marginal
higher exposures [at or above 19 ppm-hrs] in isolated, rare instances to be indicative of adverse
effects to the public welfare" (80 FR 65407, October 26, 2015). In the comments from the
current CASAC in the context of its review of the draft PA, it expressed the view that the
strategy described by the prior Administrator for the secondary standard established in 2015 with
its W126 index target of 17 ppm-hrs (in terms of a 3-year average), at or below which the 2015
standard was expected to generally restrict cumulative seasonal exposure, is "still effective in
4-103
-------
particularly protecting the public welfare in light of vegetation impacts form ozone" (Cox, 2020,
p. 21).
The current evidence base and available information (qualitative and quantitative), as in
the last review, continue to support consideration of the potential for 03-related vegetation
impacts in terms of the RBL estimates from established E-R functions as a quantitative tool
within a larger framework of considerations pertaining to the public welfare significance of O3
effects. Such consideration would include effects that are associated with effects on vegetation,
and particularly those that conceptually relate to growth, and that are causally or likely causally
related to O3 in ambient air, yet for which there are greater uncertainties affecting estimates of
impacts on public welfare. This approach to weighing the available information in reaching
judgments regarding the secondary standard additionally takes into account uncertainties
regarding the magnitude of growth impact that might be expected in mature trees, and of related,
broader, ecosystem-level effects for which the available tools for quantitative estimates are more
uncertain and those for which the policy foundation for consideration of public welfare impacts
is less well established. (80 FR 65389, October 26, 2015).
In considering the quantitative analyses available in this review, we note the findings
from the analysis of recent air quality at sites across the U.S., including in or near 65 Class I
areas, and also analyses of historical air quality. Findings from the analysis of the air quality data
from the most recent period and from the larger analysis of historical air quality data extending
back to 2000 are consistent with the air quality analysis findings that were part of the basis for
the current standard. That is, in virtually all design value periods and all locations at which the
current standard was met (more than 99.9% of the observations), the 3-year average W126
metric was at or below 17 ppm-hrs, the target identified by the Administrator in establishing the
current standard and in all such design value periods and locations the W126 metric was at or
below 19 ppm-hrs as was also the case for the earlier and smaller dataset (80 FR 65404-65410,
October 26, 2015). Additionally, across the full 19-year dataset for 56 Class I areas with
monitors meeting the current standard during at least one or as many as seventeen 3-year periods
since 2000, there are no more than 15 occurrences of a single-year W126 index above 19 ppm-
hrs, with relatively fewer occurrences during the more recent part of the historical period, 2010
to 2018 (Appendix 4D, section 4D.3.2.3). Based on considerations summarized in section 4.5.1
above, the currently available information, including such infrequent single-year deviations of
this magnitude above the average, could reasonably be judged not to pose meaningful risks of
public welfare impacts to Class I areas.
In summary, the new information available is consistent with that available in the last
review for the principal effects for which the evidence is strongest (e.g., growth, reproduction,
and related larger-scale effects, as well as, visible foliar injury) and for key aspects of the
4-104
-------
decision in that review. As discussed above, the currently available information does not provide
established quantitative relationships and tools for estimating incidence and severity of visible
foliar injury in protected areas across the U.S. or provide information linking extent and severity
of injury to aesthetic values that might be useful for considering public welfare implications.
Further, the currently available evidence for forested locations across the U.S., such as studies of
USFS biosites, does not indicate widespread incidence of significant visible foliar injury.
Additionally, the evidence regarding RBL and air quality in areas meeting the current standard
does not appear to call into question the adequacy of protection. For other vegetation-related
effects that the ISA newly concludes likely to be causally related to O3, the new information does
not provide us an indication of the extent to which such effects might be anticipated to occur in
areas that meet the current standard of a significance reasonably judged significant to public
welfare. Thus, we do not find the current information for these newly identified categories to call
into question the adequacy of the current standard. Similarly, the current information regarding
O3 contribution to radiative forcing or effects on temperature, precipitation and related climate
variables is not strengthened from that available in the last review, including with regard to
uncertainties that limit quantitative evaluations. In recognizing similarities with the information
based on which the current standard was set in 2015, we additionally note that, as in the last
review, the Administrator's decision on the adequacy of public welfare protection afforded by
the secondary O3 standard from identified 03-related effects and their potential to present
adverse effects to the public welfare will be based in part on public welfare policy judgments
regarding uncertainties and limitations in the available information.
We additionally note the advice from the CASAC in this review. With regard to the
adequacy of the current secondary standard, it stated that it "finds, in agreement with the EPA,
that the available evidence does not reasonably call into question the adequacy of the current
secondary ozone standard and concurs that it should be retained" (Cox, 2020, p. 1). The CASAC
additionally stated that it "commends the EPA for the thorough discussion and rationale for the
secondary standard." (Cox, 2020, p. 2).
Based on all of the above considerations, we conclude that the currently available
evidence and quantitative exposure/risk information does not call into question the adequacy of
the current secondary standard such that it is appropriate to consider retaining the current
standard without revision. In so doing, we also recognize that, as is the case in NAAQS reviews
in general, the extent to which the Administrator judges the current secondary O3 standard to be
adequate will depend on a variety of factors, including science policy judgments and public
welfare policy judgments.
4-105
-------
4.6 KEY UNCERTAINTIES AND AREAS FOR FUTURE RESEARCH
In this section, we highlight key uncertainties associated with reviewing and establishing
the secondary O3 standard and additionally recognize that research in these areas may
additionally be informative to the development of more efficient and effective control strategies.
The list in this section includes key uncertainties and data gaps thus far highlighted in this review
of the secondary standard. Additional information in several areas would reduce uncertainty in
our interpretation of the available information and, accordingly, reduce uncertainty in our
characterization of Ch-related welfare effects. For example, the items listed below generally
include uncertainties associated with the extrapolation to plant species and environments outside
of specific experimental or field study conditions and the assessment of ecosystem-scale impacts,
such as structure and function. Additional E-R studies in different species or for responses other
than reduced growth over multiple exposure conditions over growing seasons, that include
details on exposure circumstances (e.g., hourly concentrations throughout the exposure), and
exposure history, etc. would improve on and potentially expand characterizations of the potential
for and magnitude of the identified vegetation effects under different seasonal exposures.
Accordingly, in this section, we highlight areas for future welfare effects research, model
development, and data collection activities to address these uncertainties and limitations in the
current scientific evidence. These areas are similar to those highlighted in past reviews.
• While national visible foliar injury surveys have provided an extensive dataset on the
incidence of such effects at sites across the country that experienced differing cumulative
seasonal O3 exposures and soil moisture conditions, there remain uncertainties in the
current understanding of the relationship between seasonal O3 exposures (and other
influential factors, such as relative soil moisture) and the incidence and relative severity
of visible foliar injury. Research to better characterize the relationship between O3, soil
moisture and foliar injury and specifically a quantifiable relationship between these (and
any other influential) factors. Additionally, research would assist in interpreting
connections between 03-related foliar injury and other physiological effects and
ecosystem services. For example, research is needed on the extent and severity of visible
foliar injury that might impact ecosystem services (e.g., tourism), and the extent of
impact it might have.
• Additional controlled exposure studies of effects, such as biomass impacts, that include
multiple exposure levels within the lower range of exposures associated with ambient air
quality conditions common today, extend over multiple years, and include the collection
of detailed 03 concentration data over the exposure would reduce uncertainty in
estimates of effects across multiple-year periods and at the O3 exposures common today.
• Evidence newly available since the last review includes studies on insect-plant
interactions that have established some statistically significant effects, but the evidence is
still limited with regard to discerning a pattern of responses in growth, reproduction, or
mortality, and a directionality of responses for most effects. More research is needed to
investigate the degree of response and directionalities of these relationships, and to
4-106
-------
investigate potential effects on pollination. The evidence is also limited with regard to the
species represented (i.e., currently confined to three insect orders).
• Some evidence provides for linkages of effects on tree seedlings with larger trees and
similarities in results between exposure techniques. Uncertainties remain in this area as
well as uncertainties in extrapolating from O3 effects on young trees (e.g., seedlings
through a few years of age) to mature trees and from trees grown in the open versus
those within the forest canopy.
• Uncertainties that remain in extrapolating individual plant response spatially or to higher
levels of biological organization, including ecosystems, could be informed by research
that explores and better quantifies the nature of the relationship between O3, plant
response and multiple biotic and abiotic stressors, including those associated with the
affected ecosystem services (e.g., hydrology, productivity, carbon sequestration).
• Other uncertainties are associated with estimates of the effects of O3 on the ecosystem
processes of water, carbon, and nutrient cycling, particularly at the stand and community
levels. These below- and above-ground processes include interactions of roots with the
soil or microorganisms, effects of O3 on structural or functional components of soil food
webs and potential impacts on plant species diversity, changes in the water use of
sensitive trees, and if the sensitive tree species is dominant, potential changes to the
hydrologic cycle at the watershed and landscape level. Research on competitive
interactions under different O3 exposures and any associated impacts on biodiversity or
genetic diversity would improve current understanding.
• Uncertainties related to characterizing the potential public welfare significance of O3-
induced effects and impacts to associated ecosystem services could also be informed by
research. Research relating effects such as those on plant reproduction and propagation to
effects on production of non-timber forest products, and research to characterize public
preferences including valuation related to non-use and recreation for foliar injury, could
also help inform consideration of the public welfare significance of these effects.
4-107
-------
REFERENCES
Black, VJ, Stewart, CA, Roberts, JA and Black, CR (2012). Timing of exposure to ozone affects
reproductive sensitivity and compensatory ability in Brassica campestris. Environ Exp
Bot 75: 225-234.
Campbell, SJ, Wanek, R and Coulston, JW (2007). Ozone injury in west coast forests: 6 years of
monitoring - Introduction. U.S. Department of Agriculture. Portland, OR.
Cordell, H, Betz, FM, Mou, S and Green, G (2008). How do Americans View Wilderness. A
WILDERNESS Research Report in the Internet Research Information Series. Natinoal
Survey on Recreation and the Environment. This research is a collaborative effort
between the U.S. Department of Agriculture Forest Service's Southern Research Station
and its Forestry Sciences Laboratory in Athens, Georgia; the University of Georgia in
Athens: and the University in Tennessee in Knoxville, Tennessee..
Costanza, R, De Groot, R, Braat, L, Kubiszewski, I, Fioramonti, L, Sutton, P, Farber, S and
Grasso, M (2017). Twenty years of ecosystem services: How far have we come and how
far do we still need to go? Ecosyst Serv 28: 1-16.
Cox, LA. (2020). Letter from Louis Anthony Cox, Jr., Chair, Clean Air Scientific Advisory
Committee, to Administrator Andrew R. Wheeler. Re:CASAC Review of the EPA's
Policy Assessment for the Review of the Ozone National Ambient Air Quality Standards
(External Review 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/264cbl227d55e02c85257402007446a4/4713
D217BC07103485258515006359BA/$File/EPA-CASAC-20-003.pdf.
Darbah, JNT, Kubiske, ME, Neilson, N, Oksanen, E, Vaapavuori, E and Karnosky, DF (2007).
Impacts of elevated atmospheric C02 and 03 on paper birch (Betula papyrifera):
Reproductive fitness. ScientificWorldJournal 7: 240-246.
Darbah, JNT, Kubiske, ME, Nelson, N, Oksanen, E, Vapaavuori, E and Kamosky, DF (2008).
Effects of decadal exposure to interacting elevated C02 and/or 0-3 on paper birch
(Betula papyrifera) reproduction. Environ Pollut 155(3): 446-452.
Davis, DD and Orendovici, T (2006). Incidence of ozone symptoms on vegetation within a
National Wildlife Refuge in New Jersey, USA. Environ Pollut 143(3): 555-564.
Diaz-de-Quijano, M, Kefauver, S, Ogaya, R, Vollenweider, P, Ribas, A and Penuelas, J (2016).
Visible ozone-like injury, defoliation, and mortality in two Pinus uncinata stands in the
Catalan Pyrenees (NE Spain). Eur J Forest Res 135(4): 687-696.
Dietze, MC and Moorcroft, PR (2011). Supplemental information 1: Tree mortality in the eastern
and central United States: Patterns and drivers. Global Change Biol 17.
Frey, HC. (2014). Letter from Dr. H. Christopher Frey, Chair, Clean Air Scientific Advisory
Committee to Honorable Gina McCarthy, Administrator, US EPA. Re: CASAC Review
4-108
-------
of the EPA's Second Draft Policy Assessment for the Review of the Ozone National
Ambient Air Quality Standards. June 26, 2014. EPA-CASAC-14-004. Office of the
Administrator, Science Advisory Board Washington, DC. Available at:
https://nepis. epa.gov/Exe/ZyPURL. cgi?Dockey=Pl 00JR6F. txt.
Haefele, M, Kramer, RA and Holmes, TP (1991). Estimating the Total Value of a Forest Quality
in High-Elevation Spruce-Fir Forests. The Economic Value of Wilderness: Proceedings
of the Conference, Southeastern For Exper. Station. Asheville, NC, USDA Forest
Service.
Heck, WW and Cowling, EB (1997). The need for a long term cumulative secondary ozone
standard - An ecological perspective. Environ Manager January: 23-33.
Henderson, R. (2006). Letter from Dr. Rogene Henderson, Chair, Clean Air Scientific Advisory
Committee to Honorable Stephen L. Johnson, Administrator, US EPA. Re: CASAC Peer
Review of the Agency's 2nd Draft Ozone Staff Paper October 24, 2006. EPA-CASAC-
07-001. Office of the Administrator, Science Advisory Board U.S. EPA HQ, Washington
DC. Available at: https://nepis.epa.gov/Exe/ZyPURL.cgi?Dockey=P1000W07.txt.
Hildebrand, E, Skelly, JM and Fredericksen, TS (1996). Foliar response of ozone-sensitive
hardwood tree species from 1991 to 1993 in the Shenandoah National Park, Virginia. Can
J For Res 26(4): 658-669.
Hogsett, WE, Weber, JE, Tingey, D, Herstrom, A, Lee, EH and Laurence, JA (1997).
Environmental auditing: An approach for characterizing tropospheric ozone risk to
forests. J Environ Manage 21(1): 105-120.
King, JS, Kubiske, ME, Pregitzer, KS, Hendrey, GR, McDonald, EP, Giardina, CP, Quinn, VS
and Karnosky, DF (2005). Tropospheric 03 compromises net primary production in
young stands of trembling aspen, paper birch and sugar maple in response to elevated
atmospheric C02. New Phytol 168(3): 623-635.
Kohut, R (2007). Handbook for Assessment of Foliar Ozone Injury on Vegetation in the National
Parks: Revised Second Edition. Kohut, R.
Kubiske, ME, Quinn, VS, Heilman, WE, McDonald, EP, Marquardt, PE, Teclaw, RM, Friend,
AL and Karnoskey, DF (2006). Interannual climatic variation mediates elevated C02 and
03 effects on forest growth. Global Change Biol 12(6): 1054-1068.
Kubiske, ME, Quinn, VS, Marquardt, PE and Karnosky, DF (2007). Effects of elevated
atmospheric C02 and/or 03 on intra- and interspecific competitive ability of aspen. Plant
Biol 9(2): 342-355.
Landesmann, JB, Gundel, PE, Martrnez-Ghersa, MA and Ghersa, CM (2013). Ozone exposure of
a weed community produces adaptive changes in seed populations of Spergula arvensis.
PLoS ONE 8(9): e75820.
4-109
-------
Lee, EH and Hogsett, WE (1996). methodology for calculating inputs for ozone secondary
standard benefits analysis part II. Office of Air Quality Planning and Standards. Research
Triangle Park, NC.
Lefohn, AS, Jackson, W, Shadwick, DS and Knudsen, HP (1997). Effect of surface ozone
exposures on vegetation grown in the southern Appalachian Mountains: Identification of
possible areas of concern. Atmos Environ 31(11): 1695-1708.
Moran, EV and Kubiske, ME (2013). Can elevated C02 and ozone shift the genetic composition
of aspen (Populus tremuloides) stands? New Phytol 198(2): 466-475.
Myhre, G, Shindell, D, Breon, FM, Collins, W, Fuglestvedt, J, Huang, J, Koch, D, Lamarque, JF,
Lee, D, Mendoza, B, Nakajima, T, Robock, A, Stephens, G, Takemura, T and Zhang, H,
Eds. (2013). Anthropogenic and natural radiative forcing. Cambridge University Press
Cambridge, UK.
Oksanen, E and Holopainen, T (2001). Responses of two birch (Betula pendula Roth) clones to
different ozone profiles with similar AOT40 exposure. Atmos Environ 35(31): 5245-
5254.
Rosenberger, RS, Bell, LA, Champ, PA and White, EM (2013). Estimating the economic value
of recreation losses in Rocky Mountain National Park due to a mountain pine beetle
outbreak. Western Economics Forum 12(1): 31-39.
Samet, JM. (2010). Letter from Jonathan Samet, Chair, Clean Air Scientific Advisory
Committee, to Administrator Lisa Jackson. Re: CASAC Review of EPA's Proposed
Ozone National Ambient Air Quality Standard (Federal Register, Vol. 75, Nov. 11,
January 19, 2010).. February 19, 2010. EPA-CASAC-10-007. Office of the
Administrator, Science Advisory Board U.S. EPA HQ, Washington DC. Available at:
https://nepis.epa.gov/Exe/ZyPURL.cgi?Dockey=P10072Tl.txt.
Smith, G (2012). Ambient ozone injury to forest plants in Northeast and North Central USA: 16
years of biomonitoring. Environ Monit Assess(184): 4049-4065.
Smith, G, Coulston, J, Jepsen, E and Prichard, T (2003). A national ozone biomonitoring
program: Results from field surveys of ozone sensitive plants in northeastern forests
(1994-2000). Environ Monit Assess 87(3): 271-291.
Smith, GC, Morin, RS and McCaskill, GL (2012). Ozone injury to forests across the Northeast
and North Central United States, 1994-2010. General Technical Report NRS-103. United
States Department of Agriculture, US Forest Service, Northern Research Station.
Smith, GC, Smith, WD and Coulston, JW (2007). Ozone bioindicator sampling and estimation.
General Technical Report NRS-20. United States Department of Agriculture, US Forest
Service, Northern Research Station.
Smith, JT and Murphy, D. (2015). Memorandum to Ozone NAAQS Review Docket (EPA-HQ-
OAR-2008-0699). Additional Observations from WREA Datasets for Visible Foliar
4-110
-------
Injury. September 24, 2015.. Docket ID No. EPA-HQ-OAR-2008-0699. Office of Air
Quality Planning and Standards Research Triangle Park, NC. Available at:
https://www.regulations.gov/contentStreamer?documentld=EPA-HQ-OAR-2008-0699-
4250&contentType=pdf.
U.S. DHEW (1970). Air Quality Criteria for Photochemical Oxidants. National Air Pollution
Control Administration,. Washington, DC. U.S. DHEW. publication no. AP-63. NTIS,
Springfield, VA; PB-190262/BA.
U.S. EPA (2007). Review of the National Ambient Air Quality Standards for Ozone: Policy
Assessment of Scientific and Technical Information: OAQPS Staff Paper. Office of Air
Quality Planning and Standards. Research Triangle Park, NC. U.S. EPA. EPA-452/R-07-
003. January 2007. Available at:
https.Y/nepis. epa.gov/Exe/ZyPURL. cgi?Dockey=Pl0083VX. txt.
U.S. EPA (2018). Integrated science assessment for oxides of nitrogen, oxides of sulfur and
particulate matter -Ecological criteria (2nd external review draft). EPA/600/R-18/097.
U.S. Environmental Protection Agency, Office of Research and Development, National
Center for Environmental Assessment. Research Triangle Park, NC.
http://cfint. rtpnc. epa.gov/ncea/prod/recordisplay. cfm ?deid=340671.
U.S. Forest Service, NPS, and U.S. Fish and Wildlife Service (2010). Federal land managers' air
quality related values work group (FLAG): phase I report—revised (2010). National Park
Service, Denver, CO.
USFS (2013). Forest Inventory and Analysis: Fiscal Year 2012 Business Report. United States
Department of Agriculture, http://www.fia.fs.fed.us/library/bus-org-
documents/docs/FIA_Annual_Report_2013.pdf.
USFS (2017). Forest Inventory and Analysis: Fiscal Year 2016 Business Report. United State
Department of Agriculutre.
https://www.fs.fed. us/sites/default/files/fs_media/fs_document/publication-l 5817-usda-
forest-service-fia-annual-report-508.pdf.
van Goethem, TM, Azevedo, LB, van Zelm, R, Hayes, F, Ashmore, MR and Huijbregts, MA
(2013). Plant species sensitivity distributions for ozone exposure. Environ Pollut 178: 1-
6.
Wang, P, Baines, A, Lavine, M and Smith, G (2012). Modelling ozone injury to U.S. forests.
Environ Ecol Stat 19(4): 461-472.
Wells, B. (2015). Memorandum to Ozone NAAQS Review Docket (EPA-HQ-OAR-2008-0699).
Expanded Comparison of Ozone Metrics Considered in the Current NAAQS Review.
September 28, 2015.. Docket ID No. EPA-HQ-OAR-2008-0699. Office of Air Quality
Planning and Standards Research Triangle Park, NC. Available at:
https://www.regulations.gov/contentStreamer?documentld=EPA-HQ-OAR-2008-0699-
4325&contentType=pdf.
4-111
-------
Yun, S-C and Laurence, JA (1999). The response of clones of Populus tremuloides differing in
sensitivity to ozone in the field. New Phytol 141 (3): 411-421.
4-112
-------
APPENDIX 2A
ADDITIONAL DETAILS ON DATA ANALYSIS
PRESENTED IN PA SECTION 2.4
TABLE OF CONTENTS
2A.1 Analyses of 8-Hour Concentrations 2A-2
2A.2 Analyses of 1-Hour Concentrations 2A-3
TABLE OF FIGURES
Figure 2A-1. Boxplots comparing the distribution of MDA1 concentrations for 2000-2004
(red) to the distribution of MDA1 concentrations for 2014-2018 (blue), binned
by the 8-hour design value at each monitoring site. The boxes represent the 25th,
50th and 75th percentiles and the whiskers represent the 1st and 99th percentiles.
Outlier values are represented by circles 2A-5
Figure 2A-2. Map showing the average number of days with MDA1 > 100 ppb,
2000-2004 2A-7
Figure 2A-3. Map showing the average number of days with MDA1 > 100 ppb,
2014-2018 2A-7
Figure 2A-4. Number of days in 2016-2018 at each monitoring site with a MDA1
concentration greater than or equal to 100 ppb and an 8-hour design value less
than 98 ppb. Sites with higher design values had more days, up to a maximum
of 164 (at a site in southern CA) 2A-8
TABLE OF TABLES
Table 2A-1. Summary of criteria describing the sites for which 8-hour metrics are presented
in section 2.4 of main document 2A-3
Table 2A-2. Summary statistics for MDA1 concentrations at sites with differing design
values for 2016-2018 2A-4
Table 2A-3. Summary statistics for MDA1 concentrations at differing design values for
2000-2004 2A-6
Table 2A-4. Summary statistics for MDA1 concentrations at differing design values for
2014-2018 2A-6
2A-1
-------
2A.1 ANALYSES OF 8-HOUR CONCENTRATIONS
The analyses presented in section 2.4 of the main document are based on hourly O3
concentration data from the EPA's Air Quality System (AQS) database (retrieved on August 14,
2019) for the years 2000 to 2018 for the sites meeting data completeness criteria as summarized
in Table 2A-1 below. The daily maximum 8-hour (hr) average (MDA8) values, annual fourth
highest MDA8 values, and design values (DVs) for the current standards were calculated
according to Appendix U to 40 CFR Part 50. Those steps are generally as follows.
- 8-hr average concentrations are derived as the average of concentrations during eight
consecutive hours for the:
o 8-hr periods which have at least six hourly concentrations1; and
o 8-hr periods which have fewer than six hourly concentrations and the sum of
concentrations divided by eight, after truncation of the digits after the third
decimal place, is greater than 0.070 parts per million (ppm)2
- The digits for the resultant 8-hr average concentration are truncated after the third
decimal place.
- MDA8 concentrations are derived as the highest of the consecutive 8-hr averages
beginning with the 8-hr period from 7am to 3pm and ending with the period from
1 lpm to 7am the following day for those days with:
o 8-hr concentrations for at least 13 of the 17 8-hr periods that begin with the
7am-to-3pm period and end with the 1 lpm-to-7am (next day) period, or
o 8-hr concentrations for fewer than 13 of the 17 8-hr periods if the maximum
8-hr concentration, after truncation of the digits after the third decimal place,
is greater than 0.070 ppm.
Design Values in ppm are derived as average of the annual 4th highest MDA8
concentrations in three consecutive years, with digits after the third decimal place
truncated.
1 When there are at least six hours with a concentration reported, the 8-hr average is the average calculated using the
number of hours with concentrations in the denominator.
2 When there are fewer than six hours with a concentration reported, the 8-hr average is the average calculated using
eight in the denominator and substituting zero for the missing hourly concentrations.
2A-2
-------
Table 2A-1. Summary of criteria describing the sites for which 8-hour metrics are
presented in section 2.4 of main document.
Presentation of 8-hour
metrics in section 2.4
Time
Period
Data included
Fiqure 2-8, DVs
2016-2018
Design values are presented for all sites with valid design values,
which are sites having at least 75% data completeness in each of the
three years and at least 90% completeness on average across the
three years (per Appendix U)
Figure 2-9, DVs
2000-2018
Figure 2-10, Trends
1980-2018
Annual fourth highest MDA8 values are based on all sites with at least
75% annual data completeness for at least 30 of the 39 years, with no
more than two consecutive years having less than 75% complete data
(n = 196 sites)
Fiqure 2-11, Trends
2000-2018
Annual fourth highest MDA8 values are based on all sites with at least
75% annual data completeness for at least 15 of the 19 years, with no
more than two consecutive years having less than 75% complete data
(n = 870 sites)
Design values are presented for sites with valid DVs for at least 13 of
the 17 3-year periods, with no more than two consecutive periods
havinq invalid DVs (n = 629 sites)
Figure 2-12, Trends
2000-2018
Figure 2-13, Diurnal
Patterns
2015-2017
All hourly concentrations are presented for 2015-2017 for these four
monitorinq sites
Figure 2-14, Seasonal
Pattern
2015-2017
All valid MDA8 values are presented for 2015-2017 for these four
monitorinq sites
2A.2 ANALYSES OF 1-HOUR CONCENTRATIONS
Figure 2-15 of Chapter 2 presents hourly concentrations available in AQS (at the time of
the data query on August 14, 2019) from any site with such data during the 2016-2018 period.
The daily maximum 1-hr (MDA1) values presented in section 2.4.5 and (summary statistics
shown in Table 2A-2 below) were calculated according to Appendix H to 40 CFR Part 50 for all
sites with valid 2016-2018 design values for the current 8-hour standards. Generally, MDA1
values are derived (as the maximum 1-hr concentration during a day) for days for which at least
18 1-hr concentrations are available in AQS or for which a 1-hr concentration greater than 0.12
ppm has been reported in AQS. For this most recent design value period, the mean number of
observations per site at or above 100 parts per billion (ppb) was well below one (0.19) for sites
meeting the current standards compared to well above one (8.09) for sites not meeting the current
standards.
2A-3
-------
Table 2A-2. Summary statistics for MDA1 concentrations at sites with differing design
values for 2016-2018.
Statistic
Design Va
ue (ppb)
41-60
61-70
71-84
85-111
Number of observations (obs)
137,443
635,822
226,876
44,059
Number of sites
149
695
241
42
25th percentile concentration (ppb)
33
37
38
45
Median concentration (ppb)
40
45
48
56
Mean concentration (ppb)
40.1
45.0
49.0
59.9
75th percentile concentration (ppb)
47
53
58
73
95th percentile concentration (ppb)
58
65
76
96
99th percentile concentration (ppb)
66
75
89
115
# of obs (# of sites) > 240 ppb
1(1)
0(0)
0(0)
0(0)
# of obs (# of sites) > 200 ppb
1(1)
1(1)
0(0)
0(0)
# of obs (# of sites) > 160 ppb
1(1)
2(2)
0(0)
1(1)
# of obs (# of sites) > 120 ppb
3(2)
17(11)
26 (23)
316(27)
# of obs (# of sites) > 100 ppb
14(9)
149 (94)
561 (165)
1,729 (42)
Mean # of obs > 100 ppb per siteA
0.09
0.21
2.33
41.17
A This is the number of obs at or above 100 ppb divided by t
lowest bins combined (i.e., all sites with a design value < 70
the two highest bins combined (i.e., all sites with a design va
ie number of sites in this bin (column). For the two
ppb), the mean is 0.19 obs > 100 ppb per site, and for
lue > 70 ppb), the mean is 8.09 obs > 100 ppb per site.
The figures and tables presented below contain additional analyses based on the MDA1
concentrations for years 2000-2004 and 2014-2018. Figure 2A-1 compares the distribution of
MDA1 concentrations for each 8-hour design value bin between the earlier (2000-2004; red
boxes) and latter (2014-2018; blue boxes) periods. The comparison shows a slight upward shift
in the mid-range concentrations for the highest (> 85 ppb) and lowest (< 60 ppb) DV bins, while
the two middle bins show little change. The range between the 1st and 99th percentiles as
represented by the whiskers shrinks slightly between the earlier and latter periods in all four bins.
Finally, the very highest concentrations (shown as dots above the top whisker) are reduced in the
two highest DV bins. This is also reflected in Table 2A-3 and Table 2A-4, which show summary
statistics similar to Table 2A-2 for the 2000-2004 and 2014-2018 periods, respectively. These
tables show, as might be expected, that sites with higher design values have a larger number of
days with MDA1 values at or above 100 ppb than sites with lower design values. This statistic is
nearly 40 times higher in both periods for sites not meeting the current standards compared to
sites meeting the current standards. Across the three design value periods in 2014 to 2018, sites
not meeting the current standards have on average nearly 20 observations at or above 100 ppb
per 3-year period, while the average for sites meeting the current standards is less than 0.5.
Figure 2A-2 and Figure 2A-3 show maps of the average number of days where the
MDA1 concentrations were greater than or equal to 100 ppb (also known as the N100 metric) for
the 2000-2004 and 2014-2018 periods, respectively. These maps show that nearly all sites in the
2A-4
-------
U.S. have seen a large reduction in the number of days with high MDA1 concentrations since the
beginning of the century. This is also reflected in the final rows of Table 2A-3 and Table 2A-4,
which indicate a decrease of 84% in the total number MDA1 values greater than or equal to 100
ppb between 2000-2004 and 2014-2018.
200
160'
_Ci
Q_
Cl.
£=
O
-j—•
00
i_
~£=
U
o
£=
120 -
o
O
CO
O
L_
=5
O
E
80-
Z5
E
X
ra
gf
>-
're
Q
40-
2000-2004
2014-2018
!
I
<60
61-70
71-84
> 84
8-hour 03 Design Value (ppb)
Figure 2A-1. Boxplots comparing the distribution of MDA1 concentrations for 2000-2004
(red) to the distribution of MDA1 concentrations for 2014-2018 (blue),
binned by the 8-hour design value at each monitoring site. The boxes
represent the 25th, 50th and 75th percentiles and the whiskers represent the 1st and
99th percentiles. Outlier values are represented by circles.
2 A-5
-------
Table 2A-3. Summary statistics for MDA1 concentrations at differing design values for
2000-2004.
Statistic
Desiqn Value (ppb)
39-60
61-70
71-84
85-131
Number of observations (obs)
115,061
286,026
1,410,915
964,577
Number of desiqn values (DVs)A
127
313
1,660
1,229
25th percentile concentration (ppb)
29
35
37
39
Median concentration (ppb)
36
44
48
52
Mean concentration (ppb)
36.4
44.4
49.4
54.9
75th percentile concentration (ppb)
44
53
60
68
95th percentile concentration (ppb)
56
68
79
95
99th percentile concentration (ppb)
68
79
94
117
# of obs (# of DVsA) > 240 ppb
0(0)
0(0)
0(0)
0(0)
# of obs (# of DVsA) > 200 ppb
0(0)
0(0)
0(0)
5(5)
# of obs (# of DVsA) > 160 ppb
0(0)
0(0)
15(9)
270 (97)
# of obs (# of DVsA) > 120 ppb
0(0)
8(6)
720 (373)
7,967 (977)
# of obs (# of DVsA) > 100 ppb
24 (15)
162 (77)
7,838 (1,360)
34,587 (1,229)
Mean # of obs > 100 ppb per site6
0.19
0.52
4.72
28.14
A Since this table covers three design va
B This is the number of obs at or above 1
lowest bins combined (i.e., sites with a d
two highest bins combined (i.e., sites wit
ue periods, individual sites may be counted up to three times.
00 ppb divided by the number of site-DVs in this bin (column). For the two
esign value < 70 ppb), the mean is 0.42 obs > 100 ppb per site, and for the
h a design value > 70 ppb), the mean is 14.69 obs > 100 ppb per site.
Table 2A-4. Summary statistics for MDA1 concentrations at differing design values for
2014-2018.
Statistic
Desiqn Value (ppb)
27-60
61-70
71-84
85-112
Number of observations (obs)
473,542
1,903,711
620,637
114,656
Number of desiqn values (DVs)A
523
2,120
667
110
25th percentile concentration (ppb)
33
37
39
45
Median concentration (ppb)
40
45
48
57
Mean concentration (ppb)
40.3
45.2
49.4
60.2
75th percentile concentration (ppb)
47
53
59
74
95th percentile concentration (ppb)
58
66
76
97
99th percentile concentration (ppb)
66
75
89
115
# of obs (# of DVsA) > 240 ppb
1(1)
2(2)
0(0)
0(0)
# of obs (# of DVsA) > 200 ppb
3(3)
5(5)
0(0)
0(0)
# of obs (# of DVsA) > 160 ppb
4(4)
6(6)
0(0)
3(3)
# of obs (# of DVsA) > 120 ppb
18(13)
40 (29)
88 (61)
788 (78)
# of obs (# of DVsA) > 100 ppb
47 (31)
404 (242)
1,540 (445)
4,822 (110)
Mean # of obs > 100 ppb per site6
0.09
0.19
2.31
43.84
A Since this table covers three design va
B This is the number of obs at or above 1
lowest bins combined (i.e., sites with a d
two highest bins combined (i.e., sites wit
ue periods, individ
00 ppb divided by
esign value < 70 p
h a design value >
ual sites may be counted up to three times,
the number of site-DVs in this bin (column). For the two
3b), the mean is 0.17 obs > 100 ppb per site, and for the
70 ppb), the mean is 8.19 obs > 100 ppb per site.
2A-6
-------
Average Number of Days with MDA1 a 100 ppb, 2000 - 2004
• 0 o 0.1-1.0 o 1.1-3.0 ® 3.1-10.0 • >10.0
Figure 2A-2. Map showing the average number of days with MDA1 > 100 ppb, 2000-2004.
••
• o
Average Number of Days with MDA1 > 100 ppb, 2014 - 2018
• 0 o 0.1-1.0 o 1.1-3.0 © 3.1-10.0 • >10.0
Figure 2A-3. Map showing the average number of days with MDA1 > 100 ppb, 2014-2018.
2A-7
-------
Figure 2A-4 below shows the number of days in 2016-2018 with an VI DAI concentration
at or above 100 ppb and 8-hour design values (similar to Figure 2-16), for all sites with a 2016-
2018 design value less than 98 ppb. All sites meeting the current standard had six or fewer (i.e.,
two or fewer per year) MDA1 values at or above 100 ppb, and all but three sites meeting the
current standard had three or fewer (i.e., one or fewer per year) MDA1 values at or above 100
ppb.
60'
50-
o
Q
40-
30
~
20
10-
o-
Number of Monitors
• 1
2-3
4-9
*
•
0 or more
*
*
*
•
*
*
*
*
A
*
*kl *
•
~
*
**
* * i
*
*
* *
i <
Jh.
*4
. W
• * <
>*•*
»* »
* *
* *
V
*
ji
4
~
*
. * .
~ ~'
* * * *
1
• . t ~~ * V
* * * *•
40
50
60 70 80
-hour Ozone Design Value (ppb)
90
100
Figure 2A-4. Number of days in 2016-2018 at each monitoring site with a MDA1
concentration greater than or equal to 100 ppb and an 8-hour design value
less than 98 ppb. Sites with higher design values had more days, up to a
maximum of 164 (at a site in southern CA).
2A-8
-------
APPENDIX 2B
ADDITIONAL DETAILS ON BACKGROUND OZONE
MODELING AND ANALYSIS
TABLE OF CONTENTS
2B.1 Photochemical Modeling Methodology 2B-2
2B.1.1 Modeling Platform Overview 2B-4
2B.1.2 Emissions Overview 2B-5
2B1.2.1 Natural Emission Inventory 2B-3
2B1.2.2 Anthropogenic Emission Inventory 2B-4
2B.2 Evaluation 2B-8
2B.3 International Contributions 2B-40
References 2B-44
TABLE OF FIGURES
Figure 2B-1. NOAA U.S. climate regions 2B-9
Figure 2B-2. (a) Normalized Mean Bias (%) and (b) Mean Bias (ppb) of maximum daily
average 8-hr ozone (MDA8) by NOAA climate region (y-axis) and by season
(x-axis) at AQS monitoring sites 2B-17
Figure 2B-3. NMB (a) and MB (b) of MDA8 O3 greater than or equal to 60 ppb from the
12km resolution CONUS simulation by NOAA climate region (y-axis) and
by season (x-axis) at AQS monitoring sites 2B-17
Figure 2B-4. Density scatter plots of observed versus predicted MDA8 O3 from the 12km
resolution CONUS simulation for the Northeast region by season 2B-18
Figure 2B-5. Density scatter plots of observed versus predicted MDA8 O3 from the 12km
resolution CONUS simulation for the Central region by season 2B-19
Figure 2B-6. Density scatter plots of observed versus predicted MDA8 O3 from the 12km
resolution CONUS simulation for the EastNorthCentral region by season.... 2B-20
Figure 2B-7. Density scatter plots of observed versus predicted MDA8 O3 from the 12km
resolution CONUS simulation for the Southeast region by season 2B-21
Figure 2B-8. Density scatter plots of observed versus predicted MDA8 O3 from the 12km
resolution CONUS simulation for the South region by season 2B-22
2B-1
-------
Figure 2B-9. Density scatter plots of observed versus predicted MDA8 O3 from the 12km
resolution CONUS simulation for the Southwest region by season 2B-23
Figure 2B-10. Density scatter plots of observed versus predicted MDA8 O3 from the 12km
resolution CONUS simulation for the WestNorthCentral region by season.. 2B-24
Figure 2B-11. Density scatter plots of observed versus predicted MDA8 O3 from the 12km
resolution CONUS simulation for the Northwest region by season 2B-25
Figure 2B-12. Density scatter plots of observed versus predicted MDA8 O3 from the 12km
resolution CONUS simulation for the West region by season 2B-26
Figure 2B-13. Mean Bias (ppb) from the 12km resolution CONUS simulation of MDA8 O3
greater than or equal to 60 ppb over the period May through September 2016
at AQS and CASTNET monitoring sites in the continental U.S. modeling
domain 2B-27
Figure 2B-14. Mean Error (ppb) from the 12km resolution CONUS simulation of MDA8 O3
greater than or equal to 60 ppb over the period May through September 2016 at
AQS and CASTNET monitoring sites in the continental U.S. modeling domain.
2B-27
Figure 2B-15. NMB (%) from the 12km resolution CONUS simulation of MDA8 O3 greater
than or equal to 60 ppb over the period May through September 2016 at AQS
and CASTNET monitoring sites in the continental U.S. modeling domain... 2B-28
Figure 2B-16. NME (%) from the 12km resolution CONUS simulation of MDA8 O3 greater
than or equal to 60 ppb over the period May through September 2016 at AQS
and CASTNET monitoring sites in the continental U.S. modeling domain... 2B-28
Figure 2B-17. WOUDC sonde locations and sampling frequency used in evaluation of
hemispheric model simulation 2B-29
Figure 2B-18. WOUDC sonde releases averaged by release location over 2016; observations
(left), predictions from the hemispheric CMAQ simulation (middle), ratio
(right). Observations are ordered with increasing latitude (South to North).. 2B-30
Figure 2B-19. WOUDC sonde releases averaged by day with a 20-point moving average;
observations (left), predictions from the hemispheric CMAQ simulation
(middle), ratio (right) 2B-31
Figure 2B-20. WOUDC sonde releases averaged by release location over March, April, May
in 2016; observations (left), predictions from the hemispheric CMAQ simulation
(middle), ratio (right) 2B-32
Figure 2B-21. WOUDC sonde releases averaged by release location over June, July, August in
2016; observations (left), predictions from the hemispheric CMAQ simulation
(middle), ratio (right) 2B-33
Figure 2B-22. OMI O3 (OMPROFOZ v003, left) compared to simulated (hemispheric CMAQ
simulation, center), and ratios (right) of vertical column densities for January
(top) and April (bottom) 2B-34
2B-2
-------
Figure 2B-23
Figure 2B-24
Figure 2B-25
Figure 2B-26
Figure 2B-27
Figure 2B-28
Figure 2B-29
Figure 2B-30
Table 2B-1.
OMI O3 (OMPROFOZ v003, left) compared to simulated (hemispheric CMAQ
simulation, center), and ratios (right) of vertical column densities for July (top),
and October (bottom) 2B-35
OMI Nitrogen Dioxide (0MN02D_HR v003, left) compared to simulated
(hemispheric CMAQ simulation, center), and ratios (right) of vertical column
densities for January (top) and April (bottom) 2B-36
OMI Nitrogen Dioxide (0MN02D_HR v003, left) compared to simulated
(hemispheric CMAQ simulation, center), and ratios (right) of vertical column
densities for July (top) and and October (bottom) 2B-37
OMI Formaldehyde (OMHCHO v003, left) compared to simulated (hemispheric
CMAQ simulation, center), and ratios (right) of vertical column densities for
January (top) and April (bottom) 2B-38
OMI Formaldehyde (OMHCHO v003, left) compared to simulated (hemispheric
CMAQ simulation, center), and ratios (right) of vertical column densities for
July (top), and October (bottom) 2B-39
Total predicted MDA8 O3 and contributions (see legend) over time in the West
(top), and all East (bottom) averaged over all grid cells and days in the U.S. 2B-41
International contribution (black line) to predicted MDA8 O3 and components
(see legend) over time in the West (top), and all East (bottom) averaged over
all grid cells and days in the U.S 2B-42
International contribution (black line) to predicted MDA8 O3 and components
(see legend) over time averaged over all grid cells in the West at high elevation
(top), near-border sites (middle), and Low/Interior sites (bottom) 2B-43
TABLE OF TABLES
Summary of 12km resolution CONUS CMAQ 2016 model performance
statistics for MDA8 O3 by NOAA climate region, by season and monitoring
Network 2B-15
2B-3
-------
This appendix for the background ozone (O3) modeling and analysis includes a
description of the methodology for photochemical modeling, an evaluation of the modeling, and
a more detailed analysis of the predicted contributions from international anthropogenic
emissions. The methodology section includes a description of the modeling platform and
emissions. The evaluation section includes comparisons against surface, sondes and satellite
measurements. The international component analysis separately estimates O3 impacts from
China, India, Canada/Mexico, and global shipping at the hemispheric scale.
2B.1 PHOTOCHEMICAL MODELING METHODOLOGY
2B.1.1 Modeling Platform Overview
A multiscale modeling system is applied at both hemispheric and regional scales with
consistent methodologies for emissions inputs, meteorological inputs, model chemistry, and
photochemical models. Consistency across spatial scales reduces the number of assumptions that
have to be made in integrating predictions from the global and the regional modeling. However,
methodological consistency does not address sources of uncertainty associated with individual
inputs used by the modeling system.
The modeling system uses one emission model, one meteorological model, and one
chemical transport model. The meteorological model is the Weather Research and Forecasting
model (WRF v3.8). The emissions model is the Sparse Matrix Operating Kernel for Emissions
(SMOKE v4.5). The chemical transport model is the Community Multiscale Air Quality model
(CMAQ) version 5.2.1 with the Carbon Bond mechanism (CB6r3) and the non-volatile aerosol
option (AE6). Each of these models is applied at hemispheric and regional scales. The regional
meteorology components of the modeling system are described in more detail in section 3C.4.1.4
of Appendix 3C, while emissions inputs are summarized here.
The models identified above are configured differently for the hemispheric and regional
scales as appropriate for the intended purpose. The hemispheric scale model uses a polar
stereographic projection at 108 kilometer (km) resolution to completely and continuously cover
the Northern Hemisphere. At the regional scale, the model employs a Lambert conic conformal
projection at 36 km resolution to cover North America and at 12 km resolution to cover the
lower 48 contiguous states. The hemispheric scale allows for long-range free tropospheric
transport with 44 layers between the surface and 50 hPa (-20 km asl). The 36 km and 12 km
regional modeling has 35 vertical layers between the surface and 50 hPa. The hemispheric
modeling system was initiated on May 1, 2015 and run continuously through December 31,
2016. The regional model was initialized using the hemispheric result on December 21, 2015 and
run continuously through December 31, 2016.
2B-4
-------
2B. 1.2 Emissions Overview
The emissions inventories are summarized here and more information is available in the
Emissions Technical Support Documents (U.S. EPA, 2019a, U.S. EPA, 2019b) and in Appendix
3C. The emissions model inputs are discussed separately for natural and anthropogenic
emissions. The stratospheric fluxes (section 2.5.1.1 of main document) are not discussed here
because, although they are a source of ozone, they are not emissions. The regional inventories
over North America are based on the Inventory Collaborative 2016 emissions modeling platform
(http://views.cira.colostate.edu/wiki/wiki/9169), which was developed through the summer of
2019. Three versions of the 2016 inventory developed: "alpha" (also known as the 2016v7.1
platform) - which consisted of data closely related to the 2014 National Emissions Inventory
(NEI) version 2 and 2016-specific data for some sectors; "beta" (also known as the 2016v7.2
platform) - which incorporated data from state and local agencies and adjustments to better
represent the year 2016; and "version 1" (also known as the 2016v7.3 platform) - which has the
completed representation of 2016 and some elements from the 2017 NEI. For any regional
inventories, this analysis used the 2016 "alpha release" (specifically the modeling case
abbreviated 2016fe) that is publicly available from https://www.epa.gov/air-emissions-
modeling/2016-alpha-platform. Any changes in the 2016 "beta" or "version 1" platforms are not
included in this modeling and therefore are not captured in the subsequent analysis.
2B.1.2.1 Natural Emission Inventory
The natural emission inventory databases cover all the sources discussed in section 2.5.1
except the International Anthropogenics. The databases that are available depend upon the scale.
At the global scale, lightning NOx emissions are based on monthly climatological data; biogenic
VOC emissions have hourly and day-specific (MEGAN v2.1, Guenther et al., 2012) temporal
scales; soil NOx also has hourly and day-specific temporal scales (Berkeley Dalhousie Soil NOx
Parameterization, as implemented by Hudman et al., 2012); and fire emissions are based on day-
specific data (FINN vl.5, Wiedinmyer et al., 2011). Over our regional domain, regional
inventories supersede the biogenic VOCs, soil NOx, and fire emissions using estimates
consistent with the 2016 collaborative emissions modeling platform (https://www.epa.gov/air-
emissions-niodeling/2016-alpha-platform). The regional biogenic VOCs and soil NOx are
derived from the Biogenic Emission Inventory System (BEIS v3.61). Of the natural inventories,
only fires are expected to change significantly in future versions of the 2016 emissions platform.
The biogenic VOC and NOx changes will be minor due to small changes to the land use data
input to BEIS 3.
2B-5
-------
Emissions of NOx are of particular importance to this study and the natural inventory is
summarized here. The total natural NOx emissions3 in this platform is 56 megatons NOx
(reported as equivalent NO2 mass) which is approximately 15.5 TgN. The contributors in order
of magnitude are lightning (55%), soil (33%), and wildfires (12%). Lightning is treated as a
climatological monthly mean contribution, while soils and wildfires are day-specific. It is
important to note that outside North America, prescribed fires are not identified distinctly from
wildfires. Therefore, all wildland fires outside North America are treated as natural. Though not
directly comparable, the lightning and soil magnitudes are consistent with the ranges reported by
(Lamarque et al., 2012). Consistent with previous regional modeling platforms, the lightning
emissions are not included in the emissions inputs to the regional modeling platform. At the
regional scale, the representation of lightning as a monthly mean rate would add lightning on
days where it may not have occurred. At the hemispheric scale, omitting lightning would remove
an important contribution to the well-mixed background O3.
2B.1.2.2 Anthropogenic Emission Inventory
Anthropogenic emissions inputs include both domestic and international sources. The
domestic inventory includes a high-level of detail that is consistent with previous EPA emissions
platforms such as those used to model the year 2011 (https://www.epa.gov/air-emissions-
modeling/201 l-version-6-air-emissions-niodeling-platforms). For the hemispheric emissions
modeling platform, there are over thirty anthropogenic sector of emission files. The traditional
regional platform covers North America including the U.S. sectors, Canadian sectors, and
Mexican sectors. In addition to the typical regional platform sectors, there are nine sectors based
on the Hemispheric Transport of Air Pollution Version 2 (EDGAR-HTAPv2) inventory and 15
sectors that represent emissions in China which together comprise the anthropogenic emissions
outside of North America. The international emission inventories are synthesized from the
EDGAR-HTAP v2 harmonized emission inventory and country specific databases where updates
were likely to be influential. Previous assessments like HTAP (2010, Phase 1) and HTAP (Phase
2) have shown that the anthropogenic portion of USB is most sensitive to emissions in Mexico,
Canada, and China. For Mexico and Canada, the hemispheric platform relies on the same
country-specific databases as the regional platform. For China, as mentioned above, the
hemispheric platform uses a new country specific database. The sources are detailed further
below.
The EDGAR-HTAP v2 inventories were projected to represent the year 2014. Projection
factors were calculated from the Community Emissions Data System (CEDS) inventory at a
3 We refer to wildfires and soil NOx as natural for the purposes of this section even though both may be impacted to
various degrees by human activity.
2B-6
-------
country-sector level. This allowed our inventory to evolve without the risks associated with
transitioning to a new inventory system. Especially because EDGAR-HTAP v2 is superseded for
critical counties, this was the optimal approach. Details of scaling factor development are
described in Section 2.1.5 of the 2016v7.1 Hemispheric Modeling Platform Technical Support
Document (U.S. EPA, 2019a).
Emissions estimates over Mexico are a combination of emissions supplied by the
Mexican government and emissions developed by the EPA. For the 2016 platform, emissions for
point, nonpoint, and nonroad sources were developed based on projections of Secretariat of
Environment and Natural Resources (SEMARNAT)-supplied data for the year 2008. For the
onroad mobile sources, the EPA developed year-specific inventories for 2014 and 2017 by
applying the MOVES-Mexico model and interpolating to the year 2016. More details are
available in the 2016v7.1 emissions platform TSD (U.S. EPA, 2019b).
Emissions for Canada were supplied by Canadian agencies and reprocessed by the EPA
for the domains and model years used in this analysis. Environment and Climate Change Canada
(ECCC) supplied data for four broad inventory sectors (point, on-road mobile, fugitive dust, and
area and non-road mobile sources, the latter including commercial marine vessels). The ECCC
emissions were interpolated to 2016 based on inventories from the years 2013 and 2025.
The China emission inventory was developed at Tsinghua University (THU) and
documented in Zhao et al., 2018 (see supplement). This inventory was extensively compared to
the EDGAR-HTAP v2 and EDGAR v4.3 inventories before use. The largest differences for NOx
in 2016 occurred in individual emissions sectors rather than inventory totals. The SO2 emissions
were more different than NOx emissions between the two inventories because the THU
inventory applies controls to the metal industry that have been adopted by China. The difference
between emissions, primarily NOx emissions, causes small decrease in the spring time surface
O3 over the U.S. compared to using EDGAR-HTAP v2. Comparisons of this update are
summarized by Henderson et al.(2019).
Emissions for the United States representing the year 2016 were developed using the
2014 National Emissions Inventory version 2 (2014NEIv2) as the starting point, although
emissions for some data categories were updated to better represent the year 2016. The point
source emission inventories for the platform are partially updated to represent 2016. Because
2016 is not a year for which a full NEI is compiled, states are only required to submit emissions
for their larger point sources. For units without 2016-specific emissions, the emissions were
carried forward from the 2014 NEIv2. For electric generating units, 2016-specific Continuous
Emissions Monitoring System (CEMS) data are used where the data can be matched to units in
the NEI. Point and nonpoint oil and gas emissions were projected from 2014 to 2016 using
factors based on historic production levels.
2B-7
-------
Other sectors are briefly summarized here and the reader is directed to the TSD for more
details (U.S. EPA, 2019a). Agricultural and wildland (including prescribed) fire emissions were
developed for the year 2016 using methods similar to those used to develop the 2014 NEI, except
that the input data relied on nationally-available data sets and did not benefit from state-
submitted data as are used for NEI year emissions. The assignment of wildland fires to wild or
prescribed is a complex process that is documented in the regional platform emissions TSD (U.S.
EPA, 2019b). Most area source sectors for this platform use unadjusted 2014 NEIv2 emissions
estimates except for commercial marine vehicles (CMV), fertilizer emissions, oil and gas
emissions, and onroad and nonroad mobile source emissions. For CMV, SO2 emissions were
updated to reflect new rules for the North American Emission Control Area (regulation 13.6.1
and appendix VII of MARPOL Annex VI) on sulfur emissions that took effect in the year 2015.
For fertilizer ammonia emissions, a 2016-specific emissions inventory is used in this platform,
while animal ammonia emissions were the same as those in 2014 NEIv2. Onroad and nonroad
emissions were developed based on MOVES2014a outputs for the year 2016, and the activity
data used to compute the onroad emissions were projected from 2014 to 2016 based on distinct
state-specific factors for urban and rural roads. Emissions from 2014 NEIv2 were used directly
for residential wood combustion, fugitive dust, and other nonpoint sources, although
meteorological-based adjustments for dust sources and temporal allocation for residential wood
and agricultural ammonia sources were based on 2016 meteorology. Additional details on the
development of the U.S., Canada, and Mexico emissions are provided in the 2016v7.1 (U. S.
EPA, 2019b).
2B.2 EVALUATION
An operational model performance evaluation for O3 was conducted for the 2016fe
simulation (as referred to in Section 2.5.2.2) using monitoring data, ozone sonde data, and
satellite data in order to estimate the ability of the CMAQv5.2.1 modeling system to replicate the
2016 base year O3 concentrations for the 12 km continental U.S. domain and the 108 km
Northern Hemispheric domain. The purpose of this evaluation is to examine the ability of the
2016 air quality modeling platform to represent the magnitude and spatial and temporal
variability of measured (i.e., observed) O3 concentrations within the modeling domain. The
model evaluation for O3 focuses on comparisons of model-predicted 8-hour daily maximum
concentrations (MDA8) to the corresponding concentrations from monitoring data (for 2016)
collected at monitoring sites in the AQS. The evaluation divided these data into two datasets, one
limited to only CASTNET sites (described in section 2.3.1), and the second comprised of all
other sites. We refer to this second dataset as "AQS."
2B-8
-------
Included in the evaluation are statistical measures of model performance based upon
model-predicted versus observed MDA8 O3 concentrations that were paired in space and time.
Statistics were generated for each of the nine National Oceanic and Atmospheric Administration
(NOAA) climate regions of the 12-km U.S. modeling domain (Figure 2B-1). The regions include
the Northeast, Central, EastNorthCentral, Southeast, South, Southwest, WestNorthCentral,
Northwest and West as were originally identified in Karl and Koss (1984). Note that most
monitoring sites in the West region are located in California, therefore statistics for the West will
be mostly representative of California O3 model performance.
H NorthWest
¦ WestNorthCentral ~ EastNorthCentral ¦ Central
Hi NorthEast
¦ West ~ Southwest ~ South ~ SouthEast
Source: http://wfJVJ,ncdc,noaa,gov/monitoring-references/maps/us-climate-regions,php#references
Figure 2B-1. NOAA U.S. climate regions.
For MDA8 O3, model performance statistics were calculated for each climate region by
season and for the May through September Os season of 2016. Seasons were defined as: winter
(December-January-February), spring (March-April-May), summer (June-July-August), and fall
(September-October-November). Observational data were excluded from the analysis and model
evaluations for sites that did not meet a 75% completeness criterion.4 In addition to the
performance statistics, several graphical presentations of model performance were prepared for
MDA8 O3 concentrations. These graphical presentations include:
4 Each monitoring site had to have 75% of MDA8 values within any seasonal subset to be included in that subset.
Thus individual monitors may be included in one evaluation of season, but not another.
2B-9
-------
(1) density scatter plots of observations obtained from the AQS system excluding CASTNET
(hereafter AQS) and predicted MDA8 O3 concentrations for May through September;
(2) regional maps that show the mean bias and error as well as normalized mean bias and
error calculated for MDA8 > 60 ppb for May through September at individual AQS and
CASTNET monitoring sites;
(3) tile plots that show normalized mean bias (%) and mean bias (ppb) of MDA8 and MDA8
> 60 ppb by NOAA climate region (y-axis) and by season (x-axis) at AQS monitoring
sites;
(4) O3 sonde evaluations comparing vertically resolved ozone model predictions to ozone
sondes measurements from the World Ozone and Ultraviolet Data Centre (woudc.org).
(5) satellite evaluation comparing simulated tropospheric vertical column densities of O3,
nitrogen dioxide, and formaldehyde to OMI retrievals.
The Atmospheric Model Evaluation Tool (AMET) was used to calculate the model
performance statistics used in this evaluation (Gilliam et al., 2005). For this evaluation of the O3
predictions in the 2016fe CMAQ modeling platform, we have selected the mean bias, mean
error, normalized mean bias, and normalized mean error to characterize model performance,
statistics which are consistent with the recommendations in Simon et al. (2012) and the
photochemical modeling guidance (U.S. EPA, 2018).
Mean bias (MB) is used as average of the difference (predicted - observed) divided by
the total number of replicates (n). Mean bias is defined as:
MB = -£1 (P - O) , where P = predicted and O = observed concentrations for every site
and day included in the evaluation.
Mean error (ME) calculates the absolute value of the difference (predicted - observed)
divided by the total number of replicates (n). Mean error is defined as:
ME = ±Z?|P-0|
Normalized mean bias (NMB) is used as a normalization to facilitate a range of
concentration magnitudes. This statistic averages the difference (predicted - observed) over the
sum of observed values. NMB is a useful model performance indicator because it avoids
overinflating the observed range of values, especially at low concentrations. Normalized mean
bias is defined as:
2B-10
-------
!(/>- 0)
NMB = — *100, where P = predicted concentrations and 0 = observed
I(O)
1
Normalized mean error (NME) is also similar to NMB, where the performance statistic is
used as a normalization of the mean error. NME calculates the absolute value of the difference
(model - observed) over the sum of observed values. Normalized mean error is defined as
t\p-a
NME = —n *100
i(o)
i
As described in more detail below, the model performance statistics indicate that the
MDA8 Oa concentrations predicted by the 2016 CMAQ modeling platform closely reflect the
corresponding monitoring data-based MDA8 Oa concentrations in space and time in each region
of the U.S. modeling domain. The acceptability of model performance was judged for the 2016
CMAQ Oa performance results considering the range of performance found in recent regional Oa
model applications (NRC, 2002; Phillips et al., 2008; Simon et al., 2012; U.S. EPA, 2009; U.S.
EPA, 2018). These other modeling studies represent a wide range of modeling analyses that
cover various models, model configurations, domains, years and/or episodes, chemical
mechanisms, and aerosol modules. Overall, the 2016 CMAQ Oa model performance results are
within the range found in other recent peer-reviewed and regulatory applications. The model
performance results, as described in this document, demonstrate the predictions from the 2016
modeling platform closely replicate the corresponding observed concentrations in terms of the
magnitude, temporal fluctuations, and spatial differences for 8-hour daily maximum Oa.
The model performance bias and error statistics for MDA8 Oa predictions in each of the
nine NO A A climate regions and each season are provided in Table 2B-1. As noted above, seasons
were defined as: winter (December-January-February), spring (March-April-May), summer
(June-July-August), and fall (September-October-November). As indicated by the statistics in
Table 2-7, mean bias and error for 8-hour daily maximum Oa are relatively low in each
subregion, not only in the summer when concentrations are highest, but also during other times of
the year. Generally, MB for MDA8 Oa > 60 ppb is less than + 10 ppb. Generally, MDA8 Oa at the
AQS sites in the summer and fall is over predicted except in the Southwest, with the greatest over-
prediction in the EastNorthCentral and WestNorthCentral. Likewise, MDA8 Oa at the
CASTNET sites in the summer and fall is typically over predicted except in the West, Southwest
and WestNorthCentral where the bias shows an under-prediction. In the winter and spring,
2B-11
-------
MDA8 Oa is under predicted at AQS and CASTNET sites in all the climate regions (with NMBs
less than approximately + 25 percent in each subregion).
Figure 2B-2 and Figure 2B-3 are tile plots that summarize to provide an overview of
model performance by region and by season. Figure 2B-2 shows NMB (%) and MB (ppb) of
MDA8 by NOAA climate region (y-axis) and by season (x-axis) at AQS monitoring sites.
Likewise, Figure 2B-3 shows the NMB (%) and MB (ppb) of MDA8 > 60 ppb by NOAA climate
region (y-axis) and by season (x-axis) at AQS monitoring sites. Figure 2B-2 shows that for the
majority of the nine climate regions throughout each year the NMB is within ±10 percent. There
is greater over-prediction (<20%) during the fall in the South, EastNorthCentral (aka Upper
Midwest), and Central (aka Ohio Valley) regions and during the summer in the South, Southeast
and Central (aka Ohio Valley) regions. However, there is greater under-prediction (up to 30
percent) during the winter in the Northwest, Southwest, WestNorthCentral (aka
NRockiesPlains), EastNorthCentral (aka Upper Midwest), Central (aka Ohio Valley), and
Northeast regions as well during the spring in the Northwest.
The density scatterplots in Figure 2B-4 to Figure 2B-12 provide a qualitative comparison
of model-predicted and observed MDA8 Oa concentrations for each climate region by season. In
these plots the intensity of the colors indicates the density of individual observed/predicted
paired values. The greatest number of individual paired values is denoted by locations in the plot
denoted in warmer colors. The plots indicate that the predictions correspond closely to the
observations in that a large number of observed/predicted paired values lie along or close to the
1:1 line shown on each plot. The model is more likely to over-predict the observed values at low
and mid-range concentrations generally < 60 ppb in each of the regions. There are some
relatively infrequent very large over predictions at high concentrations. Preliminary review of
these biases finds that some are related to fire impacts.
Spatial plots of the MB, ME, NMB and NME for individual monitors are shown in Figure
2B-13 through Figure 2B-16, respectively. The statistics shown in these two figures were
calculated over the May through September period, using data pairs on days with observed 8-hr
Oa of greater than or equal to 60 ppb. Model bias at individual sites during the Oa season is
similar to that seen on a sub-regional basis for the summer. Figure 2B-13 shows the mean bias
for 8-hr daily maximum Oa greater than 60 ppb is under predicted overall, but generally within
±10 ppb across the AQS and CASTNET sites. The greatest exceptions are most evident at certain
near-coastal sites where, on average, the model over predicts MDA8 observed Oa > 60 ppb.
Likewise, the information in Figure 2B-15 indicates that the normalized mean bias for days with
observed 8-hr daily maximum Oa greater than 60 ppb is within ± 10% at the vast majority of
monitoring sites across the U.S. domain. Model error, as seen from Figure 2B-14 and Figure 2B-
16, is generally 2 to 10 ppb and 20 percent or less at most of the sites across the U.S. modeling
2B-12
-------
domain. Somewhat greater error is evident at sites in several areas most notably in the West,
WestNorthCentral, Northeast, EastNorthCentral, Southeast, and along portions of the Gulf Coast
and Great Lakes coastlines.
Sonde evaluations are shown for the 108 km Northern Hemisphere domain in Figure 2B-
18 through Figure 2B-21. The sondes used in this analysis and their release frequencies are
shown in Figure 2B-17. Figure 2B-18 shows that the annual mean prediction is generally within
20% of the measured sonde data, except for near the tropopause. Figure 2B-19 shows that the
performance of all sites is generally not as good in the spring (March, April, May) than in the
summer (June, July, August). The seasonal performance of each monitor is shown in Figure 2B-
20 for spring and Figure 2B-21 for summer. By comparison, Figure 2B-20 shows that low biases
extend deeper into the troposphere in spring than in summer. The structure of the bias seems to
suggest a stratospheric causal mechanism because the bias is near the tropopause.
Satellite evaluations in this analysis include tropospheric vertical columns of O3, nitrogen
dioxide (an ozone precursor as described in chapter 2), and formaldehyde (a VOC reaction
product which is an indicator of VOCs and total reactivity of the atmosphere). At this time, only
formaldehyde comparison includes the application of the scattering weights and air mass factor
to the model, which are often used to create an averaging kernel. Similar processing for O3 and
NO2 was not available at the time this appendix was completed. Satellite evaluations focus
exclusively on the 108 km results over the Northern Hemisphere.
Simulated O3 tropospheric vertical column densities are compared to the O3 product
described and evaluated by Huang et al. (2017). Figure 2B-22 and Figure 2B-23 compares the
model to the retrieved column data without application of the averaging kernel. Omitting the
averaging kernel introduces some error into the comparison (Huang et al., 2017; see Figure 9 for
details). Even so, the comparison shows reasonable performance within the mid-latitudes. There
is a notable low bias in January mid-latitudes and near the north pole in April. In addition, high
biases are consistently seen near the corners of the domain in January and April. This cause of
this high-bias pattern will require further analysis. Within the mid latitudes, the model is
performing well with notable low biases in January and scattered high biases in Asia in July.
Given the limitations of the comparison, the performance is quite good.
Simulated nitrogen dioxide (NO2) vertical columns are compared is the OMN02d
(Krotkov et al., 2017, as processed by Lok Lamsal called OMN02D_HR). Similar to O3, the
averaging kernel is not being applied for NO2. Figure 2B-24 and Figure 2B-25 show larger
relative biases for NO2 than O3, particularly in low NO2 regions like over the oceans. Best
performance was over land during July. Model comparisons to NO2 have commonly shown
biases and research in the broader community continues to resolve this issue.
2B-13
-------
Formaldehyde retrieval comparisons are shown in Figure 2B-26 and Figure 2B-27 using
the OMHCHO files, but using the recommended product described by Gonzalez Abad et al.
(2015). The formaldehyde retrievals show a seasonal cycle in the evaluation with a low bias for
the northern-most retrievals in January and October. During April there are high biases that seem
to migrate northward by July. Though we note this bias feature, the main result is reasonable
spatial consistency between the satellite product and the model results. Future work should
explore this evaluation further.
2B-14
-------
Table 2B-1. Summary of 12km resolution CONUS CMAQ 2016 model performance
statistics for MDA8 O3 by NOAA climate region, by season and monitoring
Network.
Climate region
Monitor
Network
Season
No. of
Obs
MB (ppb)
ME (ppb)
NMB (%)
NME (%)
Northeast
AQS
Winter
11,462
-5.9
6.9
-18.1
21.2
Spring
15,701
-4.3
6.7
-9.8
15.2
Summer
16,686
4.6
7.7
10.0
17.0
Fall
13,780
3.3
5.8
9.5
16.9
CASTNET
Winter
1,195
-6.7
7.3
-19.6
21.3
Spring
1,246
-5.0
6.9
-11.0
15.2
Summer
1,224
2.9
6.5
6.7
15.1
Fall
1,215
3.4
5.6
9.9
16.5
Central
AQS
Winter
4,178
-3.8
5.7
-12.5
18.8
Spring
15,498
-1.1
5.5
-2.5
12.1
Summer
20,501
5.5
8.1
12.1
17.9
Fall
14,041
4.9
6.1
12.6
15.7
CASTNET
Winter
1,574
-3.1
5.4
-9.6
16.3
Spring
1,600
-2.2
5.5
-4.8
12.0
Summer
1,551
3.9
7.1
9.0
16.2
Fall
1,528
2.7
5.1
6.9
12.8
EastNorthCentral
AQS
Winter
1,719
-8.5
9.2
-27.3
29.5
Spring
6,892
-3.8
6.8
-8.4
15.2
Summer
9,742
3.2
6.9
7.7
16.3
Fall
6,050
5.6
3.4
17.6
20.2
CASTNET
Winter
435
-9.6
10.1
-28.6
30.1
Spring
434
-6.5
7.8
-14.4
17.4
Summer
412
0.2
5.5
0.5
13.4
Fall
426
2.9
5.1
9.2
16.0
Southeast
AQS
Winter
7,196
-1.4
5.0
-3.9
14.0
Spring
14,569
-1.5
5.3
-3.2
11.3
Summer
15,855
5.1
7.1
12.9
17.9
Fall
12,589
3.4
5.4
8.4
13.3
CASTNET
Winter
887
-3.5
5.3
-9.3
14.3
Spring
947
-3.6
5.6
-7.5
11.7
Summer
926
3.9
6.2
9.9
16.0
Fall
928
1.7
5.0
4.0
11.9
South
AQS
Winter
11,342
-1.0
5.0
-3.1
15.0
Spring
13,093
1.3
6.1
2.8
13.9
Summer
12,819
6.0
7.8
15.7
20.4
Fall
12,443
4.8
6.3
12.1
16.0
2B-15
-------
Climate region
Monitor
Network
Season
No. of
Obs
MB (ppb)
ME (ppb)
NMB (%)
NME (%)
CASTNET
Winter
516
-1.7
5.0
-4.8
13.7
Spring
532
-1.2
5.6
-2.6
12.3
Summer
508
2.6
6.1
6.7
15.8
Fall
520
3.5
5.0
9.0
12.9
Southwest
AQS
Winter
9,695
-4.2
6.2
-11.0
16.1
Spring
10,608
-4.8
6.5
-9.4
12.7
Summer
10,549
-1.2
6.0
-2.3
11.2
Fall
10,298
2.5
4.9
6.0
12.0
CASTNET
Winter
757
-8.1
8.5
-18.0
18.9
Spring
810
-6.9
7.6
-13.1
14.5
Summer
812
-2.8
5.5
-5.3
10.3
Fall
791
-0.1
3.6
-0.3
8.3
WestNorthCentral
AQS
Winter
4,740
-9.3
9.6
-24.9
25.9
Spring
5,066
-3.1
5.9
-7.2
13.5
Summer
5,134
0.7
4.9
1.4
10.6
Fall
4,940
3.3
5.2
9.8
15.3
CASTNET
Winter
568
-9.1
9.8
-23.1
25.0
Spring
607
-5.8
7.3
-12.4
15.6
Summer
600
-1.8
4.6
-3.7
9.4
Fall
505
1.7
4.8
4.4
12.8
Northwest
AQS
Winter
677
-5.7
7.5
-17.5
23.1
Spring
1,288
-4.3
7.3
-10.5
18.2
Summer
2,444
1.2
6.6
3.3
17.5
Fall
1,236
2.8
5.9
9.0
18.7
CASTNET
Winter
--
-
-
-
-
Spring
--
-
-
-
-
Summer
--
-
-
-
-
Fall
--
-
-
-
-
West
AQS
Winter
14,550
-2.1
5.3
-6.1
15.3
Spring
17,190
-4.0
6.1
-8.8
13.3
Summer
18,046
0.6
8.1
1.2
15.2
Fall
16,163
0.4
5.5
0.9
12.8
CASTNET
Winter
506
-3.4
5.6
-8.7
14.1
Spring
519
-5.7
6.6
-11.8
13.7
Summer
526
-5.3
8.1
-8.7
13.3
Fall
530
-2.2
4.7
-4.6
10.0
2B-16
-------
(a) NMB
Northeasts
Ohio Valley-
Upper Midwest-i
Southeast -
South-
NRockiesPlains-
Southwest-
West-
Northwest-S
Fall Wtr Spr Sum
°/
/o
¦
-50 to -40
-40 to -30
-30 to -20
-20 to-10
-10 to 10
10 to 20
20 to 30
il
30 to 40
m
40 to 50
(b) MB
Northeast4
Ohio Valley
Upper Midwest
Southeast-
South -
NRockiesPlains-,
Southwest
West
Northwest-
ppb
¦
-10 to -
-8 to -6
-6 to -4
-4 to -2
-2 to 2
2 to 4
4 to 6
I '
6 to 8
¦
8 to 10
Figure 2B-2, (a) Normalized Mean Bias (%) and (b) Mean Bias (ppb) of maximum daily average 8-hr ozone (MDA8) by
NO A A climate region (y-axis) and by season (x-axis) at AQS monitoring sites. In the text, alternative names are
used: Ohio Valley is Central, Upper Midewest is EastNorthCentral, and NRockiesPlains is NorthWestCentral.
Northeast
Ohio Valley
Upper Midwest
Southeast
South
NRockiesPlains
Southwest
West
Northwest
Fall Wtr Spr Sum
(a) NMB
|
-
I
s
%
¦
-50 to -40
m
-40 to -30
-30 to -20
-20 to-10
-10 to 10
10 to 20
20 to 30
30 to 40
m
40 to 50
(b) MB
Northeast
Ohio Valley
Upper Midwest
Southeast
South
NRockiesPlains
Southwest
West
Northwest
Fall Wtr Spr Sum
ppb
|< -10 tO 20
-8 to -6
-6 to -4
-4 to -2
-2 to 2
2 to 4
4 to 6
• 6 to 8
8 to 10
Figure 2B-3. NMB (a) and MB (b) of MDA8 O3 greater than or equal to 60 ppb from the 12km resolution CONUS simulation
by NOAA climate region (y-axis) and by season (x-axis) at AQS monitoring sites. Dark grey cells indicate
missing values (i.e., no monitored days with MDA8 >= 60 ppb in that region). In the text, alternative names are
used: Ohio Valley is Central, Upper Midewest is EastNorthCentral, and NRockiesPlains is North WestCentral.
2B-17
-------
Spring
Summer
CMAQ 2016fe cb6r3 I6j 12US2 03 8hrmax for Spring Northeast
CM AO 2016fe cb6r3 16j 12US2 03 Shrmax for Summer Northeast
Y«9,3 + 069*X
§ -
40
AOS Oaiy 03 ippto!
V z J4-0,7fl-X
40 60 80 100
AOS Duty 03 Ippti!
Fall
CMAQ 2016fe cb6r3 16] 12US2 03 Shrmax for Fall Northeast
Winter
CMAO 2016fe cb6r3 16j_12US2 03 8hrmax for Winter Northeast
V= tl -0 7ft*X
V « 10-05* x
60 60 100
AOS Daiy 03!ppb}
20 30 40
AOS Duty 03 (ppb)
Figure 2B-4. Density scatter plots of observed versus predicted MDA8 O3 from the 12km
resolution CONUS simulation for the Northeast region by season. Each plot
has a separate scale that is shared for the x and y axes. The dashed line represents
the best fit linear regression line.
2B-18
-------
Spring
Summer
CMAQ 2016fe eb6r3 16j12US2 03 Shrma* for Spring Central
20 40 60 30
AOS Daily 03 Ipph]
Fall
CMAQ 2016fe cb6r3_l6j. 12US2Q3 Shrmax for Fall Central
20 40 60 80 100 120
AOS Dady 03 |ppb|
Winter
CMAQ 2016te cb6f3 I6j 12US2 Q3 Shrmax for Winter Central
CMAQ 2016fe cb6r3 I6j 12US2 03 8hrmax for Summer Central
8
V • 7 3 ~ 0 63 ' X
20 30 40
AQS Daily 03ippb}
Y » 15 - 073 "X
1 1
20 40 60 80
AOS Dnity 03 ipptn
Figure 2B-5. Density scatter plots of observed versus predicted MDA8 O3 from the 12km
resolution CONUS simulation for the Central region by season. Each plot has
a separate scale that is shared for the x and y axes. The dashed line represents the
best fit linear regression line.
2B-19
-------
Spring
Summer
CMAQ 20l6fe cb6r3_t6j12US2 03_8hrmax for SpringENCentral CMAQ 20l6fe cb6r3 16L12US2 03 Sbrmax for Summer EN Central
1.0
o.e
0j6
0.4
V- 14-0S9-X
AOS Dnity 03 ippb] AOS Datfy 03 :ppb!
Fall Winter
CMAQ 20l6fe eb6r3_16j_l2US2 03 Shrmax for Fall EN Central CMAQ_2016fe cb6r3 J6U2US2 03 8hrmax for Winter EN Central
20 40 60 ao
AOS Oafy 03lppb)
20 30 40
AOS Daily Q3tppb)
Figure 2B-6. Density scatter plots of observed versus predicted MDA8 O3 from the 12km
resolution CONUS simulation for the EastNorthCentral region by season.
Each plot has a separate scale that is shared for the x and y axes. The dashed line
represents the best fit linear regression line.
2B-20
-------
Spring
CM AO 2016f e cb6r3 2US2 03 Shrrnax tar Spring _South east
v- te-osa-x
Summer
CMAO 2016fe_cb6r3_l6j_12US2 03 Shrmax for Summer Southeast
8 -
40
AOS Daiy 03ippbl
AOS Daiy 03 Ippbl
Fal
CMAQ 2016re_cb6r3_i6jJ2US2 OS Shrmax lor Fall_Soulheasl
V » 111-OJf, -x
Winter
CMAQ _2016fe_cb6r3_16L12US2 03 Shrmax (or Winter Southeast
s -
« 3
0,5 o „
60 80
AOS Qaiy 03 Ippbl
10 20 30 40
AOS Daiy 03 fppbi
Figure 2B-7. Density scatter plots of observed versus predicted MDA8 O3 from the 12km
resolution CONUS simulation for the Southeast region by season. Each plot
has a separate scale that is shared for the x and y axes. The dashed line represents
the best fit linear regression line.
2B-21
-------
Spring
CMAG 2016le cb6r3 1 Gj 12US2 03 Shrinax for Spring South
AOS Onty OS !ppb|
Fall
CMAQ 2G16fe_cb6r3 16}_12US2 03 Shrmax tor Fall_South
V- 15-0.74" X
Summer
CMAQ 20l6fe cb6r3 16| 12US2 03 Bhrmax (or Summer Soulh
¥ * T9 —0 67 " X
J A.
40 60
AOS Daity 03 ippb]
Winter
CMAQ 2016fe cb6r3J6j_12US2 03_ Shrmax (or Winter Soulh
8 -
8 -
20 40
AOS Daffy 03 Ippb]
&0 60
AOS Dafy 03!ppm
Figure 2B-8. Density scatter plots of observed versus predicted MDA8 O3 from the 12km
resolution CONUS simulation for the South region by season. Each plot has a
separate scale that is shared for the x and y axes. The dashed line represents the
best fit linear regression line.
2B-22
-------
Spring
CMAO 20i6fe cb6r3 16j I2US2 03 Shrmax for Spring Southwest
Summer
CMAQ 2016fe cb6r3 i6j 12US2 03 Shrmax for Summer Southwest
V»2D-QS2*X
U ft
- rj.ti
30 40 50
AOS Daly 03 Ipjptt]
V # 16 - Q 69 ' X
Fall
CMAQ 2016fe cbSr3 16J.12US2 03 Shrmax for Fall Southwest
40 60
AOS Daily 03 ippbl
Winter
CMAQ 2016fe eb6r3_l6i_l2US2 Q3 8hrmax for Winter Southwest
y» »r-045*x
y , 14 - 0 71 * X
30 40 50 60
AOS OnHy 03fppfc}
40 eo 80
AOS Daiy 03 !ppb>
Figure 2B-9. Density scatter plots of observed versus predicted MDA8 O3 from the 12km
resolution CONUS simulation for the Southwest region by season. Each plot
has a separate scale that is shared for the x and y axes. The dashed line represents
the best fit linear regression line.
2B-23
-------
Spring
CMAQ 2016fe cb6r3 I6j 12US2 03 Shrmax for Spring WN Central
V *21 -Q 44 -X
- 0,6
10 20 30 40 50
AOS Daily 03 ippbl
Fall
CMAQ 2016fe cb6r3 16j 12US2Q3 Bhrmax for Fail WN Central
8 -
8 -
8 -
V * t?-0S9 -X
Summer
CMAQ 2016fe cb6r3 16J 12US2 Q3 8hrmax for Summer WN Central
V • 19 — Q 5s * X
AOS Daily 03 Ippbl
Winter
CMAQ 2016fe cb6r3 16L12US2Q3 8hrmax for Winter WN Central
10 20 30 40
AOS Daily Q3{ppb!
3JJ
10 20 30 40
AOS Daily 03ippb}
Figure 2B-10. Density scatter plots of observed versus predicted MDA8 O3 from the 12km
resolution CONUS simulation for the WestNorthCentral region by season.
Each plot has a separate scale that is shared for the x and y axes. The dashed line
represents the best fit linear regression line.
2B-24
-------
Spring
Summer
CMAQ 2016fe cb6r3 16j12US2 03 Shrmax tor Spring Northwest
CMAQ 20l6fe^cb6r3 16j 12US2 03 Shrmax for Summer Northwest
10 £0 30 40 50
AOS Daly 03 Ippbl
Fall
CMAQ 20t6fe cb6r3 16j_l2US2 03 Shrmax for Fall Northwest
40 60
AOS Daily 03 |ppb|
Winter
CMAQ 20l6fecb6r3 J6j 12US2 03 8hrmax for Winter Norlhwesl
y* tr-ft.sa**
» »' I S -* 0 S3 * X
- u.e
- 0.4
I >
4QS OnHtf 03 Jppta!
20 30 40
AOS Da4y 03!ppb}
Figure 2B-11. Density scatter plots of observed versus predicted MDA8 O3 from the 12km
resolution CONUS simulation for the Northwest region by season. Each plot
has a separate scale that is shared for the x and y axes. The dashed line represents
the best fit linear regression line.
2B-25
-------
Spring
CMAQ 20l6fe cb6r3 16j^l2US2 03 Shrrnax for Spring West
8 -
V- 12 . 0.64-X
40 60
AOS Daily 03 Ippti]
Fall
CMAQ 2016fe_eb6r3 16JJ2US2 03 8hrmax for Fa»_West
v = f A >0.71 -X
Summer
CMAQ 2016fe cb6r3 16j l2US2 03 Shrrnax for Summer West
40 60 80 100
AOS party 03{pph!
V- ta-068-X
AOS Daly 03 |ppb!
Winter
CMAQ 2016fe cb6r3 I6j 12US2 03 8hrmax for Winler West
40 60
AOS Daiiy 03 !ppb!
Figure 2B-12. Density scatter plots of observed versus predicted MDA8 O3 from the 12km
resolution CONUS simulation for the West region by season. Each plot has a
separate scale that is shared for the x and y axes. The dashed line represents the
best fit linear regression line.
2B-26
-------
03_8hrmax MB (ppb) for run CMAQ_2016fe_cb6r3_16j_12US2 for 20160501 to 20160930
units = ppb
coverage limit = 75%
TRIANGLE=CASTNET_Daily; CIRCLE=AQS_Daily_03;
Figure 2B-13. Mean Bias (ppb) from the 12km resolution CON US simulation of MDA8 O3
greater than or equal to 60 ppb over the period May through September
2016 at AQS and CASTNET monitoring sites in the continental U.S.
modeling domain.
03_8hrmax ME (ppb) for run CMAQ_2016fe_cb6r3_16j_12US2 for 20160501 to 20160930
units = ppb
coverage limit = 75%
TRIANGLE=CASTNET_Daily; CIRCLE=AQS_Daily_03;
Figure 2B-14. Mean Error (ppb) from the 12kin resolution CONUS simulation of MDA8 O3
greater than or equal to 60 ppb over the period May through September 2016
at AQS and CASTNET monitoring sites in the continental U.S. modeling
domain.
2B-27
-------
03 Shrmax NMB (%) for run CMAQ 2Q16fe cb6r3 16j J2US2 lor 20160501 to 20160930
- 75%
TRIANGLE=CASTNET_Daily; CIRCLE=AQS_Daily_03;
Figure 2B-15. NMB (%) from the 12km resolution CONUS simulation of MDA8 O3 greater
than or equal to 60 ppb over the period May through September 2016 at AQS
and CASTNET monitoring sites in the continental U.S. modeling domain.
units = %
coverage limit = 75%
>20
18
16
14
12
10
8
6
4
2
0
03_8hrmax NME (%) tor run CMAQ_20161e_cb6r3_16j_12US2 for 20160501 to 20160930
TRIANGLE=CASTNET_Daily; ClRCLE=AQS_Daily_03;
Figure 2B-16. NME (%) from the 12km resolution CONUS simulation of MDA8 O3 greater
than or equal to 60 ppb over the period May through September 2016 at AQS
and CASTNET monitoring sites in the continental U.S. modeling domain.
2B-28
-------
Alert (STN018, -62.340, 82.490}
Eureka (STN315. -85.940. 79.980)
Resolute (STN024. -94.960, 74.700)
Lerwick (STN043, -1.180, 60.130)
Churchill (STN077, -94.070, 58.740)
Stony plain (STN021, -114.100, 53.540)
GooseBay (STN076, -60.360, 53.310)
Legionowo (STN221, 20.960, 52.410)
De Bilt (STN316, 5.177, 52.100)
Valentia O (STN318, -10.250, 51.940)
Praha (STN242, 14.440, 50.000)
Kelowna (STN457. -119.400, 49.940)
Hohenpeiss (STN099, 11.000, 47.800)
Pay erne (STN 156, 6.570, 46.490)
Yarmouth (STN4S8, -66.100, 43.870)
SAPPORO (STN012. 141.330, 43.060)
earajas (STN308, -3.580, 40.470)
Boutder ES (STN067, -105.197, 39.949)
Wallops IS (STN 10 7. -75.470, 37.930)
TSUKUBA {STN014, 140.130, 36.060)
Mew Delhi (STN010, 77.127, 28.482)
NAHA (5TN190, 127.690, 26.210)
Hong Kong (STN344. 114.170, 22.310)
Hanoi (STN330, 105.800, 21.020)
Hilo (HI) (STN109, -155.040, 19.430)
San Pedro (STN524, -84.042. 9.940)
Sepang Air (STN443, 101.270, 2.730)
Petaling J (STN322. 101.270, 2.730)
Nairobi (STN175, 36.800, -1.270)
2016-01 2016-03 2016-05 2016-07 2016-09 2016-11 2017-01
¦ MM IB* Mil IIM III I IIIIM MM M Ml It I
I flii flll fll IIIMMIIMI II I It Ml II IMI • I I I I
Ml lllli III I | llMll M I
M III I tl I 11 I I I M t I « 11 I I I 11 I I I I I 11 I 11 I I I I I I II II I I
I Mill) I tl IIIM I I I
II I II M I It I II t It I I I IIM I t I II I I I I II I I 11 I II I
I I I I I 1 It 11I II IIII I I I I 11ItI I II I I I I I II I 1 1 I IIIII I I
I I II II II t I It I I It I tl II I 11 1 II II I I I
I M II I I I 11 II I II I II II I I I I t I I I II I IMIII I II I I I I I I II I II
» II III I II II I t II I I I I I I t I I I MII
I I I I II t II I I I M II I I lit II III M 1 II I I I I I I
nawiMiiMimuii ¦¦•¦•••¦¦•¦•a • i*¦ ¦¦ ¦ • «•¦¦¦¦¦¦ r
iH*iiiu11ill in•nm»¦ • i mi ¦¦¦¦¦ ¦¦>¦¦¦¦¦!aa>i
t i i t i i 11 t i 11 i i i i i 11 i t i t i t i i i it t i
III I II II I I I M It III It I Ml 11 II II If III lllll I llllll
II I I I I I I I 11 II I II I I I M I I I I t I I I I II I I II I 11 11 I I 11 t II I I I
II 11 I t It t It 0 11 11 I 11 I 11 I I I I I I t I I I 11 M 11 III 11 I I I I
•III II III III III I III III III H|
lllll ii i 11 i i i i I < 11 it i i i ii i a I 11 ill II i 11 iv 11 i i 11 i
t I l I I l l l I
I I I I I I I t II I I I 1111 I I I III IIII llllll ft 11 11 I I II I
II I I I 111 11 11 I I 111 11 I I I I I 11 11 I I CI llllll III I I I I I I I
I I I lllll I I | ai ill II
iiiii ii i ii 11 11 ill iiiii i n i a i > i i 11 i 11 11 ii 11 11 iiiii
I I I I I I M II I IIIMIIIIIIIIVIIIIMIII | I I I I I | I I | III
III I till II I III
lllll I III I
I I 11 llllllll I I I I I II I I I I I 11 I I 11 I I I I I I I I I I I I I I I III
i 1 1 t i 1 r
Figure 2B-17. WOUDC sonde locations and sampling frequency used in evaluation of
hemispheric model simulation.
STN443 (101.27, 2.73)=14 I
STN322 (101.27, 2.73) = 10
STN330 (105.80,21.02)=21
STN344 (114.17,22.3l)=48
STN 190 (127.69,26.21)=44
STN014 (140.13,36.06)=47
STN012 (141.33,43.06)=47
STN315 (-85.94,79.98)=61
STN024 (-94,96,74.70)=24
STN077 (-94.07,58.741 = 13
STN021 (-114.10,53.54)=42
STN457 (-119.40.49.94)=37
STN109 (-155,04,19.43)=5Q
STN067 (-105.20.39.951=49
i,80,-1.27)=48
STN010 ( 77.13.28.481=9
STN221 ( 20.96.52.411 = 32
STN242 ( 14.44.50,001 = 51
STN099 ( 11.00,47.801 = 129
STN 156 ( 6.57,46.491=146
STN316 ( 5.18.52.101 = 56
STN308 (-3.58,40.47)=51
STN043 (-1.18.60.131=49
STN318 (-10,25,51.94)=29
STN018 (-62,34.82.49)=47
STN076 <-60.36,53.31)=48
STN458 (-66,10,43.87)=32
STN 107 (-75.47,37.93)=29
STN524 (-84.04, 9.94)=47
2B-29
-------
Sonde (ppb)
zoo
¦2 400
600
800 ¦{
CMAQ (ppb)
Ratio
. ! '¥ t T T T , 0.50
rorN&-«r-tc0^royV^,X>r>~crvO'ro*tO'-''-'':Trrrr>OCOCTv
rsj:r^^c^^O'*ysi^cKr»cr«^pcq«rcocriOiT»'-*i«T«*^»n^*-f«^>cTi*t f>;r^f^o^^p<^fV'«»'o^JNOcbi£^o^Q^'1^r^o^QnMrgr^fnfoodc>*T(Tio4 rHrNirsjcH(yvr4vbr^i|^W(po^a>rNi
--VNjr^rNjr^rAfnrn»r •ffrfiTw^LTiirKninirhor^r^co * *TiniAtnkninkOcn,|t o co^raxn o cr**-* rrav-*N (OtOiDNdidro ro*pr^a» o^fM r^tr^n^coCD *arcvN
f-4rswr*r#+*rrrf\iGrt rf&^rr i/wyti i/*K/wy/"K0r>^aD
Figure 2B-18. WOUDC sonde releases averaged by release location over 2016; observations (left), predictions from the
hemispheric CMAQ simulation (middle), ratio (right). Observations are ordered with increasing latitude (South
to North).
2B-30
-------
Sonde (ppb)
CMAQ (ppb)
Ratio
200
c
o
It
t:
Q.
400
eoo
800
I 1 1 III. 1
m
2.00
1.50
1.20
0.80
0.66
0.50
00099099000
OOOOOOOOfHi-^fH
ID IO lO 10 WD *0 kO ig
rH H H H rH fH r—~ H H H H
OOOOO OOO OOO
cxcNfMnjrsifM fNrNirM rMfsi
rMm*ti/">
-------
Sonde (ppb)
CMAQ (ppb)
Ratio
200
¦2 400
600
800
2.00
1.50
1.20
0.80
0.66
0.S0
1-n. m ro rr» r+ #\ in •-« r*. 0"> «*
^fsir\i<^crv^rsiȣ>u>f^ ^QMi^vdr^o^o^rsrsifnmo^rmfvj
' «*«T«r*r uMnmmw* into r-r*. oo
r^mrri^frjfM^iHm>inr>mr^*Of^o>OTO,,t OHH*jmoco
o» "t poo «*;ooa>o o*»-* *r roin r^r-,
r^rsirNC^ o»-h ojr\i rnr»S Or*
~HrMfNrsir^mrn^ ** i^m/iinuMn t0t*»r*>co
r* o*TO'-,'-»',:T" ^ OC0
oi o rt jrnmO
-------
Sonde (ppb) CMAQ (ppb)
rgr~r-oi *^0^ rsjrj-oCT>«Tv,tTOOO'roo«TicT>'--<*? ^tnr>.»^r^crv'iJ'>^Ocq,«TOOcT> —oci^iCo%Q»cgfnmm,T,trri,«y*r,«t mm irnnun in vor^r^oo
Ratio
rsMfWfrvNHHicoipfnirifs^^O'r
-------
CMAQ 2016fe 2016-01
NMB (CMAQ / Sat - 1)
OMPROFOZ 2016-01
-0.50
NMB (CMAQ / Sat - 1}
OMPROFOZ 2016-04
CMAQ 201 fife 2016-04
Figure 2B-22. OMI O3 (OMPROFOZ v003, left) compared to simulated (hemispheric CMAQ simulation, center), and ratios
(right) of vertical column densities for January (top) and April (bottom).
2B-34
-------
CMAQ 2016fe 2016-07
NMB (CMAQ / Sat -1)
OMPROFOZ 2016-07
OMPROFOZ 2016-10
CMAQ 2016fe 2016-10
NMB (CMAQ / Sat - 1}
-0.25°
Figure 2B-23. OMI O3 (OMPROFOZ v003, left) compared to simulated (hemispheric CMAQ simulation, center), and ratios
(right) of vertical column densities for July (top), and October (bottom).
2B-35
-------
CMAQ 2016fe 2016-01
0MN02D HR 2016-01
0MN02D.HR 2016-04
I0,s t
CMAQ 2016fe 2016-04
NMB (CMAQ / Sat -1}
Figure 2B-24. OMI Nitrogen Dioxide (0MN02D_HR v003, left) compared to simulated (hemispheric CMAQ simulation,
center), and ratios (right) of vertical column densities for January (top) and April (bottom).
NMB (CMAO / Sat -1)
1G15 1
2B-36
-------
CMAQ 2016fe 2016-07
NMB (CMAQ / Sat -1)
0MN02D HR 2016-07
-0.50
NMB (CMAQ / Sat -1}
0MN020 HR 2016-10
CMAO 2016fe 2016-10
Figure 2B-25. OMI Nitrogen Dioxide (0MN02D_HR v003, left) compared to simulated (hemispheric CMAQ simulation,
center), and ratios (right) of vertical column densities for July (top) and and October (bottom).
2B-37
-------
CMAQ 2016fe 2016-01
OMI HCHO 2016-01
-0.25 g
2,00
NMB {Mod / Sat -1)
OMI HCHO 2016-04
CMAQ 2016fe 2016-04
Figure 2B-26. OMI Formaldehyde (QMHCHQ v003, left) compared to simulated (hemispheric CMAQ simulation, center),
and ratios (right) of vertical column densities for January (top) and April (bottom).
2B-38
-------
OMI HCHO 2016-07
CMAQ 2016fe 2016-07
Figure 2B-27. OMI Formaldehyde (QMHCHQ v003, left) compared to simulated (hemispheric CMAQ simulation, center),
and ratios (right) of vertical column densities for July (top), and October (bottom).
NMB (Mod / Sat -1)
£
I
-0.25 O
OMI HCHO 2016-10
CMAQ 2016fe 2016-10
NMB (Mod / Sat - 1)
2B-39
-------
2B.3 INTERNATIONAL CONTRIBUTIONS
This section characterizes the components of predicted international anthropogenic
contributions to local O3 concentrations and the sensitivities to model resolution. The main
characterization of predicted O3 contributions focused on results, based on simulations at a 12
km grid cell resolution, that separated Natural, International, and USA contributions to O3. In
this appendix, the International component is further characterized into some of its component
parts. The component parts are only analyzed at the 108 km hemispheric resolution. First, the
108 km results are compared to the 12 km results to ensure general consistency to build
confidence that, for large scale transport contributions, the 108 km characterization is relevant to
the 12 km results.
Figure 2B-28 shows the 108 km modeling averaged to the West (<97W) and East
(>97W), which can be compared to the 12 km results in the main body. The results from the two
modeling resolution are very consistent with very high correlation coefficients (r) for total O3
(rw est— 0.987; rEast=0.989), USA (rwest=0.987; rEast=0.993), International (rwest=0.981; rEast=0.990),
and Natural (rwest=0.959; rEast=0.814). Within International, the Canada/Mexico component was
separately estimated at both resolutions and agrees well for all grid cells (rwest=0.966;
rEast=0.935), for high-elevation (rwest=0.961, rEast=N/A), and near-border (rwest=0.961,
rEast=0.947). Since the coarser resolution model cannot resolve urban locations, the urban area
weighted results have lower r (-0.8). While any particular grid cell may deviate due to local
conditions, the averages across these large regions are quite consistent. The analysis is restricted
to large scale averages when drawing conclusions from the 108 km analysis for the 12 km
results.
Figure 2B-29 shows the predicted International contribution and some of its component
parts: Canada/Mexico, China, India, and global shipping. This analysis did not attempt to
quantify all International components separately, so the stacked bars generally account for only a
portion of the total. However, the global shipping component of international is an overestimate
as this sector includes some U.S. emissions. Global shipping includes O3 produced within the
U.S. Federal waters, which are also included in the USA contribution. As a result, the sum of
components overstates shipping contributions to the total International contribution, but
generally does not fully account for all components of the International contribution. The partial
accounting is most obvious in the Winter and Spring when large-scale transport is most
important. This suggests that during the summer, the selected components (China, India, Ships,
Canada, Mexico) are a larger fraction of total International contribution. In both the East and the
West, the International contribution peaks in Spring. The same seasonal signal can be seen for
each International component except for Canada/Mexico. As a result, areas where
2B-40
-------
Canada/Mexico are more important will have a later peak of International than those influenced
by the long-range components (e.g., India, China). The 108 km results cannot resolve the border
well and will likely not fully capture the "near- border" effect.
Figure 2B-30 demonstrates the effect of International contribution on seasonality. Figure
2B-30 shows the West broken out into high-elevation, near-border, and Low/Interior sites. The
near-border areas have a larger Canada/Mexico component. The combination of long-range
sources and Canada/Mexico create a peak International contribution at near-border sites that is
one to two months later than at high-elevation or Low/Interior sites. Note that "near-border" sites
are not well resolved by the 108 km simulations.
C West 97W HEMIS All >0 ppb Natural ¦ Res-Anth ¦¦ Intl ¦ USA
60
20I6-OI 2016-03 2016-05 2016-07 2016-09 2016-11 2017-01
C East 97W HEMIS All >0 ppb ¦¦ Natural mm Res-Anth Intl Mi USA
60
2016-01 2016-03 2016-05 2016-07 2016-09 2016-11 2017-01
Average across all grid cells derived as C =4
Figure 2B-28. Total predicted MDA8 O3 and contributions (see legend) over time in the
West (top), and all East (bottom) averaged over all grid cells and days in the
U.S.
2B-41
-------
20.0
C West 97W HEMIS All >0 ppb
India
China
Ships
CN/MX
a 15.0
* 12,5
x 100
2016-01
2016-03 2016-05
C East 97W HEMIS All >0 ppb
2016-07
2016-09 2016-11 2017-01
China ¦¦ Ships H CN/MX
* 12.5
x 10.0
2016-01
2016*03
2016*05
2016-07
2016-09
2016*11
2017-01
Average across all grid cells derived as C = — Cx
N y
Figure 2B-29. International contribution (black line) to predicted MDA8 O3 and
components (see legend) over time in the West (top), and all East (bottom)
averaged over all grid cells and days in the U.S.
2B-42
-------
C West 97W HEMIS >1500m >0 ppb ¦ India Mi China Mi Ships ¦¦ CN/MX
20,0 T
17.5
_
a 15,01
2016 07
2016-11
2017 01
C West 97W HEMIS MX/CAN < 100km >0 ppb ¦¦ India M Oiina wm Ships Mm CN/MX
20.01 .
— C West 97 W HEMIS Low/Interior >0 ppb ¦¦ India ¦¦ China ¦¦ Ships ¦¦ CN/MX
20.0
17.5
3 15.0
a
d
2016 07
2016-09
2017-01
Average across all grid cells derived as C = ~YJXCX
Figure 2B-30. International contribution (black line) to predicted MDA8 O3 and
components (see legend) over time averaged over all grid cells in the West at
high elevation (top), near-border sites (middle), and Low/Interior sites
(bottom).
2B-43
-------
REFERENCES
Gilliam, RC, Appel, KW and Phillips, S (2005). The Atmospheric Model Evaluation Tool
(AMET): Meteorology Module. 4th Annual CMAS Models-3 Users Conference.
Gonzalez Abad, G, Liu, X, Chance, K, Wang, H, Kurosu, TP and Suleiman, R (2015). Updated
Smithsonian Astrophysical Observatory Ozone Monitoring Instrument (SAO OMI)
formaldehyde retrieval. Atmospheric Measurement Techniques 8(1): 19-32.
Guenther, AB, Jiang, X, Heald, CL, Sakulyanontvittaya, T, Duhl, T, Emmons, LK and Wang, X
(2012). The Model of Emissions of Gases and Aerosols from Nature version 2.1
(MEGAN2.1): an extended and updated framework for modeling biogenic emissions.
Geosci Model Dev 5(6): 1471-1492.
Henderson, BH, Dolwick, PD, Jang, CJ, Eyth, A, Vukovich, J, Mathur, R, Hogrefe, C, Pouliot,
G, Possiel, N, Timin, B and Appel, W (2019). Meteorological and Emission Sensitivity
of Hemispheric Ozone and PM2.5. 9th International GEOS-Chem Conference Boston,
MA.
Huang, G, Liu, X, Chance, K, Yang, K, Bhartia, PK, Cai, Z, Allaart, M, Ancellet, G, Calpini, B,
Coetzee, GJR, Cuevas-Agullo, E, Cupeiro, M, De Backer, H, Dubey, MK, Fuelberg, HE,
Fujiwara, M, Godin-Beekmann, S, Hall, TJ, Johnson, B, Joseph, E, Kivi, R, Kois, B,
Komala, N, Konig-Langlo, G, Laneve, G, Leblanc, T, Marchand, M, Minschwaner, KR,
Morris, G, Newchurch, MJ, Ogino, S-Y, Ohkawara, N, Piters, AJM, Posny, F, Querel, R,
Scheele, R, Schmidlin, FJ, Schnell, RC, Schrems, O, Selkirk, H, Shiotani, M,
Skrivankova, P, Stiibi, R, Taha, G, Tarasick, DW, Thompson, AM, Thouret, V, Tully,
MB, Van Malderen, R, Vomel, H, von der Gathen, P, Witte, JC and Yela, M (2017).
Validation of 10-year SAO OMI Ozone Profile (PROFOZ) product using ozonesonde
observations. Atmospheric Measurement Techniques 10(7): 2455-2475.
Hudman, RC, Moore, NE, Mebust, AK, Martin, RV, Russell, AR, Valin, LC and Cohen, RC
(2012). Steps towards a mechanistic model of global soil nitric oxide emissions:
implementation and space based-constraints. Atmos Chem Phys 12(16): 7779-7795.
Karl, T and Koss, WJ (1984). Regional and national monthly, seasonal, and annual temperature
weighted by area, 1895-1983. 4-3. National Environmental Satellite and Data
Information Service (NESDIS). Asheville, NC.
Krotkov, NA, Lamsal, LN, Celarier, EA, Swartz, WH, Marchenko, SV, Bucsela, EJ, Chan, KL
and Wenig, M (2017). The version 3 OMI N02 standard product. Atmospheric
Measurement Techniques Discussions: 1-42.
Lamarque, J-F, Emmons, LK, Hess, PG, Kinnison, DE, Tilmes, S, Vitt, F, Heald, CL, Holland,
EA, Lauritzen, PH, Neu, J, Orlando, JJ, Rasch, PJ and Tyndall, GK (2012). CAM-chem:
description and evaluation of interactive atmospheric chemistry in the Community Earth
System Model. Geosci Model Dev 5(2): 369-411.
2B-44
-------
NRC (2002). National Research Council Committee on Estimating the Health-Risk-Reduction
Benefits of Proposed Air Pollution Regulations. National Academies Press (US).
Washington (DC).
Phillips, S, Wang, K, Jang, C, Possiel, N, Strum, M and Fox, T (2008). Evaluation of 2002
Multi-pollutant Platform: Air Toxics, Ozone, and Particulate Matter. 7th Annual CMAS
Conference.
Simon, H, Baker, KR and Phillips, S (2012). Compilation and interpretation of photochemical
model performance statistics published between 2006 and 2012. Atmos Environ 61: 124-
139.
U.S. EPA (2009). Technical Support Document for the Proposal to Designate an Emissions
Control Area for Nitrogen Oxides, Sulfur Oxides, and Particulate Matter. U.S.
Environmental Protection Agency. Research Triangle Park, NC. U.S. EPA. EPA-420-R-
007. Available at: http://www.epa.gov/otaq/regs/nonroad/marine/ci/420r09007.pdf.
U.S. EPA (2018). Modeling Guidance for Demonstrating Attainment of Air Quality Goals for
Ozone, PM2.5, and Regional Haze. U.S. Environmental Protection Agency. Research
Triangle Park, NC. EPA 454/R-18-009. Available at:
https://www3.epa.gov/ttn/scram/guidance/guide/03-PM-RH-Modeling_Guidance-
2018.pdf.
U.S. EPA (2019a). Technical Support Document: Preparation of Emissions Inventories for the
Version 7.1 2016 Hemispheric Emissions Modeling Platform. Office of Air Quality
Planning and Standards, U.S. Environmental Protection Agency. Research Triangle Park,
NC. Available at: https://www.epa.gov/sites/production/files/2019-
12/documents/2016fe_hemispheric_ tsd. pdf.
U.S. EPA (2019b). Techical Support Document: Preparation of Emissions Inventories for the
Version 7.1 2016 North American Emissions Modeling Platform. Office of Air Quality
Planning and Standards, U.S. Environmental Protection Agency. Research Triangle Park,
NC. Available at: https://www.epa.gov/sites/production/files/2019-
08/documents/2016v7. l_northamerican_emismod_tsd. pdf.
Wiedinmyer, C, Akagi, SK, Yokelson, RJ, Emmons, LK, Al-Saadi, JA, Orlando, JJ and Soja, AJ
(2011). The Fire INventoiy from NCAR (FINN): a high resolution global model to
estimate the emissions from open burning. Geosci Model Dev 4(3): 625-641.
Zhao, B, Zheng, H, Wang, S, Smith, KR, Lu, X, Aunan, K, Gu, Y, Wang, Y, Ding, D, Xing, J,
Fu, X, Yang, X, Liou, K-N and Hao, J (2018). Change in household fuels dominates the
decrease in PM2.5 exposure and premature mortality in China in 2005-2015. Proc Natl
Acad Sci USA 115(49): 12401-12406.
2B-45
-------
APPENDIX 3A
DETAILS ON CONTROLLED HUMAN EXPOSURE
STUDIES
3A-1
-------
3A.1. OVERVIEW
This appendix gives further study-specific details of the range of respiratory effects (with
a particular focus on pulmonary function) in controlled human O3 exposures during exercise. In
these studies, the magnitude or severity of the respiratory effects induced by O3 was influenced
by ventilation rate, exposure duration, and exposure concentration. Because ventilation rates
increase with increased physical activity level, the exposure concentrations eliciting a significant
response in exercising subjects are lower than in subjects exposed while at rest (ISA, Appendix
3, section 3.1.4.2.1).
Table 3A-1 presents the O3 induced change in forced expiratory volume in one second
(FEVi) in 6.6 to 8-hour controlled human exposure studies (involving quasi-continuous or
intermittent exercise). The FEVi values presented are derived by subtracting the percent changes
in mean FEVi in response to filtered air exposure with exercise from the corresponding percent
changes in FEVi in response to O3 exposure with exercise. The controlled human exposure
studies presented involve exposures, with intermittent exercise, of duration 6 to 8 hours and
target exposure concentrations ranging from 0.04 to 0.16 ppm O3. Study design variables are also
described in Table 3A-1 and include mode of exposure (chamber or facemask), whether the
exposure concentration is constant or varying, exposure duration, exercise duration, and minute
ventilation rate normalized by body surface area during exercise (equivalent ventilation rate,1 or
EVR). Table 3A-2 provides further details of individual study design protocols and subject
characteristics for the studies summarized in Table 3A-1.
Table 3A-3 summarizes studies of controlled human exposure to O3 for shorter durations
(1 to 3 hours) during continuous or intermittent exercise in contrast to similar exposure durations
at rest. The table presents reported effects related to pulmonary function, airway responsiveness,
respiratory symptoms, inflammation and/or host defense. Key study design variables are also
described and include exposure concentrations (ranging from 0.07 to 0.40 ppm O3 for studies
during exercise and 0.10 to 1.00 ppm for studies of subjects at rest), ventilation characteristics
during exercise and subject characteristics (sex and health status). This table was adapted from
Tables 7-1, 7-2 and 7-10 in the 1996 AQCD (U.S. EPA, 1996) and Table AX6-1 in the 2006
AQCD (U.S. EPA, 2006), with additional studies from Tables AX6-8 through AX6-13 in the
2006 AQCD, as well as more recent studies from the 2013 ISA (U.S. EPA, 2013) and 2020 ISA
(U.S. EPA, 2020).
1 The EVR is derived by dividing the minute ventilation rate (Ve in L/min) by body surface area in m2. Values
reflect the study mean EVR across the six exercise periods except for R11, as described below.
3A-2
-------
Table 3A-1. Cross-study comparison of mean Ch-induced FEVi decrements in 6.6 to 8-
hour controlled human exposure studies (that include periods of exercise).
Exposure Design0
RefD
EVRe
(L/min
-m2)
AFEViA B (%)
Average Target Ozone Concentration Durin
g Exercise Periods (ppm)F
0.04
0.06
0.07
0.08
0.087
0.10
0.12
0.16
6.6 Hour
Chamber: Six 50-
min exercise
periods, each
followed by 10 min
rest; 35 min rest-
lunch after 3rd
hour.
[constant]
R1
20
-2.85*
-6.06*
R2
20
-1.71*
-3.46*
R3
20
-7.45*
-8.45*
-13.14*
R4
20
-6.17*
R5
19
-15.65*
R6
22
-14.92*
R7
20
-7.71*
-13.88*2
R8
20
-12.79*
[varying]
R1
20
-0.17
-2.78
-6.99*
R4
20
-5.77*
R9
20
-3.52
-6.14*
-7.82*
-12.23*
6.6 Hour with 6-
hour facemask
exposure: Six 60-
min periods each
consisting of 50
min of exercise
and 10 min of rest,
each followed by 3
min testing period
without exposure;
24 min lunch
without exposure
after 3rd hour.
[constant]
R4
20
-6.14*
R5
20
-1.24
-6.35*
-15.41*
R10
17
-11.28*
R10
20
-13.69*
R10
23
-15.88*
R11
18#
-11.00*
R11
2015-23
-13.68*
R4
20
-5.45*
[varying]
R11
109-12
0.80
R11
"127-11
-3.50
R11
18#'x
-13.96*
R11
18#'Y
-10.31*
7.6 Hour
Chamber:
additional hour
onto 6.6 hr
protocol above.
[constant]
R12
15
-9.8*
R12as
14
-19.4*
8-Hour Chamber:
Eight 30-min
exercise periods,
each followed by
30 min rest
[constant]
R13
20
-8.13*
R14
20
-4.07*
1—
O
E
£*
TO
R13
20
-6.73*
R14
20
-5.62*
AValues reflect 03-induced percent change in FEVi at the group mean level, based on subtraction of the filtered air percent
change (post-pre exposure) from the O3 % change in FEVi. For studies R1, R2, R4, R5 and R9, AFEV1 values were calculated
from individual subject data provided by author. AFEV1 values for R3, R6, R12 were calculated from individual subject data in
publication; R7, R8, R10, R11, R13 and R14 AFEV1 values were derived from group mean response provided in publication.
Statistically significant findings are indicated by asterisk (*). A lack of statistical testing is indicted by (A). Unless indicated
otherwise, all studies were in healthy adults.
B In addition to AFEV1, some studies reported respiratory symptoms scores (e.g. cough and pain on deep inspiration). The
exposures with statistically significant increase in respiratory symptoms scores are indicated by orange shading ( ). Blue
shading ( ) indicates symptom scores that were not statistically significant from filtered air.
c Exposure designs with nonvarying exposure concentrations are indicated by [constant], while studies involving different O3
concentrations for different periods of exposures are indicated by [varying], [varying]1 denotes triangular wave exposure
concentrations (0.07 ppm->0.16 ppm->0.10 ppm). Further details on concentrations are provided in Table 3A-2.
D R1=Adams (2006b) and Brown et al. (2008); R2=Kim et al. (2011) and McDonnell et al., 2012; R3=Horstman et al. (1990);
R4=Adams (2003); R5=Adams (2002); R6=Folinsbee et al. (1988); R7=McDonnell et al. (1991); R8=Folinsbee et al. (1994);
R9=Schelegle et al. (2009). R10=Adams (2000); R11=Adams and Ollison (1997); R12=Horstman et al. (1995). R12As refers to
subjects with asthma; R13=Adams (2006a); R14=Hazucha etal. (1992).
EThe average mean EVR during exercise periods (calculated from study-reported information, see also Table 3A-2).
3A-3
-------
* indicates value derived as average of reported mean hourly EVR (which included 50 minutes exercise and 10 minutes rest)
(although study protocol indicated EVR of 20 L/min-m2).
15 23 indicates hourly ventilation rate varied from 15-23 L/min-m2; value presented is the average mean EVR across the entire
experimental period (including both exercise and rest periods).
912 indicates hourly ventilation rate varied from 9-12 L/min-m2; value presented is average mean EVR across the entire
experimental period (including both exercise and rest periods).
7-11 indicates hourly ventilation rate varied from 7-11 L/min-m2; value presented is the average mean EVR across the entire
experimental period (including both exercise and rest periods).
x andY refer to two different varying concentration protocols (Details on concentrations are provided in Table 3A-2.)
F Author's target for average O3 concentrations across the six exercise periods. This differs from the time-weighted average
concentration (based on target or measurements) for full exposure period. For example, as shown in Table 3A-2 in chamber
studies implementing a varying concentration protocol with targets of 0.03, 0.07, 0.10, 0.15, 0.08 and 0.05 ppm, the exercise
period average concentration is 0.08 ppm while the TWA for the full exposure period (based on targets) is 0.82 ppm due to the
0.6 hour lunchtime exposure to 0.10 ppm between periods 3 and 4.
G Results at 0.08 ppm for a subset of the study subjects that were exposed to 0.10 ppm.
3A-4
-------
Table 3A-2. Study-specific details of O3 exposure protocols for 6.6 to- 8-hour controlled
human exposure studies (that include periods of exercise).
RefA
EVRB
during
exercise
(L/min-m2)
Target Exposure Concentration0 (ppm)
Number of
SubjectsE
Avg.
Age
(Range)
Reference
Constant,
(6.6-hr TWA)*
Varying (hourly concentrations),
(6.6-hr TWA)D
6.6-Hour Chamber Study: 50m+10m, 50m+10m, 50m+10m, 35m, 50m+10m, 50m+10m, 50m+10m
Face Mask Exposure (FM): 50m+10m, 3m, 50m+10m, 3m, 50m+10m, 24m, 50m+10m, 3m, 50m+10m, 3m, 50m+10m
red=Os exposure, black = no exposure (i.e., no facemask) bold =exercise periods, non-bold=restperiods
R1
20
0.06
0.08
0.04 (0.03, 0.04, 0.05, 0.05, 0.04, 0.03), (0.041)
0.06 (0.04, 0.07, 0.09, 0.07, 0.05, 0.04), (0.063)
0.08 (0.03, 0.07, 0.10, 0.15, 0.08, 0.05), (0.082)
30 (15M.15F)
23
(21-29)
Adams (2006b)
Brown et al. (2008)
R2
20
0.06
0.08
59 (27M.32F)
30 (15M.15F)
25
(19-35)
Kimetal. (2011)F
R3
20
0.08
0.10
0.12
22 (M)
25
(18-35)
Horstman et al.
(1990)
R4
20
0.08
0.08™, (0.073)
0.08(0.03, 0.07, 0.10, 0.15, 0.08, 0.05), (0.082)
0.08™ (0.03, 0.07, 0.10, 0.15, 0.08, 0.05), (0.073)
30
(15M.15F)
22
Adams (2003)
R5
19-20
0.04™, (0.036)
0.08™, (0.073)
0.12™, (0.109)
0.12
30
(15M.15F)
22
Adams (2002)
R6
22
0.12
10 (M)
25
(18-33)
Folinsbee et al.
(1988)
R7
20
0.08
0.08+0.10
38 (M)
10 (M)
25
(18-30)
McDonnell et al.
(1991)
R8
18, 20
0.12
17 (M)
25
Folinsbee et al.
(1994)
R9
20
0.06 (0.04, 0.07, 0.07, 0.09, 0.05, 0.04), (0.061p
0.07 (0.05, 0.07, 0.08, 0.09, 0.08, 0.05), (0.071p
0.08 (0.03, 0.07, 0.10, 0.15, 0.08, 0.05), (0.082)G
0.087 (0.04, 0.08, 0.09, 0.12, 0.10, 0.09), (0.087P
31 (15M.16F)
21
(18-25)
Schelegle et al.
(2009)
R10
17, 20, 23
0.12™, (0.109)
30(15M, 15F)
22
Adams (2000)
R11
109-12,
117-11,
18#, 2015-
23
0.08™, (0.073)
0.12™, (0.109)
0.12fm (0.07, 0.16, 0.10), (0.109)
0.12™ (0.115, 0.115, 0.130, 0.130, 0.115, 0.115),
(0.109)
12
(6M, 6F)
22
Adams and Ollison
(1997)
7.6-hour Chamber: Additional hour on 6.6 hr chamber protocol above.
R12
15-17
0.16
13 (NR)
17As(7M,10F)
25
(18-35)
Horstman et al.
(1995)
8-hour Chamber: Eight 30-min exercise periods, each followed by 30 min rest
R13
20
0.12
0.12 triangular* (0->0.24->0)
30 (15M.15F)
23
(21-29)
Adams (2006a)
R14
20
0.12
0.12 triangular* (0->0.24->0)
23 (M)
26
(20-35)
Hazucha et al.
(1992)
A R1-R14 matches study codes in Table 3A-1.
B EVR values are the study means during exercise periods except for R11, for which the EVRs are described below.
912 indicates the study protocol varied the hourly ventilation rate from 9-12 L/min-m2 and value reflects the average mean EVR
across the 6-hr experimental period which includes 50-min of exercise and 10 min of rest.
711 indicates the study protocol varied the hourly ventilation rate from 7-11 L/min-m2 and the value reflects the average mean
EVR across the 6-hr experimental period which includes 50-min of exercise and 10 min of rest.
3A-5
-------
#The study protocol describes the target exercise EVR as 20 L/min-m2 but the actual mean EVR during exercise was not
reported and could not be calculated from study data presented. The value was derived from the average of the mean hourly
EVR which consisted of 50-min of exercise and 10-min of rest resulting in an EVR somewhat lower than the target of 20 L/min-
m2.
15 23 indicates the study varied the hourly ventilation rate from 15-23 L/min-m2; and the value reflects the average mean EVR
across the 6-hr experimental period which includes 50-min of exercise and 10 min of rest.
c Unless marked by "F" (for face mask exposure), exposures were conducted in exposure chamber.
D TWA (time weighted average) was calculated taking into account all exposure concentrations during experiment, including lunch
and rest periods. The TWA concentrations for facemask exercise protocols (whether the exposure concentration was constant or
varying) are lower than the target exposure concentrations because the subjects were not exposed to O3 during the 3 minute rest
and 24 minute lunch periods. Conversely, the TWA concentrations for varying exposure chamber protocols were higher than the
targeted average exposure because of the sequence of concentrations, and their relative magnitude during the 35 minute lunch
period.
E All subjects were healthy adults unless marked by "As" for subjects with asthma.M=male, F=female, NR=sex not reported.
F The 0.08 ppm data for the Kim study were reported in McDonnell et al., 2012.
* Triangular = steadily increasing concentration from 0 ppm to 0.24 ppm at hour 4, then back to 0 ppm.
G While Schelegle et al. (2009) reported measured O3 concentrations, the TWA target concentrations listed in the table for the four
protocols are 0.061,0.071,0.082 and 0.087. Based on the O3 concentration measurements taken during the 6 exercise periods, the
average O3 concentrations for the four protocols are 0.063 ppm, 0.072 ppm, 0.081 ppm and 0.088 ppm, while the 6.6-hourTWA
concentrations are 0.063 ppm, 0.073 ppm, 0.083 ppm and 0.088 ppm
3A-6
-------
Table 3A-3. Summary of controlled human exposures to O3 for 1 to 3 hours during exercise or at rest.
03a
(ppm)
Exposure and Ventilation
Characteristics During Exercisew
Subject
Characteristics5
Reported Effects on Pulmonary Function (PF), Airway
Resistance and/or Responsiveness (AR), Respiratory
Symptoms (SY), Inflammation (IF)E and Host Defense (HD)
Reference
AQCD/ISA
Popc
nD
Adult Subjects During Moderate to Heavy Exercise
0.07
3 Ihir IE (6 *15 min, EVR=1S-17
II 'mini m?"
HNS
35M and 52IF
PF: No significant change in IFEV1
SY: fi.• 4jnifie,infdviii'>-
If: II Ho :i''!iiiiiikviiil dnnfr..
Arjomandi et all.., 2018
ri.implon • I .ill ?Q 17
0.08
1 hr CE (mean Ve-57 L/min)
H«
42 M and 8F
(mean 26 yrs)
PF: No significant change in FEV1
SY: No significant change
Avol etal., 1984F
0.08
2 hr IE (4x15 min, Ve=68 L/min)
H
24M
(18-33 yrs)
PF: No significant change in FEV1
SY: No significant change
Linn etal., 1986;
1996 AQCD, Table 7-1
ii iii
i" Ihi i."'.km hnv. iii.ill .il Ill :.ii
10jiid j r'Cj
II!"
inKdiii j-i jii-j
If-: IIII 1\i nun p-«-1.11^h> • ^' in-^ ¦ 'i 1 f-hi¦ no* in iniLinim.il"iv
i&purcc Iheal •jnll-i on Oz onll> complied L< conUd,
significantly increased in nasal Club cells and glutathione after high-
temperature O3 relative to lower temperature FA control.
, r. ,
0.10
1 hr IE (2 x 15 min, VE=27 L/min)
AsM
12M and 9F
(19-40 yrs)
PF/AR: No significant differences in FEV1 or FVC compared to FA and
no exacerbation of exercise-induced asthma in a postexposure
exercise challenge
SY: No significant change
Weymeretal., 1994;
1996 AQCD, Table 7-2
0.10
2 hr Mild IE
H
12M and 10F;
(mean 30 yrs)
PF: No significant change in FEV1
IF: Markers of exposure in exhaled breath condensate including
markers of inflammation (8-isoprostane, TBARS and LTB4), and
markers of oxidative stress (ROS-DNA interaction: 8-OHdG),
increased in a sub-set of NQ01 wildypes and GSTM1 null subjects
Corradi et al., 2002;
2006 AQCD, Table AX6-12
0.10
2 hr IE (4x15 min, \/E=68 L/min)
H
12M and 10F;
(mean 30 yrs)
PF: No significant change in FEV1
SY: No significant change
Linn etal., 1986;
1996 AQCD, Table 7-1
0.10
2 hr IE (4x15 min at either Ve=30
L/min, Ve=50 L/min or Ve=70 L/min)
H
30 M
(19-28 yrs)
PF: No significant change at any ventilation rate
Folinsbee et al., 1978;
1996 AQCD, p. 7-9
0.10
2 hr IE (4x14 min, VE=70 L/min)
HNS
20M
(mean 25 yrs)
PF: No significant change
AR: No significant change in sRAW
SY: No significant change
Kulle etal., 1985;
1996 AQCD, Table 7-1
0.10
3 hr IE (6*15 min,, EVR=25 L/min-m2)
HNS
1SIIVI and 9IF
(18-40 yrs)
IFF: No significant change
SY: No significant change
IFrampton etal., 2015:
2020 ISA, p. 3-15, Table 3-4
0.12
45 min IE (VE=40-46 L/min)
(two sequential 10 min exposures to
0.1 and 0.25 ppm SO2); +/- 4 wk pre-
treatment with antioxidant
As302
5M and12F
(19- 38 yrs)
PF: i FEV1* with no significant differences due to O3 between
placebo and antioxidant supplement
AR: No significant differences due to O3 in placebo vs, antioxidant
pretreatment in bronchial hyperresponsiveness to 0.1 ppm SO2.
Trenga etal., 2001
2006 AQCD, p. 6-67, Table
AX6-7
3A-7
-------
03a
(ppiTl)
Exposure and Ventilation
Characteristics During Exercisew
Subject
Characteristics6
Reported Effects on Pulmonary Function (PF), Airway
Resistance and/or Responsiveness (AR), Respiratory
Symptoms (SY), Inflammation (IF)E and Host Defense (HD)
Reference
AQCD/ISA
Popc
nD
0.12
1 hr CE (30 min warm up Ve=54
L/min, 30 min competitive Ve=120
L/min; overall mean Ve=87 L/min)
HAth
10M
(19-29 yrs)
PF: No significant change in pulmonary function compared to FA
SY: No significant symptoms
Schelegle and Adams, 1986;
1996 AQCD, p. 7-11, Table
7-1
0.12
1 hr CE (mean Ve=89 L/min)
HAlh
15M and2F
(19-30 yrs)
PF: i FEVi J
AR: > 20% increase in histamine responsiveness in one subject
SY: Mild respiratory symptoms
Gong et al., 1986;
1996 AQCD, Tables 7-1,7-
10
0.12
1.5 hr IE (3x15 min, VE=20 L/min)
AsA
|—|NAs
5M and 5F
4M and 4F
(18-41 yrs)
NL immediately and 24 hr after exposure
PF: No change in lung or nasal function.
IF: No change in PMN number
McBride et al., 1994;
2006 AQCD, Table AX6-12
0.12
3 hr IE (8*15 min, EVR=15-17
L/min/m2)
HNS
35M and 52F
(55-70 yrs)
PF: Small statistically significant attenuation of exercise-related
increases FEVi and FVC
SY: No significant change
IF: Significant increase in PMN independent of GSTM1 phenotype
and significant increase in plasma CC16 (marker of airway epithelial
injury) 4 hr and 22hr postexposure
Arjomandi et al., 2018
Frampton et al., 2017;
2020 ISA, p.3-30, Table 3-4
0.12
2 hr IE (4x15 min, EVR=20 L/min-m2)
HNS
9M and 3F
(mean 28 yrs)
PF: No changes in FEVi or FVC
IF: Increased percentage of vessels expressing P-selectin in bronchial
biopsies 1.5 hr postexposure; no change in BAL markers, PMNs or
expression of VCAM-1, E-selectin or ICAM-1 in vessel biopsies
Krishna et al., 1997;
2006 AQCD, Table AX6-12
0.12
2 hr IE (4x15 min, tfE=68 L/min)
H
24M
(18-33 yrs)
PF: No significant change in FEVi
SY: No significant change in respiratory symptoms
Linn et al., 1986;
1996 AQCD, Table 7-1
0.12
2.5 hr IE (4x15 min, Ve=65 L/min)
H
22M
(18-30 yrs)
PF: | FVC*, | FEVi* and j FEF25-75*
AR: No significant change in sRaw
SY: Increased respiratory symptoms
McDonnell et al., 1983;
1996 AQCD, p. 7-164, Table
7-1
0.12
2.5 hr IE (4x15 min, EVR=25L/min-
m2)
H
30M and 31F
(18-35 yrs)
PF: i FEVi* compared with FA
AR: No significant change in sRaw
SY: No significant change
Seal et al., 1993;
1996 AQCD, p. 7-164, Table
7-1
0.125
3 hr IE (4x15 min, Ve=30 L/min);
3 hr IE (4x15 min, VE=30 L/min) x 4
days; Challenged with allergen 20 hr
following the last exposure and
sputum collected 6-7 hr later
<
CO
< <
6M and 5F
(20-53 yrs)
16M and 6F
(19-48 yrs)
PF: Incidence and magnitude of early-phase FEVi decrements to
allergen were significantly greater in Al subjects exposed for 4 days.
IF: Significant increase in sputum eosinophils in AsA and Al subjects
exposed for 4 days: increased sputum lymphocytes, mast cell
tryptase, histamine, and LDH only in AsA subjects exposed for 4 days.
Holzetal.,2002;
2006 AQCD, Tables AX6-3,
AX6-11
0.14
2 hr IE (4x15 min, Ve=68 L/min)
H
24M
(18-33 yrs)
PF: No significant change in FEVi
SY: No significant change in respiratory symptoms
Linn et al., 1986;
1996 AQCD, Table 7-1
3A-8
-------
03a
(ppiTl)
Exposure and Ventilation
Characteristics During Exercisew
Subject
Characteristics6
Reported Effects on Pulmonary Function (PF), Airway
Resistance and/or Responsiveness (AR), Respiratory
Symptoms (SY), Inflammation (IF)E and Host Defense (HD)
Reference
AQCD/ISA
Popc
nD
0.15
2 hr IE (4x14 min, VE=70 L/min)
HNS
20M
(mean 25 yrs)
PF: | FEVi*
AR: 6 subjects with >15% decrease in sGaw
SY: No significant change in respiratory symptoms
Kulle et al., 1985;
1996 AQCD, Table 7-1
0.15
3 hr IE (4*15 min, Ve=26 L/min) and
H
AsM
10M and 11F
(mean 28 yrs)
5M and10F
(mean 30 yrs)
PF: No significant change in pulmonary function.
IF: Small but significant neutrophil increases in AsM subjects
Holzetal., 1999;
2006 AQCD, Table AX6-3
0.16
1 hr CE (mean Ve -57 L/min)
HAt
42 M and 8F
(mean 26 yrs)
PF: Small j FEVi*
SY: t in mild respiratory symptoms*
Avol et al., 1984;
1996 AQCD, Table 7-1
0.16
2 hr IE (4x15 min, \/E=68 L/min)
H
24M
(18-33 yrs)
PF: Small j FEVi*
SY: No significant change in respiratory symptoms
Linn et al., 1986; 1996
AQCD, p. 7-10, Table 7-1
0.18
1 hr CE (30 min warm up Ve=54
L/min, 30 min competitive Ve=120
L/min; overall mean Ve=87 L/min)
HAt
10M
(19-29 yrs)
PF: i FVC* and J, FEV1* compared to FA; J, exercise time for
subjects unable to complete simulation
SY: | respiratory symptoms*
Schelegle and Adams, 1986;
1996 AQCD, p. 7-11, Table
7-1
0.18
2 hr IE (4x15 min, EVR=35 L/min-m?)
Al
26M with
(18-30 yrs)
PF: | FVC*, | FEVi*, j FEF25-75*
AR: | sRaw* and increased reactivity to histamine*
SY: t respiratory symptoms*
McDonnell et al., 1987;
1996 AQCD, Table 7-2
0.18
2.5 hr IE (4x15 min, EVR=25L/min-
m2)
H
32 M and 32F
(18-35 yrs)
PF: i FEVi* compared with FA
AR: t sRaw* compared with FA
SY: t respiratory symptoms* compared with FA
Seal et al., 1993;
1996 AQCD, p. 7-164, Table
7-1
0.18
2.5 hr IE (4x15 min, Ve=65 L/min)
H
20M
(18-30 yrs)
PF: | FVC*, | FEVi* and j FEF25-75*
AR: No significant change in sRaw
SY: t respiratory symptoms*
McDonnell et al., 1983;
1996 AQCD, p. 7-164, Table
7-1
0.20
30 to 80 min CE (Ve=33 or 66 L/min)
H
8M
(22-46 yrs)
PF: O3 effective dose significantly related to pulmonary function
decrements (threshhold for significant responses > 0.2 ppm) and
exercise ventilatory pattern changes; O3 concentration accounted for
the majority of the pulmonary function variance
Adams et al., 1981;
1996 AQCD, Table 7-1
0.20
1 hrCE(VE=80 L/min); 1 hr
competitive simulation (30 min at
Ve=52 L/min, 30 min at Ve=100
L/min; overall mean Ve =77.5 L/min)
HAt
10M
(19-31 yrs)
PF: I FVC*, I FEVi* and J, FEF25-75* compared to FA with both
protocols; j Vt* and t fR* with CE
SY: f respiratory symptoms*
Adams and Schelegle, 1983;
1996 AQCD, Table 7-1
0.20
1 hrCE (Ve=89 L/min)
HAlh
15M and2F
(19-30 yrs)
PF: i Vsmax*, i V02max*, j Vimax*, j work load*, j ride time*, j FVC*,
and i FEVi* compared with FA
AR: > 20% increase in histamine responsiveness in nine subjects
SY: t respiratory symptoms*
Gong et al., 1986;
1996 AQCD, Tables 7-1,7-
10
3A-9
-------
03a
(ppiTl)
Exposure and Ventilation
Characteristics During Exercisew
Subject
Characteristics6
Reported Effects on Pulmonary Function (PF), Airway
Resistance and/or Responsiveness (AR), Respiratory
Symptoms (SY), Inflammation (IF)E and Host Defense (HD)
Reference
AQCD/ISA
Popc
nD
0.20
1 hr CE (mean Ve=60 L/min);
2 exposures x 24 hr apart
HNS
15M
(mean 25 yrs)
PF: Consecutive days of exposure produced similar J, FVC* and J,
FEVi* on each day compared to FA
SY: Consecutive days of exposure produced similar | respiratory
symptoms*
Brookes etal., 1989;
2006 AQCD. Table AX6-9
0.20
2 hr IE (4x15 min, 2 x resting Ve)
HNS
12M and 7F
(21-32 yrs)
AR: No change in sRaw to a 10-breath histamine (1.6%) aerosol
challenge after O3 exposure.
Dimeo etal., 1981;
2006 AQCD, Table AX6-11
0.20
2 hr IE (4x15 min, Ve=20 L/min)
AsA
4M and 5F
(21-42 yrs)
PF: I FEVi*butnotFVC
AR: No change in sRaw
IF: 6 hr postexposure | PMNs* with no change in permeability
markers; 24 hr postexposure PMNs decreased while albumin, total
protein, myeloperoxidase and eosinophil cationic protein increased.
Newson et al., 2000;
2006 AQCD, Tables AX6-3,
AX6-13
0.20
2 hr IE (4x15 min, Ve=30 L/min)
HNS
10M and 2F
(mean 28 yrs)
IF: Significant increase in PMNs and epithelial cells, IL-8, Gro-a,and
total protein in BAL fluid; % PMNs correlated positively with
chemokine levels; significant decrease in the CD4+/CD8+ ratio and
% of activated CD4+ and CD8+ T cells in BAL fluid.
Krishna etal., 1998
2006 AQCD, Table AX6-13
0.20
2 hr IE (4x15 min, EVR=20 L/min-m2)
HNS
8M and 5F
(20-31 yrs)
PF: J,FVC*, | FEV1*, and J, FEF25-75*
IF: Spirometry responses did not predict inflammatory responses;
increased adhesion molecule expression, submucosal mast cell
numbers and alterations in lining fluid redox status; increase in human
leukocyte antigen+ alveolar macrophages in BAL 1.5 hr postexposure.
Blomberg etal., 1999;
2006 AQCD, Tables AX6-1
and AX6-12
0.20
2 hr IE (4x15 min, EVR=20 L/min-m2)
H
10M and12F
(mean 24 yrs)
PF: i FEV1* immediately postexposure but not significantly different
from baseline 2 hr later.
IF: Elevated CC16 levels remained high 6 hr postexposure but
returned to baseline by 18 hr postexposure. No correlation between
CC16 and FEV1 decrement.
Blomberg et al., 2003;
2006 AQCD, Table AX6-1
0.20
2 hr IE (4x15 min, EVR=20 L/min-m2)
chronic inhaled corticosteroid
As
8M and 5F
(mean 33 yrs)
PF: 4FEV1* and jFVC*
AR: Significant increase sRaw
IF: Significant increase in BAL neutrophils, but not eosinophils 18 hr
postexposure; significant increase in mast cells in bronchial biopsy
Stenforset al., 2010;
2013 ISA, p. 6-21
0.20
2 hr IE (4x15 min, EVR=20 L/min-m2)
HNAs
AsM
6M and9F
(19-32 yrs);
9M and 6F
(21-48 yrs)
PF: 4FEV1* (8%, HNAs; 3% AsM) and J, FVC* in both groups with no
significant difference between HNAs and AsM
IF: Significant increase in PMN in both groups with no significant
difference between AsM and HNAs 6 hr postexposure; no relationship
between antioxidant levels and spirometric or cellular responses
Mudway etal., 2001;
Stenforset al., 2002;
2006 AQCD, Table AX6-1
3 A-10
-------
03a
(ppiTl)
Exposure and Ventilation
Characteristics During Exercisew
Subject
Characteristics6
Reported Effects on Pulmonary Function (PF), Airway
Resistance and/or Responsiveness (AR), Respiratory
Symptoms (SY), Inflammation (IF)E and Host Defense (HD)
Reference
AQCD/ISA
Popc
nD
0,20
- fir IF (4-15 min, EVR=20 L/min-m2)
H
8M and 5F
(19-31 yrs)
6M and 9F
(19-32 yrs)
16M and 15F
(19-32 yrs)
IF; Postexposure bronchoscopy was performed at 1.5 hr, 6 fir, and 18-
hr; significant correlations between lung PMNs and blood PMNs
postexposure; significant increase in PMN at 6 hr in bronchial wash
and BAL-fluid as well as in bronchial epithelium and submucosa
biopsies; 18 hr, PMN increase persisted in both bronchial wash and
BAL while PMN in biopsies tended slightly lower; significant decrease
in blood PMNs in subjects 15 hr postexposure compared to FA that
rebounded above FA levels at 6 hr and at 18 hr postexposure, there
was no difference in PMN levels when compared to FA
Bosson et a!., 2013;
2020 ISA, p. 3-29, p. 4-28
0.20
2 hr IE (4x15 min, EVR=20 L/min-m2)
|—|NAs
AsM
6M and 6F
(19-31 yrs)
9M and 6F
(21-48 yrs)
IF: Significantly higher baseline expression of IL-4 and IL-5 in
bronchial mucosal biopsies from AsM vs. HNAs subjects 6 hr
postexposure. Epithelial expression of IL-5, GM-CSF, ENA-78, and IL-
8 increased significantly in AsM vs. HNAs subjects.
Bosson et al., 2003;
2006 AQCD, Table AX6-12
0.20
2 hr IE (4x15 min, EVR=20 L/min-m2)
HNS
8M and 5F
(20-31 yrs)
IF: No neutrophils in NL 1.5 hr postexposure. 30% depletion of uric
acid in NL during hr 2 of exposure with increase in plasma uric acid
levels. No depletion of ascorbic acid, reduced glutathione, or
extracellular superoxide dismutase.
Mudwayetal., 1999;
2006 AQCD, Table AX6-12
0.20
2 hr IE (4x14 min, VE=70 L/min)
HNS
20M
(mean 25 yrs)
PF: | FVC*, | FEVi*, j FEF25-75*, j IC* and j TLC*
AR: i sGaw
SY: t respiratory symptoms*
Kulleetal., 1985;
1996 AQCD, Table 7-1
0.20
3 hr IE (6x15 min, EVR=25 L/min-m2)
HNS
15M and 9F
(18-40 yrs)
PF: l FEW"' and.[. FVC*
SY: t respiratory symptoms*
Frampton et al., 2015;
2020 ISA, p. 3-15, Table 3-4
0.21
1 hrCE (75%V02max)
HAth
6M and 1F
(18-27 yrs)
PF: | FVC*, | FEV1*, j FEF25-75*, and j MW* compared to FA
SY: t respiratory symptoms*
Folinsbee et al., 1984; 1996
AQCD, p. 7-52, Table 7-1
0.21
1 hr CE (Ve =80 L/min) followed by
maximal sprint (peak Ve >140 L/min)
Pre-treatment with albuterol or
placebo
HAth
14M and 1F
(16-34 yrs)
PF: i FVC*, i FEV1*, J, FEF25-75*, and J,Ve™x in both treatment
groups. No difference in the effects of albuterol on exercise
performance vs. placebo.
AR: No significant differences in the effects of albuterol on airway
reactivity to histamine challenge vs placebo.
Gong et al., 1988;
1996 AQCD, Table 7-1
0.22
2.25 hr IE (4x15 min, 6-8xresting Ve)
H
83M and 55F
(mean 22 yrs)
PF: | FVC* and j FEV1*
AR: Increased airway responsiveness 1 day postexposure
IF: Increased epithelial permeability 1 day postexposure; airway
responsiveness and epithelial permeability 1 day postexposure did not
correlate with FEV1 responses immediately following the O3 exposure
Que et al., 2011;
2013 ISA, p. 6-74
0.24
1 hr CE (mean Ve=57 L/min)
HAth
42 M and 8F
(mean 26 yrs)
PF: | FEV1*
SY: t respiratory symptoms*
Avol et al., 1984;
1996 AQCD, Table 7-1
3 A-11
-------
03a
(ppiTl)
Exposure and Ventilation
Characteristics During Exercisew
Subject
Characteristics6
Reported Effects on Pulmonary Function (PF), Airway
Resistance and/or Responsiveness (AR), Respiratory
Symptoms (SY), Inflammation (IF)E and Host Defense (HD)
Reference
AQCD/ISA
Popc
nD
0.24
1 hr competitive simulation at mean
Ve=87 L/min; (30 min at Ve=54
L/min, 30 min at Ve=1 20 L/min)
HAth
10M
(19-29 yrs)
PF: i FVC*, i FEVi* and J, FEF25-75* compared to FA; J, exercise
time* for subjects unable to complete simulation
SY: | respiratory symptoms*
Schelegle and Adams, 1986;
1996 AQCD, p. 7-11, Table
7-1
0.24
1.5 hr IE (3*15 min, Ve=20 L/min)
AsA
|—|NAs
5M and 5F
4M and 4F
(18-41 yrs)
NL immediately and 24 hr after exposure
PF: No change in pulmonary or nasal function.
IF: Significant increase in PMNs (at both time points) and in epithelial
cells (immediately after exposure) only in AsA subjects
McBrideetal., 1994;
2006 AQCD, Table AX6-12
0.24
2.5 hr IE (4x15 min, EVR=25L/min-
m2)
H
31M and 33F
(18-35 yrs)
PF: I FEV1* compared with FA
AR: | sRaw* compared with FA
SY: t respiratory symptoms* compared with FA
Seal etal., 1993; 1996
AQCD, p. 7-164, Table 7-1
0.24
2.5 hr IE (4x15 min, Ve=65 L/min)
H
21M
(18-30 yrs)
PF: | FVC*, | FEV1*, j FEF25-75* and j VT* and | f*
AR: t sRaw*
SY: t respiratory symptoms*
McDonnell etal., 1983;
1996 AQCD, Table 7-1
0.25
1 hr IE (2x15 min, Ve=27 L/min)
AsM
12M and 9F
(19-40 yrs)
PF/AR: No significant differences in FEV1 or FVC compared to FA and
no exacerbation of exercise-induced asthma in a postexposure
exercise challenge
Weymeretal., 1994; 2006
AQCD, Table AX6-11
0.25
1 hr CE (EVR=30 L/min-m2)
HNS
5M and 2F
(22-30 yrs)
PF: | FEV1*
IF: t substance P* and | 8-epi-PGF2a* in segmental washing but not
BAL fluid
Hazbun etal., 1993;
1996 AQCD, Table 7-1
0.25
1 hr CE (VE=30L/min);
Facemask exposure
HNS
32 M and 28F
(mean 23 yrs)
PF: i FEV1*; sex differences in FEV1 decrements not significant;
Uptake of O3 greater in M vs. F, but uptake not correlated with
significant differences in spirometric responses between M and F.
Ultman et al., 2004;
2006 AQCD, Table AX6-1
0.25
1 hr CE (mean Ve =63 L/min)
H
19M and7F
(mean 21 yrs)
PF: | FVC*, | FEV1*, j FEF25-75* and j MW* compared to FA
Folinsbee et al., 1986; 1996
AQCD, Table 7-1
0.25
2 hr IE (2x30 min at VE=39 L/min)
4 consecutive days
HNS
5M and 3F
25-31 yrs
PF: Maximal mean j FEV1* and J, FVC* on day 2, negligible by day 4.
AR/IF: Significant small airway function depression accompanied by
significant PMN in BAL fluid one day following the end of O3 exposure;
PMN number in BAL fluid on day 5 were significantly higher following
O3, compared to air exposures
Frank et al., 2001;
AQCD 2006 Tables AX6-9,
AX6-12
0.25
2 hr IE (4x14 min, Ve=70 L/min)
HNS
20M
(mean 25 yrs)
PF: | FVC*, | FEV1*, j FEF25-75*, IIC* and j TLC*
AR: I SGaw*
SY: t respiratory symptoms*
Kulleetal., 1985;
1996 AQCD, Table 7-1
0.25
3 hr IE (4x15 min, EVR=14 L/min-m2)
H
15M and 3F
(mean 43yrs)
IF: significant increase in 3 hr postexposure sputum PMN compared
to pre-exposure sputum; Bimosiamose pretreatmeni reduced PMN
after O3 exposure to approximately the pre-exposure baseline
Kirsten etal., 2011;
2020 ISA, p. 3-30, Table 3-9
3A-12
-------
03a
(ppiTl)
Exposure and Ventilation
Characteristics During Exercisew
Subject
Characteristics6
Reported Effects on Pulmonary Function (PF), Airway
Resistance and/or Responsiveness (AR), Respiratory
Symptoms (SY), Inflammation (IF)E and Host Defense (HD)
Reference
AQCD/ISA
Popc
nD
0.25
3 hr IE (4*15 min, Ve=30 L/min)
AsA
Al
HNS
13M and 11F
(mean 26 yrs)
6M and 6F
(mean 25 yrs)
5M and5F
(mean 23 yrs)
PF: 03-induced FEV1* decrements of 12.5,14.1, and 10.2% in AsM, Al
and HNS, respectively (group differences not significant)
AR: Methacholine responsiveness increased in AsA subjects; allergen
responsiveness increased significantly after O3 exposure in both AsA
and Al subjects; no change in FINS subjects; allergen or methacholine
response not correlated with each other or lung function
Jorresetal., 1996;
2006 AQCD, Table AX6-11
0.25
3 hr IE (4*15 min, Ve=30 L/min)
"challenged with allergen 20 hr
following the last exposure and
sputum collected 6-7 hr later
AsM
Al
6M and 5F
(20-53 yrs);
16M and 6F
(19-48 yrs)
PF/AR: Significantly greater mean early-phase allergen FEV1
response and number of >20% reductions in FEV1 in Al subjects
IF: Significant increase in sputum eosinophils (AsM and Al) and
lymphocytes, mast cell tryptase, histamine, and LDH (AsM only).
Holzetal.,2002;
2006 AQCD, Tables AX6-3,
AX6-11
0.25
3 hr IE (4x15 min, EVR=20 L/min-m2)
four O3 exposures: screening,
placebo, and two treatments (inhaled
or oral corticosteroids)
HNS
14M and 4F
(20-48 yrs)
PF: Postexposure spirometry not significantly different from baseline.
IF: Screening and placebo O3 exposures caused > 9-fold increase in
sputum neutrophils relative to baseline levels; relative to placebo,
inhaled or oral corticosteroids significantly reduced neutrophil levels
Holzetal.,2005
2006 AQCD, p. AX6-123,
Table AX6-13
0.25
3 hr IE (4x15 min, EVR=20 L/min-m2)
H
12M and 12F
(20-48 yrs)
IF/HD: Sputum neutrophils, sputum CD14+ cells, as well as
concentrations of IL1B, IL8, IL8, MMP9, and TNFa in sputum
supernatant significantly increased 3 hr postexposure
Holzetal., 2015;
2020 ISA, p.3-29, Table 3-9
0.25
3 hr IE (4*15 min, E\/R=20 L/min-m2)
H
11M and 3F
(mean 33 yrs)
IF: Increase in blood neutrophils, neutrophil activation and total
leukocytes at 5 and 7 hr postexposure, but not 24 hr.
Billeret al., 2011;
2020 ISA, p. 4-28, Table 3-4
0.25
3 hr IE (4x15 min, EVR=20 L/min-m2)
H
11M and 3F
(22-47 yrs)
PF: I FVC*, and J, FEV1*
IF: PMN increased in the blood 5 hr after the start of a 3-hr exposure
and returned to baseline 21 hr postexposure
Tanketal., 2011;
2020 ISA, p,3-29, Table 3-4
0.25
3 hr IE (4x15 min, Ve=26 L/min) and
repeated 1 week later
HNS
AsM
10M and 11F
(mean 28 yrs)
5M and10F
(mean 30 yrs)
PF/SY: Significant J, FVC* and j FEV1 that tended to be greater in the
AsM; no significant group differences in symptoms or spirometry.
IF: Significant | neutrophils that did not differ between groups.
Holzetal., 1999; 2006
AQCD, p, AX6-35, Table
AX6-3
0.27
2 hr IE (3x20 min, EVR=25 L/min-m2)
AsA
12-sex not
indicated
(18-37 yrs)
PF/SY: i FVC*, j FEV1* and j VC* and significant increase in
symptom scores 24 hr following allergen challenge compared to FA
IF: Percentage of eosinophils, but not neutrophils, in induced sputum
was higher 6 hr after O3 vs. FA exposure
Vagaggini et al., 2002;
AQCD 2006 Table AX6-12
0.27
2 hr CE (EVR=25 L/min-m2)
FA and to O3 exposures before and
after 4 wk of treatment with
budesonide
AsM
7M and 7F
(20-50 yrs)
PF/SY: Significant J, FEV1 and symptom scores; no change in FEV1
decrements or symptom scores with budesonide
IF: Significant 03-induced increase in sputum PMN and IL-8 was
significantly reduced by budesonide 6 hr postexposure.
Vagaggini et al., 2001;
AQCD 2006 Table AX6-13
3 A-13
-------
03a
(ppiTl)
Exposure and Ventilation
Characteristics During Exercisew
Subject
Characteristics6
Reported Effects on Pulmonary Function (PF), Airway
Resistance and/or Responsiveness (AR), Respiratory
Symptoms (SY), Inflammation (IF)E and Host Defense (HD)
Reference
AQCD/ISA
Popc
nD
0.27
2 hr IE (3x20 min, EVR=25 L/min-m2)
repeated 4 days after prednisone or
placebo
AsA
8M and 1F
(mean 25 yrs)
PF: Corticosteroid pretreatment did not prevent J, FEVi * vs placebo.
IF: Significant inflammatory response (PMN influx) was prevented by
corticosteroid pretreatment in induced sputum 6 hr postexposure.
Vagaggini et al., 2007;
2013 ISA, p. 6-78
0.30
30 to 80 min CE (Ve=33 or 66 L/min)
H
8M
(22-46 yrs)
PF: Significant pulmonary function decrements and exercise
ventilatory pattern changes; multiple regression analysis showed O3
effective dose is a better predictor of response than concentration, Ve,
or duration of exposure, and O3 concentration accounted for the
majority of the pulmonary function variance
Adams et al., 1981;
1996 AQCD, Table 7-1
0.30
1 hr CE (EVR=15 L/min-m2)
HNS
S
17M and 13F
(mean 25 yrs)
19M and 11F
(mean 24 yrs)
PF: | FEW was similar in both groups; based on exhaled CO2, only
smokers showed a reduction in dead space (-6.1 ± 1.2%) and an
increase in the alveolar slope
Bates et al., 2014;
2020 ISA, p. 3-18, Table 3-4
0.30
1 hr CE (Ve -60 L/min)
H
5M
PF: i FVC* and J, FEV1* 1 hr postexposure
AR: | sRaw* 1 hr postexposure
IF: t PMNs* at 1 hr, 6 hr, and 24 hr postexposure compared with FA
in first aliquot "bronchial" sample (peaked at 6 hr); | PMNs* at 6 and
24 hr in pooled aliquots.
Schelegle et al., 1991;
1996 AQCD, Table 7-1
0.30
1 hr CE (VE =60 L/min) or
2hr IE (Ve =45-47 L/min)
H
12M
(mean 24 yrs)
PF: i FEV1* was equivalent for both protocols
SY: Significant symptom scores only in CE protocol
McKittrick and Adams, 1995;
1996 AQCD, Table 7-1
0.30
2 hr CE (EVR=25 L/min-m2)
As
13M and 10F
(mean 33 yrs);
PF: 4% group mean FEV1 decrement; no baseline difference between
responders (8 subjects with >10% FEV1 decrements) and
nonresponders
IF: Significant correlation between changes in FEV1 and changes in
sputum neutrophils 6 hr postexposure compared to FA in responders;
significant increase in eosinophils in nonresponders only; NQ01
wildtype and GSTM1 null genotypes (6 subjects) not associated with
the changes in lung function or inflammatory responses
Vagaggini et al., 2010;
2013 ISA, p. 6-79-80
0.30
2 hr IE (4*15 min, EVR=25 L/min-m2)
at 22°C and 32.5°C
HNS
14M and 2F
(20-36 yrs)
PF: J. FVC* and j FEV1* compared to FA; no significant effect of
temperature or Os-temperature interaction
IF: Significant decrease in PAI-1 and plasminogen levels 24 hr
postexposure at 22°C, but a significant increase in these coagulation
markers 24 hr postexposure at 32.5°C
Kahle et al., 2015;
2020 ISA, p. 4-28, Table 3-4
0.30
2 hr IE (4*15 min, EVR=25 L/min-m2)
H
14M artd5F
(18-35 yrs)
PF: { FVC* and j, FEV1*
IF: Significant relationship between FEV1 and plasma ferritin (larger
FEV1 decrements in subjects with lower baseline plasma ferritin)
Ghioet al., 2014;
2020 ISA, p. 3-15, Table 3-4
3A-14
-------
03a
(ppiTl)
Exposure and Ventilation
Characteristics During Exercisew
Subject
Characteristics6
Reported Effects on Pulmonary Function (PF), Airway
Resistance and/or Responsiveness (AR), Respiratory
Symptoms (SY), Inflammation (IF)E and Host Defense (HD)
Reference
AQCD/ISA
Popc
nD
n.-n
r fir IF i,4- !5 rnin, EvR-2ri Lmiin-m")
M
20M end 3F
if' viyi;-'!
IF; Ciynifiesnl increase in CRF. 11-1, rit! IL-C bui ii'lli iF-a
'-iijnifirsnt denwe in PAl ) iininedir.ioly-nd "71 In po^pooi'p:
ri>"ijL"jlon]i",? c'lir'lvri^ of EALF Mnpl-^ concluded fer 1 In
refleifo:l o>idr,ike aid c-i 7,4 hr lefevied
iic sue- lepaif
!>,v|in 2ft 12;
Cheng et ,i_, 2f,l7>;
20 Is*-1,, p. 4-2G.
0.30
2 hr IE (4*15 min at either Ve-30
L/min, Ve=50 L/min or Ve=70 L/min)
H
30 M
(19-26 yrs)
PF: I FEVi* and J, FVC* at all ventilation rates; J, MW* only at
highest Ve. Note: additional exposure at 0.50 ppb resulted in J, FEVi*,
1 FVC*, i MW*, i IC*, and J, TLC* at all ventilation rates.
Folinsbeeetal., 1978;
1996 AQCD p. 7-9
0.30
2.5 hr IE (4*15 min, Ve=65 L/min)
H
20M
(18-30 yrs)
PF: | FVC*, | FEVi*, 1 FEF25-75* and |VT*; and t fR*
AR: t sRaw*
SY: t respiratory symptoms*
McDonnell etal., 1983;
1996 AQCD, p. 7-164, Table
7-1
0.30
2.5 hr IE (4x15 min, EVR=25L/min-
m2)
H
30 M and 30 F
(18-35 yrs)
PF: I FEVi* compared with FA
AR: | sRaw* compared with FA
SY: t respiratory symptoms* compared with FA
Seal etal., 1993; 1996
AQCD, p. 7-164, Table 7-1
o.?n
" fin IE > I- ir. min, p-'Prlf' Lj'min-m-)
2 consecutive tf.v-
H
11M end 4F
'2? 3r» vr >
PF; 2 ronsycuiive divs f'/pc^up leaked in greoler j FFv -"
ilvsn thedHici^im-ni ininie^fi-.r lie fust d:»vo10; ^-posuiv.
Madden et cl. ?r> 14,
7 itr IF »4--15 bin F7R~:Tr' L/min-m*-s
fci c '.idV£.
H
11f,1c-.ncMF
(2vi'b!
PF'IF, I FEv'.' p05ilit>1y'f.-ifGi.:iedw.-iHi donifiwni decree in rbe
irifl^niniaii-i t'C'"tol<-ii>- IH'4 in iliohlood
•Sfegel 2017:
2--2n|.SA p. 7-]h Trblo ;-4
? hr IF niiii. Fv'P=2r' L'min-nrT-i
%
06M.Tid7.4F
s mean IZ jh-i
Ff'AR: tegninjde of 0„-indi vx! FEv7 -n^o inc re: r-ed with
decr^^inq lv;c>J[ie FEv1;; nd Lri- -1 inhf.M rorlicoifKoi'! trernmenl-
FT1,': re^pon.'e vivs mirel-i-d to m=,ihr20 I7/-\, p. 7-17. p >47,
l.^ble 7-"m
0.32
1 hr CE (mean Ve =57 L/min)
HAt
42 M and 8F
(mean 26 yrs)
PF: | FEVi*
SY: t respiratory symptoms*
Avol etal., 1984;
1996 AQCD, Table 7-1
0.33
2 hr IE (4x15 min, bicycle at 600
kpm/min)
HNS
9M
(mean 27 yrs)
PF: i FVC*; post FA, normal gradient in ventilation which increased
from apex to the base of the lung; post-03, ventilation shifted away
from the lower-lung into middle and upper-lung regions; post-03
increase in ventilation to mid-lung region correlated with decrease in
midmaximal expiratory flow (r = 0.76, p < 0.05).
Foster etal., 1993;
2006 AQCD, Table AX6-1
0.35
50 min CE (Ve=60 L/min)
repeat exposures over 4 days
HNS
8M
(19-26 yrs)
(some known
03-sensitive)
PF: i FVC*, i FEVi*, j FEF25-75* and |Vt* compared to FA on days
1-4; largest |FEVi* on day 2; j exercise performance time* on day 1
significantly less after the 4th day; | fR*, and J, V02max* on day 1,
recovered by day 4.
Foxcroft and Adams, 1986;
2006 AQCD, Tables AX6-9,
AX6-10
0.35
1 hr CE (VE=80L/min) or 1 hr
competitive simulation (30 min at
Ve=52 L/min, 30 at min Ve=100
L/min; overall mean Ve =77.5 L/min)
HAt
10M
(19-31 yrs)
PF: I FVC*, I FEVi* and J, FEF25-75* compared to FA with both
protocols; j Vt* and | fR* with CE; reduced exercise time in 3 subjects
who were unable to complete CE and competitive protocols
SY: | respiratory symptoms*
Adams and Schelegle, 1983;
1996 AQCD, Table 7-1
3 A-15
-------
03a
(ppiTl)
Exposure and Ventilation
Characteristics During Exercisew
Subject
Characteristics6
Reported Effects on Pulmonary Function (PF), Airway
Resistance and/or Responsiveness (AR), Respiratory
Symptoms (SY), Inflammation (IF)E and Host Defense (HD)
Reference
AQCD/ISA
Popc
nD
0.35
1 hr CE (mean Ve=60 L/min)
Pretreatment: no drug, placebo, or
indomethacin
H
14M
(18-34 yrs)
PF: i FVC* and J, FEVi*; indomethacin significantly attenuated
decreases in FVC and FEVi compared to no drug and placebo;
AR: | sRaw* not affected by indomethacin
Schelegle etal., 1987;
1996 AQCD, Table 7-1
0.35/
0.20
1 hr CE (mean Ve=60 L/min);
2 exposures 24 hr apart
HNS
15M
(mean 25 yrs)
PF: i FVC*, i FEVi* responses on each day compared to FA with an
increased response to 0.20 ppm on the second day
SY: Consecutive exposures produced similar t respiratory symptoms*
Brookes etal., 1989;
2006 AQCD. Table AX6-9
0.35
1 hr CE (mean Ve=60 L/min);
2 exposures 24 hr apart
HNS
15M
(mean 25 yrs)
PF: Significant J, FVC*, J, FEVi* responses on each day compared to
FA with an increased response to 0.35 ppm on the second day
SY: Signifciant symptom responses were worse after second day of
exposure to 0.35 ppm
Brookes etal., 1989;
2006 AQCD. Table AX6-9
0.35
1 hr CE (Ve=60 L/min);
two exposures for each subject
separated by 24, 48, 72, or 120 hr
HNS
40M, 4 groups
of 10
(19-35 yrs)
PF/AR: i FVC*, j FEVi*, j FEF25-75* and | sRaw* for all exposures.
Enhanced FEVi* response after 24 hr repeat exposure and a trend
toward an enhanced response at 48 hr. No differences between
responses to exposures separated by 72 or 120 hr. Similar trends
observed for sRaw.
Schonfeld etal., 1989;
2006 AQCD. Table AX6-9
0.35
70 min IE (Ve=40 L/min)
HNS
18F
(19-28 yrs)
PF: | FVC*, | FEVi*, j FEF25-75* and j MW* immediately
postexposure.
AR: t sRaw* at 1 hr and 18 hr postexposure.
Folinsbee and Hazucha,
1989; 2006 AQCD, Table
AX6-11
0.35
1.25 hr IE (2x30 min, VE=40 L/min)
H
19F
(mean 22 yrs)
PF: I FVC*, I FEVi* and J, FEF25-75* 1 hr postexposure; Persistence
of small effects on both inspired and expired spirometry past 18 hr.
AR: t sRaw* 1 hr and 18 hr postexposure but not 42 hr postexposure.
Folinsbee and Hazucha,
2000;
2006 AQCD, Table AX6-6
0.35
2.2 hr IE (2 x 30 min, Ve=50 L/min;
final 10 min rest)
HNS
15M
(mean 25 yrs)
PF: i FVC* and j FEVi*; pronounced slow phase in multi-breath
nitrogen washouts post O3 exposure; washout delays not related to
changes in ventilatory pattern or lung volume at FRC.
Foster etal., 1997;
2006 AQCD, Table AX6-1
0.37
2 hr IE (Ve=2.5 x rest)
H
20M and 8F
(19-29 yrs)
PF: I FEF25* and J, FEF50* compared to FA
Note: additional exposure at 0.50 and 0.75 ppb resulted in J, FVC*, J,
FEVi*, J, FEF25* and 1 FEF50* compared to FA
Silverman etal., 1976;
1996 AQCD, Table 7-1
0.40
1 hr IE (2.x 15 min, VE=27 L/min)
AsM
6M and 6F
(19-40 yrs)
PF: i FEVi* but no exacerbation of exercise-induced asthma in a
postexposure exercise challenge
SY: Significant increase in respiratory symptoms regardless of
exercise induced asthma status (7 subjects)
Weymeretal., 1994;
2006 AQCD, Table AX6-11
0.40
1 hr CE (EVR=20 L/min-m?)
H
22M
(18-35 yrs)
PF: | FVC*, | FEVi*, j FEWFVC*, and j FEF25-75; half-width of an
expired aerosol bolus was significantly increased, suggesting an O3-
induced change in small airway function.
Keefeetal., 1991;
1996 AQCD, Table 7-1
0.40
1 hr CE (EVR=20 L/min-m2)
H
20M
(18-35 yrs)
PF: 25% i Vt and 9% J, O3 uptake efficiency in the lower respiratory
tract
Gerrity et al., 1994;
1996 AQCD, Table 7-1
3 A-16
-------
03a
(ppiTl)
Exposure and Ventilation
Characteristics During Exercisew
Subject
Characteristics6
Reported Effects on Pulmonary Function (PF), Airway
Resistance and/or Responsiveness (AR), Respiratory
Symptoms (SY), Inflammation (IF)E and Host Defense (HD)
Reference
AQCD/ISA
Popc
nD
0.40
1 hr CE (EVR=30 L/min-m2)
HNS
4 subjects
(sex and age
not indicated)
IF: Apoptotic cells in BAL fluid 6 hr postexposure
HD: Alveolar macrophages from BAL fluid showed the presence of 4-
HNE, protein adduct, 72-kD heat shock protein and ferritin.
Hamilton etal., 1998;
2006 AQCD, Table AX6-12
0.40
2 hr IE (4*15 min, cycle ergometry:
100Wfor M and 83Wfor F)
HNS
7Mand3F
(23-41 yrs)
AR: Increase in airway responsiveness to methacholine challenge
IF: Increase in percentage of PMN and PGF2a; increased TBX2, and
PGE2 concentrations in BAL fluid 3 hr postexposure vs FA
Seltzer etal., 1986;
1996 AQCD, Tables 7-1,7-
11
0.40
2 hr IE (4x15 min, VE=30 L/min)
3 day indomethacin pretreatment
|—| NAs
AsM
5M and 4F
6Mand7F
(18-28 yrs)
PF: i FVC* and J, FEV1* in both groups; significant reductions in mid-
flows in both groups but were greater in AsM vs. HNAs subjects;
indomethacin pretreatment attenuated j FVC* and j FEV1*
responses to O3 in HNAs but not AsM subjects.
Alexis et al., 2000;
2006 AQCD, Table AX6-1,
AX6-13
0.40
2 hr IE (4x15 min, VE=30-40 L/min)
|-|GSTM+
|—|GSTM-
6M and13F
9Mand7F
(mean 24 yrs)
PF: i FVC* and J, FEV1* from baseline across groups; no difference
in lung function response between groups
IF: | PMN* and increased expression of HLA-DR on airway
macrophages and dendritic cells in GSTM1- subjects 24 hr
postexposure; decreased macrophages in GSTM1-sufficient subjects
4-24 hr postexposure. Note: no FA control
Alexis et al., 2009;
2013 ISA, p. 6-80, p. 6-125
0.40
2 hr IE (4x15 min, VE=30-40 L/min)
HNS
4M and 5F
(21-30 yrs)
IF/HD: Significant increase in sputum neutrophils; activation of
monocytes and upregulation of cell surface molecules associated with
antigen presentation (HLA-DR and CD86)
Lay et al., 2007;
2013 ISA, p. 5-44
0.40
: (4*15 min, Ve=30-40 L/min)
||-|NAs
A|NAs
AsA
14M and 20IF
: mean 24 yrs)
?M and 7 IF
[mean 25 yirs)
m and 10F
'mean 24 yirs)
IF/HD: Enhanced inflammatory response in AsA with greater numbers
of neutrophils, higher levels of cytokines (IIIL-6, IL-8, IL-18, and TNF-
cs) and greater macrophage cell-surface expression of T1LJR4 and IgE
receptors in induced sputum compared with HNAs; increase
hyalluironan in AIINAs and AsA compared with HNAs
Note: no FA control
Hernandez etal., 2010;
Hernandez etal., 2012;
2013 ISA, p. 6-130, p. 8-13:
2020 ISA, p. 3-29 p. 3-52,
Table 3-20
0.40
2 hr IE (4x15 min, VE=40 L/min);
Mouthpiece exposure
H
5M and 5F
(mean 30 yrs)
IF: Significant increase in PMNs and decrease in macrophages in
sputum 4 hr postexposure; IL-6, IL-8, and myeleperoxidase increased;
possible relationship of IL-8 and PMN levels.
Fahy etal., 1995;
2006 AQCD, Table AX6-12
0.40
2 hr IE (4x15 min, EVR=18 L/min-m2)
Postexposure, HWR treated with
naxloxone or saline and HSR treated
with sufentanil or saline
|—| WR
HSR
7M and13F
21M and 21F
(20-59 yrs)
PF/SY: i spirometric lung function* across groups, young adults (<35
yrs) significantly more responsive that older individuals (>35 yrs).
Sufentanil, a narcotic analgesic, largely abolished symptom
responses and improved FEV1 in strong responders. Naloxone, an
opioid antagonist, did not affect O3 effects in weak responders.
Passannanteetal., 1998;
2006 AQCD, Table AX6-13
0.40
2 hr IE (4x15 min, EVR=20 L/min-m2)
HNAs
AsA
5M and 1F
(mean 29 yrs)
6M
(mean 24 yrs)
PF: Similar j FEV1* in both groups
AR: Maximal FEV1 response to methacholine increased similarly in
both groups 12 hr postexposure
IF: Significant increase in PMN in both groups
Hiltermann etal., 1995;
2006 AQCD, Table AX6-3
3 A-17
-------
03a
(ppiTl)
Exposure and Ventilation
Characteristics During Exercisew
Subject
Characteristics6
Reported Effects on Pulmonary Function (PF), Airway
Resistance and/or Responsiveness (AR), Respiratory
Symptoms (SY), Inflammation (IF)E and Host Defense (HD)
Reference
AQCD/ISA
Popc
nD
0.40
2 hr IE (4x15 min, EVR=20 L/min-m2)
AsM
1Mand5F
(18-27 yrs)
PF: | FEVi*
AR: Increased airway responsiveness to methacholine 16 hr
postexposure; no effect of proteinase inhibitor (rALP)
Hiltermann etal., 1998;
2006 AQCD, Table AX6-12
0.40
2 hr IE (4x15 min, EVR=20 L/min-m2)
As
10M and 6F
(19-35 yrs)
IF: Levels of eosinophil cationic protein, IL-8 and percentage
eosinophils highly correlated in sputum and BAL 16 hr postexposure.
Hiltermann etal., 1999;
2006 AQCD, Table AX6-12
0.40
2 hr IE (4x15 min, EVR=20 L/min-m2)
Apocynin or placebo
AsM
1Mand6F
(19-26 yrs)
AR/IF: Increased bronchial responsiveness to methacholine 16 hr
postexposure; inhaled apocynin (an inhibitor of NADPH oxidase
present in inflammatory cells) treatment significantly reduced O3-
induced airway responsiveness
Peters etal., 2001;
2006 AQCD, Table AX6-11,
0.40
2 hr IE (4x15 min, EVR=20 L/min-m2)
HNS
HNS
Placebo: 15M
and 1F
Antioxidant:
13M and2F
(mean 27 yrs)
AR: i FVC*, and j FEV1* in both groups
IF: no difference in PMNs and IL-6 levels in BAL fluid 1 hr
postexposure between treatment groups.
Sametetal., 2001; Steck-
Scottet al., 2004;
2006 AQCD, Tables AX6-1,
AX6-13
0.40
i (4*15 min, VE=25 IL/inin)
IHwt
Ob
19F
19F
(18-35 yirs)
PF: IIFVC* and J, IFIEV1* in both groups; J. IFVC* was greater in obese
women than in normal-weight women.
AR/IF: Increase in airway responsiveness oir increase in PIVIN after O3
exposure did not differ between normal-weight and obese women.
SY: Symptoms in response to exposure did not differ between groups
Bennett etal., 2016;
2020 ISA. p. 3-57. p. 3-59,
Tables 3-4, 3-8, 3-9, 3-31
0.40
2 hr IE (4x20 min of mild-moderate
exercise)
2 wk pretreatment with budesonide or
placebo
|—|NAs
6M and 9F
(mean 31 yrs)
PF:j FVC* and j FEV1* immediately postexposure; FVC and FEV1
decrements recovered 4 hr postexposure;
AR: Small increased bronchial reactivity to methacholine
IF: Increased PMNs and myeloperoxidase in 4 hr postexposure
sputum; no protection from inhaled corticosteroid, budesonide.
Nightingale et al., 2000;
2006 AQCD, Table AX6-13
0.40
2 hr IE (4x20 min, 50W cycle
ergometry, 10 min rest)
2 wk pretreatment with budesonide or
placebo
HNS
4M and 5F
(mean 30 yrs)
PF: Placebo-control: Immediately postexposure significant J, FVC and
FEV1 relative to pre-exposure values; 3 hr postexposure FVC and
FEV1 recovered to preexposure values.
IF: Significant increases in 8-isoprostane at 4 hr postexposure;
Budesonide for 2 wk prior to exposure did not affect responses.
Montuschi et al., 2002;
2006 AQCD, Table AX6-1
0.40
: (4*15 min, Ve=50-75 L/niin)
|-|NAs
AIN/te
AsA
5M and 8F
4M and 11F
3M and 8IF
(21-35 yrs)
PIF/IIIF: IFIEV1 responses to O3 not differentiated by asthma; pirecent
predicted IFIEV1 both before and after O3 exposure did not differ
between inflammatory irespondeirs (>10% increase in PIVIN) and
nonirespondeirs
Fry etal., 2012;
2020 ISA, p. 3-29, p. 3-36,
Table 3-17
0.40
2 hr IE (4x15 min, Ve=50-75 L/min)
Pretreatment: saline or atropine
HNS
8M
(18-27yrs)
PF: | FVC*, | FEV1*, |VT*, and j TLC*; and | fe*. Atropine
pretreatment attenuated FEV1 and FEF25-75 response.
AR: t sRaw*; Atropine pretreatment abolished increase in sRaw
Beckett etal., 1985;
1996 AQCD, Table 7-1
3 A-18
-------
03a
(ppiTl)
Exposure and Ventilation
Characteristics During Exercisew
Subject
Characteristics6
Reported Effects on Pulmonary Function (PF), Airway
Resistance and/or Responsiveness (AR), Respiratory
Symptoms (SY), Inflammation (IF)E and Host Defense (HD)
Reference
AQCD/ISA
Popc
nD
0.40
2 hr IE (4*15 min, Ve=53-55 L/min)
|—| NAs
AsM
4M and 5F
4M and 5F
(18-34 yrs)
PF: i FVC*, i FEVi*, and J, FEF25-75 in both groups with a significantly
greater percent j in As compared to HNAs subjects
AR: t sRaw* in As; airway responsiveness (methacholine challenge)
was not statistically different between HNAs and AsM subjects
Kreitetal., 1989;
2006 AQCD, Table AX6-11
0.40
2 hr IE (4x15 min, EVR=30 L/min-m?)
4 day pretreatment with indomethacin
or placebo
HNS
13M
(18-31 yrs)
PF: Indomethacin pretreatment resulted in a significantly smaller FVC
and FEV1 decrements than with O3 alone
AR: airway hyperresponsiveness was not significantly affected by
indomethacin pretreatment.
Ying etal., 1990;
1996 AQCD, Table 7-1
0.40
2 hr IE (4x15 min, Ve=66 L/min)
HNS
8M
(18-35 yrs)
IF: BAL fluid at 1 hr postexposure vs. 18 hr postexposure. At 1 hr,
PMN's, total protein, LDH, a1-antitrypsin, fibronectin, PGE2,
thromboxane B2, C3a, tissue factor, and clotting factor VII were
increased; IL-6 and PGE2 were higher after 1 hr than 18 hr; fibronectin
and tissue plasminogen activator higher after 18 hr. No time
differences for PMN and protein.
Devlin etal., 1996;
2006 AQCD, Table AX6-12
0.40
2 hr IE (4x15 min, VE=70 L/min);
H
11M
(18-35 yrs)
IF/HD: Macrophages 18 hr postexposure had changes in the rate of
synthesis of 123 different proteins as assayed by computerized
densitometry of two-dimensional gel protein profiles
Devlin and Koren, 1990;
2006 AQCD, Table AX6-12
0.40
2 hr IE (4x15 min, VE=70 L/min);
H
11M
(18-35 yrs)
IF/HD: BAL fluid 18 hr postexposure contained increased levels of the
coagulation factors, tissue factor, and factor VII; macrophages in the
BAL fluid had elevated tissue factor mRNA
McGeeetal., 1990;
2006 AQCD, Table AX6-12
0.40
2 hr IE (4x15 min, VE=70 L/min);
H
11M
(18-35 yrs)
IF: NL done immediately before, immediately after, and 22 hr after
exposure; increased PMNs at both postexposure times; increased
levels of tryptase (marker of mast cell degranulation) immediately
postexposure; increased levels of albumin 22 hr postexposure.
Graham and Koren, 1990;
Koren etal., 1990;
2006 AQCD, Table AX6-12
0.40
2 hr IE (4x15 min, EVR=35 L/min-m?)
H
11M
(18-35 yrs)
PF/IF: Significant increase in PMNs, total protein, albumin, IgG, PGE2,
plasminogen activator, neutrophil elastase complement C3a, and
fibronectin; no correlation between pulmonary function and
inflammatory endpoints in BAL fluid 18 hr postexposure
HD: decrease in percentage of macrophages compared to FA
Koren etal., 1989a;
Koren etal., 1989b;
1996 AQCD, Table 7-1;
2006 AQCD, Table AX6-12
0.40
2 hr IE (4x15 min, EVR=35 L/min-m2)
H
10M
(18-35 yrs)
PF/IF: Increased PMN, protein, PGE2, LDH, TXB2, IL-6 a-1 anti-
trypsin, and tissue factor in BAL fluid 1 hr postexposure compared to
18 hr; fibronectin and urokinase-type plasminogen activator higher 18
hr postexposure than 1 hr
HD: Decreased phagocytosis of yeast by alveolar macrophages.
Koren etal., 1991;
1996 AQCD, Table 7-1;
2006 AQCD, Table AX6-12
3 A-19
-------
03a
(ppiTl)
Exposure and Ventilation
Characteristics During Exercisew
Subject
Characteristics6
Reported Effects on Pulmonary Function (PF), Airway
Resistance and/or Responsiveness (AR), Respiratory
Symptoms (SY), Inflammation (IF)E and Host Defense (HD)
Reference
AQCD/ISA
Popc
nD
0.40
2 hr IE (4x15 min, EVR=35 L/min-m2)
H
8M
(20-30 yrs)
PF: | FVC*
AR: t SRaw*
IF: Significantly increased clearance of 99mTc-DTPA from the lung
indicating epithelial damage, and changes in permeability.
Kehrl etal., 1987;
2006 AQCD, Table AX6-13
0.40
2.5 hr IE (4x15 min, EVR=25L/min-
m2)
H
30 M and 30F
(18-35 yrs)
PF: I FEVi* compared with FA
AR: | sRaw* compared with FA
SY: t Respiratory symptoms* compared with FA
Seal etal., 1993; 1996
AQCD, p. 7-164, Table 7-1
0.40
2.5 hr IE (4*15 min at Ve=65 L/min)
H
29M
(18-30 yrs)
PF: | FVC*, | FEVi*, j FEF25-75*, |VT* and | f*
AR: t sRaw*
SY: t Respiratory symptoms*
McDonnell etal., 1983;
1996 AQCD, p. 7-164 Table
7-1
0.40
2 Hr IE (4x15 min, 2 x resting Ve)
2 Hr IE (4x15 min, 2 x resting Ve) x 3
days
HNS
12M and 7F
(21-32 yrs)
AR: Significant increase in histamine airway responsiveness with
progressive adaptation of the effect; after day 3 histmine
responsiveness was not different from sham exposures
Dimeo etal., 1981;
2006 AQCD, Table AX6-11
0.40
IE (2x15 min, Ve=40 L/min-m2)
2 h/day for 5 days,
2 h either 10 or 20 days later
HNS
16M
(18-35 yrs)
PF: i FEVi* at each time point; FEVi decrement was greatest on day
2 and was significantly attenuated by days 4 and 5.
IF: BAL immediately after day 5 of exposure and again after exposure
10 or 20 days later. Most markers of inflammation (PMNs, IL-6, PGE2,
fibronectin) showed complete attenuation; markers of damage (LDH,
IL-8, protein, 1-antitrypsin, elastase) did not. Reversal of attenuation
was not complete for some markers, even after 20 days.
Devlin etal., 1997;
2006 AQCD, Tables AX6-9,
6-12
0.40
3 hr/day (2 hr resting followed by 1 hr
CE at 4-5 resting Ve ) for 5 days,
HNS
13M and 11F
(19-46 yrs)
AR: Enhanced airway response to methacholine after the first 3 days
which normalized by day 5
Kulle etal., 1982;
1996 AQCD, Table 7-10
0.40
3 hr/day for 5 days: IE (6x15 min
mild-moderate exercise, Ve=32
L/min)
AsM
8M and 2F
(mean 31 yrs)
PF/SY: Significant j FEVi and increase in symptom response on O3
exposure days 1 and 2 that diminished with continued exposure;
tolerance partially lost 4 and 7 days postexposure
AR: bronchial reactivity to methacholine peaked after O3 exposure on
day 1, but remained elevated with continued exposure
Gong etal., 1997;
2006 AQCD, Table AX6-11
Children During Moderate Exercise
0.12
2.5 hr IE (4x15 min, EVR=35 L/min-
m2)
H
23 M
(8-11 yrs)
PF: i FEVi* compared with clean air which persisted for 16-20 hr
SY: No significant increase in severity of respiratory symptoms
McDonnell etal. (1985);
2006 AQCD
Adult Subjects at Rest
0.10
2 hr
H
10M
(18-28 yrs)
PF: No significant change in pulmonary function
Folinsbeeetal., 1978°
3A-20
-------
03a
(ppiTl)
Exposure and Ventilation
Characteristics During Exercisew
Subject
Characteristics6
Reported Effects on Pulmonary Function (PF), Airway
Resistance and/or Responsiveness (AR), Respiratory
Symptoms (SY), Inflammation (IF)E and Host Defense (HD)
Reference
AQCD/ISA
Popc
nD
0.10
2 hr
HNS
13M and 1F
(mean 24 yrs)
AR: No increased airway responsiveness to methacholine
immediately after exposure.
Konig etal., 1980;
1996 AQCD, Table 7-10
0.12
1 hr
Ai r-antigen/03-antigen
AsA
4M and 3F
(21-64 yrs)
PF: No change in baseline pulmonary function.
AR: Increased allergen-specific airway responsiveness to inhaled
ragweed or grass after O3 exposure compared to FA
Molfinoetal., 1991;
1996 AQCD, Tables 7-2, 7-
10
0.12
1 hr
AsA
10M and 5F
(19-34 yrs)
PF: No significant change in pulmonary function to O3 alone.
AR: No significant change in sRaw to O3 alone; no significant effect
on airway responsivness to grass allergen
Ball etal., 1996
2006 AQCD, Table AX6-11
0.12
1 hr
(Air-Antigen)
AsA
9M and6F
(18-49 yrs)
AR: No effect of O3 on airway responsiveness to grass or ragweed
allergen.
Hananiaetal., 1998;
2006 AQCD, Table AX6-11
0.12
1 hr O3 at rest followed by 6 min
maximal exercise
AsA
7M and 8F
(19-45 yrs)
PF: No significant change in FEV1
AR: O3 pre-exposure did not affect the magnitude or time course of
exercise-induced bronchoconstriction.
Fernandes etal., 1994
1996 AQCD, Table 7-2
0.20
2 hr
HNS
15 subjects
IF/HD: Increased numbers of CD3+, CD4+, and CD8+ T lymphocyte
subsets, in addition to neutrophils, in BAL fluid 6 hr postexposure.
Blomberg etal., 1997;
2006 AQCD, Table AX6-12
0.25
2 hr
H
8M and 5F
(21-22 yrs)
PF: No significant change in FVC compared with FA
Horvath etal., 1979;
1996 AQCD, Table 7-1
0.30
2 hr
H
10M
(18-28 yrs)
PF: No significant change in pulmonary function
Folinsbeeetal., 1978°
n 'in
2 hr
I s
n mr-
0.32
2 hr
HNS
13M and 1F
(mean 24 yrs)
AR: Increased airway responsiveness to methacholine immediately
after exposure.
Konig etal., 1980;
1996 AQCD, Table 7-10
0.37
2 hr
H
20M and 8F
(19-29 yrs)
PF: No significant change in FEV1, FEF25, and FEF50 compared with
FA
Silverman etal., 1976;
1996 AQCD, Table 7-1
0.40
2 hr
AsA
12 subjects
(18-35 yrs)
IF: Release of early-onset mast cell-derived mediators into NL in
response to allergen not enhanced after O3 exposure. No increase in
neutrophil and eosinophil inflammatory mediators after O3 exposure or
enhancement after allergen challenge. O3 increased eosinophil influx
following allergen exposure.
Michelson etal., 1999
2006 AQCD, Table AX6-12
0.40
2 hr
AsM
10 subjects
(18-35 yrs)
IF: Increased response to allergen; significant increase in PMN and
eosinophils after O3 plus allergen challenge; O3 alone increased nasal
inflammation (PMN).
Peden etal., 1995;
2006 AQCD, Table AX6-12
3A-21
-------
03a
(ppiTl)
Exposure and Ventilation
Characteristics During Exercisew
Subject
Characteristics6
Reported Effects on Pulmonary Function (PF), Airway
Resistance and/or Responsiveness (AR), Respiratory
Symptoms (SY), Inflammation (IF)E and Host Defense (HD)
Reference
AQCD/ISA
Popc
nD
0.4m
Zhr • 2 -It; - dui in? .-nd nyi -j .j-rc,
j. r'll'jn
Hi
C.U rTi<;! T
ni>'nn yrM
IF. Ciyiifie.-nl incr^rr- in nrc.,:l inucip. I' >l-I protein rlbuiiiin, PWf b,
eosinophil" j. duray polled ;ear.on. buun
i[.irc,r,n ?y,,rf?Rij the iHhiriBfoi. cnmei [> coniluc^d
[Xl-pUc,(j M^fNcol i-.-iS ,101 peifclffiod IllO
D'-Hc riii'lTpjl-e oko-C'oHc.
0.50
2 hr
H
10M
(18-28 yrs)
PF: | FEV1*, | FVC* but no change in MW
Folinsbee et al., 1978;
1996 AQCD, Table 7-1
0.50
2 hr
H
8M and 5F
(21-22 yrs)
PF: i FVC* compared with FA
Horvath et al., 1979;
1996 AQCD, Table 7-1
0.50
2 hr
H
20M and 8F
(19-29 yrs)
PF: No significant change in FEV1, FEF25, and FEF50 compared with
FA
Silverman et al., 1976;
1996 AQCD, Table 7-1
0.60
2 hr
HNS
5M and 3F
(22-30 yrs)
AR: 300% increase in histamine-induced ARaw 5 min after O3
exposure; 84 and 50% increases 24 hr and 1 week after exposure (p
> 0.05), respectively. Two subjects had an increased response to
histamine 1 week after exposure.
Golden et al., 1978;
1996 AQCD, Table 7-10;
0.75
2 hr
H
8M and 5F
(21-22 yrs)
PF: i FVC* compared with FA
Horvath et al., 1979;
1996 AQCD, Table 7-1
0.75
2hr
H
20M and 8F
(19-29 yrs)
PF: i FEV1*, i FEF25*, and J, FEF50* compared with FA
Silverman et al., 1976;
1996 AQCD, Table 7-1
1.00
2 hr
HNS
13M and 1F
(mean 24 yrs)
AR: Increased airway responsiveness to methacholine immediately
after exposure.
Konig et al., 1980
1996 AQCD, Table 7-10
Note: Newly added studies for this review are in blue font.
A Reported target mean O3 concentrations
w Focused on O3 exposures below 0.4 ppm during exercise and below 1.00 ppm at rest
B Subject Characteristics are subdivided into subject population (Pop) and number (n) subjects.
c Subject population included: healthy subjects (H), athletes included competitive endurance cyclists and runners (HAt), nonsmokers (HNS), nonasthmatics (HNAs), nonasthmatics
with allergies (AINAs), asthmatics (As), mild asthmatics (AsM), S02-sensitive asthmatics (As302), asthmatics with allergies (AsA), subjects with allergies (Al), smokers (S), healthy
subjects with the GSTM1 genotype (HGSTM+) or null for the GSTM1 genotype (HGS™-), healthy subjects that have a weak O3 response (HWR) or have a strong O3 response
(HSR), healthy weight subjects (Hwt) and obese subjects (Ob).
D Number is further characterized by sex, male (M) and female (F), and age range or mean age of the subjects.
E For the purposes of this table the "IF" category includes reported effects on inflammation (the most commonly tested endpoint) as well as injury and oxidative stress
responses because injury, inflammation, and oxidative stress responses are difficult to disentangle. Inflammation generally occurs as a consequence of injury and oxidative
stress, but it can also lead to further oxidative stress and injury due to secondary production of reactive oxygen species (ROS) by inflammatory cells (2020 ISA section 3.1.3).
* Indicates statistical significance
F Avol et al., 1984 reported 03-induced effects for 0.08, 0.16, 0.24 and 0.32 ppm but only effects from 0.16, 0.24 and 0.32 ppm was referenced in 1996 AQCD, Table 7-1.
3A-22
-------
G Folinsbeeetal., 1978 reported data for subjects exposed to O3 during exercise at 0.1 ppm and 0.3 ppm at 3 different ventilation rates and at rest at 0.1 ppm, 0.3 ppm and
0.5 ppm. Only the 0.5 ppm O3 exposure to subjects at rest was referenced in 1996 AQCD, Table 7-1.
J Subtracted from FA, the group mean decrement in FEV1 was 9.7% (2006 AQCD and 2013 ISA).
Abbreviations: BAL, bronchoalveolar lavage; C3a, complement protein fragment; CC16, protein secreted by Clara cells in the non-ciliated respiratory epithelium; CD86,
surface costimulatory marker for T-cell activation; CE, continuous exercise; CRP, C-reactive protein; ENA-78, epithelial cell-derived neutrophil-activating peptide; FA, filtered
air; FEF25, (formerly designated as V25%vc) instantaneous forced expiratory flow after 25% of forced vital capacity; FEF25-75, forced expiratory flow over the middle half of forced
vital capacity; FEF50, (formerly designated as Vso%vc) instantaneous forced expiratory flow after 50% of forced vital capacity; FEV1, forced expiratory volume in one second;
fR, respiratory frequency (also abbreviated as f); FRC, functional reserve capacity; FVC, forced vital capacity; GM-CSF, granulocyte-macrophage colony-stimulating factor;
GSTM1, glutathione S-transferase M1 polymorphism; HLA-DR, human leukocyte antigens; 4-HNE, 4-hydroxynonenal; IC, inspiratory capacity; IE, intermittent exercise;; IgE,
immunoglobulin E; IgG, immunoglobulin G antibody; IL-6, IFN-y, interferon-gamma; IL-1, interleukin 1 pro-inflammatory cytokine interleukin 6 pro-inflammatory cytokine; IL-8,
interleukin 8 pro-inflammatory cytokine; IL-18, interleukin 18 pro-inflammatory cytokine; ISA, Integrated Science Assessment; LDH, lactate dehydrogenase; LTB4, leukotriene;
MMP9, metallopeptidase 9; MW, maximal voluntary ventilation; NQ01, NAD(P)H:quinoneoxidoreductase; NL, nasal lavage; 8-OHdG, 8-hydroxy-2'-deoxyguanosine; PAI-1,
plasminogen activator fibrinogen inhibitor-1; PGE2, prostaglandin E2 a mediator of inflammation; PGE2, bronchodilatory prostaglandin; PGF2a, prostaglandin 2 alpha; PMN,
polymorphonuclear neutrophils; ROS, reactive oxygen species; sGaw, specific airway conductance; sGaw, specific airway conductance; sRaw, specific airway resistance;
substance P, neuropeptide that act as a neurotransmitter and neuromodulator; TBARS, Thiobarbituric acid reactive substance, 99mTc-DTPA, radiolabled diethylene triamine
pentaacetic acid; TBX2, thromboxane B2; TLC, total lung capacity; TLR4, Toll-like receptor protein 4; TNF-a, tumor necrosis factor alpha; tPA, tissue plasminogen activator;
VC, vital capacity; VEmax, maximal expiratory volume; V02max, maximum rate of oxygen consumption during exercise; Vt, tidal volume; Vimax, peak tidal volume during
exercise; W, watts; 8-epi-PGF2a, prostaglandin 2 alpha; 99mTc-DTPA, technetium 99m-labelled diethylenetriamine penta-acetic acid used aerosol ventilation studies
3A-23
-------
REFERENCES
Adams, WC (2000). Ozone dose-response effects of varied equivalent minute ventilation rates. J
Expo Anal Environ Epidemiol 10(3): 217-226.
Adams, WC (2002). Comparison of chamber and face-mask 6.6-hour exposures to ozone on
pulmonary function and symptoms responses. Inhal Toxicol 14(7): 745-764.
Adams, WC (2003). Comparison of chamber and face mask 6.6-hour exposure to 0.08 ppm
ozone via square-wave and triangular profiles on pulmonary responses. Inhal Toxicol
15(3): 265-281.
Adams, WC (2006a). Human pulmonary responses with 30-minute time intervals of exercise and
rest when exposed for 8 hours to 0.12 ppm ozone via square-wave and acute triangular
profiles. Inhal Toxicol 18(6): 413-422.
Adams, WC (2006b). Comparison of chamber 6.6-h exposures to 0.04-0.08 PPM ozone via
square-wave and triangular profiles on pulmonary responses. Inhal Toxicol 18(2): 127-
136.
Adams, WC and Ollison, WM (1997). Effects of prolonged simulated ambient ozone dosing
patterns on human pulmonary function and symptomatology Air & Waste Management
Association Pittsburgh, PA.
Adams, WC, Savin, WM and Christo, AE (1981). Detection of ozone toxicity during continuous
exercise via the effective dose concept. J Appl Physiol (1985) 51(2): 415-422.
Adams, WC and Schelegle, ES (1983). Ozone and high ventilation effects on pulmonary
function and endurance performance. J Appl Physiol (1985) 55(3): 805-812.
Alexis, N, Urch, B, Tarlo, S, Corey, P, Pengelly, D, O'Byrne, P and Silverman, F (2000).
Cyclooxygenase metabolites play a different role in ozone-induced pulmonary function
decline in asthmatics compared to normals. Inhal Toxicol 12(12): 1205-1224.
Alexis, NE, Zhou, H, Lay, JC, Harris, B, Hernandez, ML, Lu, TS, Bromberg, PA, Diaz-Sanchez,
D, Devlin, RB, Kleeberger, SR and Peden, DB (2009). The glutathione-S-transferase Mu
1 null genotype modulates ozone-induced airway inflammation in human subjects. J
Allergy Clin Immunol 124(6): 1222-1228.
Arjomandi, M, Balmes, JR, Frampton, MW, Bromberg, P, Rich, DQ, Stark, P, Alexis, NE,
Costantini, M, Hollenbeck-Pringle, D, Dagincourt, N and Hazucha, MJ (2018).
Respiratory responses to ozone exposure: the multicenter ozone study in older subjects
(moses). Am J Respir Crit Care Med 197: 1319-1327.
Avol, EL, Linn, WS, Venet, TG, Shamoo, DA and Hackney, JD (1984). Comparative respiratory
effects of ozone and ambient oxidant pollution exposure during heavy exercise. J Air
Waste Manage Assoc 34(8): 804-809.
3A-24
-------
Ball, BA, Folinsbee, LJ, Peden, DB and Kehrl, HR (1996). Allergen bronchoprovocation of
patients with mild allergic asthma after ozone exposure. J Allergy Clin Immunol 98(3):
563-572.
Bartoli, ML, Vagaggini, B, Malagrino, L, Bacci, E, Cianchetti, S, Dente, FL, Novelli, F, Costa, F
and Paggiaro, P (2013). Baseline airway inflammation may be a determinant of the
response to ozone exposure in asthmatic patients. Inhal Toxicol 25(3): 127-133.
Bates, ML, Brenza, TM, Ben-Jebria, A, Bascom, R, Eldridge, MW and Ultman, JS (2014).
Pulmonary function responses to ozone in smokers with a limited smoking history.
Toxicol Appl Pharmacol 278(1): 85-90.
Beckett, WS, McDonnell, WF, Horstman, DH and House, DE (1985). Role of the
parasympathetic nervous system in acute lung response to ozone. J Appl Physiol (1985)
59(6): 1879-1885.
Bennett, WD, Ivins, S, Alexis, NE, Wu, J, Bromberg, PA, Brar, SS, Travlos, G and London, SJ
(2016). Effect of obesity on acute ozone-induced changes in airway function, reactivity,
and inflammation in adult females. PLoS ONE 11(8): e0160030.
Biller, H, Holz, O, Windt, H, Koch, W, Miiller, M, Jorres, RA, Krug, N and Hohlfeld, JM
(2011). Breath profiles by electronic nose correlate with systemic markers but not ozone
response. RespirMed 105(9): 1352-1363.
Blomberg, A, Helleday, R, Pourazar, J, Stenfors, N, Kelly, FJ, Frew, AJ, Holgate, ST and
Sandstrom, T (1997). Early airway and peripheral blood cell responses to 020 ppm ozone
in healthy human subjects. Eur Respir J 10: 274S.
Blomberg, A, Mudway, I, Svensson, M, Hagenbjork-Gustafsson, A, Thomasson, L, Helleday, R,
Dumont, X, Forsberg, B, Nordberg, G and Bernard, A (2003). Clara cell protein as a
biomarker for ozone-induced lung injury in humans. Eur Respir J 22(6): 883-888.
Blomberg, A, Mudway, IS, Nordenhall, C, Hedenstrom, H, Kelly, FJ, Frew, AJ, Holgate, ST and
Sandstrom, T (1999). Ozone-induced lung function decrements do not correlate with
early airway inflammatory or antioxidant responses. Eur Respir J 13(6): 1418-1428.
Bosson, J, Stenfors, N, Bucht, A, Helleday, R, Pourazar, J, Holgate, ST, Kelly, FJ, Sandstrom, T,
Wilson, S, Frew, AJ and Blomberg, A (2003). Ozone-induced bronchial epithelial
cytokine expression differs between healthy and asthmatic subjects. Clin Exp Allergy
33(6): 777-782.
Bosson, JA, Blomberg, A, Stenfors, N, Helleday, R, Kelly, FJ, Behndig, AF and Mudway, I
(2013). Peripheral blood neutrophilia as a biomarker of ozone-induced pulmonary
inflammation. PLoS ONE 8(12): e81816.
Brookes, KA, Adams, WC and Schelegle, ES (1989). 035 ppm 03 exposure induces
hyperresponsiveness on 24-h reexposure to 020 ppm 03. J Appl Physiol (1985) 66(6):
2756-2762.
3A-25
-------
Brown, JS, Bateson, TF and McDonnell, WF (2008). Effects of exposure to 0.06 ppm ozone on
FEV1 in humans: a secondary analysis of existing data. Environ Health Perspect 116(8):
1023-1026.
Cheng, W, Duncan, KE, Ghio, AJ, Ward-Caviness, C, Karoly, ED, Diaz-Sanchez, D, Conolly,
RB and Devlin, RB (2018). Changes in metabolites present in lung lining fluid following
exposure of humans to ozone. Toxicol Sci 163(2): 430-439.
Corradi, M, Alinovi, R, Goldoni, M, Vettori, M, Folesani, G, Mozzoni, P, Cavazzini, S,
Bergamaschi, E, Rossi, L and Mutti, A (2002). Biomarkers of oxidative stress after
controlled human exposure to ozone. Toxicol Lett 134(1-3): 219-225.
Devlin, RB, Duncan, KE, Jardim, M, Schmitt, MT, Rappold, AG and Diaz-Sanchez, D (2012).
Controlled exposure of healthy young volunteers to ozone causes cardiovascular effects.
Circulation 126(1): 104-111.
Devlin, RB, Folinsbee, LJ, Biscardi, F, Hatch, G, Becker, S, Madden, MC, Robbins, M and
Koren, HS (1997). Inflammation and cell damage induced by repeated exposure of
humans to ozone. Inhal Toxicol 9(3): 211-235.
Devlin, RB and Koren, HS (1990). The use of quantitative two-dimensional gel electrophoresis
to analyze changes in alveolar macrophage proteins in humans exposed to ozone. Am J
Respir Cell Mol Biol 2(3): 281-288.
Devlin, RB, McDonnell, WF, Becker, S, Madden, MC, McGee, MP, Perez, R, Hatch, G, House,
DE and Koren, HS (1996). Time-dependent changes of inflammatory mediators in the
lungs of humans exposed to 0.4 ppm ozone for 2 hr: A comparison of mediators found in
bronchoalveolar lavage fluid 1 and 18 hr after exposure. Toxicol Appl Pharmacol 138(1):
176-185.
Dimeo, MJ, Glenn, MG, Holtzman, MJ, Sheller, JR, Nadel, JA and Boushey, HA (1981).
Threshold concentration of ozone causing an increase in bronchial reactivity in humans
and adaptation with repeated exposures. Am Rev Respir Dis 124(3): 245-248.
Dokic, D and Trajkovska-Dokic, E (2013). Ozone exaggerates nasal allergic inflammation.
Makedonska Akademija na Naukite i Umetnostite. Oddelenie za Bioloski i Medicinski
Nauki. Prilozi 34(1): 131-141.
Fahy, JV, Wong, HH, Liu, JT and Boushey, HA (1995). Analysis of induced sputum after air and
ozone exposures in healthy subjects. Environ Res 70(2): 77-83.
Fernandes, ALG, Molfino, NA, McClean, PA, Silverman, F, Tarlo, S, Raizenne, M, Slutsky, AS
and Zamel, N (1994). The effect of pre-exposure to 012 ppm of ozone on exercise-
induced asthma. Chest 106(4): 1077-1082.
Folinsbee, LJ, Bedi, JF and Horvath, SM (1984). Pulmonary function changes after 1 h
continuous heavy exercise in 0.21 ppm ozone. J Appl Physiol (1985) 57(4): 984-988.
3A-26
-------
Folinsbee, LJ, Drinkwater, BL, Bedi, JF and Horvath, SM, Eds. (1978). The influence of
exercise on the pulmonary function changes due to exposure to low concentrations of
ozone. Academic Press New York, NY.
Folinsbee, LJ and Hazucha, MJ (1989). Atmospheric ozone research and its policy implications:
Persistence of ozone-induced changes in lung function and airway responsiveness.
Elsevier. Amsterdam, The Netherlands.
Folinsbee, LJ and Hazucha, MJ (2000). Time course of response to ozone exposure in healthy
adult females. Inhal Toxicol 12(3): 151-167.
Folinsbee, LJ, Horstman, DH, Kehrl, HR, Harder, S, Abdul-Salaam, S and Ives, PJ (1994).
Respiratory responses to repeated prolonged exposure to 0.12 ppm ozone. Am J Respir
Crit Care Med 149(1): 98-105.
Folinsbee, LJ, Horstman, DH, Vorona, RD, Prince, JM and Berry, M (1986). Determinants of
endurance performance during ozone inhalation. Vancouver, Canada.
Folinsbee, LJ, McDonnell, WF and Horstman, DH (1988). Pulmonary function and symptom
responses after 6.6-hour exposure to 0.12 ppm ozone with moderate exercise. JAPCA
38(1): 28-35.
Foster, WM, Silver, JA and Groth, ML (1993). Exposure to ozone alters regional function and
particle dosimetry in the human lung. J Appl Physiol (1985) 75(5): 1938-1945.
Foster, WM, Weinmann, GG, Menkes, E and Macri, K (1997). Acute exposure of humans to
ozone impairs small airway function. Ann Occup Hyg 1 (inhaled particles VIII): 659-666.
Foxcroft, WJ and Adams, WC (1986). Effects of ozone exposure on four consecutive days on
work performance and V02max. J Appl Physiol (1985) 61(3): 960-966.
Frampton, MW, Balmes, JR, Bromberg, PA, Stark, P, Arjomandi, M, Hazucha, MJ, Rich, DQ,
Hollenbeck-Pringle, D, Dagincourt, N, Alexis, N, Ganz, P, Zareba, W and Costantini,
MG (2017). Multicenter Ozone Study in oldEr Subjects (MOSES: Part 1. Effects of
exposure to low concentrations of ozone on respiratory and cardiovascular outcomes.
Research Report 192, Pt 1. Health Effects Institute. Boston, MA.
Frampton, MW, Pietropaoli, A, Dentler, M, Chalupa, D, Little, EL, Stewart, J, Frasier, L, Oakes,
D, Wiltshire, J, Vora, R and Utell, MJ (2015). Cardiovascular effects of ozone in healthy
subjects with and without deletion of glutathione-S-transferase Ml. Inhal Toxicol 27(2):
113-119.
Frank, R, Liu, MC, Spannhake, EW, Mlynarek, S, Macri, K and Weinmann, GG (2001).
Repetitive ozone exposure of young adults: Evidence of persistent small airway
dysfunction. Am J Respir Crit Care Med 164(7): 1253-1260.
3A-27
-------
Fry, RC, Rager, JE, Zhou, H, Zou, B, Brickey, JW, Ting, J, Lay, JC, Peden, DB and Alexis, NE
(2012). Individuals with increased inflammatory response to ozone demonstrate muted
signaling of immune cell trafficking pathways. Respir Res 13: 89.
Gerrity, TR, McDonnell, WF and House, DE (1994). The relationship between delivered ozone
dose and functional responses in humans. Toxicol Appl Pharmacol 124(2): 275-283.
Ghio, AJ, Soukup, JM, Dailey, LA, Richards, JH, Duncan, KE and Lehmann, J (2014). Iron
decreases biological effects of ozone exposure. Inhal Toxicol 26(7): 391-399.
Golden, JA, Nadel, JA and Boushey, HA (1978). Bronchial hyperirritability in healthy subjects
after exposure to ozone. Am Rev Respir Dis 118(2): 287-294.
Gomes, EC, Stone, V and Florida-James, G (2011). Impact of heat and pollution on oxidative
stress and CC16 secretion after 8 km run. Eur J Appl Physiol 111(9): 2089-2097.
Gong, H, Jr., Bedi, JF and Horvath, SM (1988). Inhaled albuterol does not protect against ozone
toxicity in nonasthmatic athletes. Arch Environ Occup Health 43(1): 46-53.
Gong, H, Jr., Bradley, PW, Simmons, MS and Tashkin, DP (1986). Impaired exercise
performance and pulmonary function in elite cyclists during low-level ozone exposure in
a hot environment. Am J Respir Crit Care Med 134(4): 726-733.
Gong, H, Jr., McManus, MS and Linn, WS (1997). Attenuated response to repeated daily ozone
exposures in asthmatic subjects. Arch Environ Occup Health 52(1): 34-41.
Graham, DE and Koren, HS (1990). Biomarkers of inflammation in ozone-exposed humans:
Comparison of the nasal and bronchoalveolar lavage. Am J Respir Crit Care Med 142(1):
152-156.
Hamilton, RF, Li, L, Eschenbacher, WL, Szweda, L and Holian, A (1998). Potential involvement
of 4-hydroxynonenal in the response of human lung cells to ozone. Am J Physiol 274(1
Pt 1): L8-L16.
Hanania, NA, Tarlo, SM, Silverman, F, Urch, B, Senathirajah, N, Zamel, N and Corey, P (1998).
Effect of exposure to low levels of ozone on the response to inhaled allergen in allergic
asthmatic patients. Chest 114(3): 752-756.
Hazbun, ME, Hamilton, R, Holian, A and Eschenbacher, WL (1993). Ozone-induced increases
in substance P and 8-epi-prostaglandin F2 alpha in the airways of human subjects. Am J
Respir Cell Mol Biol 9(5): 568-572.
Hazucha, MJ, Folinsbee, LJ and Seal, E, Jr. (1992). Effects of steady-state and variable ozone
concentration profiles on pulmonary function. Am Rev Respir Dis 146(6): 1487-1493.
Hernandez, M, Brickey, WJ, Alexis, NE, Fry, RC, Rager, JE, Zhou, B, Ting, JP, Zhou, H and
Peden, DB (2012). Airway cells from atopic asthmatic patients exposed to ozone display
3A-28
-------
an enhanced innate immune gene profile. J Allergy Clin Immunol 129(1): 259-261.e251-
252.
Hernandez, ML, Lay, JC, Harris, B, Esther, CR, Brickey, WJ, Bromberg, PA, Diaz-Sanchez, D,
Devlin, RB, Kleeberger, SR, Alexis, NE and Peden, DB (2010). Atopic asthmatic
subjects but not atopic subjects without asthma have enhanced inflammatory response to
ozone. J Allergy Clin Immunol 126(3): 537-544.
Hiltermann, JTN, Lapperre, TS, Van Bree, L, Steerenberg, PA, Brahim, JJ, Sont, JK, Sterk, PJ,
Hiemstra, PS and Stolk, J (1999). Ozone-induced inflammation assessed in sputum and
bronchial lavage fluid from asthmatics: A new noninvasive tool in epidemiologic studies
on air pollution and asthma. Free Radical Biol Med 27(11-12): 1448-1454.
Hiltermann, TJN, Peters, EA, Alberts, B, Kwikkers, K, Borggreven, PA, Hiemstra, PS, Dijkman,
JH, van Bree, LA and Stolk, J (1998). Ozone-induced airway hyperresponsiveness in
patients with asthma: Role of neutrophil-derived serine proteinases. Free Radical Biol
Med 24(6): 952-958.
Hiltermann, TJN, Stolk, J, Hiemstra, PS, Fokkens, PHB, Rombout, PJA, Sont, JK, Sterk, PJ and
Dijkman, JH (1995). Effect of ozone exposure on maximal airway narrowing in non-
asthmatic and asthmatic subjects. Clinical Science 89(6): 619-624.
Holz, O, Biller, H, Mueller, M, Kane, K, Rosano, M, Hanrahan, J, Hava, DL and Hohlfeld, JM
(2015). Efficacy and safety of inhaled calcium lactate PUR118 in the ozone challenge
model—a clinical trial. BMC Pharmacol Toxicol 16: 21.
Holz, O, Jorres, RA, Timm, P, Mucke, M, Richter, K, Koschyk, S and Magnussen, H (1999).
Ozone-induced airway inflammatory changes differ between individuals and are
reproducible. Am J Respir Crit Care Med 159(3): 776-784.
Holz, O, Mucke, M, Paasch, K, Bohme, S, Timm, P, Richter, K, Magnussen, H and Jorres, RA
(2002). Repeated ozone exposures enhance bronchial allergen responses in subjects with
rhinitis or asthma. Clin Exp Allergy 32(5): 681-689.
Holz, O, Tal-Singer, R, Kanniess, F, Simpson, KJ, Gibson, A, Vessey, RSJ, Janicki, S,
Magnussen, H, Jorres, RA and Richter, K (2005). Validation of the human ozone
challenge model as a tool for assessing anti-inflammatory drugs in early development. J
Clin Pharmacol 45(5): 498-503.
Horstman, DH, Ball, BA, Brown, J, Gerrity, T and Folinsbee, LJ (1995). Comparison of
pulmonary responses of asthmatic and nonasthmatic subjects performing light exercise
while exposed to a low level of ozone. Toxicol Ind Health 11(4): 369-385.
Horstman, DH, Folinsbee, LJ, Ives, PJ, Abdul-Salaam, S and McDonnell, WF (1990). Ozone
concentration and pulmonary response relationships for 6.6-hour exposures with five
hours of moderate exercise to 0.08, 0.10, and 0.12 ppm. Am Rev Respir Dis 142(5):
1158-1163.
3A-29
-------
Horvath, SM, Gliner, JA and Matsen-Twisdale, JA (1979). Pulmonary function and maximum
exercise responses following acute ozone exposure. Aviat Space Environ Med 50(9):
901-905.
Jorres, R, Nowak, D, Magnussen, H, Speckin, P and Koschyk, S (1996). The effect of ozone
exposure on allergen responsiveness in subjects with asthma or rhinitis. Am J Respir Crit
Care Med 153(1): 56-64.
Kahle, JJ, Neas, LM, Devlin, RB, Case, MW, Schmitt, MT, Madden, MC and Diaz-Sanchez, D
(2015). Interaction effects of temperature and ozone on lung function and markers of
systemic inflammation, coagulation, and fibrinolysis: a crossover study of healthy young
volunteers. Environ Health Perspect 123(4): 310-316.
Keefe, MJ, Bennett, WD, Dewitt, P, Seal, E, Strong, AA and Gerrity, TR (1991). The effect of
ozone exposure on the dispersion of inhaled aerosol boluses in healthy human subjects.
Am J Respir Crit Care Med 144(1): 23-30.
Kehrl, HR, Vincent, LM, Kowalsky, RJ, Horstman, DH, O'Neil, JJ, McCartney, WH and
Bromberg, PA (1987). Ozone exposure increases respiratory epithelial permeability in
humans. Am Rev Respir Dis 135(5): 1124-1128.
Kim, CS, Alexis, NE, Rappold, AG, Kehrl, H, Hazucha, MJ, Lay, JC, Schmitt, MT, Case, M,
Devlin, RB, Peden, DB and Diaz-Sanchez, D (2011). Lung function and inflammatory
responses in healthy young adults exposed to 0.06 ppm ozone for 6.6 hours. Am J Respir
Crit Care Med 183(9): 1215-1221.
Kirsten, A, Watz, H, Kretschmar, G, Pedersen, F, Bock, D, Meyer-Sabellek, W and Magnussen,
H (2011). Efficacy of the pan-selectin antagonist Bimosiamose on ozone-induced airway
inflammation in healthy subjects - A double blind, randomized, placebo-controlled,
cross-over clinical trial. Pulm Pharmacol Ther 24(5): 555-558.
Konig, G, Rommelt, H, Kienle, H, Dirnagl, K, Polke, H and Fruhmann, G (1980). Anderung der
bronchomotorischen Reagibilitat des Menschen durch Einwirkung von Ozon [Changes in
the bronchial reactivity of humans caused by the influence of ozone], Arbeitsmed
Sozialmed Umweltmed 151: 261-263.
Koren, G, Sharav, T, Pastuszak, A, Garrettson, LK, Hill, K, Samson, I, Rorem, M, King, A and
Dolgin, JE (1991). A multicenter, prospective study of fetal outcome following accidental
carbon monoxide poisoning in pregnancy. Reprod Toxicol 5(5): 397-403.
Koren, HS, Devlin, RB, Graham, DE, Mann, R and McDonnell, WF (1989a). Atmospheric ozone
research and its policy implications: The inflammatory response in human lung exposed
to ambient levels of ozone. Amsterdam, The Netherlands.
Koren, HS, Devlin, RB, Graham, DE, Mann, R, McGee, MP, Horstman, DH, Kozumbo, WJ,
Becker, S, House, DE, McDonnell, WF and Bromberg, PA (1989b). Ozone-induced
inflammation in the lower airways of human subjects. Am J Respir Crit Care Med
139(2): 407-415.
3A-30
-------
Koren, HS, Hatch, GE and Graham, DE (1990). Nasal lavage as a tool in assessing acute
inflammation in response to inhaled pollutants. Toxicology 60(1-2): 15-25.
Kreit, JW, Gross, KB, Moore, TB, Lorenzen, TJ, D'Arcy, J and Eschenbacher, WL (1989).
Ozone-induced changes in pulmonary function and bronchial responsiveness in
asthmatics. J Appl Physiol (1985) 66(1): 217-222.
Krishna, MT, Blomberg, A, Biscione, GL, Kelly, F, Sandstrom, T, Frew, A and Holgate, S
(1997). Short-term ozone exposure upregulates P-selectin in normal human airways. Am
J Respir Crit Care Med 155(5): 1798-1803.
Krishna, MT, Madden, J, Teran, LM, Biscione, GL, Lau, LCK, Withers, NJ, Sandstrom, T,
Mudway, I, Kelly, FJ, Walls, A, Frew, AJ and Holgate, ST (1998). Effects of 02 ppm
ozone on biomarkers of inflammation in bronchoalveolar lavage fluid and bronchial
mucosa of healthy subjects. Eur Respir J 11(6): 1294-1300.
Kulle, TJ, Sauder, LR, Hebel, JR and Chatham, MD (1985). Ozone response relationships in
healthy nonsmokers. Am Rev Respir Dis 132(1): 36-41.
Kulle, TJ, Sauder, LR, Kerr, HD, Farrell, BP, Bermel, MS and Smith, DM (1982). Duration of
pulmonary function adaptation to ozone in humans. Am Ind Hyg Assoc J 43(11): 832-
837.
Lay, JC, Alexis, NE, Kleeberger, SR, Roubey, RA, Harris, BD, Bromberg, PA, Hazucha, MJ,
Devlin, RB and Peden, DB (2007). Ozone enhances markers of innate immunity and
antigen presentation on airway monocytes in healthy individuals. J Allergy Clin Immunol
120(3): 719-722.
Linn, WS, Avol, EL, Shamoo, DA, Spier, CE, Valencia, LM, Venet, TG, Fischer, DA and
Hackney, JD (1986). A dose-response study of healthy, heavily exercising men exposed
to ozone at concentrations near the ambient air quality standard. Toxicol Ind Health 2(1):
99-112.
Madden, MC, Stevens, T, Case, M, Schmitt, M, Diaz-Sanchez, D, Bassett, M, Montilla, TS,
Berntsen, J and Devlin, RB (2014). Diesel exhaust modulates ozone-induced lung
function decrements in healthy human volunteers. Part Fibre Toxicol 11(1): 37.
McBride, DE, Koenig, JQ, Luchtel, DL, Williams, PV and Henderson, WR, Jr. (1994).
Inflammatory effects of ozone in the upper airways of subjects with asthma. Am J Respir
Crit Care Med 149(5): 1192-1197.
McDonnell, WF, 3rd, Chapman, RS, Leigh, MW, Strope, GL and Collier, AM (1985).
Respiratory responses of vigorously exercising children to 0.12 ppm ozone exposure. The
American review of respiratory disease 132(4): 875-879.
McDonnell, WF, Horstman, DH, Abdul-Salaam, S, Raggio, LJ and Green, JA (1987). The
respiratory responses of subjects with allergic rhinitis to ozone exposure and their
relationship to nonspecific airway reactivity. Toxicol Ind Health 3(4): 507-517.
3A-31
-------
McDonnell, WF, Horstman, DH, Hazucha, MJ, Seal, E, Jr., Haak, ED, Salaam, SA and House,
DE (1983). Pulmonary effects of ozone exposure during exercise: Dose-response
characteristics. J Appl Physiol (1985) 54(5): 1345-1352.
McDonnell, WF, Kehrl, HR, Abdul-Salaam, S, Ives, PJ, Folinsbee, LJ, Devlin, RB, O'Neil, JJ
and Horstman, DH (1991). Respiratory response of humans exposed to low levels of
ozone for 6.6 hours. Arch Environ Health 46(3): 145-150.
McDonnell, WF, Stewart, PW, Smith, MV, Kim, CS and Schelegle, ES (2012). Prediction of
lung function response for populations exposed to a wide range of ozone conditions.
Inhal Toxicol 24(10): 619-633.
McGee, MP, Devlin, R, Saluta, G and Koren, H (1990). Tissue factor and factor VII messenger
RNAs in human alveolar macrophages: effects of breathing ozone. Blood 75(1): 122-127.
McKittrick, T and Adams, WC (1995). Pulmonary function response to equivalent doses of
ozone consequent to intermittent and continuous exercise. Arch Environ Occup Health
50(2): 153-158.
Michelson, PH, Dailey, L, Devlin, RB and Peden, DB (1999). Ozone effects on the immediate-
phase response to allergen in the nasal airways of allergic asthmatic subjects. Otolaryngol
Head Neck Surg 120(2): 225-232.
Molfino, NA, Wright, SC, Katz, I, Tarlo, S, Silverman, F, McClean, PA, Szalai, JP, Raizenne,
M, Slutsky, AS and Zamel, N (1991). Effect of low concentrations of ozone on inhaled
allergen responses in asthmatic subjects. The Lancet 338(8761): 199-203.
Montuschi, P, Nightingale, JA, Kharitonov, SA and Barnes, PJ (2002). Ozone-induced increase
in exhaled 8-isoprostane in healthy subjects is resistant to inhaled budesonide. Free
Radical Biol Med 33(10): 1403-1408.
Mudway, IS, Krishna, MT, Frew, AJ, Macleod, D, Sandstrom, T, Holgate, ST and Kelly, FJ
(1999). Compromised concentrations of ascorbate in fluid lining the respiratory tract in
human subjects after exposure to ozone. Occup Environ Med 56(7): 473-481.
Mudway, IS, Stenfors, N, Blomberg, A, Helleday, R, Dunster, C, Marklund, SL, Frew, AJ,
Sandstrom, T and Kelly, FJ (2001). Differences in basal airway antioxidant
concentrations are not predictive of individual responsiveness to ozone: A comparison of
healthy and mild asthmatic subjects. Free Radical Biol Med 31(8): 962-974.
Newson, EJ, Krishna, MT, Lau, LCK, Howarth, PH, Holgate, ST and Frew, AJ (2000). Effects
of short-term exposure to 0.2 ppm ozone on biomarkers of inflammation in sputum,
exhaled nitric oxide, and lung function in subjects with mild atopic asthma. J Occup
Environ Med 42(3): 270-277.
Nightingale, JA, Rogers, DF, Chung, KF and Barnes, PJ (2000). No effect of inhaled budesonide
on the response to inhaled ozone in normal subjects. Am J Respir Crit Care Med 161(2 Pt
1): 479-486.
3A-32
-------
Passannante, AN, Hazucha, MJ, Bromberg, PA, Seal, E, Folinsbee, L and Koch, G (1998).
Nociceptive mechanisms modulate ozone-induced human lung function decrements. J
Appl Physiol (1985) 85(5): 1863-1870.
Peden, DB, Setzer, RW, Jr. and Devlin, RB (1995). Ozone exposure has both a priming effect on
allergen-induced responses and an intrinsic inflammatory action in the nasal airways of
perennially allergic asthmatics. Am J Respir Crit Care Med 151(5): 1336-1345.
Peters, EA, Hiltermann, JT and Stolk, J (2001). Effect of apocynin on ozone-induced airway
hyperresponsiveness to methacholine in asthmatics. Free Radical Biol Med 31(11): 1442-
1447.
Que, LG, Stiles, JV, Sundy, JS and Foster, WM (2011). Pulmonary function, bronchial
reactivity, and epithelial permeability are response phenotypes to ozone and develop
differentially in healthy humans. J Appl Physiol (1985) 111(3): 679-687.
Samet, JM, Hatch, GE, Horstman, D, Steck-Scott, S, Arab, L, Bromberg, PA, Levine, M,
McDonnell, WF and Devlin, RB (2001). Effect of antioxidant supplementation on ozone-
induced lung injury in human subjects. Am J Respir Crit Care Med 164(5): 819-825.
Schelegle, ES and Adams, WC (1986). Reduced exercise time in competitive simulations
consequent to low level ozone exposure. Med Sci Sports Exerc 18(4): 408-414.
Schelegle, ES, Adams, WC and Siefkin, AD (1987). Indomethacin pretreatment reduces ozone-
induced pulmonary function decrements in human subjects. Am Rev Respir Dis 136(6):
1350-1354.
Schelegle, ES, Morales, CA, Walby, WF, Marion, S and Allen, RP (2009). 6.6-hour inhalation of
ozone concentrations from 60 to 87 parts per billion in healthy humans. Am J Respir Crit
Care Med 180(3): 265-272.
Schelegle, ES, Siefkin, AD and McDonald, RJ (1991). Time course of ozone-induced
neutrophilia in normal humans. Am J Respir Crit Care Med 143(6): 1353-1358.
Schonfeld, BR, Adams, WC and Schelegle, ES (1989). Duration of enhanced responsiveness
upon re-exposure to ozone. Arch Environ Occup Health 44(4): 229-236.
Seal, E, Jr., McDonnell, WF, House, DE, Salaam, SA, Dewitt, PJ, Butler, SO, Green, J and
Raggio, L (1993). The pulmonary response of white and black adults to six
concentrations of ozone. Am J Respir Crit Care Med 147(4): 804-810.
Seltzer, J, Bigby, BG, Stulbarg, M, Holtzman, MJ, Nadel, JA, Ueki, IF, Leikauf, GD, Goetzl, EJ
and Boushey, HA (1986). 03-induced change in bronchial reactivity to methacholine and
airway inflammation in humans. J Appl Physiol (1985) 60(4): 1321-1326.
Silverman, F, Folinsbee, LJ, Barnard, J and Shephard, RJ (1976). Pulmonary function changes in
ozone - interaction of concentration and ventilation. J Appl Physiol (1985) 41(6): 859-
864.
3A-33
-------
Speen, AM, Kim, HH, Bauer, RN, Meyer, M, Gowdy, KM, Fessler, MB, Duncan, KE, Liu, W,
Porter, NA and Jaspers, I (2016). Ozone-derived oxysterols affect liver X receptor (LXR)
signaling: a potential role for lipid-protein adducts. J Biol Chem 291(48): 25192-25206.
Steck-Scott, S, Arab, L, Craft, NE and Samet, JM (2004). Plasma and lung macrophage
responsiveness to carotenoid supplementation and ozone exposure in humans. Eur J Clin
Nutr 58(12): 1571-1579.
Stenfors, N, Bosson, J, Helleday, R, Behndig, AF, Pourazar, J, Tornqvist, H, Kelly, FJ, Frew, AJ,
Sandstrom, T, Mudway, IS and Blomber, A (2010). Ozone exposure enhances mast-cell
inflammation in asthmatic airways despite inhaled corticosteroid therapy. Inhal Toxicol
22(2): 133-139.
Stenfors, N, Pourazar, J, Blomberg, A, Krishna, MT, Mudway, I, Helleday, R, Kelly, FJ, Frew,
AJ and Sandstrom, T (2002). Effect of ozone on bronchial mucosal inflammation in
asthmatic and healthy subjects. Respir Med 96(5): 352-358.
Stiegel, MA, Pleil, JD, Sobus, JR, Stevens, T and Madden, MC (2017). Linking physiological
parameters to perturbations in the human exposome: Environmental exposures modify
blood pressure and lung function via inflammatory cytokine pathway. J Toxicol Environ
Health, A: Currlss 80(9): 485-501.
Tank, J, Biller, H, Heusser, K, Holz, O, Diedrich, A, Framke, T, Koch, A, Grosshennig, A,
Koch, W, Krug, N, Jordan, J and Hohlfeld, JM (2011). Effect of acute ozone induced
airway inflammation on human sympathetic nerve traffic: a randomized, placebo
controlled, crossover study. PLoS ONE 6(4): el8737.
Trenga, CA, Koenig, JQ and Williams, PV (2001). Dietary antioxidants and ozone-induced
bronchial hyperresponsiveness in adults with asthma. Arch Environ Occup Health 56(3):
242-249.
U.S. EPA (1996). Air Quality Criteria for Ozone and Related Photochemical Oxidants. Volumes
I to III. . U.S. EPA. Research Triangle Park, NC. EPA/600/P-93/004aF, EPA/600/P-
93/004bF, and EPA/600/P-93/004cF.
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 (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 (2020). Integrated Science Assessment for Ozone and Related Photochemical
Oxidants. U.S. Environmental Protection Agency. Washington, DC. Office of Research
3A-34
-------
and Development. EPA/600/R-20/012. Available at: https://www.epa.gov/isa/integrated-
science-assessment-isa-ozone-and-related-photochemical-oxidants.
Ultman, JS, Ben-Jebria, A and Arnold, SF (2004). Uptake distribution of ozone in human lungs:
Intersubject variability in physiologic response. HEI Research Report 125. Health Effects
Institute. Boston, MA. http://pubs.healtheffects.org/view.php?id=70.
Vagaggini, B, Bartoli, MLE, Cianchetti, S, Costa, F, Bacci, E, Dente, FL, Di Franco, A,
Malagrino, L and Paggiaro, P (2010). Increase in markers of airway inflammation after
ozone exposure can be observed also in stable treated asthmatics with minimal functional
response to ozone. Respir Res 11:5.
Vagaggini, B, Cianchetti, S, Bartoli, M, Ricci, M, Bacci, E, Dente, FL, Di Franco, A and
Paggiaro, P (2007). Prednisone blunts airway neutrophilic inflammatory response due to
ozone exposure in asthmatic subjects. Respiration 74(1): 61-58.
Vagaggini, B, Taccola, M, Cianchetti, S, Carnevali, S, Bartoli, ML, Bacci, E, Dente, FL, Di
Franco, A, Giannini, D and Paggiaro, PL (2002). Ozone exposure increases eosinophilic
airway response induced by previous allergen challenge. Am J Respir Crit Care Med
166(8): 1073-1077.
Vagaggini, B, Taccola, M, Conti, I, Carnevali, S, Cianchetti, S, Bartoli, ML, Bacci, E, Dente,
FL, Di Franco, A, Giannini, D and Paggiaro, PL (2001). Budesonide reduces neutrophilic
but not functional airway response to ozone in mild asthmatics. Am J Respir Crit Care
Med 164(12): 2172-2176.
Weymer, AR, Gong, H, Jr., Lyness, A and Linn, WS (1994). Pre-exposure to ozone does not
enhance or produce exercise-induced asthma. Am J Respir Crit Care Med 149(6): 1413-
1419.
Ying, RL, Gross, KB, Terzo, TS and Eschenbacher, WL (1990). Indomethacin does not inhibit
the ozone-induced increase in bronchial responsiveness in human subjects. Am Rev
Respir Dis 142(4): 817-821.
3A-35
-------
APPENDIX 3B
AIR QUALITY INFORMATION FOR LOCATIONS OF
EPIDEMIOLOGIC STUDIES OF RESPIRATORY EFFECTS
3B-1
-------
This appendix provides summary information about the O3 concentrations in locations
and time periods of epidemiologic studies of associations between O3 in ambient air and
respiratory health outcomes. Included here are studies conducted in the U.S. and Canada that
found associations between O3 exposure and respiratory health effects such as emergency
department visits and hospital admissions, including studies that are newly available in this
review, as well as those that were available at the time of the last review, and that are identified
in the ISA. Information for studies identified in the ISA1 as short-term are summarized in Table
3B-1 and a subset of studies identified as long-term are summarized in Table 3B-2.
Air quality information for U.S.-based studies was obtained from the EPA's Air Quality
System (AQS) database.2 For Canada-based studies, air quality information was obtained from
the National Air Pollutant Surveillance (NAPS) program.3 In Table 3B-1 and Table 3B-2, design
values (DVs)4 are presented as a range across all locations and time periods in the study.5
Detailed information about designs values for individual study locations and time periods are
available in the Attachment.6
1 Single- and multi-city studies are included. Given the purpose of describing the air quality conditions in the cities
studied, meta-analysis studies are not included; rather, the relevant underlying studies would be.
2 Available at: https://www.epa.gov/aqs.
3 Available at: https://www.canada.ca/en/environment-climate-change/services/air-pollution/monitoring-networks-
data/national-air-pollution-program.html.
4 The design value for the current standard is the 3-year average of the annual 4th highest daily maximum 8-hour
average O3 concentration.
5 For those locations with more than one monitor, the design values presented in Table 3B-1, Table 3B-2, and in the
attachment for that location are for the highest monitor in that area.
6 In the attachment tables, blank cells indicate one of two situations: (1) monitoring data are unavailable for the
specific time period or the entire period for the city, or (2) the available data do not meet the data requirements for
the calculations.
3B-2
-------
Table 3B-1. Epidemiologic studies of associations between short-term ozone concentrations and respiratory effects.
Study Information
Ambient Air
Quality
Study Area
Health
Study
Time
Period
Air
Quality
Time
Period
Study
Reference A
Health Outcome
O3 Concentration
Metric
Associated with
Health Outcome
Assignment of Monitors to
Study Subjects
Study-reported 03
Concentrations, in terms of
study metric (ppb)
Design Values
for Current
NAAQS,
across cities
and study
years (ppb)B
Mean/
median
Range
U.S. Studies
Single City Studies
Indianapolis,
IN
2007-
2011
2007-
2011
Byers etal.,
2015
ED Visits for Asthma
8-hr daily
maximums,
moving average of
lag 0-2
Distance and population-
weighted daily average O3
concentration of 11 monitor
values for the Indianapolis MSA
(9 counties)
8-hr (WS): 48.5
NA
73-77
Atlanta, GA
1993-
2004
1993-
2004
Darrow et
al„ 2011
ED Visits for
Aggregate
Respiratory Diseases
1-hrand 8-hr daily
maximums,
previous day lag
(lag 1)
Daily O3 concentration of single
centrally located monitor in the
Atlanta MSA
1-hr (WS): 62.0
8-hr (WS): 53.0
1-h Max:
180.0
8-hr Max:
148.0
91-121
Atlanta, GA
1993-
2010
1993-
2010
Darrow et
al„ 2014
ED Visits for
Respiratory Infection
8-hr daily
maximum, 3-day
moving average of
lag 0-2
Population-weighted daily
average O3 concentration of 5
monitor values for the Atlanta
MSA (20 counties)
8-hr (YR):
45.9
3.0-127.1
80-121
New Jersey
2004-
2007
2004-
2007
Gleason et
al„ 2014
ED Visits for Asthma
8-hr daily
maximum, same
day lag (lag 0)
Daily O3 concentration obtained
from Bayesian spatio-temporal
model assigned to study
participants based on
corresponding grid cells for
geocoded residential addresses
NA
NA
92-93
New York,
NY
1999-
2009
1999-
2009
Goodman et
al., 2017a
HA for Asthma
8-hr daily
maximum,
average of lag 0-1
Daily average O3 concentrations
of all monitors within 20-mile of
the geographic center of NY city
8-hr (YR):
30.7
2.0-105.4
84-115
New York,
NY
1999-
2002
1999-
2002
Ito etal.,
2007
ED Visits for Asthma
8-hr daily
maximum,
average of lag 0-1
Average of 16 monitors within
20 miles of the geographic city
center of NY city
8-hr (YR):
30.4
8-hr (WS): 42.7
5th and 95th
percentiles:
YR: 6.0-68.0
WS: 18.0-77.0
109-115
Atlanta, GA
1998-
2007
1998-
2007
Klemm etal.,
2011
Respiratory Mortality
8-hr daily
maximum,
average of lag 0-1
Daily average O3 concentration of
all monitors in four counties in
Atlanta
8-hr (YR):
35.5
0.0-109.1
90-121
3B-3
-------
Study Information
Ambient Air
Quality
Health
Air
O3 Concentration
Study-reported 03
Concentrations, in terms of
Design Values
for Current
Study Area
Study
Quality
Study
Health Outcome
Metric
Assignment of Monitors to
study metric (ppb)
NAAQS,
Time
Time
Reference A
Associated with
Study Subjects
Mean/
median
across cities
Period
Period
Health Outcome
Range
and study
years (ppb)B
Atlanta, GA
2002-
2008
2002-
2008
O'Lenick et
al„ 2017
ED Visits for Asthma
8-hr daily
maximum, 3-day
moving average of
lag 0-2
Daily O3 concentration obtained
from spatio-temporal model
assigned to study participants
based on corresponding ZCTA
for residential ZIP code
NA
NA
90-95
Little Rock,
AR
2002-
2012
2002-
2012
Rodopoulou
etal., 2015
ED Visits for
Respiratory Infection
8-hr daily
maximum, lag 2
Daily O3 concentration from one
monitor in Little Rock, AR
8-hr (YR):
40.0
NA
70-83
Atlanta, GA
1999-
2002
1999-
2002
Sarnatetal.,
2013
ED Visits for Asthma
24-hr daily
average
Spatially resolved daily O3
concentration at ZIP code
centroid assigned to participants
based on residential ZIP code
8-hr (YR):
41.9
3.5-132.7
99-107
St. Louis,
MO
2001-
2003
2001-
2004
Sarnatetal.,
2015
ED Visits for Asthma
8-hr daily
maximum,
distributed lags
(lags 0-2)
Daily O3 concentration from one
monitor in St. Louis, MO.
8-hr (YR):
36.2
NA
92
New York,
NY
2005-
2011
2005-
2012
Sheffield et
al„ 2015
ED Visits for Asthma
24-hr daily
average
Daily average O3 concentration of
seven monitors in NYC.
NA
NA
82-94
New York,
NY
2005-
2011
2005-
2011
Shmool et
al„ 2016
ED Visits for Asthma
24-hr daily
average, case-day
Near-residence exposure was
determined by combining data
from temporally- and spatially-
refined estimates
Temporal
estimates
(my. 30.4
Spatiotemporal
estimates: 29.0
Temporal
estimates:
5.0-60.0
Spatiotempor
al estimates:
4.6-60.3
82-94
New York,
NY
1999-
2006
1999-
2006
Silverman
and Ito, 2010
HA for Asthma
8-hr daily
maximum,
average of lag 0-1
Average of 13 monitors within
20 miles of the geographic city
center of NY city
8-hr (WS): 41.0
10th and 90th
percentiles:
18.0-77.0
93-115
Atlanta, GA
2002-
2010
2002-
2010
Strickland et
al„ 2014
ED Visits for Asthma
8-hr daily
maximum, 3-day
moving average
lag 0-2
Distance and population-
weighted daily average of five
monitor values for the Atlanta
MSA (20 counties)
8-hr (YR):
42.2
NA
80-95
Atlanta, GA
1993-
2004
1993-
2004
Tolbert et al.,
2007
ED Visits for
Aggregate
Respiratory Diseases
8-hr daily
maximum,
average of lag 0-1
Average of monitors in Atlanta
city
8-hr (EC): 53.0
2.9-147.5
91-121
3B-4
-------
Study Information
Ambient Air
Quality
Study Area
Health
Study
Time
Period
Air
Quality
Time
Period
Study
Reference A
Health Outcome
O3 Concentration
Metric
Associated with
Health Outcome
Assignment of Monitors to
Study Subjects
Study-reported 03
Concentrations, in terms of
study metric (ppb)
Design Values
for Current
NAAQS,
across cities
and study
years (ppb)B
Mean/
median
Range
St. Louis,
MO
2001-
2007
2001-
2007
Winquist et
al„ 2012
HA for Asthma
ED Visits for Asthma
ED Visits for
Respiratory Infection
HA for Aggregate
Respiratory Diseases
ED Visits Aggregate
Respiratory Diseases
8-hr daily
maximum,
distributed lags
(lags 0-4)
Daily O3 concentration from one
monitor in St. Louis, MO.
NA
NA
86-92
Atlanta, GA
1998-
2004
1998-
2004
Winquist et
al„ 2014
ED Visits for Asthma
8-hr daily
maximum, 3-day
moving average of
laq 0-2
Population-weighted daily
average of five monitor values for
the Atlanta MSA (20 counties)
8-hr (WS):
53.9
NA
91-121
Multi-city Studies
3 U.S. cities
1993-
2009
1993-
2009
Alhanti etal.,
2016
ED Visits for Asthma
8-hr maximum, 3-
day moving
average of lag 0-2
Population-weighted daily
average of monitor values for
each city
8-hr (YR) for 3
cities
mean range:
37.3-43.7
NA
86-121
5 U.S. cities
2002-
2008
2002-
2008
Barry etal.,
2018
ED Visits for Asthma
ED Visits for
Respiratory Infection
ED Visits Aggregate
Respiratory Diseases
8-hr maximum, 3-
day moving
average of lag 0-2
Daily O3 concentration obtained
model simulations and monitor
measurements were spatially
averaged for each metropolitan
area usinq population weiqhtinq
8-hr (YR) for 5
cities
mean range:
37.5-42.2
Min Range:
3.9-9.4
Max Range:
80.2-106.3
83-95
3 metro
areas in TX
2003-
2011
2003-
2011
Goodman et
al., 2017b
HA for Asthma
8-hr maximum,
same day lag (lag
0)
City-specific daily O3
concentrations were calculated
using all monitors within each
city: Dallas (8 monitors), Houston
(44 monitors), Austin (6
monitors), then were averaged to
obtain area-specific daily
maximum 8-hr concentrations
8-hr (YR): 41.8
2.0-107.0
74-103
Nationwide
(U.S.)
1987-
1996
1987-
1996
Katsouyanni
etal., 2009
Respiratory Mortality
1-hr maximum, 2-
day average of lag
0-1
Daily average of O3
concentrations from all monitors
in each city
NA
NA
18-192
3B-5
-------
Study Information
Ambient Air
Quality
Study Area
Health
Study
Time
Period
Air
Quality
Time
Period
Study
Reference A
Health Outcome
O3 Concentration
Metric
Associated with
Health Outcome
Assignment of Monitors to
Study Subjects
Study-reported 03
Concentrations, in terms of
study metric (ppb)
Design Values
for Current
NAAQS,
across cities
and study
years (ppb)B
Mean/
median
Range
California
2005-
2008
2005-
2009
Malig etal.,
2016
ED Visits for Asthma
ED Visits for
Respiratory Infection
ED Visits Aggregate
Respiratory Diseases
1-hr maximum, 2-
day average of lag
0-1
Daily O3 concentration from
nearest monitor within 20 km of
population-weighted ZIP code
centroid assigned to participants
based on residential ZIP code
8-hr for 16
climatic zones
mean range:
(YR): 33.0-55.0
(WS): 31.0-75.0
NA
119-122
3 U.S. cities
2002-
2008
2002-
2008
O'Lenick et
al„ 2017
ED Visits Aggregate
Respiratory Diseases
8-hr daily
maximum, 3-day
moving average of
lag 0-2
Daily O3 concentration obtained
from spatio-temporal model
assigned to study participants
based on corresponding ZCTA
for residential ZIP code
8-hr (YR) for 3
cities
mean ranges
from 40.0-42.2
Min Range:
0.15-2.21
Max Range:
115-125
85-96
North
Carolina
2006-
2008
2006-
2008
Sacks etal.,
2014
ED Visits for Asthma
8-hr daily
maximum, 3-day
moving average of
lag 0-2
O3 estimates from CMAQ model
with Bayesian space-time
approach assigned to census
tract centroids and aggregated to
county-level using area-weighted
average of census tract centroids
8-hr (YR): 43.6
8-hr (WS): 50.1
Max:108.1
94
Georgia
2002-
2008
2002-
2008
Xiao etal.,
2016
ED Visits for Asthma
ED Visits for
Respiratory Infection
8-hr daily
maximum, 3-day
moving average of
lag 0-2
Daily O3 concentration obtained
from spatio-temporal model
assigned to study participants
based on residential ZIP code
8-hr (YR): 42.1
5.4-106.1
91-95
48 U.S.
cities
1989-
2000
1989-
2000
Zanobetti
and
Schwartz,
2008
Respiratory Mortality
8-hr daily average,
same day lag (lag
0)
Daily average of O3
concentrations from all monitors
in each city
8-hr (WS) for
40 U.S. cities
mean range:
15.1-62.8
Min Range:
0.9-23.6
Max Range:
34.3-146.2
45-179
6 cities in TX
2001-
2013
2001-
2013
Zuetal.,
2017
HA for Asthma
8-hr daily
maximum, lag 0-3
City specific daily O3
concentrations were calculated
using all monitors within each
city: Dallas (15 monitors),
Houston (44 monitors), Austin (6
monitors), El Paso (6 monitors),
Fort Worth (9 monitors); then
were averaged to obtain area-
specific daily maximum 8-hr
concentrations.
8-hr (YR):
32.2
1.0-82.8
71-103
3B-6
-------
Study Information
Ambient Air
Quality
Study Area
Health
Study
Time
Period
Air
Quality
Time
Period
Study
Reference A
Health Outcome
O3 Concentration
Metric
Associated with
Health Outcome
Assignment of Monitors to
Study Subjects
Study-reported 03
Concentrations, in terms of
study metric (ppb)
Design Values
for Current
NAAQS,
across cities
and study
years (ppb)B
Mean/
median
Range
Canadian Studies
Single City Studies
Edmonton,
Canada
1992-
2002
1992-
2002
Kousha and
Rowe, 2014
EDVisitsfor
Respiratory Infection
8-hr daily
maximum, same
day lag (lag 0).
Daily average of O3
concentrations from three
monitors in Edmonton, Canada
8-hr (YR):
18.6
NA
56-65
Windsor,
Canada
2004-
2010
2004-
2010
Kousha and
Castner,
2016
ED Visits for
Respiratory Infection
8-hr daily
maximum, same
day lag (lag 0).
Daily average of O3
concentrations from monitors in
Windsor, Canada
8-hr (YR):
25.3
NA
73-87
Alberta,
Canada
1992-
2002
1992-
2002
Villeneuve et
al„ 2007
ED Visits for Asthma
8-hr daily
maximum, lag 1.
Daily average of three monitors in
census metropolitan of
Edmonton, Alberta
8-hr (WS): 38.0
(Median)
NA
60-69
Multi-city Studies
7 Canadian
cities
1992-
2003
1992-
2003
Stieb et al„
2009
ED visits for Asthma
24-hr average, lag
1
Daily average of O3
concentrations from monitors in
each city
24-hr (YR):
Mean range:
10.3-22.1
NA
51-85
9 Canadian
cities
2004-
2011
2004-
2011
Szyszkowicz
etal., 2018
ED Visits for Asthma
ED Visits for
Respiratory Infection
24-hr daily
average, lag 1.
Daily average of O3
concentrations from all monitors
within 35 km of participants
residential 3-digit postal codes
24-hr (YR) for 9
urban
areas/districts
mean range:
22.5-29.2
Min Range:
1.0-3.0
Max Range:
60.7-80.0
57-79
10 Canadian
cities
1981-
1999
1981-
1999
Vanos etal.,
2014
Respiratory Mortality
24-hr daily
average, lag 1.
Daily average O3 concentrations
from all monitors either downtown
or at city airports located within
27 km of downtown
24-hr (YR):
19.3
NA
51-94
ED - emergency department; HA - hospital admission; WS - warm season; YR - year round; ZCTA - ZIP code tabulation area
A Studies investigating associations between short-term O3 exposure and respiratory mortality are summarized in the following tables and figures of Appendix 3 of the ISA (U.S. EPA, 2020): HA for
asthma: Table 3-13, Figure 3-4; EDvisitsfor asthma: Table 3-14, Figure 3-5; ED visits for respiratory infection: Table 3-39, Figure 3-6; Respiratory-related HA and ED: Figure 3-7; HA for
aggregate respiratory diseases: Table 3-41; ED visits for aggregate respiratory diseases: Table 3-42.
B For those studies available at the time of the last review, design values were drawn from (Wells, 2012) and are presented in units of ppm. For those studies available since the time of the last
review, design values were calculated based on data available from the EPA's Air Quality System (AQS) for U.S. studies and the National Air Pollutant Surveillance (NAPS) program for Canadian
studies.
3B-7
-------
Table 3B-2. Subset of epidemiologic studies of associations between long-term, ozone and respiratory effects.
Study Information
Ambient Air
Quality Data
Study
Area
Health
Study
Time
Period
Air Quality
Time
Period
Study
Reference A
Health
Outcome
O3 Concentration Metric
Associated with Health
Outcome
Assignment of Monitors to
Study Subjects
Study-reported O3
Concentrations, in terms
of study metric (ppb)
Design Values
for Current
NAAQS,
across cities
and study
years (ppb)B
Mean/
median
Range
U.S. Studies, multi-city
Nationwide
1982-2000
1977-2000
Jerrettetal.,
2009
Respiratory
Mortality
Long-term warm-season
average O3 value including
year 1977-2000
Study participants assigned long-
term O3 concentrations for MSA
of residence c
Mean range
for MSAs:
33.33-104.0
NA
59-248
California
1982-2000
1988-2002
Jerrettetal.,
2013
Respiratory
Mortality
Monthly average O3 value
calculated using IDW from
year 1988-2002
Study participants were assigned
O3 concentration based on their
residential address
corresponding to the study siteD
50.35
17.11-
89.33
128-186
California
(9 areas)
1993-2001,
1996-2004,
2006-2014
1993 -2014
Garcia etal.,
2019
Asthma
diagnosis
Areawide annual mean O3
concentration (10am-6pm)
Community-specific annual mean
concentrations for each year of
each of the three cohorts.
-
26-76
65-165
[for 1993-2014]
Canadian Studies, multi-cit
/
Nationwide
1991-2011
2002-2009
Weichenthal
etal., 2017
Respiratory
Mortality
Monthly average O3 value
calculated using pollutant-
specific interpolation
techniques to generate 21
km2 grid cell concentrations
Study participants were assigned
O3 concentration from
interpolation surface based on
their residential postal codeE
38.29
<1-60.46
35-98
Quebec
1999-2010
1999-2010
Tetreaultet
al„ 2016
Asthma
incidence
Average summer (June-Aug)
concentration [8hr midday
concentration per ISA]
Study participants were assigned
concentration estimated for
postal code centroid using
interpolation based approach.
Mean: 32.07
Median:
32.19
12.18-
43.12
49-79
A Studies investigating associations between long-term O3 exposure and respiratory mortality are summarized in the ISA, Appendix 6, Table 6-8 and Figure 6-9 (U.S. EPA, 2020).
B For those studies available at the time of the last review, design values were drawn from (Wells, 2012). For those studies available since the time of the last review, design values were
calculated based on data available from the EPA's Air Quality System (AQS) for U.S. studies and the National Air Pollutant Surveillance (NAPS) program for Canadian studies.
c Data for monitors were obtained for 1977-2000. Daily maximum 1-hour O3 concentrations were used to calculate quarterly averages for each monitor. Averages for quarters 2 and 3 were then
averaged to create a warm-season average O3 concentration for each monitor. The warm-season O3 concentrations for the time period 1977-2000 were computed for each year to form a single
annual time series of O3 measurements for 96 metropolitan areas.
D Inverse distance weighted monthly average O3 concentrations for all sites within a 50 km radius of operating monitors were calculated for the years 1988-2002.
E A surface for average daily 8-hour maximum O3 concentrations was generated for the months of May-October for years 2002-2009 using an air pollution-specific interpolation technique to
generate a 21 km2 grid value. The interpolation method incorporates modeled O3 from the Canadian Hemispheric Regional Ozone and NOx (CFIRONOS) air quality forecast model with
observations from Canada and the U.S.
3B-8
-------
ATTACHMENT
DESIGN VALUES FOR LOCATIONS AND TIME PERIODS ANALYZED IN EPIDEMIOLOGIC
STUDIES
NOTE: Design values generally provided in parts per billion (ppb) rather than parts per million (ppm) in tables below for simplicity of
presentation.
Alhanti et al., 2016 (3019562) - ED Visits for Asthma
Three U.S. cities. O3: Atlanta (1993-2009), Dallas (2006-2009), St. Louis (2001-2007)
City
Census Area
Name
dv.1993
1995
dv.1994
.1996
dv.1995
.1997
dv.1996
.1998
dv.1997
.1999
dv.1998
.2000
dv.1999
.2001
dv.2000
.2002
dv.2001
.2003
dv.2002
.2004
dv.2003
.2005
dv.2004
.2006
dv.2005
.2007
dv.2006
.2008
dv.2007
.2009
Atlanta, GA
Atlanta-Sandy
Springs-
Roswell, GA
109
105
110
113
118
121
107
99
91
93
90
91
95
95
87
City
Census Area Name
dv.2006.2008
dv.2007.2009
Dallas, TX
Dallas-Fort Worth, TX-OK
91
86
City
Census Area Name
dv.2001.
2003
dv.2002.
2004
dv.2003.
2005
dv.2004.
2006
dv.2005.
2007
St. Louis
St. Louis-St. Charles-Farmington, MO-IL
92
89
86
86
89
Barry et al., 2018 (4829120) - ED Visits for Asthma, ED Visits for Aggregate Respiratory Diseases, ED Visit - Respiratory Infection
Five U.S. Cities: 20-co Atlanta (2002-2008), 7-co Birmingham (2002-2008), 12-co Dallas-Ft Worth (2006-2008), 3-co Pittsburgh
(2002-2008), 16-co St Louis (2002-2007)
City
Census Area Name
dv.2002.
2004
dv.2003.
2005
dv.2004.
2006
dv.2005.
2007
dv.2006.
2008
Atlanta, GA
Atlanta-Sandy Springs-Roswell, GA
93
90
91
95
95
3B-9
-------
City
Census Area Name
dv.2002.
2004
dv.2003.
2005
dv.2004.
2006
dv.2005.
2007
dv.2006.
2008
Birmingham, AL
Birmingham-Hoover-Talladega, AL
85
84
85
89
87
City
Census Area Name
dv.2006.2008
Dallas-Ft Worth
Dallas-Fort Worth, TX-OK
91
City
Census Area Name
dv.2002.2004
dv.2003.2005
dv.2004.2006
dv.2005.2007
dv.2006.2008
Pittsburgh, PA
Pittsburgh-New Castle-Weirton, PA-OH-WV
90
84
83
87
86
City
Census Area Name
dv.2002.2004
dv.2003.2005
dv.2004.2006
dv.2005.2007
St. Louis
St. Louis-St. Charles-Farmington, MO-IL
89
86
86
89
Byers et al., 2015 (3019032) - ED Visits for Asthma
Indianapolis MSA (Marion and 8 surrounding counties), IN, U.S. O3: 2007-2011
City
Census Area Name
dv.2007.2009
dv.2008.2010
dv.2009.2011
Indianapolis, IN
Indianapolis-Carmel-Muncie, IN
77
73
74
Cakmak et al., 2017 (4167344) - Long-Term Ozone and Respiratory Mortality
Nationwide, Canada. O3: 2002-2009
Air quality data are not described for this study as it relied on O3 concentrations for the years 2002-2009 as surrogates for study
population annual O3 concentrations during the 1984 to 2011 period (Cakmak, 2017).
Crouse et al., 2015 (3019335) - Long-Term Ozone and Respiratory Mortality
Nationwide, Canada. O3: 2002-2009
Air quality data are not described for this study as it relied on O3 concentrations for the years 2002-2009 as surrogates for study
population annual O3 concentrations during the 1984 to 2006 period (Crouse, 2015).
3B-10
-------
Darrow et al., 2011 (202800) - ED Visits for Aggregate Respiratory Diseases
20-county Atlanta area, GA, U.S. O3: 1993-2004
City
Census Area Name
dv1993_
1995
dv1994_
1996
dv1995_
1997
dv1996
1998
dv1997_
1999
dv1998_
2000
dv1999_
2001
dv2000_
2002
dv2001_
2003
dv2002_
2004
Atlanta, GA
Atlanta-Sandy Springs-Marietta,
GA
0.109
0.105
0.110
0.113
0.118
0.121
0.107
0.099
0.091
0.093
Note: Design values for this study were available in the last review (see Wells, 2012) and are presented in units of ppm, rather than ppb.
Darrow et al., 2014 (2526768) - ED Visit - Respiratory Infection
20-county Atlanta area, GA, U.S. O3: 1993-2010
City
Census Area Name
dv.1993.
1995
dv.1994.
1996
dv.1995.
1997
dv.1996.
1998
dv.1997.
1999
dv.1998.
2000
dv.1999.
2001
dv.2000.
2002
Atlanta, GA
Atlanta-Sandy Springs-Roswell, GA
109
105
110
113
118
121
107
99
dv.2001.
2003
dv.2002.
2004
dv.2003.
2005
dv.2004.
2006
dv.2005.
2007
dv.2006.
2008
dv.2007.
2009
dv.2008.
2010
91
93
90
91
95
95
87
80
Eckel et al., 2016 (3426159) - Long-Term Ozone and Respiratory Mortality
California, U.S. Os: 1988-2011
State
dv.1988.
1990
dv.1989.
1991
dv.1990.
1992
dv.1991.
1993
dv.1992.
1994
dv.1993.
1995
dv.1994.
1996
dv.1995.
1997
dv.1996.
1998
dv.1997.
1999
dv.1998.
2000
California
186
182
180
177
171
165
161
148
154
147
146
dv.1999.
2001
dv.2000.
2002
dv.2001.
2003
dv.2002.
2004
dv.2003.
2005
dv.2004.
2006
dv.2005.
2007
dv.2006.
2008
dv.2007.
2009
dv.2008.
2010
dv.2009.
2011
129
128
131
127
127
121
122
119
118
112
107
3B-11
-------
Garcia et al., 2019 (5119704) - Asthma Incidence
Nine communities in Southern California, U.S. O3: 1993-2014
City
Census Area Name
dv.1993
.1995
dv.1994
.1996
dv.1995
.1997
dv.1996
.1998
dv.1997
.1999
dv.1998
.2000
dv.1999.
2001
dv.2000
.2002
dv.2001
.2003
dv.2002
.2004
Long Beach, San
Dimas
Los Angeles-Long
Beach-Anaheim, CA
(CBSA)
156
145
135
133
118
115
105
113
126
125
Lake Elsinore,
Lake Gregory,
Mira Loma,
Riverside, Upland
Riverside-San
Bernardino-Ontario,
CA (CBSA)
165
161
148
154
147
146
129
128
131
127
Alpine
San Diego-Carlsbad,
CA (CBSA)
108
104
99
102
99
100
94
95
93
89
Santa Maria
Santa Maria-Santa
Barbara, CA (CBSA)
90
94
89
87
82
81
80
82
84
82
City
Census Area Name
dv.2003
.2005
dv.2004
.2006
dv.2005
.2007
dv.2006
.2008
dv.2007
.2009
dv.2008
.2010
dv.2009.
2011
dv.2010
.2012
dv.2011
.2013
dv.2012
.2014
Long Beach, San
Dimas
Los Angeles-Long
Beach-Anaheim, CA
(CBSA)
120
112
110
107
108
103
97
96
99
97
Lake Elsinore,
Lake Gregory,
Mira Loma,
Riverside, Upland
Riverside-San
Bernardino-Ontario,
CA (CBSA)
127
121
122
119
118
112
107
106
107
102
Alpine
San Diego-Carlsbad,
CA (CBSA)
86
88
89
92
89
88
82
91
80
79
Santa Maria
Santa Maria-Santa
Barbara, CA (CBSA)
78
75
75
73
77
76
73
68
65
68
3B-12
-------
Gleason et al., 2014 (2369662) - ED Visits for Asthma
New Jersey (statewide), U.S. O3: April-September, 2004-2007
State
dv.2004.2006
dv.2005.2007
New Jersey
93
92
Goodman et al., 2017a (3859548) - Hospital Admissions for Asthma,
New York City (20-mi radius from center), NY, U.S. O3: 1999-2009
City
Census Area Name
dv.1999
.2001
dv.2000
.2002
dv.2001
.2003
dv.2002
.2004
dv.2003
.2005
dv.2004
.2006
dv.2005
.2007
dv.2006
.2008
dv.2007
.2009
New York, NY
New York-Newark, NY-NJ-CT-PA
109
115
109
102
94
93
94
89
84
Goodman et al., 2017b (4169406) - Hospital Admissions for Asthma
Houston, Dallas, and Austin, TX metro areas, U.S. O3: 2003-2011
City
Census Area Name
dv.2003.
2005
dv.2004.
2006
dv.2005.
2007
dv.2006.
2008
dv.2007.
2009
dv.2008.
2010
dv.2009.
2011
Houston
Houston-The Woodlands, TX
103
103
96
91
84
84
89
Dallas
Dallas-Fort Worth, TX-OK
95
96
95
91
86
86
90
Austin
Austin-Round Rock, TX (CBSA ONLY)
82
82
80
77
75
74
75
Ito et al., 2007 (156594) - Emergency Department Visits for Asthma
New York City, NY. Os: 1999-2002
City
Census Area Name
dv.1999.2001
dv.2000.2002
New York, NY
New York-Northern New Jersey-Long Island, NY-NJ-PA
0.109
0.115
Note: Design va
than ppb.
ues for this study were available in the last review (see Wells, 2012) and are presented in units of ppm, rather
3B-13
-------
Jerrett et al., 2009 (194160) - Long-Term Ozone and Respiratory Mortality
Nationwide, U.S. O3: 1977-2000
City
Census Area Name
dv1977_
1979
dv1978_
1980
dv1979_
1981
dv1980_
1982
dv1981_
1983
dv1982_
1984
dv1983_
1985
Charleston, SC
Charleston-North Charleston-Summerville, SC
0.088
0.08
0.074
0.072
0.076
0.077
Charleston, WV
Charleston, WV
0.077
0.077
0.075
0.078
0.082
0.086
0.087
Charlotte, NC
Charlotte-Gastonia-Concord, NC-SC
0.1
0.099
0.097
0.098
Chattanooga, TN
Chattanooga, TN-GA
0.09
0.094
0.097
0.097
0.093
0.091
Chicago, IL
Chicago-Naperville-Joliet, IL-IN-WI
0.112
0.112
0.1
0.096
0.103
0.103
0.106
Cincinnati, OH
Cincinnati-Middletown, OH-KY-IN
0.119
0.109
0.104
0.101
0.1
0.1
0.097
Cleveland, OH
Cleveland-Elyria-Mentor, OH
0.108
0.101
0.094
0.092
0.096
0.098
0.1
Colorado Springs, CO
Colorado Springs, CO
0.06
0.06
0.063
0.062
Columbia, SC
Columbia, SC
0.078
0.109
0.091
0.087
0.088
0.084
0.082
Columbus, OH
Columbus, OH
0.098
0.103
0.091
0.093
0.092
0.094
0.093
Corpus Christi, TX
Corpus Christi, TX
0.079
0.086
0.084
Dallas/Ft Worth, TX
Dallas-Fort Worth-Arlington, TX
0.109
0.111
0.108
0.108
0.11
0.118
Dayton, OH
Dayton, OH
0.122
0.108
0.102
0.103
0.104
0.1
0.092
Denver, CO
Denver-Aurora-Broomfield, CO
0.091
0.089
0.088
0.084
0.089
0.087
0.082
Detroit, Ml
Detroit-Warren-Livonia, Ml
0.101
0.097
0.092
0.097
0.103
0.098
0.094
El Paso, TX
El Paso, TX
0.079
0.084
0.089
Evansville, IN
Evansville, IN-KY
0.096
0.094
0.092
Flint, Ml
Flint, Ml
0.082
0.086
0.082
0.085
0.088
0.087
0.08
Fresno, CA
Fresno, CA
0.101
0.103
0.123
0.123
0.116
0.114
0.11
Ft. Lauderdale, FL
Broward County, FL
0.074
0.075
0.071
0.069
0.069
City
Census Area Name
dv1977_
1979
dv1978_
1980
dv1979_
1981
dv1980_
1982
dv1981_
1983
dv1982_
1984
dv1983_
1985
Gary, IN
Lake County, IN
0.105
0.098
0.087
0.09
0.095
0.097
0.095
Greely, CO
Greeley, CO
0.059
0.071
0.069
Greensboro, NC
Greensboro-High Point, NC
0.086
0.09
0.087
0.089
0.087
Greenville, SC
Greenville-Mauldin-Easley, SC
0.094
0.094
0.093
0.089
0.088
Harrisburg, PA
Harrisburg-Carlisle, PA
0.095
0.087
0.096
0.098
0.1
0.098
Houston, TX
Houston-Sugar Land-Baytown, TX
0.099
0.14
0.132
0.124
0.139
0.128
0.124
3B-14
-------
Huntington, WV
Huntington-Ashland, WV-KY-OH
0.088
0.09
0.095
0.097
0.097
Indianapolis, IN
Indianapolis-Carmel, IN
0.076
0.09
0.087
0.103
0.101
0.101
0.096
Jackson, MS
Jackson, MS
0.098
0.09
0.084
0.081
0.079
0.076
0.078
Jacksonville, FL
Jacksonville, FL
0.086
0.086
0.087
0.085
0.08
0.076
0.075
Jersey City, NJ
Hudson County, NJ
0.111
Johnstown, PA
Johnstown, PA
0.1
0.107
0.1
0.097
0.087
0.087
0.087
Kansas City, MO
Kansas City, MO-KS
0.074
0.081
0.097
0.089
0.089
0.094
0.096
Kenosha, Wl
Kenosha County, Wl
0.095
0.103
0.097
0.1
Knoxville, TN
Knoxville, TN
0.09
0.088
0.083
Lancaster, PA
Lancaster, PA
0.088
0.096
0.092
0.096
0.101
0.1
0.098
Lansing, Ml
Lansing-East Lansing, Ml
0.086
0.073
0.08
0.08
0.076
Las Vegas, NV
Las Vegas-Paradise, NV
0.074
0.085
0.085
0.08
0.079
Lexington, KY
Lexington-Fayette, KY
0.091
0.087
0.085
0.086
0.091
0.092
Little Rock, AR
Little Rock-North Little Rock-Conway, AR
0.098
0.107
0.1
0.085
0.082
0.083
0.087
Los Angeles, CA
Los Angeles-Long Beach-Santa Ana, CA
0.174
0.248
0.229
0.21
0.204
0.225
0.226
Madison, Wl
Madison, Wl
0.096
0.102
0.095
0.088
0.078
0.076
0.078
Memphis, TN
Memphis, TN-MS-AR
0.102
0.103
0.085
0.096
0.097
0.092
0.092
Milwaukee, Wl
Milwaukee-Waukesha-West Allis, Wl
0.114
0.11
0.11
0.106
0.111
0.104
0.105
Minneapolis, MN
Minneapolis-St. Paul-Bloomington, MN-WI
0.08
0.079
0.076
0.073
0.073
Nashville, TN
Nashville-Murfreesboro-Franklin, TN
0.092
0.085
0.077
0.083
0.083
0.09
0.095
Nassau, NY
Nassau County, NY
New Haven, CT
New Haven-Milford, CT
0.135
0.127
0.118
0.121
0.13
0.136
0.128
New Orleans, LA
NewOrleans-Metairie-Kenner, LA
0.087
0.087
0.085
0.083
0.099
0.089
City
Census Area Name
dv1977
1979
dv1978
1980
dv1979
1981
dv1980
1982
dv1981
1983
dv1982
1984
dv1983
1985
New York City, NY
New York-Northern New Jersey-Long Island, NY-
NJ-PA
0.124
0.118
0.116
0.12
0.121
0.12
0.128
Newark, NJ
Essex County, NJ
Norfolk, VA
Virginia Beach-Norfolk-Newport News, VA-NC
0.1
0.101
0.091
0.091
0.096
0.095
0.093
Oklahoma City, OK
Oklahoma City, OK
0.089
0.093
0.084
0.084
0.087
0.085
0.089
Orlando, FL
Orlando-Kissimmee, FL
0.078
0.08
0.077
0.078
0.078
0.078
0.074
3B-15
-------
Philadelphia, PA
Philadelphia-Camden-Wilmington, PA-NJ-DE-MD
0.126
0.136
0.127
0.125
0.114
0.122
0.119
Phoenix, AZ
Phoenix-Mesa-Scottsdale, AZ
0.076
0.078
0.081
0.085
0.09
0.093
0.096
Pittsburgh, PA
Pittsburgh, PA
0.111
0.123
0.109
0.104
0.106
0.099
0.099
Portland, ME
Portland-South Portland-Biddeford, ME
0.107
0.11
0.116
Portland, OR
Portland-Vancouver-Beaverton, OR-WA
0.084
0.088
0.082
0.082
0.081
0.074
0.076
Portsmouth, NH
Rockingham County, NH
0.097
0.094
0.082
0.077
Providence, Rl
Providence-New Bedford-Fall River, RI-MA
0.121
0.124
0.124
0.121
0.115
0.121
0.121
Racine, Wl
Racine, Wl
0.093
0.112
0.108
0.109
0.113
0.112
0.111
Raleigh, NC
Raleigh-Cary, NC
0.088
0.091
0.089
0.085
0.087
Reading, PA
Reading, PA
0.098
0.105
0.109
0.114
0.106
0.102
0.1
Richmond, VA
Richmond, VA
0.084
0.098
0.098
0.099
Riverside, CA
Riverside-San Bernardino-Ontario, CA
0.239
0.245
0.235
0.217
0.21
0.209
0.211
Roanoke, VA
Roanoke, VA
0.083
0.086
0.084
Rochester, NY
Rochester, NY
0.093
0.091
0.084
0.086
0.09
0.091
0.09
Sacramento, CA
Sacramento-Arden Arcade-Roseville, CA
0.102
0.112
0.114
0.115
0.118
Salinas, CA
Salinas, CA
0.066
0.061
0.061
0.057
0.065
0.074
San Antonio, TX
San Antonio, TX
0.086
0.089
0.092
0.09
0.087
0.086
San Diego, CA
San Diego-Carlsbad-San Marcos, CA
0.115
0.118
0.141
0.137
0.13
0.126
0.132
San Francisco, CA
San Francisco-Oakland-Fremont, CA
0.085
0.092
0.086
0.091
0.089
0.091
0.096
San Jose, CA
San Jose-Sunnyvale-Santa Clara, CA
0.093
0.101
0.102
0.094
0.095
0.1
0.103
Seattle, WA
Seattle-Tacoma-Bellevue, WA
0.088
0.081
0.084
0.085
0.08
0.069
0.069
Shreveport, LA
Shreveport-Bossier City, LA
0.08
0.081
0.077
0.079
South Bend, IN
South Bend-Mishawaka, IN-MI
0.093
0.093
0.102
0.095
0.09
0.088
City
Census Area Name
dv1977
1979
dv1978
1980
dv1979
1981
dv1980
1982
dv1981
1983
dv1982
1984
dv1983
1985
Springfield, MA
Springfield, MA
0.1
0.112
St Louis, MO
St. Louis, MO-IL
0.122
0.117
0.109
0.101
0.107
0.111
0.113
Steubenville, OH
Weirton-Steubenville, WV-OH
0.098
0.099
0.088
0.083
0.073
0.071
0.064
Syracuse, NY
Syracuse, NY
Tacoma, WA
Seattle-Tacoma-Bellevue, WA
0.088
0.081
0.084
0.085
0.08
0.069
0.069
Tampa, FL
Tampa-St. Petersburg-Clearwater, FL
0.09
0.088
0.087
0.087
0.089
0.09
0.087
3B-16
-------
Toledo, OH
Toledo, OH
0.108
0.104
0.102
0.1
0.101
0.09
0.087
Trenton, NJ
Trenton-Ewing, NJ
0.116
0.117
0.12
Tucson, AZ
Tucson, AZ
0.07
0.074
0.074
0.082
0.081
0.082
0.079
Vallejo, CA
Vallejo-Fairfield, CA
0.068
0.069
0.063
0.074
0.072
0.074
0.075
Ventura, CA
Ventura County, CA
0.13
0.13
0.109
0.104
0.098
0.112
0.113
Washington, DC
Washington-Arlington-Alexandria, DC-VA-MD-WV
0.112
0.101
0.101
0.113
0.113
0.112
0.11
Wichita, KS
Wichita, KS
0.074
0.078
0.079
0.081
Wilmington, DE
New Castle County, DE
0.083
0.088
0.093
0.106
0.112
0.116
Worcester, MA
Worcester, MA
0.102
0.092
0.096
0.099
York, PA
York-Hanover, PA
0.105
0.107
0.098
0.096
0.097
0.098
0.099
Youngstown, OH
Youngstown-Warren-Boardman, OH-PA
0.097
0.093
0.089
Note: Design values for this study were available in the last review (see Wells, 2012) and are presented in units of ppm, rather than ppb.
Jerrett et al., 2009 (194160) - Long-Term Ozone and Respiratory Mortality (Continued)
City
Census Area Name
dv1984
1986
dv1985
1987
dv1986
1988
dv1987
1989
dv1988
1990
dv1989
1991
dv1990
1992
Charleston, SC
Charleston-North Charleston-Summerville, SC
0.081
0.085
0.09
0.087
0.083
0.076
0.074
Charleston, WV
Charleston, WV
0.084
0.087
0.099
0.094
0.089
0.081
0.074
Charlotte, NC
Charlotte-Gastonia-Concord, NC-SC
0.094
0.102
0.112
0.104
0.101
0.092
0.091
Chattanooga, TN
Chattanooga, TN-GA
0.089
0.089
0.094
0.092
0.09
0.086
0.083
Chicago, IL
Chicago-Naperville-Joliet, IL-IN-WI
0.098
0.101
0.112
0.114
0.114
0.104
0.099
Cincinnati, OH
Cincinnati-Middletown, OH-KY-IN
0.093
0.098
0.109
0.106
0.107
0.102
0.095
Cleveland, OH
Cleveland-Elyria-Mentor, OH
0.094
0.092
0.104
0.105
0.104
0.093
0.09
Colorado Springs, CO
Colorado Springs, CO
0.062
0.06
0.061
0.063
0.065
0.066
0.063
Columbia, SC
Columbia, SC
0.081
0.084
0.069
0.091
0.091
0.081
0.084
Columbus, OH
Columbus, OH
0.089
0.089
0.093
0.097
0.095
0.089
0.092
Corpus Christi, TX
Corpus Christi, TX
0.078
0.083
0.086
0.089
0.085
0.079
0.077
Dallas/Ft Worth, TX
Dallas-Fort Worth-Arlington, TX
0.113
0.108
0.101
0.1
0.105
0.105
0.099
Dayton, OH
Dayton, OH
0.088
0.09
0.095
0.096
0.092
0.086
0.082
Denver, CO
Denver-Aurora-Broomfield, CO
0.079
0.081
0.088
0.087
0.086
0.08
0.074
Detroit, Ml
Detroit-Warren-Livonia, Ml
0.089
0.093
0.1
0.099
0.099
0.096
0.091
El Paso, TX
El Paso, TX
0.096
0.096
0.092
0.088
0.083
0.08
0.079
3B-17
-------
City
Census Area Name
dv1984
1986
dv1985
1987
dv1986
1988
dv1987
1989
dv1988
1990
dv1989
1991
dv1990
1992
Evansville, IN
Evansville, IN-KY
0.09
0.094
0.099
0.1
0.099
0.091
0.088
Flint, Ml
Flint, Ml
0.077
0.079
0.09
0.091
0.09
0.085
0.081
Fresno, CA
Fresno, CA
0.117
0.118
0.121
0.115
0.11
0.108
0.108
Ft. Lauderdale, FL
Broward County, FL
0.073
0.073
0.077
0.076
0.079
0.075
0.073
Gary, IN
Lake County, IN
0.088
0.087
0.093
0.096
0.092
0.087
0.083
Greely, CO
Greeley, CO
0.067
0.068
0.07
0.072
0.074
0.075
0.072
Greensboro, NC
Greensboro-Hiqh Point, NC
0.089
0.089
0.1
0.097
0.1
0.088
0.085
Greenville, SC
Greenville-Mauldin-Easley, SC
0.085
0.089
0.091
0.09
0.085
0.075
0.075
Harrisburq, PA
Harrisburq-Carlisle, PA
0.091
0.096
0.103
0.103
0.098
0.094
0.091
Houston, TX
Flouston-Suqar Land-Baytown, TX
0.127
0.127
0.118
0.117
0.119
0.119
0.116
Huntinqton, WV
Huntinqton-Ashland, WV-KY-OFI
0.09
0.093
0.099
0.103
0.103
0.092
0.096
Indianapolis, IN
Indianapolis-Carmel, IN
0.09
0.091
0.096
0.098
0.095
0.091
0.089
Jackson, MS
Jackson, MS
0.077
0.076
0.077
0.075
0.079
0.076
0.076
Jacksonville, FL
Jacksonville, FL
0.075
0.081
0.084
0.086
0.084
0.081
0.079
Jersey City, NJ
Hudson County, NJ
0.104
0.109
0.117
0.118
0.115
0.107
0.104
Johnstown, PA
Johnstown, PA
0.085
0.087
0.097
0.097
0.093
0.086
0.083
Kansas City, MO
Kansas City, MO-KS
0.089
0.084
0.088
0.088
0.086
0.082
0.083
Kenosha, Wl
Kenosha County, Wl
0.089
0.098
0.111
0.114
0.114
0.104
0.099
Knoxville, TN
Knoxville, TN
0.094
0.087
0.097
0.093
0.094
0.086
0.089
Lancaster, PA
Lancaster, PA
0.09
0.091
0.097
0.097
0.093
0.09
0.09
Lansinq, Ml
Lansinq-East Lansinq, Ml
0.073
0.077
0.09
0.089
0.087
0.081
0.082
Las Veqas, NV
Las Veqas-Paradise, NV
0.08
0.083
0.082
0.081
0.078
0.078
0.076
Lexinqton, KY
Lexinqton-Fayette, KY
0.092
0.094
0.099
0.099
0.096
0.085
0.078
Little Rock, AR
Little Rock-North Little Rock-Conway, AR
0.087
0.089
0.09
0.085
0.082
0.079
0.08
Los Anqeles, CA
Los Anqeles-Lonq Beach-Santa Ana, CA
0.222
0.217
0.205
0.192
0.186
0.179
0.177
Madison, Wl
Madison, Wl
0.075
0.079
0.09
0.091
0.079
0.081
0.079
Memphis, TN
Memphis, TN-MS-AR
0.093
0.096
0.1
0.095
0.095
0.089
0.091
Milwaukee, Wl
Milwaukee-Waukesha-West Allis, Wl
0.095
0.105
0.113
0.117
0.105
0.101
0.095
Minneapolis, MN
Minneapolis-St. Paul-Bloominqton, MN-WI
0.071
0.073
0.077
0.08
0.079
0.075
0.071
Nashville, TN
Nashville-Murfreesboro-Franklin, TN
0.097
0.098
0.106
0.104
0.104
0.096
0.096
3B-18
-------
City
Census Area Name
dv1984
1986
dv1985
1987
dv1986
1988
dv1987
1989
dv1988
1990
dv1989
1991
dv1990
1992
Nassau, NY
Nassau County, NY
New Haven, CT
New Haven-Milford, CT
0.115
0.108
0.112
0.113
0.116
0.116
0.113
New Orleans, LA
NewOrleans-Metairie-Kenner, LA
0.089
0.088
0.094
0.09
0.085
0.077
0.08
New York City, NY
New York-Northern New Jersey-Lonq Island, NY-
0.119
0.122
0.129
0.129
0.128
0.122
0.116
Newark, NJ
Essex County, NJ
0.086
0.092
0.105
0.098
0.088
0.086
Norfolk, VA
Virqinia Beach-Norfolk-Newport News, VA-NC
0.087
0.089
0.095
0.093
0.091
0.084
0.086
Oklahoma City, OK
Oklahoma City, OK
0.087
0.084
0.085
0.087
0.087
0.086
0.084
Orlando, FL
Orlando-Kissimmee, FL
0.075
0.078
0.082
0.082
0.082
0.08
0.079
Philadelphia, PA
Philadelphia-Camden-Wilminqton, PA-NJ-DE-MD
0.119
0.123
0.132
0.123
0.12
0.113
0.107
Phoenix, AZ
Phoenix-Mesa-Scottsdale, AZ
0.09
0.086
0.081
0.077
0.082
0.083
0.091
Pittsburqh, PA
Pittsburqh, PA
0.09
0.093
0.104
0.107
0.098
0.092
0.088
Portland, ME
Portland-South Portland-Biddeford, ME
0.112
0.112
0.112
0.117
0.115
0.109
0.105
Portland, OR
Portland-Vancouver-Beaverton, OR-WA
0.085
0.086
0.085
0.077
0.085
0.082
0.091
Portsmouth, NH
Rockinqham County, NH
0.078
0.087
0.094
0.104
0.1
0.098
0.092
Providence, Rl
Providence-New Bedford-Fall River, RI-MA
0.114
0.107
0.113
0.108
0.108
0.107
0.105
Racine, Wl
Racine, Wl
0.102
0.107
0.12
0.124
0.11
0.098
0.088
Raleiqh, NC
Raleiqh-Cary, NC
0.087
0.092
0.104
0.099
0.093
0.089
0.086
Readinq, PA
Readinq, PA
0.092
0.096
0.104
0.105
0.102
0.096
0.094
Richmond, VA
Richmond, VA
0.095
0.097
0.104
0.103
0.097
0.087
0.087
Riverside, CA
Riverside-San Bernardino-Ontario, CA
0.21
0.2
0.188
0.188
0.185
0.182
0.18
Roanoke, VA
Roanoke, VA
0.083
0.087
0.095
0.092
0.085
0.076
0.074
Rochester, NY
Rochester, NY
0.09
0.091
0.099
0.099
0.095
0.092
0.09
Sacramento, CA
Sacramento-Arden Arcade-Roseville, CA
0.118
0.114
0.114
0.114
0.107
0.105
0.105
Salinas, CA
Salinas, CA
0.071
0.071
0.068
0.072
0.07
0.07
0.071
San Antonio, TX
San Antonio, TX
0.085
0.083
0.084
0.085
0.085
0.082
0.079
San Dieqo, CA
San Dieqo-Carlsbad-San Marcos, CA
0.125
0.124
0.121
0.125
0.129
0.125
0.118
San Francisco, CA
San Francisco-Oakland-Fremont, CA
0.093
0.089
0.087
0.089
0.087
0.084
0.082
San Jose, CA
San Jose-Sunnyvale-Santa Clara, CA
0.097
0.092
0.092
0.097
0.088
0.082
0.083
Seattle, WA
Seattle-Tacoma-Bellevue, WA
0.075
0.077
0.074
0.076
0.079
0.078
0.086
Shreveport, LA
Shreveport-Bossier City, LA
0.082
0.085
0.086
0.087
0.088
0.084
0.086
3B-19
-------
City
Census Area Name
dv1984
1986
dv1985
1987
dv1986
1988
dv1987
1989
dv1988
1990
dv1989
1991
dv1990
1992
South Bend, IN
South Bend-Mishawaka, IN-MI
0.081
0.088
0.092
0.093
0.087
0.08
0.083
Sprinqfield, MA
Sprinqfield, MA
0.102
0.096
0.106
0.109
0.115
0.107
0.105
St Louis, MO
St. Louis, MO-IL
0.103
0.102
0.114
0.111
0.102
0.098
0.098
Steubenville, OH
Weirton-Steubenville, WV-OH
0.062
0.069
0.086
0.09
0.088
0.085
0.083
Syracuse, NY
Syracuse, NY
0.083
0.096
0.092
0.088
0.083
0.083
Tacoma, WA
Seattle-Tacoma-Bellevue, WA
0.075
0.077
0.074
0.076
0.079
0.078
0.086
Tampa, FL
Tampa-St. Petersburq-Clearwater, FL
0.088
0.091
0.09
0.086
0.085
0.079
0.081
Toledo, OH
Toledo, OH
0.079
0.083
0.097
0.102
0.099
0.086
0.082
Trenton, NJ
Trenton-Ewinq, NJ
0.11
0.114
0.124
0.123
0.117
0.111
0.112
Tucson, AZ
Tucson, AZ
0.076
0.074
0.069
0.071
0.075
0.074
0.075
Vallejo, CA
Vallejo-Fairfield, CA
0.073
0.077
0.079
0.082
0.075
0.074
0.074
Ventura, CA
Ventura County, CA
0.116
0.114
0.131
0.132
0.13
0.126
0.117
Washinqton, DC
Washinqton-Arlinqton-Alexandria, DC-VA-MD-WV
0.104
0.11
0.116
0.115
0.107
0.1
0.1
Wichita, KS
Wichita, KS
0.077
0.076
0.08
0.08
0.081
0.075
0.074
Wilminqton, DE
New Castle County, DE
0.102
0.106
0.114
0.114
0.115
0.107
0.101
Worcester, MA
Worcester, MA
0.091
0.086
0.088
0.091
0.091
0.089
0.091
York, PA
York-Hanover, PA
0.093
0.094
0.1
0.099
0.099
0.094
0.093
Younqstown, OH
Younqstown-Warren-Boardman, OH-PA
0.085
0.089
0.101
0.103
0.099
0.09
0.091
Note: Desiqn values for this study were available in the last review (see Wells, 2012) and are presentee
in units of ppm, rather than ppb
Jerrett et al., 2009 (194160) - Long-Term Ozone and Respiratory Mortality (Continued)
City
Census Area Name
dv1991_
1993
dv1992_
1994
dv1993_
1995
dv1994_
1996
dv1995_
1997
dv1996_
1998
dv1997_
1999
dv1998_
2000
Charleston, SC
Charleston-North Charleston-Summerville,
SC
0.074
0.075
0.074
0.072
0.076
0.077
0.079
0.082
Charleston, WV
Charleston, WV
0.069
0.064
0.076
0.081
0.081
0.081
0.09
0.093
Charlotte, NC
Charlotte-Gastonia-Concord, NC-SC
0.091
0.092
0.094
0.094
0.097
0.103
0.104
0.104
Chattanooqa, TN
Chattanooqa, TN-GA
0.082
0.086
0.091
0.091
0.09
0.093
0.094
0.097
Chicaqo, IL
Chicaqo-Naperville-Joliet, IL-IN-WI
0.1
0.093
0.099
0.097
0.096
0.091
0.095
0.093
Cincinnati, OH
Cincinnati-Middletown, OH-KY-IN
0.091
0.091
0.098
0.099
0.095
0.092
0.095
0.094
3B-20
-------
Cleveland, OH
Cleveland-Elyria-Mentor, OH
0.092
0.093
0.098
0.1
0.099
0.098
0.099
0.095
Colorado Springs,
CO
Colorado Springs, CO
0.062
0.061
0.061
0.059
0.056
0.059
0.062
0.065
Columbia, SC
Columbia, SC
0.085
0.087
0.086
0.081
0.08
0.087
0.092
0.096
Columbus, OH
Columbus, OH
0.09
0.086
0.09
0.092
0.092
0.093
0.097
0.095
Corpus Christi, TX
Corpus Christi, TX
0.078
0.079
0.082
0.083
0.083
0.08
0.081
0.083
Dallas/Ft Worth, TX
Dallas-Fort Worth-Arlington, TX
0.095
0.096
0.106
0.104
0.104
0.098
0.101
0.102
Dayton, OH
Dayton, OH
0.084
0.086
0.092
0.093
0.091
0.093
0.093
0.09
Denver, CO
Denver-Aurora-Broomfield, CO
0.071
0.074
0.081
0.081
0.079
0.084
0.083
0.086
Detroit, Ml
Detroit-Warren-Livonia, Ml
0.089
0.088
0.093
0.094
0.092
0.093
0.095
0.089
El Paso, TX
El Paso, TX
0.078
0.081
0.084
0.089
0.08
0.082
0.078
0.08
Evansville, IN
Evansville, IN-KY
0.087
0.089
0.094
0.095
0.093
0.093
0.094
0.091
Flint, Ml
Flint, Ml
0.077
0.071
0.075
0.082
0.084
0.086
0.089
0.086
Fresno, CA
Fresno, CA
0.111
0.107
0.108
0.107
0.111
0.115
0.113
0.111
Ft. Lauderdale, FL
Broward County, FL
0.076
0.079
0.074
0.069
0.069
0.072
0.075
0.075
Gary, IN
Lake County, IN
0.08
0.077
0.084
0.091
0.095
0.09
0.091
0.088
Greely, CO
Greeley, CO
0.068
0.066
0.068
0.071
0.07
0.071
0.071
0.071
Greensboro, NC
Greensboro-High Point, NC
0.083
0.084
0.088
0.086
0.085
0.089
0.092
0.094
Greenville, SC
Greenville-Mauldin-Easley, SC
0.082
0.081
0.082
0.081
0.083
0.087
0.09
0.09
Harrisburg, PA
Harrisburg-Carlisle, PA
0.091
0.089
0.092
0.087
0.088
0.088
0.094
0.093
Houston, TX
Houston-Sugar Land-Baytown, TX
0.104
0.11
0.114
0.116
0.117
0.116
0.118
0.112
Huntington, WV
Huntington-Ashland, WV-KY-OH
0.092
0.09
0.096
0.091
0.088
0.092
0.095
0.094
Indianapolis, IN
Indianapolis-Carmel, IN
0.087
0.09
0.094
0.098
0.097
0.098
0.097
0.095
City
Census Area Name
dv1991_
1993
dv1992_
1994
dv1993_
1995
dv1994_
1996
dv1995_
1997
dv1996_
1998
dv1997_
1999
dv1998_
2000
Jackson, MS
Jackson, MS
0.074
0.075
0.076
0.077
0.077
0.08
0.081
0.083
Jacksonville, FL
Jacksonville, FL
0.079
0.081
0.08
0.078
0.081
0.088
0.088
0.085
Jersey City, NJ
Hudson County, NJ
0.103
0.096
0.1
0.095
0.098
0.093
0.1
0.092
Johnstown, PA
Johnstown, PA
0.084
0.08
0.085
0.085
0.088
0.091
0.093
0.091
Kansas City, MO
Kansas City, MO-KS
0.082
0.082
0.09
0.092
0.094
0.093
0.091
0.089
Kenosha, Wl
Kenosha County, Wl
0.1
0.093
0.099
0.097
0.096
0.09
0.095
0.093
3B-21
-------
Knoxville, TN
Knoxville, TN
0.088
0.089
0.093
0.093
0.095
0.1
0.104
0.104
Lancaster, PA
Lancaster, PA
0.093
0.091
0.096
0.093
0.096
0.096
0.101
0.097
Lansing, Ml
Lansing-East Lansing, Ml
0.081
0.079
0.082
0.084
0.083
0.08
0.082
0.082
Las Vegas, NV
Las Vegas-Paradise, NV
0.075
0.079
0.079
0.08
0.079
0.08
0.077
0.085
Lexington, KY
Lexington-Fayette, KY
0.077
0.079
0.087
0.087
0.085
0.085
0.087
0.085
Little Rock, AR
Little Rock-North Little Rock-Conway, AR
0.078
0.077
0.08
0.08
0.081
0.08
0.082
0.087
Los Angeles, CA
Los Angeles-Long Beach-Santa Ana, CA
0.177
0.168
0.156
0.145
0.135
0.133
0.118
0.115
Madison, Wl
Madison, Wl
0.073
0.072
0.072
0.08
0.081
0.078
0.08
0.078
Memphis, TN
Memphis, TN-MS-AR
0.09
0.09
0.091
0.094
0.095
0.093
0.095
0.097
Milwaukee, Wl
Milwaukee-Waukesha-West Allis, Wl
0.09
0.084
0.092
0.097
0.098
0.093
0.097
0.092
Minneapolis, MN
Minneapolis-St. Paul-Bloomington, MN-WI
0.07
0.07
0.072
0.074
0.072
0.07
0.074
0.074
Nashville, TN
Nashville-Murfreesboro-Franklin, TN
0.095
0.096
0.099
0.099
0.099
0.101
0.102
0.1
Nassau, NY
Nassau County, NY
New Haven, CT
New Haven-Milford, CT
0.108
0.097
0.105
0.101
0.107
0.1
0.103
0.096
New Orleans, LA
NewOrleans-Metairie-Kenner, LA
0.081
0.086
0.084
0.085
0.083
0.084
0.086
0.091
New York City, NY
New York-Northern New Jersey-Long
Island, NY-NJ-PA
0.108
0.1
0.106
0.104
0.108
0.104
0.107
0.107
Newark, NJ
Essex County, NJ
0.084
0.081
0.088
0.088
0.092
0.088
0.093
0
Norfolk, VA
Virginia Beach-Norfolk-Newport News, VA-
NC
0.09
0.088
0.087
0.083
0.087
0.09
0.094
0.089
Oklahoma City, OK
Oklahoma City, OK
0.081
0.081
0.084
0.085
0.083
0.085
0.086
0.084
Orlando, FL
Orlando-Kissimmee, FL
0.078
0.082
0.079
0.079
0.078
0.084
0.085
0.085
Philadelphia, PA
Philadelphia-Camden-Wilmington, PA-NJ-
DE-MD
0.106
0.099
0.104
0.101
0.11
0.107
0.11
0.106
Phoenix, AZ
Phoenix-Mesa-Scottsdale, AZ
0.088
0.086
0.089
0.09
0.092
0.091
0.088
0.088
City
Census Area Name
dv1991_
1993
dv1992_
1994
dv1993_
1995
dv1994_
1996
dv1995_
1997
dv1996_
1998
dv1997_
1999
dv1998_
2000
Pittsburgh, PA
Pittsburgh, PA
0.095
0.096
0.105
0.103
0.105
0.099
0.101
0.096
Portland, ME
Portland-South Portland-Biddeford, ME
0.102
0.095
0.096
0.092
0.096
0.092
0.092
0.084
Portland, OR
Portland-Vancouver-Beaverton, OR-WA
0.076
0.078
0.071
0.083
0.078
0.08
0.071
0.072
Portsmouth, NH
Rockingham County, NH
0.096
0.093
0.096
0.094
0.095
0.091
0.09
0.08
Providence, Rl
Providence-New Bedford-Fall River, RI-MA
0.099
0.092
0.097
0.094
0.097
0.09
0.092
0.088
3B-22
-------
Racine, Wl
Racine, Wl
0.086
0.082
0.088
0.089
0.092
0.088
0.091
0.085
Raleigh, NC
Raleigh-Cary, NC
0.087
0.086
0.087
0.087
0.089
0.096
0.103
0.101
Reading, PA
Reading, PA
0.094
0.086
0.088
0.089
0.092
0.091
0.096
0.092
Richmond, VA
Richmond, VA
0.091
0.092
0.093
0.087
0.09
0.092
0.099
0.091
Riverside, CA
Riverside-San Bernardino-Ontario, CA
0.177
0.171
0.165
0.161
0.148
0.154
0.147
0.146
Roanoke, VA
Roanoke, VA
0.077
0.08
0.082
0.078
0.078
0.085
0.09
0.089
Rochester, NY
Rochester, NY
0.088
0.08
0.085
0.081
0.083
0.08
0.086
0.081
Sacramento, CA
Sacramento-Arden Arcade-Roseville, CA
0.11
0.104
0.106
0.106
0.099
0.103
0.103
0.107
Salinas, CA
Salinas, CA
0.069
0.07
0.069
0.067
0.065
0.066
0.062
0.064
San Antonio, TX
San Antonio, TX
0.079
0.082
0.087
0.087
0.087
0.085
0.088
0.086
San Diego, CA
San Diego-Carlsbad-San Marcos, CA
0.112
0.109
0.108
0.104
0.099
0.102
0.099
0.1
San Francisco, CA
San Francisco-Oakland-Fremont, CA
0.081
0.082
0.087
0.093
0.09
0.089
0.086
0.087
San Jose, CA
San Jose-Sunnyvale-Santa Clara, CA
0.08
0.08
0.083
0.088
0.085
0.086
0.082
0.082
Seattle, WA
Seattle-Tacoma-Bellevue, WA
0.077
0.074
0.071
0.076
0.078
0.081
0.074
0.075
Shreveport, LA
Shreveport-Bossier City, LA
0.085
0.086
0.083
0.08
0.082
0.084
0.089
0.092
South Bend, IN
South Bend-Mishawaka, IN-MI
0.089
0.087
0.089
0.094
0.094
0.092
0.092
0.088
Springfield, MA
Springfield, MA
0.1
0.095
0.094
0.092
0.097
0.096
0.099
0.089
St Louis, MO
St. Louis, MO-IL
0.091
0.091
0.098
0.104
0.1
0.095
0.095
0.094
Steubenville, OH
Weirton-Steubenville, WV-OH
0.085
0.08
0.087
0.086
0.085
0.084
0.087
0.083
Syracuse, NY
Syracuse, NY
0.087
0.081
0.082
0.079
0.079
0.077
0.082
0.08
Tacoma, WA
Seattle-Tacoma-Bellevue, WA
0.077
0.074
0.071
0.076
0.078
0.081
0.074
0.075
Tampa, FL
Tampa-St. Petersburg-Clearwater, FL
0.08
0.08
0.08
0.081
0.082
0.088
0.09
0.088
Toledo, OH
Toledo, OH
0.085
0.086
0.09
0.091
0.089
0.089
0.086
0.084
Trenton, NJ
Trenton-Ewing, NJ
0.111
0.105
0.104
0.1
0.101
0.097
0.104
0.102
City
Census Area Name
dv1991_
1993
dv1992_
1994
dv1993_
1995
dv1994_
1996
dv1995_
1997
dv1996_
1998
dv1997_
1999
dv1998_
2000
Tucson, AZ
Tucson, AZ
0.077
0.078
0.081
0.079
0.079
0.077
0.075
0.073
Vallejo, CA
Vallejo-Fairfield, CA
0.074
0.073
0.077
0.079
0.078
0.082
0.085
0.085
Ventura, CA
Ventura County, CA
0.115
0.112
0.117
0.119
0.115
0.112
0.106
0.105
Washington, DC
Washington-Arlington-Alexandria, DC-VA-
MD-WV
0.101
0.096
0.098
0.094
0.1
0.101
0.106
0.101
3B-23
-------
Wichita, KS
Wichita, KS
0.068
0.065
0.07
0.072
0.074
0.078
0.08
0.08
Wilmington, DE
New Castle County, DE
0.098
0.099
0.103
0.098
0.099
0.095
0.1
0.097
Worcester, MA
Worcester, MA
0.095
0.095
0.089
0.087
0.087
0.094
0.088
York, PA
York-Hanover, PA
0.091
0.085
0.086
0.083
0.087
0.09
0.094
0.093
Youngstown, OH
Youngstown-Warren-Boardman, OH-PA
0.091
0.089
0.091
0.092
0.093
0.096
0.096
0.092
Note: Design values for this study were available in the last review (see Wells, 2012) and are presentee
in units of ppm, rather than ppb.
Jerrett et al., 2013 (2094363) - Long-Term Ozone and Respiratory Mortality
California, U.S. Os: 1988-2002
State
dv.1988
.1990
dv.1989
.1991
dv.1990
.1992
dv.1991
.1993
dv.1992
.1994
dv.1993
.1995
dv.1994
.1996
dv.1995
.1997
dv.1996
.1998
dv.1997
.1999
dv.1998
.2000
dv.1999
.2001
dv.2000
.2002
California
186
182
180
177
171
165
161
148
154
147
146
129
128
Katsouyanni et al., 2009 (199899) - Short-Term Ozone and Respiratory Mortality
Nationwide, U.S. O3: 1987-1996
City
Census Area Name
dv1987
1989
dv1988
1990
dv1989
1991
dv1990
1992
dv1991
1993
dv1992
1994
dv1993
1995
dv1994
1996
Honolulu, HI
Honolulu, HI
0.020
0.018
Lincoln, NE
Lincoln, NE
0.058
0.061
0.058
0.061
0.058
0.059
0.057
0.058
Colorado Springs,
Colorado Springs, CO
0.063
0.066
0.063
0.062
0.061
0.061
0.056
Des Moines, IA
Des Moines-West Des Moines, IA
0.062
Spokane, WA
Spokane, WA
0.064
0.066
Omaha, NE
Omaha-Council Bluffs, NE-IA
0.077
0.078
0.072
0.071
0.065
0.062
0.062
0.067
Albuquerque, NM
Albuquerque, NM
0.073
0.073
0.071
0.071
0.069
0.070
0.071
0.074
Wichita, KS
Wichita, KS
0.080
0.081
0.075
0.073
0.067
0.065
0.065
0.072
Mobile, AL
Mobile, AL
0.078
0.080
0.062
0.064
0.070
0.074
0.075
0.077
Minneapolis, MN
Minneapolis-St. Paul-Bloomington, MN-
0.080
0.079
0.068
0.070
0.069
0.070
0.072
0.074
Tucson, AZ
Tucson, AZ
0.068
0.074
0.069
0.072
0.077
0.078
0.081
0.079
Jackson, MS
Jackson, MS
0.075
0.079
0.076
0.076
0.074
0.075
0.076
0.077
Seattle, WA
Seattle-Tacoma-Bellevue, WA
0.076
0.079
0.078
0.086
0.077
0.074
0.071
0.076
3B-24
-------
City
Census Area Name
dv1987_
1989
dv1988_
1990
dv1989_
1991
dv1990_
1992
dv1991_
1993
dv1992_
1994
dv1993_
1995
dv1994_
1996
Tacoma, WA
Seattle-Tacoma-Bellevue, WA
0.076
0.079
0.078
0.086
0.077
0.074
0.071
0.076
Miami, FL
Miami-Fort Lauderdale-Pompano Beach,
0.083
0.079
0.075
0.073
0.076
0.080
0.080
0.074
Las Vegas, NV
Las Vegas-Paradise, NV
0.081
0.078
0.078
0.076
0.075
0.079
0.079
0.080
Madison, Wl
Madison, Wl
0.091
0.079
0.081
0.079
0.073
0.072
0.072
0.080
Portland, OR
Portland-Vancouver-Beaverton, OR-WA
0.077
0.085
0.082
0.091
0.076
0.078
0.058
0.083
Denver, CO
Denver-Aurora-Broomfield, CO
0.087
0.086
0.080
0.074
0.071
0.074
0.081
0.081
Little Rock, AR
Little Rock-North Little Rock-Conway, AR
0.085
0.082
0.079
0.080
0.078
0.077
0.080
0.080
Orlando, FL
Orlando-Kissimmee, FL
0.082
0.082
0.080
0.079
0.078
0.082
0.079
0.079
Salt Lake City, UT
Salt Lake City, UT
0.085
0.082
0.078
0.075
0.076
0.079
0.082
0.089
Jacksonville, FL
Jacksonville, FL
0.086
0.084
0.081
0.079
0.079
0.081
0.080
0.078
Corpus Christi, TX
Corpus Christi, TX
0.089
0.085
0.079
0.077
0.078
0.079
0.082
0.083
St. Petersburg, FL
Tampa-St. Petersburg-Clearwater, FL
0.086
0.085
0.079
0.081
0.080
0.080
0.080
0.081
Tampa, FL
Tampa-St. Petersburg-Clearwater, FL
0.086
0.085
0.079
0.081
0.080
0.080
0.080
0.081
Huntsville, AL
Huntsville, AL
0.087
0.083
0.077
0.082
0.085
0.083
0.080
0.078
El Paso, TX
El Paso, TX
0.088
0.083
0.080
0.079
0.078
0.081
0.084
0.089
San Antonio, TX
San Antonio, TX
0.085
0.085
0.082
0.079
0.079
0.082
0.087
0.087
New Orleans, LA
NewOrleans-Metairie-Kenner, LA
0.090
0.085
0.077
0.080
0.081
0.086
0.084
0.085
Austin, TX
Austin-Round Rock, TX
0.084
0.086
0.084
0.084
0.081
0.082
0.084
0.084
Oklahoma City, OK
Oklahoma City, OK
0.087
0.087
0.086
0.084
0.081
0.081
0.084
0.085
Syracuse, NY
Syracuse, NY
0.092
0.088
0.083
0.083
0.087
0.081
0.082
0.079
Shreveport, LA
Shreveport-Bossier City, LA
0.087
0.088
0.084
0.086
0.085
0.086
0.083
0.080
San Jose, CA
San Jose-Sunnyvale-Santa Clara, CA
0.097
0.088
0.082
0.083
0.080
0.080
0.083
0.088
Kansas City, MO
Kansas City, MO-KS
0.088
0.086
0.082
0.083
0.082
0.082
0.090
0.092
Oakland, CA
San Francisco-Oakland-Fremont, CA
0.089
0.087
0.084
0.082
0.081
0.082
0.087
0.093
San Francisco, CA
San Francisco-Oakland-Fremont, CA
0.089
0.087
0.084
0.082
0.081
0.082
0.087
0.093
Phoenix, AZ
Phoenix-Mesa-Scottsdale, AZ
0.077
0.082
0.083
0.091
0.088
0.086
0.089
0.090
Lexington, KY
Lexington-Fayette, KY
0.099
0.096
0.085
0.078
0.077
0.079
0.087
0.087
Tulsa, OK
Tulsa, OK
0.089
0.090
0.087
0.087
0.082
0.083
0.088
0.091
Stockton, CA
Stockton, CA
0.093
0.090
0.087
0.088
0.088
0.087
0.086
0.085
Rochester, NY
Rochester, NY
0.099
0.095
0.092
0.090
0.088
0.080
0.085
0.081
3B-25
-------
City
Census Area Name
dv1987_
1989
dv1988_
1990
dv1989_
1991
dv1990_
1992
dv1991_
1993
dv1992_
1994
dv1993_
1995
dv1994_
1996
Dayton, OH
Dayton, OH
0.096
0.092
0.086
0.082
0.084
0.086
0.092
0.093
Greensboro, NC
Greensboro-High Point, NC
0.097
0.100
0.088
0.085
0.083
0.084
0.088
0.086
Ft. Wayne, IN
Fort Wayne, IN
0.094
0.092
0.087
0.085
0.085
0.088
0.089
0.093
Buffalo, NY
Buffalo-Niagara Falls, NY
0.100
0.095
0.089
0.088
0.086
0.083
0.087
0.086
Raleigh, NC
Raleigh-Cary, NC
0.099
0.093
0.089
0.086
0.087
0.086
0.087
0.087
Newark, NJ
Essex County, NJ
0.105
0.098
0.088
0.086
0.084
0.081
0.088
0.088
Toledo, OH
Toledo, OH
0.102
0.099
0.086
0.082
0.085
0.086
0.090
0.091
Knoxville, TN
Knoxville, TN
0.093
0.094
0.086
0.089
0.088
0.089
0.093
0.093
Columbus, OH
Columbus, OH
0.097
0.095
0.089
0.092
0.090
0.086
0.090
0.092
Birmingham, AL
Birmingham-Hoover, AL
0.094
0.093
0.084
0.088
0.089
0.092
0.096
0.096
Worcester, MA
Worcester, MA
0.091
0.091
0.089
0.091
0.095
0.095
0.089
Memphis, TN
Memphis, TN-MS-AR
0.095
0.095
0.089
0.091
0.090
0.090
0.091
0.094
Grand Rapids, Ml
Grand Rapids-Wyoming, Ml
0.105
0.103
0.096
0.090
0.085
0.081
0.086
0.089
Indianapolis, IN
Indianapolis-Carmel, IN
0.098
0.095
0.091
0.089
0.087
0.090
0.094
0.098
Madera, CA
Madera-Chowchilla, CA
0.091
0.096
0.091
0.093
0.093
Detroit, Ml
Detroit-Warren-Livonia, Ml
0.099
0.099
0.096
0.091
0.089
0.088
0.093
0.094
Baton Rouge, LA
Baton Rouge, LA
0.098
0.101
0.099
0.096
0.090
0.087
0.091
0.094
Modesto, CA
Modesto, CA
0.102
0.099
0.095
0.092
0.086
0.093
0.095
0.096
Charlotte, NC
Charlotte-Gastonia-Concord, NC-SC
0.104
0.101
0.092
0.091
0.091
0.092
0.094
0.094
Louisville, KY
Louisville/Jefferson County, KY-IN
0.098
0.099
0.096
0.092
0.094
0.094
0.100
0.094
Akron, OH
Akron, OH
0.112
0.109
0.099
0.093
0.094
0.088
0.090
0.089
Boston, MA
Boston-Cambridge-Quincy, MA-NH
0.105
0.101
0.098
0.092
0.096
0.093
0.096
0.094
Cleveland, OH
Cleveland-Elyria-Mentor, OH
0.105
0.104
0.093
0.090
0.092
0.093
0.098
0.100
Milwaukee, Wl
Milwaukee-Waukesha-West Allis, Wl
0.117
0.105
0.101
0.095
0.090
0.084
0.092
0.097
Pittsburgh, PA
Pittsburgh, PA
0.107
0.098
0.092
0.088
0.095
0.096
0.105
0.103
Cincinnati, OH
Cincinnati-Middletown, OH-KY-IN
0.106
0.107
0.102
0.095
0.091
0.091
0.098
0.099
Nashville, TN
Nashville-Murfreesboro-Franklin, TN
0.104
0.104
0.096
0.096
0.095
0.096
0.099
0.099
St Louis, MO
St. Louis, MO-IL
0.111
0.102
0.098
0.098
0.091
0.091
0.098
0.104
Dallas/Ft Worth, TX
Dallas-Fort Worth-Arlington, TX
0.100
0.105
0.105
0.099
0.095
0.096
0.106
0.104
Providence, Rl
Providence-New Bedford-Fall River, Rl-
0.108
0.108
0.107
0.105
0.099
0.092
0.097
0.094
3B-26
-------
City
Census Area Name
dv1987_
1989
dv1988_
1990
dv1989_
1991
dv1990_
1992
dv1991_
1993
dv1992_
1994
dv1993_
1995
dv1994_
1996
Washington, DC
Washington-Arlington-Alexandria, DC-VA-
MD-WV
0.115
0.107
0.100
0.100
0.101
0.096
0.098
0.094
Chicago, IL
Chicago-Naperville-Joliet, IL-IN-WI
0.114
0.114
0.104
0.099
0.100
0.093
0.099
0.097
Jersey City, NJ
Hudson County, NJ
0.118
0.115
0.107
0.104
0.103
0.096
0.100
0.095
Atlanta, GA
Atlanta-Sandy Springs-Marietta, GA
0.113
0.107
0.104
0.105
0.101
0.101
0.109
0.105
Sacramento, CA
Sacramento-Arden Arcade-Roseville, CA
0.114
0.107
0.105
0.105
0.110
0.104
0.106
0.106
Baltimore, MD
Baltimore-Towson, MD
0.125
0.115
0.104
0.106
0.107
0.103
0.107
0.105
Philadelphia, PA
Philadelphia-Camden-Wilmington, PA-NJ-
DE-MD
0.123
0.120
0.113
0.107
0.106
0.099
0.104
0.101
Fresno, CA
Fresno, CA
0.115
0.110
0.108
0.108
0.111
0.107
0.108
0.107
New York City, NY
New York-Northern New Jersey-Long
0.129
0.128
0.122
0.116
0.108
0.100
0.106
0.104
Houston, TX
Houston-Sugar Land-Baytown, TX
0.117
0.119
0.119
0.116
0.104
0.110
0.114
0.116
Bakersfield, CA
Bakersfield, CA
0.116
0.112
0.118
0.115
0.112
0.111
0.119
0.119
San Diego, CA
San Diego-Carlsbad-San Marcos, CA
0.125
0.129
0.125
0.118
0.112
0.109
0.108
0.104
Anaheim, CA
Orange County, CA
0.141
0.138
0.127
0.120
0.114
0.117
0.107
0.100
Los Angeles, CA
Los Angeles-Long Beach-Santa Ana, CA
0.192
0.186
0.179
0.177
0.177
0.168
0.156
0.145
Riverside, CA
Riverside-San Bernardino-Ontario, CA
0.188
0.185
0.182
0.180
0.177
0.171
0.165
0.161
San Bernardino, CA
San Bernardino County, CA
0.188
0.185
0.182
0.180
0.177
0.171
0.165
0.161
Anchorage, AK
Anchorage, AK
Note: Design values for this study were available in the last review (see Wells, 2012) and are presented in units 0
"ppm, rather than ppb.
Klemm et al., 2011 (1011160) - Short-Term Ozone and Respiratory Mortality
Atlanta (Fulton, DeKalb, Gwinnet & Cobb counties), GA, U.S. O3: 8/1998 - 12/2007
City
Census Area Name
dv.1998.
2000
dv.1999.
2001
dv.2000.
2002
dv.2001.
2003
dv.2002.
2004
dv.2003.
2005
dv.2004.2
006
dv.2005.
2007
Atlanta, GA
Atlanta-Sandy
Springs-Roswell, GA
121
107
99
91
93
90
91
95
3B-27
-------
Kousha and Rowe, 2014 (2443421) - ED Visit - Respiratory Infection
Edmonton, Canada. O3: 1992-2002
City
dv.1992.
1994
dv.1993.
1995
dv.1994.
1996
dv.1995.
1997
dv.1996.
1998
dv.1997.
1999
dv.1998.
2000
dv.1999.
2001
dv.2000.
2002
Edmonton
60
61
58
56
62
64
64
64
65
Kousha and Castner, 2016 (3160295) - ED Visit - Respiratory Infection
Windsor, Canada. O3: 2004-2010
City
dv.2004.2006
dv.2005.2007
dv.2006.2008
dv.2007.2009
dv.2008.2010
Windsor
80
87
84
80
73
Malig et al., 2016 (3285875) - ED Visits for Asthma, ED Visits Aggregate Respiratory Diseases, ED Visit for Respiratory Infection
California (statewide), U.S. O3: 2005-2008
State
dv.2005.2007
dv.2006.2008
California
122
119
Nishimura et al., 2013 (1632336)
Four U.S. cities (Chicago, Houston, New York, San Francisco) and Puerto Rico.
This is a case control study with study participants, aged 8-21 years, identified during 2006-2011. Associations examined for annual
average O3 concentration (1-h max; 8-h max, per ISA), averaged across first three years of life. Median birth year was 1996.
O'Lenick et al., 2017 (3421578) - ED Visits for Asthma
20-county Atlanta metro area, GA, U.S. O3: 2002-2008
City
Census Area Name
dv.2002.2004
dv.2003.2005
dv.2004.2006
dv.2005.2007
dv.2006.2008
Atlanta, GA
Atlanta- Athens-Clarke County-Sandy Springs, GA
93
90
91
95
95
O'Lenick et al., 2017 (3859553) - ED Visits Aggregate Respiratory Diseases
20-co Atlanta, GA; 12-co Dallas, TX, and 16-co St. Louis, MO, U.S. O3: 2002-2008
City
Census Area Name
dv.2002.
2004
dv.2003.
2005
dv.2004.
2006
dv.2005.
2007
dv.2006.
2008
Atlanta, GA
Atlanta-Athens-Clarke County-Sandy Springs, GA
93
90
91
95
95
3B-28
-------
City
Census Area Name
dv.2002.2004
dv.2003.2005
dv.2004.2006
dv.2005.2007
dv.2006.2008
Dallas, TX
Dallas-Fort Worth, TX-OK
98
95
96
95
91
City
Census Area Name
dv.2002.2004
dv.2003.2005
dv.2004.2006
dv.2005.2007
dv.2006.2008
St. Louis, MO
St. Louis-St. Charles-Farmington, MO-IL
89
86
86
89
85
Rodopoulou et al., 2015 (2965674) - ED Visit for Respiratory Infection
Little Rock, AK, U.S. Os: 2002-2012
City
Census Area
Name
dv.2002
.2004
dv.2003
.2005
dv.2004
.2006
dv.2005
.2007
dv.2006
.2008
dv.2007
.2009
dv.2008
.2010
dv.2009
.2011
dv.2010
.2012
Little Rock, AK
Little Rock-North
Little Rock, AR
78
77
80
83
80
73
70
74
77
Sacks et al., 2014 (2228782) - ED Visits for Asthma
North Carolina (Statewide), U.S. O3: 2006-2008
State
dv.2006.2008
North Carolina
94
Sarnat et al., 2013 (1640373) - ED Visits for Asthma
Metro Atlanta area (186 zip codes), GA, U.S. O3: 1999-2002
City
Census Area Name
dv.1999.2001
dv.2000.2002
Atlanta, GA
Atlanta-Sandy Springs-Roswell, GA
107
99
3B-29
-------
Sarnat et al., 2015 (2772940) - ED Visits for Asthma
St. Louis metro area, MO (8 MO counties, 8 IL counties), U.S. O3: 2001-2003
City
Census Area Name
dv.2001.2003
dv.2002.2004
St. Louis
St. Louis-St. Charles-Farmington, MO-IL
92
92
Sheffield et al., 2015 (3025138) - ED Visits for Asthma
New York City (all boroughs), NY, U.S. O3: May-Sept. 2005-2011
City
Census Area Name
dv.2005.2007
dv.2006.2008
dv.2007.2009
dv.2008.2010
dv.2009.2011
dv.2010.2012
New York, NY
New York-Newark, NY-NJ-CT-PA
94
89
84
82
84
85
Shmool et al., 2016 (3288326) - ED Visits for Asthma
New York City, NY, U.S. Os: June-Aug 2005-2011
City
Census Area Name
dv.2005.2007
dv.2006.2008
dv.2007.2009
dv.2008.2010
dv.2009.2011
New York, NY
New York-Newark, NY-NJ-CT-PA
94
89
84
82
84
Silverman and Ito, 2010 (386252) HA for Asthma
New York, NY. Os: 1999-2006
City
Census Area Name
dv.1999.2001
dv.2000.2002
dv.2001.2003
dv.2002.2004
dv.2003.2005
dv.2004.2006
New York
City, NY
New York-Northern New Jersey-
Long Island, NY-NJ-PA
0.109
0.115
0.109
0.102
0.094
0.093
Note: Design va
ues for this study were available in the last review (see Wells, 2012) and are presented in units of ppm, rather than ppb.
Stieb et al., 2009 (195858) - ED Visits for Asthma
7 Canadian cities
O3:1992-
2003City
dv1992
1994
dv1993
1995
dv1994
1996
dv1995
1997
dv1996
1998
dv1997
1999
dv1998
2000
dv1999
2001
dv2000
2002
dv2001
2003
Montreal
77
73
73
72
Ottawa
64
64
63
66
65
69
63
Edmonton
60
61
58
56
62
64
64
64
65
3B-30
-------
Saint John
51
54
58
Halifax
54
Toronto
79
81
85
Vancouver
52
54
57
Strickland et al., 2014 (2519636) - ED Visits for Asthma
20-county Atlanta area, GA, U.S. O3: 2002-2010
City
Census Area Name
dv.2002.
2004
dv.2003.
2005
dv.2004.
2006
dv.2005.
2007
dv.2006.
2008
dv.2007.
2009
dv.2008.
2010
Atlanta, GA
Atlanta-Sandy Springs-Roswell, GA
93
90
91
95
95
87
80
Szyszkowicz et al., 2018 (4245266) - ED Visits for Asthma, [ED Visit - Respiratory Infection]
Multicity (9), Canada. O3: 2004-2011
City
dv.2004.2006
dv.2005.2007
dv.2006.2008
dv.2007.2009
dv.2008.2010
dv.2009.2011
Algoma
67
70
68
65
62
59
Oakville
73
78
75
73
70
69
Burlington
70
74
73
70
68
66
Hamilton
73
75
73
72
70
69
London
69
72
71
68
65
64
Parkhill
Longwoods
Ottawa
64
69
66
64
62
57
Brampton
74
78
75
73
68
67
Mississauga
-
79
-
-
65
64
Toronto
74
79
76
74
73
70
Essex
-
79
74
New Market
77
79
75
75
70
69
Stouffville
Note: Some of the locations named as city in the study appear as Municipality in NAPS dataset from Canada and included few ot
boundary. In such instances, DV data (if available) were pulled for all the cities included within those municipalities, e.g., Halton ir
Burlington), Middlesex included (London, Parkhill, Longwoods), Peel included (Toronto, Brampton, Mississauga), York included (
Stouffville).
her cities within its
lcluded (Oakville,
New Market,
3B-31
-------
Tolbert et al., 2007 (90316) - ED Visits for Aggregate Respiratory Diseases
Atlanta, GA. Os: 1993-2004
City
Census Area Name
dv1993_
1995
dv1994_
1996
dv1995_
1997
dv1996
1998
dv1997_
1999
dv1998_
2000
dv1999_
2001
dv2000_
2002
dv2001_
2003
dv2002_
2004
Atlanta, GA
Atlanta-Sandy Springs-Marietta, GA
0.109
0.105
0.110
0.113
0.118
0.121
0.107
0.099
0.091
0.093
Note: Design values for this study were available in the last review (see Wells, 2012) and are presented in units of ppm, rather than ppb.
Tetreault et al., 2016 (3073711) - Asthma Incidence
Quebec Province, Canada. O3: 1996-2011
City
dv.1996.
1998
dv.1997.
1999
dv.1998.
2000
dv.1999.
2001
dv.2000.
2002
dv.2001.
2003
dv.2002.
2004
dv.2003.
2005
dv.2004.
2006
dv.2005.
2007
dv.2006.
2008
dv.2007.
2009
dv.2008.
2010
dv.2009.
2011
Montreal
69
77
72
72
72
79
72
70
66
70
67
65
61
58
Quebec
61
65
59
60
63
70
64
59
56
63
60
58
55
54
Laval
72
75
67
68
68
75
68
67
62
65
62
61
59
60
Brassard
-
-
-
-
-
-
-
-
-
-
-
-
-
-
Longueuil
70
76
70
70
68
74
71
68
64
65
62
61
60
60
Terrebonne
-
-
-
-
-
-
-
-
-
-
-
-
-
-
Gatineau
-
75
72
73
69
71
67
66
66
-
-
-
59
55
Levis
-
-
-
-
-
-
-
-
-
-
59
57
52
50
Sherbrooke
-
-
-
-
-
-
-
-
63
63
60
59
57
56
Saguenay
-
-
-
-
-
-
-
54
54
57
56
54
52
51
Rouyn-Noranda
-
-
-
-
-
-
-
-
59
63
59
58
55
54
Trois-Rivieres
-
-
-
68
65
70
64
64
59
64
59
58
55
-
St. Zephirin-de-
Courval (MUNI)
72
75
67
71
73
79
73
70
66
69
65
62
60
60
Forestville
55
-
-
-
-
-
-
-
-
-
-
-
-
-
Charette (MUNI)
70
71
62
65
64
68
63
61
58
62
61
61
58
55
Saint-Remi
67
-
-
-
-
-
-
-
-
-
-
-
-
-
Saint-Simon (MUNI)
66
71
65
65
64
70
66
62
58
59
58
56
55
55
Saint-Faustin-Lac-
Carre (MUNI)
67
71
66
69
65
68
66
69
67
68
67
65
61
56
La Peche (MUNI)
-
71
72
74
72
73
68
66
64
67
66
63
-
54
Varennes
74
75
68
69
69
75
68
67
63
65
60
58
55
56
Temiscaming (MUNI)
-
-
-
-
-
-
-
-
-
-
63
60
58
57
Auclair (MUNI)
-
-
-
-
-
-
-
-
-
-
60
57
55
53
Causapscal
-
-
-
-
-
-
-
-
-
-
-
-
-
-
Riviere-Eternite
(MUNI)
-
-
-
-
-
-
-
-
-
-
-
-
-
-
3B-32
-------
City
dv.1996.
1998
dv.1997.
1999
dv.1998.
2000
dv.1999.
2001
dv.2000.
2002
dv.2001.
2003
dv.2002.
2004
dv.2003.
2005
dv.2004.
2006
dv.2005.
2007
dv.2006.
2008
dv.2007.
2009
dv.2008.
2010
dv.2009.
2011
La Dore (MUNI)
54
58
58
62
58
62
56
57
55
61
58
57
53
52
Deschambault (MUNI)
68
70
64
67
68
72
65
61
57
61
58
57
56
55
Saint-Frangois
65
69
65
64
65
69
64
62
59
64
60
58
57
55
Notre-Dame-du-
Rosaire (MUNI)
65
66
60
62
64
67
62
60
59
60
59
57
55
53
St-Hilaire-de-Dorset
(MUNI)
67
70
66
67
69
73
71
67
65
65
65
63
59
57
Tinqwick (MUNI)
69
73
66
66
66
72
70
67
62
63
-
60
61
57
Lac-Edouard (MUNI)
-
-
62
65
60
62
58
57
55
59
58
58
54
51
Montmorency
(COUNTY)
-
-
-
-
-
-
-
-
-
-
-
-
-
-
Sutton
-
-
-
-
-
-
-
-
-
-
70
68
65
61
Chapais
-
-
-
-
-
-
-
-
-
59
56
56
56
55
Ste-Francoise (MUNI)
72
76
78
-
-
-
-
-
-
-
-
-
-
-
Saint-Anicet (MUNI)
74
79
76
75
74
79
75
73
69
70
68
-
67
63
L'Assomption
(COUNTY)
-
76
70
70
67
71
60
60
58
67
64
-
64
59
La Patrie (MUNI)
-
68
66
67
71
73
72
68
65
63
63
62
59
55
Ferme Neuve (MUNI)
58
60
58
59
56
59
54
-
-
-
59
57
53
50
Senneterre
-
-
-
-
-
-
-
-
-
-
59
57
55
54
Lemieux (MUNI)
-
-
-
-
-
64
66
63
65
65
64
-
63
59
Saint-Jean-sur-
Richelieu
-
-
-
-
68
73
67
65
61
64
62
59
57
56
Frelighsburg (MUNI)
-
-
-
-
-
-
-
-
-
68
68
66
63
59
Mingan (First Nations
Reserce)
-
-
-
-
-
-
-
-
-
-
52
51
49
47
Turner et al., 2016 (3060878) - Long-Term Ozone and Respiratory Mortality
Nationwide, U.S. O3: 2002-2004
Air quality data are not described for this study as it relied on estimated O3 concentrations for the years 2002-2004 as surrogates for
study population O3 concentrations during the 1982 to 2004 period (Turner et al., 2016).
3B-33
-------
Vanos et al., 2014 (2231512) - Short-Term Ozone and Respiratory Mortality
10 Canadian cities, Canada. O3: 1981 - 1999. The table below does not include design values prior to 1988 as data are not readily
available for years prior to 1986.
City
dv.1986.
1988
dv.1987.
1989
dv.1988.
1990
dv.1989.
1991
dv.1990.
1992
dv.1991.
1993
dv.1992.
1994
dv.1993.
1995
dv.1994.
1996
dv.1995.
1997
dv.1996.
1998
dv.1997.
1999
Saint John
65
67
68
66
61
51
54
58
60
55
55
Toronto
90
89
85
81
78
75
70
72
73
77
80
84
Montreal
66
74
77
72
73
73
69
65
63
71
68
77
Ottawa
67
68
73
71
71
69
64
64
63
66
65
69
Windsor
94
94
91
82
79
79
78
85
90
86
86
86
Quebec
60
62.5
59
57.5
Calgary
64
63
60
60
60
59
60
59
60
57
59
58
Edmonton
62
60
57
60
62
62
60
61
58
56
62
64
Winnipeg
62
64
63
58
53
53
54
54
54
56
56
62
Vancouver
73
70
74
61
60
55
55
65
63
59
61
58
Villeneuve et al., 2007 (195859) - ED Visits for Asthma
Census Metropolitan of Edmonton, Alberta, Canada. 1992-2002
City
dv.1992.
1994
dv.1993.
1995
dv.1994.
1996
dv.1995.
1997
dv.1996.
1998
dv.1997.
1999
dv.1998.
2000
dv.1999.
2001
dv.2000.
2002
Census
Metropolitan
of Edmonton
60
67
69
63
61
64
64
63
64
Weichenthal et al., 2017 (4165121) - Long-Term Ozone and Respiratory Mortality
Nationwide, Canada. O3: 2002-2009
City
dv.2002.2004
dv.2003.2005
dv.2004.2006
dv.2005.2007
dv.2006.2008
dv.2007.2009
All cities (DV range)
45-98
43-93
42-85
36-89
35-86
37-83
3B-34
-------
Winquist et al., 2012 (1668375) - Hospital Admissions for Asthma, ED Visits for Asthma, Hospital Admissions for Aggregate
Respiratory, ED Visits for Aggregate Respiratory Diseases, ED Visits for Respiratory Infection
St. Louis, MO (8 MO and 8 IL counties, 269 zip codes), U.S. O3: 2001-2007
City
Census Area Name
dv.2001.2003
dv.2002.2004
dv.2003.2005
dv.2004.2006
dv.2005.2007
St. Louis
St. Louis-St. Charles-Farmington, MO-IL
92
89
86
86
89
Winquist et al., 2014 (2347402) - ED Visits for Asthma
Atlanta metro area, GA, U.S. O3: 1998-2004
City
Census Area Name
dv.1998.2000
dv.1999.2001
dv.2000.2002
dv.2001.2003
dv.2002.2004
Atlanta, GA
Atlanta-Sandy Springs-Roswell, GA
121
107
99
91
93
Xiao et al., 2016 (3455927) - ED Visits for Asthma, ED Visit - Respiratory Infection
Georgia (statewide), U.S. O3: 2002-2008
State
dv.2002.2004
dv.2003.2005
dv.2004.2006
dv.2005.2007
dv.2006.2008
Georgia
93
93
91
95
95
Zanobetti and Schwartz, 2008 (101596) - Short-Term Ozone and Respiratory Mortality
48 U.S. cities
City
Census Area Name
dv1989
_1991
dv1990
_1992
dv1991
_1993
dv1992
_1994
dv1993
_1995
dv1994
_1996
dv1995
_1997
dv1996
_1998
dv1997
_1999
dv1998
_2000
Honolulu, HI
Honolulu, HI
0.045
0.048
0.047
Colorado Springs,
CO
Colorado Springs, CO
0.066
0.063
0.062
0.061
0.061
0.056
0.062
0.065
Spokane, WA
Spokane, WA
0.064
0.066
0.066
0.068
0.067
0.067
Albuquerque, NM
Albuquerque, NM
0.071
0.071
0.069
0.070
0.071
0.074
0.069
0.073
0.071
0.075
Ft. Lauderdale, FL
Broward County, FL
0.075
0.073
0.076
0.079
0.074
0.069
0.069
0.072
0.075
0.075
Boulder, CO
Boulder, CO
0.076
0.073
0.073
0.071
0.072
0.071
0.071
0.078
0.078
0.078
Provo/Orem, UT
Provo-Orem, UT
0.069
0.068
0.071
0.076
0.082
0.082
0.086
3B-35
-------
City
Census Area Name
dv1989
_1991
dv1990
_1992
dv1991
_1993
dv1992
_1994
dv1993
_1995
dv1994
_1996
dv1995
_1997
dv1996
_1998
dv1997
_1999
dv1998
_2000
Miami, FL
Miami-Fort Lauderdale-Pompano
Beach, FL
0.075
0.073
0.076
0.080
0.080
0.074
0.075
0.077
0.078
0.079
Seattle, WA
Seattle-Tacoma-Bellevue, WA
0.078
0.086
0.077
0.074
0.071
0.076
0.078
0.081
0.074
0.075
Denver, CO
Denver-Aurora-Broomfield, CO
0.080
0.074
0.071
0.074
0.081
0.081
0.079
0.084
0.083
0.086
Orlando, FL
Orlando-Kissimmee, FL
0.080
0.079
0.078
0.082
0.079
0.079
0.078
0.084
0.085
0.085
Salt Lake City, UT
Salt Lake City, UT
0.078
0.075
0.076
0.079
0.082
0.089
0.085
0.088
0.082
0.088
Tampa, FL
Tampa-St. Petersburg-
Clearwater, FL
0.079
0.081
0.080
0.080
0.080
0.081
0.082
0.088
0.090
0.088
New Orleans, LA
New Orleans-Metairie-Kenner, LA
0.077
0.080
0.081
0.086
0.084
0.085
0.083
0.084
0.086
0.091
Oklahoma City,
Oklahoma City, OK
0.086
0.084
0.081
0.081
0.084
0.085
0.083
0.085
0.086
0.084
Terra Haute, IN
Terre Haute, IN
0.087
0.081
0.077
0.079
0.084
0.092
0.088
0.088
0.083
0.080
Austin, TX
Austin-Round Rock, TX
0.084
0.084
0.081
0.082
0.084
0.084
0.081
0.081
0.089
0.089
San Francisco, CA
San Francisco-Oakland-Fremont,
CA
0.084
0.082
0.081
0.082
0.087
0.093
0.090
0.089
0.086
0.087
Greensboro, NC
Greensboro-High Point, NC
0.088
0.085
0.083
0.084
0.088
0.086
0.085
0.089
0.092
0.094
Tulsa, OK
Tulsa, OK
0.087
0.087
0.082
0.083
0.088
0.091
0.089
0.087
0.088
0.093
Kansas City, KS
Kansas City, MO-KS
0.082
0.083
0.082
0.082
0.090
0.092
0.094
0.093
0.091
0.089
Phoenix, AZ
Phoenix-Mesa-Scottsdale, AZ
0.083
0.091
0.088
0.086
0.089
0.090
0.092
0.091
0.088
0.088
Canton, OH
Canton-Massillon, OH
0.091
0.089
0.089
0.088
0.091
0.089
0.088
0.089
0.091
0.091
Columbus, OH
Columbus, OH
0.089
0.092
0.090
0.086
0.090
0.092
0.092
0.093
0.097
0.095
Detroit, Ml
Detroit-Warren-Livonia, Ml
0.096
0.091
0.089
0.088
0.093
0.094
0.092
0.093
0.095
0.089
Youngstown, OH
Youngstown-Warren-Boardman,
OH-PA
0.090
0.091
0.091
0.089
0.091
0.092
0.093
0.096
0.096
0.092
Birmingham, AL
Birmingham-Hoover, AL
0.084
0.088
0.089
0.092
0.096
0.096
0.095
0.095
0.097
0.102
Boston, MA
Boston-Cambridge-Quincy, MA-
NH
0.098
0.092
0.096
0.093
0.096
0.094
0.095
0.091
0.093
0.086
Milwaukee, Wl
Milwaukee-Waukesha-West Allis,
Wl
0.101
0.095
0.090
0.084
0.092
0.097
0.098
0.093
0.097
0.092
Cincinnati, OH
Cincinnati-Middletown, OH-KY-IN
0.102
0.095
0.091
0.091
0.098
0.099
0.095
0.092
0.095
0.094
Cleveland, OH
Cleveland-Elyria-Mentor, OH
0.093
0.090
0.092
0.093
0.098
0.100
0.099
0.098
0.099
0.095
Charlotte, NC
Charlotte-Gastonia-Concord, NC-
SC
0.092
0.091
0.091
0.092
0.094
0.094
0.097
0.103
0.104
0.104
3B-36
-------
City
Census Area Name
dv1989
_1991
dv1990
_1992
dv1991
_1993
dv1992
_1994
dv1993
_1995
dv1994
_1996
dv1995
_1997
dv1996
_1998
dv1997
_1999
dv1998
_2000
St Louis, MO
St. Louis, MO-IL
0.098
0.098
0.091
0.091
0.098
0.104
0.100
0.095
0.095
0.094
Chicago, IL
Chicago-Naperville-Joliet, IL-IN-
Wl
0.104
0.099
0.100
0.093
0.099
0.097
0.096
0.091
0.095
0.093
Pittsburgh, PA
Pittsburgh, PA
0.092
0.088
0.095
0.096
0.105
0.103
0.105
0.099
0.101
0.096
Nashville, TN
Nashville-Murfreesboro-Franklin,
TN
0.096
0.096
0.095
0.096
0.099
0.099
0.099
0.101
0.102
0.100
Jersey City, NJ
Hudson County, NJ
0.107
0.104
0.103
0.096
0.100
0.095
0.098
0.093
0.100
0.092
Washington, DC
Washington-Arlington-Alexandria,
DC-VA-MD-WV
0.100
0.100
0.101
0.096
0.098
0.094
0.100
0.101
0.106
0.101
Dallas/Ft Worth,
Dallas-Fort Worth-Arlington, TX
0.105
0.099
0.095
0.096
0.106
0.104
0.104
0.098
0.101
0.102
New Haven, CT
New Haven-Milford, CT
0.116
0.113
0.108
0.097
0.105
0.101
0.107
0.100
0.103
0.096
Sacramento, CA
Sacramento-Arden Arcade-
Roseville, CA
0.105
0.105
0.110
0.104
0.106
0.106
0.099
0.103
0.103
0.107
Baltimore, MD
Baltimore-Towson, MD
0.104
0.106
0.107
0.103
0.107
0.105
0.107
0.104
0.109
0.107
Philadelphia, PA
Philadelphia-Camden-Wilmington,
PA-NJ-DE-MD
0.113
0.107
0.106
0.099
0.104
0.101
0.110
0.107
0.110
0.106
San Diego, CA
San Diego-Carlsbad-San Marcos,
CA
0.125
0.118
0.112
0.109
0.108
0.104
0.099
0.102
0.099
0.100
New York City, NY
New York-Northern New Jersey-
Long Island, NY-NJ-PA
0.122
0.116
0.108
0.100
0.106
0.104
0.108
0.104
0.107
0.107
Atlanta, GA
Atlanta-Sandy Springs-Marietta,
GA
0.104
0.105
0.101
0.101
0.109
0.105
0.110
0.113
0.118
0.121
Houston, TX
Houston-Sugar Land-Baytown, TX
0.119
0.116
0.104
0.110
0.114
0.116
0.117
0.116
0.118
0.112
Los Angeles, CA
Los Angeles-Long Beach-Santa
Ana, CA
0.179
0.177
0.177
0.168
0.156
0.145
0.135
0.133
0.118
0.115
Note: Design values for this study were available in the last review (see Wells, 2012) and
are presented in units of ppm, rather than ppb.
3B-37
-------
Zu et al., 2017 (3859551) - Hospital Admissions for Asthma
6 Texas City Metro areas (Austin, Dallas, El Paso, Ft Worth, Houston, San Antonio), U.S. (pooled, not individually assessed)
Os: 2001-2013
City
Census Area
Name
dv.2001
.2003
dv.2002
.2004
dv.2003
.2005
dv.2004
.2006
dv.2005
.2007
dv.2006
.2008
dv.2007
.2009
dv.2008
.2010
dv.2009
.2011
dv.2010
.2012
dv.2011
.2013
Dallas and
Fort Worth
Dallas-Fort
Worth, TX-OK
100
98
95
96
95
91
86
86
90
87
87
El Paso
El Paso-Las
Cruces, TX-NM
79
78
76
78
79
78
75
71
71
72
75
Houston
Flouston-The
Woodlands, TX
102
101
103
103
96
91
84
84
89
88
87
Austin
Austin-Round
Rock, TX
(CBSA only)
84
85
82
82
80
77
75
74
75
74
73
San Antonio
San Antonio-
New Braunfels,
TX (CBSA only)
89
91
86
87
82
78
74
75
75
80
81
3B-38
-------
REFERENCES
Alhanti, BA, Chang, HH, Winquist, A, Mulholland, JA, Darrow, LA and Sarnat, SE (2016).
Ambient air pollution and emergency department visits for asthma: a multi-city
assessment of effect modification by age. J Expo Sci Environ Epidemiol 26(2): 180-188.
Barry, V, Klein, M, Winquist, A, Chang, HH, Mulholland, JA, Talbott, EO, Rager, JR, Tolbert,
PE and Sarnat, SE (2018). Characterization of the concentration-response curve for
ambient ozone and acute respiratory morbidity in 5 US cities. J Expo Sci Environ
Epidemiol 29(2): 267-277.
Byers, N, Ritchey, M, Vaidyanathan, A, Brandt, AJ and Yip, F (2015). Short-term effects of
ambient air pollutants on asthma-related emergency department visits in Indianapolis,
Indiana, 2007-2011. J Asthma 53(3): 1-8.
Cakmak, S, Hebbern, C, Pinault, L, Lavigne, E, Vanos, J, Crouse, DL and Tjepkema, M (2017).
Associations between long-term PM2.5 and ozone exposure and mortality in the
Canadian Census Health and Environment Cohort (CANCHEC), by spatial synoptic
classification zone. Environ Int 111: 200-211.
Crouse, DL, Peters, PA, Hystad, P, Brook, JR, van Donkelaar, A, Martin, RV, Villeneuve, PJ,
Jerrett, M, Goldberg, MS, Pope, CA, 3rd, Brauer, M, Brook, RD, Robichaud, A, Menard,
R and Burnett, RT (2015). Ambient PM2.5, 0(3), and NO(2) Exposures and Associations
with Mortality over 16 Years of Follow-Up in the Canadian Census Health and
Environment Cohort (CanCHEC). Environ Health Perspect 123(11): 1180-1186.
Darrow, LA, Klein, M, Flanders, WD, Mulholland, JA, Tolbert, PE and Strickland, MJ (2014).
Air pollution and acute respiratory infections among children 0-4 years of age: an 18-year
time-series study. Am J Epidemiol 180(10): 968-977.
Darrow, LA, Klein, M, Sarnat, JA, Mulholland, JA, Strickland, MJ, Sarnat, SE, Russell, AG and
Tolbert, PE (2011). The use of alternative pollutant metrics in time-series studies of
ambient air pollution and respiratory emergency department visits. J Expo Sci Environ
Epidemiol 21(1): 10-19.
Eckel, SP, Cockburn, M, Shu, YH, Deng, H, Lurmann, FW, Liu, L and Gilliland, FD (2016). Air
pollution affects lung cancer survival. Thorax 71 (10): 891-898.
Garcia, E, Berhane, KT, Islam, T, McConnell, R, Urman, R, Chen, Z and Gilliland, FD (2019).
Association of changes in air quality with incident asthma in children in California, 1993-
2014. J Am Med Assoc 321(19): 1906-1915.
Gleason, JA, Bieloiy, L and Fagliano, JA (2014). Associations between ozone, PM2.5, and four
pollen types on emergency department pediatric asthma events during the warm season in
New Jersey: a case-crossover study. Environ Res 132: 421-429.
3B-39
-------
Goodman, JE, Loftus, CT, Liu, X and Zu, K (2017a). Impact of respiratory infections, outdoor
pollen, and socioeconomic status on associations between air pollutants and pediatric
asthma hospital admissions. PLoS One 12(7): e0180522.
Goodman, JE, Zu, K, Loftus, CT, Tao, G, Liu, X and Lange, S (2017b). Ambient ozone and
asthma hospital admissions in Texas: a time-series analysis. Asthma Res Pract 3: 6.
Ito, A, Sillman, S and Penner, JE (2007). Effects of additional nonmethane volatile organic
compounds, organic nitrates, and direct emissions of oxygenated organic
species on global tropospheric chemistry. J Geophys Res 112: 21 PP.
Jerrett, M, Burnett, RT, Beckerman, BS, Turner, MC, Krewski, D, Thurston, G, Martin, RV, van
Donkelaar, A, Hughes, E, Shi, Y, Gapstur, SM, Thun, MJ and Pope, CA, 3rd (2013).
Spatial analysis of air pollution and mortality in California. Am J Respir Crit Care Med
188(5): 593-599.
Jerrett, M, Burnett, RT, Pope, CA, 3rd, Ito, K, Thurston, G, Krewski, D, Shi, Y, Calle, E and
Thun, M (2009). Long-term ozone exposure and mortality. N Engl J Med 360(11): 1085-
1095.
Katsouyanni, K, Samet, JM, Anderson, HR, Atkinson, R, Le Tertre, A, Medina, S, Samoli, E,
Touloumi, G, Burnett, RT, Krewski, D, Ramsay, T, Dominici, F, Peng, RD, Schwartz, J
and Zanobetti, A (2009). Air pollution and health: a European and North American
approach (APHENA). Res Rep Health Eff Inst(142): 5-90.
Klemm, RJ, Thomas, EL and Wyzga, RE (2011). The impact of frequency and duration of air
quality monitoring: Atlanta, GA, data modeling of air pollution and mortality. J Air
Waste Manag Assoc 61(11): 1281-1291.
Kousha, T and Castner, J (2016). The air quality health index and emergency department visits
for otitis media. J Nurs Scholarsh 48(2): 163-171.
Kousha, T and Rowe, BH (2014). Ambient ozone and emergency department visits due to lower
respiratory condition. Int J Occup Med Environ Health 27(1): 50-59.
Malig, BJ, Pearson, DL, Chang, YB, Broadwin, R, Basu, R, Green, RS and Ostro, B (2016). A
Time-Stratified Case-Crossover Study of Ambient Ozone Exposure and Emergency
Department Visits for Specific Respiratory Diagnoses in California (2005-2008). Environ
Health Perspect 124(6): 745-753.
Nishimura, KK, Galanter, JM, Roth, LA, Oh, SS, Thakur, N, Nguyen, EA, Thyne, S, Farber, HJ,
Serebrisky, D, Kumar, R, Brigino-Buenaventura, E, Davis, A, Lenoir, MA, Meade, K,
Rodriguez-Cintron, W, Avila, PC, Borrell, LN, Bibbins-Domingo, K, Rodriguez-Santana,
JR, Sen, S, Lurmann, F, Balmes, JR and Burchard, EG (2013). Early-life air pollution and
asthma risk in minority children: the GALA II and SAGE II studies. Am J Respir Crit
Care Med 188(3): 309-318.
3B-40
-------
O'Lenick, CR, Winquist, A, Mulholland, JA, Friberg, MD, Chang, HH, Kramer, MR, Darrow,
LA and Sarnat, SE (2017). Assessment of neighbourhood-level socioeconomic status as a
modifier of air pollution-asthma associations among children in Atlanta. J Epidemiol
Community Health 71(2): 129-136.
Rodopoulou, S, Samoli, E, Chalbot, MG and Kavouras, IG (2015). Air pollution and
cardiovascular and respiratory emergency visits in Central Arkansas: A time-series
analysis. Sci Total Environ 536: 872-879.
Sacks, JD, Rappold, AG, Davis, JA, Jr., Richardson, DB, Waller, AE and Luben, TJ (2014).
Influence of urbanicity and county characteristics on the association between ozone and
asthma emergency department visits in North Carolina. Environ Health Perspect 122(5):
506-512.
Sarnat, JA, Sarnat, SE, Flanders, WD, Chang, HH, Mulholland, J, Baxter, L, Isakov, V and
Ozkaynak, H (2013). Spatiotemporally resolved air exchange rate as a modifier of acute
air pollution-related morbidity in Atlanta. J Expo Sci Environ Epidemiol 23(6): 606-615.
Sarnat, SE, Winquist, A, Schauer, JJ, Turner, JR and Sarnat, JA (2015). Fine particulate matter
components and emergency department visits for cardiovascular and respiratory diseases
in the St. Louis, Missouri-Illinois, metropolitan area. Environ Health Perspect 123(5):
437-444.
Sheffield, PE, Zhou, J, Shmool, JL and Clougherty, JE (2015). Ambient ozone exposure and
children's acute asthma in New York City: a case-crossover analysis. Environ Health 14:
25.
Shmool, JL, Kinnee, E, Sheffield, PE and Clougherty, JE (2016). Spatio-temporal ozone
variation in a case-crossover analysis of childhood asthma hospital visits in New York
City. Environ Res 147: 108-114.
Silverman, RA and Ito, K (2010). Age-related association of fine particles and ozone with severe
acute asthma in New York City. J Allergy Clin Immunol 125(2): 367-373.
Stieb, DM, Szyszkowicz, M, Rowe, BH and Leech, JA (2009). Air pollution and emergency
department visits for cardiac and respiratory conditions: A multi-city time-series analysis.
Environ Health 8(25): 25.
Strickland, MJ, Klein, M, Flanders, WD, Chang, HH, Mulholland, JA, Tolbert, PE and Darrow,
LA (2014). Modification of the effect of ambient air pollution on pediatric asthma
emergency visits: susceptible subpopulations. Epidemiology 25(6): 843-850.
Szyszkowicz, M, Kousha, T, Castner, J and Dales, R (2018). Air pollution and emergency
department visits for respiratory diseases: A multi-city case crossover study. Environ Res
163: 263-269.
3B-41
-------
Tetreault, LF, Doucet, M, Gamache, P, Fournier, M, Brand, A, Kosatsky, T and Smargiassi, A
(2016). Childhood exposure to ambient air pollutants and the onset of asthma: an
administrative cohort study in Quebec. Environ Health Perspect 124(8): 1276-1282.
Tolbert, PE, Klein, M, Peel, JL, Sarnat, SE and Sarnat, JA (2007). Multipollutant modeling
issues in a study of ambient air quality and emergency department visits in Atlanta. J
Expo Sci Environ Epidemiol 17(Suppl 2): S29-S35.
Turner, MC, Jerrett, M, Pope, CA, 3rd, Krewski, D, Gapstur, SM, Diver, WR, Beckerman, BS,
Marshall, JD, Su, J, Crouse, DL and Burnett, RT (2016). Long-Term Ozone Exposure
and Mortality in a Large Prospective Study. Am J Respir Crit Care Med 193(10): 1134-
1142.
U.S. EPA (2020). 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.
Vanos, JK, Hebbern, C and Cakmak, S (2014). Risk assessment for cardiovascular and
respiratory mortality due to air pollution and synoptic meteorology in 10 Canadian cities.
Environ Pollut 185: 322-332.
Villeneuve, PJ, Chen, L, Rowe, BH and Coates, F (2007). Outdoor air pollution and emergency
department visits for asthma among children and adults: A case-crossover study in
northern Alberta, Canada. Environ Health 6: 40.
Weichenthal, S, Pinault, LL and Burnett, RT (2017). Impact of Oxidant Gases on the
Relationship between Outdoor Fine Particulate Air Pollution and Nonaccidental,
Cardiovascular, and Respiratory Mortality. Sci Rep 7(1): 16401.
Wells, BW, K.; Jenkins, S. (2012). Memorandum to Ozone NAAQS Review Docket (EPA-HQ-
OAR-2008-0699). Analysis of Recent U.S. Ozone Air Quality Data to Support the 03
NAAQS Review and Quadratic Rollback Simulations to Support the First Draft of the
Risk and Exposure Assessment. August 15, 2012.. Docket ID No. EPA-HQ-OAR-2008-
0699. Office of Air Quality Planning and Standards Research Triangle Park, NC.
Available at: https://www. regulations.gov/contentStreamer?documentId=EPA -HQ-OAR-
2008-0699-4253&contentType=pdf.
Winquist, A, Kirrane, E, Klein, M, Strickland, M, Darrow, LA, Sarnat, SE, Gass, K, Mulholland,
J, Russell, A and Tolbert, P (2014). Joint effects of ambient air pollutants on pediatric
asthma emergency department visits in Atlanta, 1998-2004. Epidemiology 25(5): 666-
673.
Winquist, A, Klein, M, Tolbert, P, Flanders, WD, Hess, J and Sarnat, SE (2012). Comparison of
emergency department and hospital admissions data for air pollution time-series studies.
Environ Health 11: 70.
3B-42
-------
Xiao, Q, Liu, Y, Mulholland, JA, Russell, AG, Darrow, LA, Tolbert, PE and Strickland, MJ
(2016). Pediatric emergency department visits and ambient Air pollution in the U.S. State
of Georgia: a case-crossover study. Environ Health 15(1): 115.
Zanobetti, A and Schwartz, J (2008). Mortality displacement in the association of ozone with
mortality: an analysis of 48 cities in the United States. Am J Respir Crit Care Med
177(2): 184-189.
Zu, K, Liu, X, Shi, L, Tao, G, Loftus, CT, Lange, S and Goodman, JE (2017). Concentration-
response of short-term ozone exposure and hospital admissions for asthma in Texas.
Environ Int 104: 139-145.
3B-43
-------
APPENDIX 3C
AIR QUALITY DATA USED IN POPULATION EXPOSURE AND RISK
ANALYSES
Table of Figures 3C-2
Table of Tables 3C-10
3C.1 Overview 3C-12
3C.2 Urban Study Areas 3C-13
3C.3 Ambient Air Ozone Monitoring Data 3C-15
3C.4 Air Quality Modeling Data 3C-24
3C.4.1 Comprehensive Air Quality Model with Extensions (CAMx) 3C-24
3C.4.2 Evaluation of Modeled Ozone Concentrations 3C-30
3C.5 Air Quality Adjustment to Meet Current and Alternative Air Quality Scenarios .3C-62
3C.5.1 Overview of the Higher Order Direct Decoupled Method (HDDM) .... 3C-62
3C.5.2 Using CAMx/HDDM to Adjust Monitored Ozone Concentrations 3C-65
3C.6 Interpolation of Adjusted Air Quality using Voronoi Neighbor Averaging 3C-92
3C.7 Results for Urban Study Areas 3C-94
3C.7.1 Design Values 3C-94
3C.7.2 Distribution of Hourly O3 Concentrations 3C-101
3C.7.3 Air Quality Inputs for the Exposure and Risk Analyses 3C-118
References 3C-153
-------
TABLE OF FIGURES
Figure 3C-1. Flowchart showing inputs, processes and outputs of the approach to
generate ambient air concentration estimates for use in the exposure
modeling 3C-12
Figure 3C-2. Map of the eight urban study areas analyzed 3C-14
Figure 3C-3. Map of the Atlanta study area 3C-16
Figure 3C-4. Map of the Boston study area 3C-17
Figure 3C-5. Map of the Dallas study area 3C-18
Figure 3C-6. Map of the Detroit study area 3C-19
Figure 3C-7. Map of the Philadelphia study area 3C-20
Figure 3C-8. Map of the Phoenix study area 3C-21
Figure 3C-9. Map of the Sacramento study area 3C-22
Figure 3C-10. Map of the St. Louis study area 3C-23
Figure 3C-11. Map of the CAMx modeling domain 3C-25
Figure 3C-12. Normalized mean bias for MDA8 O3 in the Northeastern U.S., winter
2016 3C-32
Figure 3C-13. Normalized mean bias for MDA8 O3 in the Northeastern U.S., spring
2016 3C-33
Figure 3C-14. Normalized mean bias for MDA8 O3 in the Northeastern U.S., summer
2016 3C-33
Figure 3C-15. Normalized mean bias for MDA8 O3 in the Northeastern U.S., fall 2016.
3C-34
Figure 3C-16. Time series of monitored (black) and modeled (red) MDA8 O3 at Boston
monitoring sites in 2016 3C-35
Figure 3C-17. Time series of monitored (black) and modeled (red) hourly O3
concentrations at Boston monitoring sites in January (top left), April (top
right), July (bottom left), and October (bottom right) 2016 3C-36
Figure 3C-18. Time series of monitored (black) and modeled (red) MDA8 O3 at
Philadelphia monitoring sites in 2016 3C-37
Figure 3C-19. Time series of monitored (black) and modeled (red) hourly O3
concentrations at Philadelphia monitoring sites for January (top left),
3C-2
-------
April (top right), July (bottom left), and October (bottom right) 2016.
.3C-38
Figure 3C-20.
Figure 3C-21.
Figure 3C-22.
Figure 3C-23.
Figure 3C-24.
Figure 3C-25.
Figure 3C-26.
Figure 3C-27.
Figure 3C-28.
Figure 3C-29.
Figure 3C-30.
Figure 3C-31.
Figure 3C-32.
Figure 3C-33.
Normalized mean bias for MDA8 O3 in the Southeastern U.S., winter
2016 3C-39
Normalized mean bias for MDA8 O3 in the Southeastern U.S., spring
2016 3C-40
Normalized mean bias for MDA8 O3 in the Southeastern U.S., summer
2016 3C-40
Normalized mean bias for MDA8 O3 in the Southeastern U.S., fall 2016.
3C-41
Time series of monitored (black) and modeled (red) MDA8 O3 at Atlanta
monitoring sites in 2016 3C-42
Time series of monitored (black) and modeled (red) hourly O3
concentrations at Atlanta monitoring sites in January (top left), April (top
right), July (bottom left), and October (bottom right) 2016 3C-43
Normalized mean bias for MDA8 O3 in the Midwest U.S., winter 2016
3C-44
Normalized mean bias for MDA8 O3 in the Midwest U.S., spring 2016
3C-45
Normalized mean bias for MDA8 O3 in the Midwest U.S., summer 2016.
3C-45
Normalized mean bias for MDA8 O3 in the Midwest U.S., fall 2016
.3C-46
Time series of monitored (black) and modeled (red) MDA8 O3 at Detroit
monitoring sites in 2016 3C-47
Time series of monitored (black) and modeled (red) hourly O3
concentrations at Detroit monitoring sites in January (top left), April (top
right), July (bottom left), and October (bottom right) 2016 3C-48
Normalized mean bias for MDA8 O3 in the Central U.S., winter 2016
3C-49
Normalized mean bias for MDA8 O3 in the Central U.S., spring 2016
3C-50
3C-3
-------
Figure 3C-34. Normalized mean bias for MDA8 O3 in the Central U.S., summer 2016
3C-50
Figure 3C-35. Normalized mean bias for MDA8 O3 in the Central U.S., fall 2016
3C-51
Figure 3C-36. Time series of monitored (black) and modeled (red) MDA8 O3 at St. Louis
monitoring sites in 2016 3C-52
Figure 3C-37. Time series of monitored (black) and modeled (red) hourly O3
concentrations at St. Louis monitoring sites in January (top left), April
(top right), July (bottom left), and October (bottom right) 2016 3C-53
Figure 3C-38. Time series of monitored (black) and modeled (red) MDA8 O3 at Dallas
monitoring sites in 2016 3C-54
Figure 3C-39. Time series of monitored (black) and modeled (red) hourly O3
concentrations at Dallas monitoring sites in January (top left), April (top
right), July (bottom left), and October (bottom right) 2016 3C-55
Figure 3C-40. Normalized mean bias for MDA8 O3 in the Western U.S., winter 2016
3C-57
Figure 3C-41. Normalized mean bias for MDA8 O3 in the Western U.S., spring 2016
3C-57
Figure 3C-42. Normalized mean bias for MDA8 O3 in the Western U.S., summer 2016.
3C-58
Figure 3C-43. Normalized mean bias for MDA8 O3 in the Western U.S., fall 2016
3C-58
Figure 3C-44. Time series of monitored (black) and modeled (red) MDA8 O3 at
Sacramento monitoring sites in 2016 3C-59
Figure 3C-45. Time series of monitored (black) and modeled (red) hourly O3
concentrations at Sacramento monitoring sites in January (top left), April
(top right), July (bottom left), and October (bottom right) 2016
3C-60
Figure 3C-46. Time series of monitored (black) and modeled (red) MDA8 O3 at Phoenix
monitoring sites in 2016 3C-61
3C-4
-------
Figure 3C-47. Time series of monitored (black) and modeled (red) hourly O3
concentrations at Phoenix monitoring sites in January (top left), April
(top right), July (bottom left), and October (bottom right) 2016 3C-62
Figure 3C-48. Flow diagram demonstrating HDDM model-based O3 adjustment
approach 3C-67
Figure 3C-49. Conceptual picture of 3-step application of HDDM sensitivities 3C-70
Figure 3C-50. Comparison of brute force and 3-step HDDM O3 estimates for 50% NOx
cut conditions in Atlanta 3C-74
Figure 3C-51. Comparison of brute force and 3-step HDDM O3 estimates for 50% NOx
cut conditions in Boston 3C-75
Figure 3C-52. Comparison of brute force and 3-step HDDM O3 estimates for 50% NOx
cut conditions in Dallas 3C-76
Figure 3C-53. Comparison of brute force and 3-step HDDM O3 estimates for 50% NOx
cut conditions in Detroit 3C-77
Figure 3C-54. Comparison of brute force and 3-step HDDM O3 estimates for 50% NOx
cut conditions in Philadelphia 3C-78
Figure 3C-55. Comparison of brute force and 3-step HDDM O3 estimates for 50% NOx
cut conditions in Phoenix 3C-79
Figure 3C-56. Comparison of brute force and 3-step HDDM O3 estimates for 50% NOx
cut conditions in Sacramento 3C-80
Figure 3C-57. Comparison of brute force and 3-step HDDM O3 estimates for 50% NOx
cut conditions in St. Louis 3C-81
Figure 3C-58. Comparison of brute force and 3-step HDDM O3 estimates for 90% NOx
cut conditions in Atlanta 3C-82
Figure 3C-59. Comparison of brute force and 3-step HDDM O3 estimates for 90% NOx
cut conditions in Boston 3C-83
Figure 3C-60. Comparison of brute force and 3-step HDDM O3 estimates for 90% NOx
cut conditions in Dallas 3C-84
Figure 3C-61. Comparison of brute force and 3-step HDDM O3 estimates for 90% NOx
cut conditions in Detroit 3C-85
Figure 3C-62. Comparison of brute force and 3-step HDDM O3 estimates for 90% NOx
cut conditions in Philadelphia 3C-86
3C-5
-------
Figure 3C-63. Comparison of brute force and 3-step HDDM O3 estimates for 90% NOx
cut conditions in Phoenix 3C-87
Figure 3C-64. Comparison of brute force and 3-step HDDM O3 estimates for 90% NOx
cut conditions in Sacramento 3C-88
Figure 3C-65. Comparison of brute force and 3-step HDDM O3 estimates for 90% NOx
cut conditions in St. Louis 3C-89
Figure 3C-66. Numerical example of the Voronoi Neighbor Averaging (VNA) technique.
3C-93
Figure 3C-67. Diurnal distribution of hourly O3 concentrations at monitoring sites in
Atlanta 3C-103
Figure 3C-68. Diurnal distribution of hourly O3 concentrations at monitoring sites in
Boston 3C-104
Figure 3C-69. Diurnal distribution of hourly O3 concentrations at monitoring sites in
Dallas 3C-105
Figure 3C-70. Diurnal distribution of hourly O3 concentrations at monitoring sites in
Detroit 3C-106
Figure 3C-71. Diurnal distribution of hourly O3 concentrations at monitoring sites in
Philadelphia 3C-107
Figure 3C-72. Diurnal distribution of hourly O3 concentrations at monitoring sites in
Phoenix 3C-108
Figure 3C-73. Diurnal distribution of hourly O3 concentrations at monitoring sites in
Sacramento 3C-109
Figure 3C-74. Diurnal distribution of hourly O3 concentrations at monitoring sites in St.
Louis 3C-110
Figure 3C-75. Monthly distribution of hourly O3 concentrations at monitoring sites in
Atlanta 3C-111
Figure 3C-76. Monthly distribution of hourly O3 concentrations at monitoring sites in
Boston 3C-112
Figure 3C-77. Monthly distribution of hourly O3 concentrations at monitoring sites in
Dallas 3C-113
Figure 3C-78. Monthly distribution of hourly O3 concentrations at monitoring sites in
Detroit 3C-114
3C-6
-------
Figure 3C-79.
Figure 3C-80.
Figure 3C-81.
Figure 3C-82.
Figure 3C-83.
Figure 3C-84.
Figure 3C-85.
Figure 3C-86.
Figure 3C-87.
Figure 3C-89.
Figure 3C-90.
Figure 3C-91.
Figure 3C-92.
Figure 3C-93.
Figure 3C-94.
Figure 3C-95.
Figure 3C-96.
Monthly distribution of hourly O3 concentrations at monitoring sites in
Philadelphia 3C-115
Monthly distribution of hourly O3 concentrations at monitoring sites in
Phoenix 3C-116
Monthly distribution of hourly O3 concentrations at monitoring sites in
Sacramento 3C-117
Monthly distribution of hourly O3 concentrations at monitoring sites in St.
Louis 3C-118
Changes
Changes
Changes
Changes
Changes
in MDA8 O3 based on HDDM adjustments in Atlanta 3C-121
in MDA8 O3 based on HDDM adjustments in Boston 3C-122
in MDA8 O3 based on HDDM adjustments in Dallas 3C-123
in MDA8 O3 based on HDDM adjustments in Detroit. 3C-124
in MDA8 O3 based on HDDM adjustments in Philadelphia
3C-125
Changes in MDA8 O3 based on HDDM adjustments in Sacramento.
.3C-127
Changes in MDA8 O3 based on HDDM adjustments in St. Louis.
.3C-128
Annual 4th highest MDA8 O3 and May-September mean MDA8 O3 based
on HDDM adjustments in Atlanta 3C-129
Changes in annual 4th highest MDA8 O3 and May-September mean
MDA8 O3 based on HDDM adjustments in Atlanta 3C-130
Annual 4th highest MDA8 O3 and May-September mean MDA8 O3 based
on HDDM adjustments in Boston 3C-131
Changes in annual 4th highest MDA8 O3 and May-September mean
MDA8 O3 based on HDDM adjustments in Boston 3C-132
Annual 4th highest MDA8 O3 and May-September mean MDA8 O3 based
on HDDM adjustments in Dallas 3C-133
Changes in annual 4th highest MDA8 O3 and May-September mean
MDA8 O3 based on HDDM adjustments in Dallas 3C-134
3C-7
-------
Figure 3C-97. Annual 4th highest MDA8 O3 and May-September mean MDA8 O3 based
on HDDM adjustments in Detroit 3C-135
Figure 3C-98. Changes in annual 4th highest MDA8 O3 and May-September mean
MDA8 O3 based on HDDM adjustments in Detroit 3C-136
Figure 3C-99. Annual 4th highest MDA8 O3 and May-September mean MDA8 O3 based
on HDDM adjustments in Philadelphia 3C-137
Figure 3C-100. Changes in annual 4th highest MDA8 O3 and May-September mean
MDA8 O3 based on HDDM adjustments in Philadelphia 3C-138
Figure 3C-101. Annual 4th highest MDA8 O3 and May-September mean MDA8 O3 based
on HDDM adjustments in Phoenix 3C-139
Figure 3C-102. Changes in annual 4th highest MDA8 O3 and May-September mean
MDA8 O3 based on HDDM adjustments in Phoenix 3C-140
Figure 3C-103. Annual 4th highest MDA8 O3 and May-September mean MDA8 O3 based
on HDDM adjustments in Sacramento 3C-141
Figure 3C-104. Changes in annual 4th highest MDA8 O3 and May-September mean
MDA8 O3 based on HDDM adjustments in Sacramento 3C-142
Figure 3C-105. Annual 4th highest MDA8 O3 and May-September mean MDA8 O3 based
on HDDM adjustments in St. Louis 3C-143
Figure 3C-106. Changes in annual 4th highest MDA8 O3 and May-September mean
MDA8 O3 based on HDDM adjustments in St. Louis 3C-144
Figure 3C-107. Annual 4th highest MDA8 O3 and May-September mean MDA8 O3 by
population based on HDDM adjustments in Atlanta 3C-145
Figure 3C-108. Annual 4th highest MDA8 O3 and May-September mean MDA8 O3 by
population based on HDDM adjustments in Boston 3C-146
Figure 3C-109. Annual 4th highest MDA8 O3 and May-September mean MDA8 O3 by
population based on HDDM adjustments in Dallas 3C-147
Figure 3C-110. Annual 4th highest MDA8 O3 and May-September mean MDA8 O3 by
population based on HDDM adjustments in Detroit 3C-148
Figure 3C-111. Annual 4th highest MDA8 O3 and May-September mean MDA8 O3 by
population based on HDDM adjustments in Philadelphia 3C-149
Figure 3C-112. Annual 4th highest MDA8 O3 and May-September mean MDA8 O3 by
population based on HDDM adjustments in Phoenix 3C-150
3C-8
-------
Figure 3C-113. Annual 4th highest MDA8 O3 and May-September mean MDA8 O3 by
population based on HDDM adjustments in Sacramento 3C-151
Figure 3C-114. Annual 4th highest MDA8 O3 and May-September mean MDA8 O3 by
population based on HDDM adjustments in St. Louis 3C-152
3C-9
-------
TABLE OF TABLES
Table 3C-1. Summary information for the eight urban study areas 3C-14
Table 3C-2. Geographic elements of domain used in the CAMx/HDDM modeling
3C-25
Table 3C-3. Vertical layer structure for 2016 WRF and CAMx simulations 3C-27
Table 3C-4. Summary of U.S. emissions totals by sector for the 12km CONUS domain
(in thousand tons) 3C-29
Table 3C-5. CAMx model performance at monitoring sites in the Northeastern U.S
3C-32
Table 3C-6. CAMx model performance at monitoring sites in the Boston area 3C-34
Table 3C-7. CAMx model performance at monitoring sites in the Philadelphia area
3C-36
Table 3C-8. CAMx model performance at monitoring sites in the Southeastern U.S
3C-39
Table 3C-9. CAMx model performance at monitoring sites in the Atlanta area 3C-41
Table 3C-10. CAMx model performance at monitoring sites in the Midwest U.S 3C-44
Table 3C-11. CAMx model performance at monitoring sites in the Detroit area 3C-46
Table 3C-12. CAMx model performance at monitoring sites in the Central U.S 3C-49
Table 3C-13. CAMx model performance at monitoring sites in the Saint Louis area
3C-51
Table 3C-14. CAMx model performance at monitoring sites in the Dallas area 3C-53
Table 3C-15. CAMx model performance at monitoring sites in the Western U.S 3C-56
Table 3C-16. CAMx model performance at monitoring sites in the Sacramento area
3C-59
Table 3C-17. CAMx model performance at monitoring sites in the Phoenix area 3C-60
Table 3C-18. X and Y cutpoints used in Equations (3C-4) through (3C-7) 3C-73
Table 3C-19. Percent emissions changes used for each urban area to just meet each of the
air quality scenarios evaluated 3C-92
Table 3C-20. 2015-2017 design values for monitors in the Atlanta area 3C-95
3C-10
-------
Table 3C-21. 2015-2017 design values for monitors in the Boston area 3C-96
Table 3C-22. 2015-2017 design values for monitors in the Dallas area 3C-97
Table 3C-23. 2015-2017 design values for monitors in the Detroit area 3C-97
Table 3C-24. 2015-2017 design values for monitors in the Philadelphia area 3C-98
Table 3C-25. 2015-2017 design values for monitors in the Phoenix area 3C-99
Table 3C-26. 2015-2017 design values for monitors in the Sacramento area 3C-100
Table 3C-27. 2015-2017 design values for monitors in the St. Louis area 3C-101
3C-11
-------
3C.1 OVERVIEW
This appendix describes the development of the ozone (O3) air quality estimates used in
the population exposure and risk modeling described in Appendix 3D. Figure 3C-1 below shows
a flowchart of the various data sources, processes and outputs involved in generating these
ambient O3 concentration surfaces. This approach was used for eight urban study areas, which
are described further in section 3C.2.
Photochemical Modeling:
2016 CAMx with HDDM
J
\
Hourly 03
Sensitivities to
NOx Emissions
Ambient 03 Concentration Data:
Hourly 2015-2017 measurements
at individual monitors
Evaluate NOx
emissions reductions
required to meet
vair quality scenarios ,~
Hourly 03 concentrations
just meeting three a ir quality scenarios
(75 ppb, 70 ppb, 65 ppb)
at monitor locations
j Hourly census-tract level
/ O3 concentration surfaces for
8 urban casestudy areas
\ justmeeting airquality
\ scenarios
Voronoi Neighbor
Averaging (VNA)
Interpolation
Figure 3C-1. Flowchart showing inputs, processes and outputs of the approach to
generate ambient air concentration estimates for use in the exposure and
risk modeling.
Generation of the O3 concentration surfaces for the exposure and risk modeling relied on
a combination of recent monitoring data and a model-based adjustment. Ambient hourly O3
monitoring data for years 2015 through 2017 in each of the eight urban study areas was adjusted
using a model-based adjustment approach to create three different air quality scenarios. These
scenarios included conditions that just meet the current O3 standard (design value of 70 ppb), as
well as conditions that just meet two alternative air quality scenarios having design values of 75
ppb and 65 ppb. Section 3C.3 provides additional information on the monitoring data. Section
3C.4 describes the air quality modeling that was used to perform the adjustments, as well as
results from the model evaluation that was performed to assess the accuracy of the modeled
concentrations. Section 3C.5 describes the model-based adjustment approach and its application
to the ambient air quality data to create the three air quality scenarios.
3C-12
-------
The final step in preparing the air quality input data for the exposure and risk modeling is
to interpolate the adjusted air quality data from the ambient air monitoring site locations to each
census tract in the eight urban study areas using Voronoi Neighbor Averaging (VNA), which is
described in section 3C.6. Finally, section 3C.7 provides various results from the model-based
adjustment procedure and the final air quality dataset used as inputs to the Air Pollutants
Exposure Model (APEX). The APEX model and its application to air quality in the eight urban
study areas is described in Appendix 3D.
The draft PA, with a draft version of this appendix was provided to the CASAC for its
review and to the public for public comment, as summarized in section 1.4 of this PA. In
consideration of the CASAC and public comments, this appendix incorporates a number of
additions and clarifications, including the following:
• Cites section in Appendix 3D for description of study area selection (section 3C.2);
• Summarizes differences in emissions between 2014 NEI and 2016 Platform used for
modeling in this assessment (section 3C.4.1.5);
• Adds clarifications regarding the model evaluation tables and figures presented in section
3C.4.2 (Figure 3C-12 to Figure 3C-47; Table 3C-5 to Table 3C-17);
• Provides rationale for choosing nitrogen oxides (NOx) reductions only instead of the
combined NOx and volatile organic compounds (VOC) reductions which were used in
the previous review (section 3C.5.2.2.3); and
• Adds a reference to a cross-validation analysis conducted in the last review, which
supports the use of the VNA technique for generating the air quality spatial fields (section
3C.6).
3C.2 URBAN STUDY AREAS
Eight urban study areas were chosen for analysis based on several criteria, including
geographic distribution, population, current air quality levels, availability of exposure model
inputs, air quality model performance, and ambient air monitoring network coverage. The
selection criteria and any other considerations in study area selection are described in Appendix
3D, section 3D.2.1, of this PA. The eight urban study areas selected were: Atlanta, GA; Boston,
MA; Dallas, TX; Detroit, MI; Philadelphia, PA; Phoenix, AZ; Sacramento, CA; and St. Louis,
MO. Figure 3C-2 shows a map of these eight study areas and 0 provides summary information
for each area. The spatial extent of each study area was determined using the Combined
Statistical Area (CSA), with the exception of the Phoenix study area, which is not in a CSA. In
that case, the Core Based Statistical Area (CBSA) was used as the area boundary.1
1 CSA and CBSA boundaries are based on delineations promulgated by the Office of Management and Budget
(OMB) in February of 2013. CBSA and CSA delineation files are available at
https://www.census.gov/geographies/reference-files/time-series/demo/metro-micro/delineation-files.htinl.
3C-13
-------
i_~
w>
•Sacramento
mTirHiiiiirtii
(l
* r E.
Detroit*
Boston*
St. Louis* Philadelphia
Phoenix
/ I
Dallas
Atlanta
¦f
¦%>
0
Figure 3C-2. Map showing the location of the eight urban study areas.
Table 3C-1. Summary information for the eight urban study areas.
Study Area
Name
CSA Name
Land Area
(km2)
Population
(2010)
Number of
03
Monitors
2015-2017
DV(ppb)
Atlanta
Atlanta-Athens-Clarke
County-Sandy Springs, GA
30,665
5,910,296
12
75
Boston
Boston-Worcester-
Providence, MA-RI-NH-CT
25,117
7,893,376
23
73
Dallas
Dallas-Fort Worth, TX-OK
42,664
6,851,398
21
79
Detroit
Detroit-Warren-Ann Arbor,
Ml
16,884
5,318,744
13
73
Philadelphia
Philadelphia-Reading-
Camden, PA-NJ-DE-MD
18,959
7,067,807
20
80
Phoenix
Phoenix-Mesa-Scottsdale,
AZA
34,799
4,192,887
30
76
Sacramento
Sacramento-Roseville, CA
18,871
2,414,783
21
86
St. Louis
St, Louis-St. Charles-
Farmington, MO-IL
23,019
2,892,497
16
72
A The Phoenix study area is not part of a CSA. The name listed in 0 is the CBSA name.
3C-14
-------
3C.3 AMBIENT AIR OZONE MONITORING DATA
Hourly O3 concentration data for all U.S. monitoring sites for 2015-2017 was retrieved
from the EPA's Air Quality System (AQS) database in July of 2018. Design values2 for 2015-
2017 were calculated for each monitoring site according to the data handling requirements in
Appendix U to 40 CFR Part 50. Monitors within the study area boundary for each urban study
area were identified. These monitors were used to determine the NOx emissions changes
necessary to meet the current standard of 70 ppb, and the two alternative air quality scenarios
having design values of 75 ppb and 65 ppb, following the model-based adjustment approach
described in section 3C.5.
Additionally, monitors within 50 km of the study area boundary were identified as
"buffer sites." Once the emissions changes required to meet the various air quality scenarios had
been determined using the monitors within the CSA, these emissions changes were applied to
both the CSA monitors and the buffer sites, as described in section 3C.5. The purpose of the
buffer sites was to provide additional data for the spatial interpolation approach described in
section 3C.6, providing improved estimates of air quality near the edges of the urban study area
domain. Figure 3C-3 through Figure 3C-10 show maps of the boundaries for each urban study
area, along with the locations of the monitoring sites used in the analysis. In each map, the
shaded counties comprise the air quality domain of the urban study area used for estimating
exposure and risk, the monitoring sites located inside the study area are denoted by black circles,
and buffer sites are denoted by black squares.
2 The design value is the 3-year average of the annual 4th highest daily maximum 8-hour average O3 concentration.
A monitoring site meets the current standard if its design value is less than or equal to 70 ppb.
3C-15
-------
Figure 3C-3. Map of the Atlanta study area. Counties in the CSA are shaded, monitoring
sites in the CSA are denoted by black circles, and buffer sites are denoted by
black squares.
3C-16
-------
_C~?
Figure 3C-4. Map of the Boston study area. Counties in the CSA are shaded, monitoring
sites in the CSA are denoted by black circles, and buffer sites are denoted by
black squares.
3C-17
-------
Figure 3C-5. Map of the Dallas study area. Counties in the CSA are shaded, monitoring
sites in the CSA are denoted by black circles, and buffer sites are denoted by
black squares.
3C-18
-------
Figure 3C-6. Map of the Detroit study area. Counties in the CSA are shaded, monitoring
sites in the CSA are denoted by black circles, and buffer sites are denoted by
black squares.
3C-19
-------
Figure 3C-7. Map of the Philadelphia study area. Counties in the CSA are shaded,
monitoring sites in the CSA are denoted by black circles, and buffer sites are
denoted by black squares.
3C-20
-------
••
• •
Figure 3C-8. Map of the Phoenix study area. Counties in the CBS A are shaded, monitoring
sites in the CBS A are denoted by black circles, and buffer sites are denoted by
black squares.
3C-21
-------
Figure 3C-9. Map of the Sacramento study area. Counties in the CSA are shaded,
monitoring sites in the CSA are denoted by black circles, and buffer sites are
denoted by black squares.
3C-22
-------
Figure 3C-10. Map of the St. Louis study area. Counties in the CSA are shaded, monitoring
sites in the CSA are denoted by black circles, and buffer sites are denoted by
black squares.
It is worth noting that for an area to show compliance with the current O3 standard, all
monitors within the urban area must have design values less than or equal to 70 ppb. According
to Appendix U to 40 CFR Part 50, air quality monitors must also meet certain data completeness
requirements to show compliance with the standard. However, any design value based on 3 years
of monitoring data that exceeds the standard is not in compliance, regardless of data
completeness. Therefore, when performing the air quality adjustments to create the three air
quality scenarios, all monitors in each urban study area with data reported for each of the 3 years
were included, regardless of data completeness.
Finally, per Appendix U to 40 CFR Part 50, data not meeting the ambient air monitoring
requirements in 40 CFR Part 58, data reported using methods other than Federal Reference or
Equivalent Methods, and data concurred by the appropriate EPA Regional Office as having been
affected by an exceptional event were excluded from design value calculations. However, once
the emissions changes required to determine compliance with the various air quality scenarios
3C-23
-------
had been determined, these values were included in the final adjustment and spatial interpolation.
In practice, fewer than 10,000 hourly concentrations out of more than 3 million (-0.3%) were
excluded from design value calculations in this manner.
3C.4 AIR QUALITY MODELING DATA
3C.4.1 Comprehensive Air Quality Model with Extensions (CAMx)
3C.4.1.1 Model Set-up and Simulation
The Comprehensive Air Quality Model with Extensions (CAMx) was used as the
modeling tool for this assessment. CAMx is a peer-reviewed model that simulates the formation
and fate of photochemical oxidants, aerosol concentrations, acid deposition, and air toxics, over
multiple scales for given input sets of meteorological conditions and emissions. CAMx is used
frequently for a range of scientific and regulatory applications related to the analysis of air
quality in the U.S. The Higher Order Direct Decoupled Method (HDDM) was implemented in
CAMx to estimate the model sensitivities to emissions changes as described in section 3C.5 of
this appendix. The CAMx-HDDM configuration tracks gas-phase species concentrations through
all modeled processes. However, HDDM implemented in CAMx does not track the effects of
aerosol and cloud processing on calculated O3 sensitivities. Differences in predicted O3
concentrations between the CAMx-HDDM configuration described here and a standard CAMx
v6.5 simulation with full treatment of aerosol-03 interactions did not influence O3 predictions in
the urban study areas examined in this assessment. CAMx v6.53 was run using the carbon bond
version 6 (CB06r4) gas-phase chemical mechanism (Yarwood et al., 2010; Gery et al., 1989) and
the AER06 aerosol module which includes ISORROPIA for gas-particle partitioning of
inorganic species (Nenes et al., 1998) and secondary organic aerosol treatment as described in
Carlton et al. (2010).
3C.4.1.2 Model Domain
For this analysis, all CAMx runs were performed for a domain that covers the 48
contiguous states including portions of southern Canada and Northern Mexico with a 12 x 12 km
resolution (Figure 3C-11). The CAMx simulations were performed with 35 vertical layers with a
top layer at about 17,600 meters, or 50 millibars (mb). Table 3C-2 and Table 3C-3 provide some
basic geographic information regarding the CAMx domain and vertical layer structure,
respectively. Results from the lowest layer of the model were used for analyses to support the
risk and exposure analyses described in Appendix 3D.
3 For more information, see: http://www.camx.com/files/camxusersguide_v6-50.pdf.
3C-24
-------
12US2 domain - ' \
x,y origin: -2412000m, >1621
col: 396 row:246 1 :
L J 1 tL
Figure 3C-11. Map of the CAMx modeling domain.
Table 3C-2. Geographic elements of domain used in the CAMx/HDDM modeling.
Domain Element
CAMx Modeling Configuration Grid
Map Projection
Lambert Conformal Projection
Grid Resolution
12 km
True Latitudes
33 deg N and 45 deg N
Grid Dimensions
396 x 246 x 35
Vertical extent
35 Layers: Surface to 50 millibar level
3C-25
-------
3C.4.1.3 Model Time Period
The CAMx/HDDM modeling was performed for January 1 - December 31 of 2016. The
simulations included a 10-day spin-up period4 from December 22-31, 2015. The spin-up days
were not considered in the analysis for the HDDM results.
3C.4.1.4 Model Inputs: Meteorology
CAMx model simulations require inputs of meteorological fields, emissions, and initial
and boundary conditions. The gridded meteorological data for the entire year of 2016 at the 12
km continental U.S. scale domain were derived from version 3.8 of the Weather Research and
Forecasting Model (WRF), Advanced Research WRF (ARW) core (Skamarock et al., 2008). The
WRF Model is a mesoscale numerical weather prediction system developed for both operational
forecasting and atmospheric research applications.5 The 2016 WRF simulation included the
physics options of the Pleim-Xiu land surface model (LSM), Asymmetric Convective Model
version 2 planetary boundary layer (PBL) scheme, Morrison double moment microphysics,
Kain-Fritsch cumulus parameterization scheme and the RRTMG long-wave radiation (LWR)
scheme (Gilliam and Pleim, 2009). Additionally, lightning data assimilation was utilized to
suppress (force) deep convection where lightning was absent (present) in observational data.
This method is described by Heath et al. (2016) and was employed to help improve precipitation
estimates generated by the WRF model.
The WRF and CAMx simulations used the same map projection, a lambert conformal
projection centered at (-97, 40) with true latitudes at 33 and 45 degrees north. The WRF and
CAMx simulations utilized 35 vertical layers with a surface layer of approximately 19 meters.
Table 3C-3 shows the vertical layer structure used in WRF to generate the CAMx meteorological
inputs.
The WRF meteorological outputs were processed to create model-ready inputs for CAMx
using the wrfcamx version 4.3 meteorological pre-processor (Ramboll Environ, 2014). The
specific meteorological inputs to CAMx include: horizontal wind components (i.e., speed and
direction), temperature, moisture, vertical diffusion rates, and rainfall rates for each grid cell in
each vertical layer.
4 It is standard practice to allow chemical transport models to run for several days to weeks prior to the time period
of interest in order to minimize the influence of initial conditions.
5 See: http://wrf-model.org
3C-26
-------
Table 3C-3. Vertical layer structure for 2016 WRF and CAMx simulations.
Layer Top
Height (m)
Pressure
(mb)
Model
Layer
17,556
50
35
14,780
97.5
34
12,822
145
33
11,282
192.5
32
10,002
240
31
8,901
287.5
30
7,932
335
29
7,064
382.5
28
6,275
430
27
5,553
477.5
26
4,885
525
25
4,264
572.5
24
3,683
620
23
3,136
667.5
22
2,619
715
21
2,226
753
20
1,941
781.5
19
1,665
810
18
1,485
829
17
1,308
848
16
1,134
867
15
964
886
14
797
905
13
714
914.5
12
632
924
11
551
933.5
10
470
943
9
390
952.5
8
311
962
7
232
971.5
6
154
981
5
115
985.75
4
77
990.5
3
38
995.25
2
19
997.63
1
A detailed meteorological model performance evaluation was conducted for the 2016
WRF simulations (U.S. EPA, 2017). The analysis included statistical evaluation of temperature,
wind speed, and water vapor mixing ratios against observational data from airports, as well as
evaluations of monthly precipitation compared to the Parameter-elevation Relationships on
Independent Slopes Model (PRISM) and shortwave radiation compared to data from the Surface
3C-27
-------
Radiation Budget Measurement Network (SURFRAD) and the Solar Radiation Network
(SOLRAD).
3C.4.1.5 Model Inputs: Emissions
The emissions data used are based on the alpha version of the Inventory Collaborative
2016 emissions modeling platform.6 The modeling case used is abbreviated "2016fe" and is
publicly available.7
Emissions were processed to photochemical model inputs with the SMOKE modeling
system version 4.5 (Houyoux et al., 2000). For this analysis, emissions from wildfires and
prescribed burns were based on year 2016 nationally available fire datasets. Electric generating
unit (EGU) emissions are temporally allocated to hourly values based on patterns derived from
year 2016 Continuous Emissions Monitoring System (CEMS) data. In addition, U.S. emissions
are included from other point sources, area sources, agricultural sources (ammonia only),
anthropogenic fugitive dust sources, nonroad mobile sources, onroad mobile sources, and
biogenic sources. Emissions for onroad mobile sources were created using the EPA's MOVES
2014a model,8 except that California emissions were adjusted to match the county total
emissions obtained directly from the California Air Resources Board. Biogenic emissions were
estimated using the Biogenic Emissions Inventory System version 3.61 (BEISv3.61) (Pouliot and
Bash, 2015). Other North American emissions from areas outside the U.S. are based on a 2013
Canadian inventory scaled to 2015, and projections of the 2008 Mexican inventory to the year
2016 along with the scaling of MOVES-Mexico emissions to year 2016 (ERG, 2017). The
construction of the emissions is described in more detail in the technical support document
Preparation of Emissions Inventories for the Version 7.1 2016 Regional Emissions Modeling
Platform (U.S. EPA, 2019). Emissions totals within the United States are summarized in Table
3C-4 for CO, NH3, NOx, PM10, PM2.5, SO2, and VOC. Anthropogenic NOx emissions in the
2016 platform are about 19% lower than those reported in the 2014 NEI due to both improved
inventory development methods and updates to specific components (e.g., cleaner vehicles
entering the onroad mobile fleet or EGUs transitioning from coal to natural gas).
6 http://views.cira.colostate.edu/wiki/wiki/9169
1 https://www.epa.gov/air-emissions-modeling/2016-alpha-platform
8 https://www.epa.gov/moves
3C-28
-------
Table 3C-4. Summary of U.S. emissions totals by sector for the 12km CONUS domain (in
thousand tons). "NA" indicates not applicable.
Sector
Abbrev.
Sector Description
CO
nh3
NOx
PM10
PM2.5
S02
voc
afdust_adj
Anthropogenic fugitive dust
NA
NA
NA
6,217
874
NA
NA
ag
Agricultural sources
NA
2,777
NA
NA
NA
NA
NA
ptagfire
Agricultural fires
593
80
18
96
68
6
36
cmv_c1c2
Category 1 and 2
Commercial Marine
Vessels
47
NA
260
6
6
NA
5
cmv_c3
Ocean-going (Category 3)
Commercial Marine
Vessels
11
NA
108
4
4
4
5
nonpt
Nonpoint (area) sources
not in other sectors
2,681
121
758
609
496
162
3,673
np_oilgas
Nonpoint oil and gas
sources
642
NA
676
18
17
39
2,986
nonroad
Nonroad (off-road)
equipment
12,189
2
1,207
122
115
2
1,465
onroad
Onroad mobile sources
20,446
101
4,046
273
130
27
1,962
ptfire
Wild and Prescribed Fires
23,642
388
333
2,415
2,046
181
5,581
ptegu
Point sources: electric
generation units
672
25
1,289
171
141
1,545
33
ptnonipm
Point sources other than
electric generating units
1,848
61
1,073
407
264
673
809
pt_oilgas
Oil and gas-related Point
Sources
178
4
360
12
11
42
133
rail
Locomotive emissions
118
NA
673
21
19
1
35
rwc
Residential Wood
Combustion emissions
2,099
15
30
314
314
8
338
Total anthro
Total US anthropogenic
emissions (including
wildfires)
65,167
3,576
10,832
10,685
4,507
2,689
17,241
beis
U.S. biogenic emissions
7,297
NA
979
NA
NA
NA
42,861
Total with
biogenic
Total US emissions
including biogenic
emissions
72,463
3,576
11,812
10,685
4,507
2,689
60,102
3C.4.1.6 Model Inputs: Boundary and Initial Conditions
Initial and lateral boundary concentrations for the 12 km US2 domain are provided by the
hemispheric version of the Community Multi-scale Air Quality model (H-CMAQ) v5.2.1. H-
CMAQ was run for 2016 with a horizontal grid resolution of 108 km and 44 vertical layers up to
50 hPa. The H-CMAQ predictions were used to provide one-way dynamic boundary conditions
at one-hour intervals. An operational evaluation against sonde and satellite observations showed
3C-29
-------
that the 2016 H-CMAQ simulation reasonably captured general patterns of O3 transport within
the northern Hemisphere that are relevant for the 12US2 domain (Henderson et al., 2018).
3C.4.2 Evaluation of Modeled Ozone Concentrations
In this section we present the results of an evaluation of the CAMx configuration used to
produce the air quality results described in Chapter 3. Specifically, we summarize the ability of
the CAMx model to reproduce the corresponding 2016 measured O3 concentrations. This
operational evaluation shows that in general for most regions and seasons, the CAMx model
predictions for 2016 generally reproduce patterns of observed O3. The notable exception to this
is a persistent underestimate in winter across almost all regions, particularly at higher latitude
sites.
In the following sections we present general model performance statistics and plots for
five regions of the U.S. We compare model predictions of maximum daily 8-hr average (MDA8)
O3 concentrations to measurements reported in EPA's AQS. We note that these comparisons are
based on MDA8 values calculated across all available modeled CAMx values and all observed
(AQS) concentrations, and that these comparisons include buffer sites. Model performance could
be different for comparisons without buffer sites, or using the modeled CAMx MDA8 values
only when the corresponding observed MDA8 values are available.
The model statistics presented here include mean bias, mean error, normalized mean bias,
and normalized mean error as calculated below, where n represents the total number of
observations:
Mean Bias: (£ modeled — observed)/n
Mean Error: (£ |modeled — observed \)/n
Normalized Mean Bias: (£ modeled — observed)/(£ observed)
Normalized Mean Error (£ \modeled — observed])/(£ observed)
Our analysis focuses on regional model evaluation statistics from five US regions as well
as evaluations of the eight urban study areas included in the exposure and risk analysis - Atlanta,
Boston, Dallas, Detroit, Philadelphia, Phoenix, Sacramento, and St. Louis.9,10 Statistics for
CAMx model performance in these regions and urban study areas are shown by season in Table
3C-5 through Table 3C-17 for observed days with MDA8 O3 values > 60 ppb, observed days
9 The five regions are defined as follows: Northeast (Connecticut, Delaware, District of Columbia, Maine,
Maryland, Massachusetts, New Hampshire, New Jersey, New York, Pennsylvania, Rhode Island, Vermont),
Southeast (Alabama, Florida, Georgia, Kentucky, Mississippi, North Carolina, South Carolina, Tennessee,
Virginia, West Virginia), Midwest (Illinois, Indiana, Michigan, Ohio, Wisconsin), Central (Arkansas, Iowa,
Kansas, Louisiana, Minnesota, Missouri, Nebraska, Oklahoma, Texas), and West (Arizona, California, Colorado,
Idaho, Nevada, New Mexico, Oregon, Utah, Wyoming).
10 Monitoring sites for each urban study area were selected based on core-based statistical area (CBSA) groupings.
3C-30
-------
with MDA8 O3 < 60 ppb, and for all observed days. For each of the five regions listed above,
spatial plots are provided for each season showing Normalized Mean Bias (NMB) for MDA8 O3
at individual sites. Summary NMB ranges are included at the bottom of each map showing the
min and max values for the season/region across all sites, as well as the 25th, 50th, and 75th
percentile values. Time series plots are provided for MDA8 O3 in each urban study area for the
period from January-December 2016. Hourly time series plots are also provided for one month in
each season (January, April, July, October).11
3C.4.2.1 Operational Evaluation in the Northeastern U.S.
Table 3C-5 shows that in the Northeast Region, model mean bias was generally less than
7 ppb and normalized mean bias was less than 15% in most cases. Errors were largest in the
winter, with underestimates also extending to the spring. Spatial maps of normalized mean bias
are shown in Figure 3C-12 through Figure 3C-15. During the O3 season performance was best on
high O3 days, particularly in the summer and fall. Two of the eight urban study areas evaluated
were in the Northeast: Boston and Philadelphia.
Model performance at the Boston study area monitoring sites (Table 3C-6) was similar to
that of the Northeast Region. The time series plots show that the model reasonably reproduces
the measured day-to-day variability in MDA8 O3 concentrations (Figure 3C-16). The
underestimate in winter-spring observed in the Northeast region statistics is particularly
pronounced in Boston, likely due to its relatively northerly location where seasonal daylight and
temperature changes are more exaggerated. Variability of hourly daytime and nighttime O3
concentrations is generally well modeled in all seasons, again noting the persistent underestimate
in January/April. Model characterization of hourly variability is particularly good in July,
although peak daytime O3 is slightly overestimated. Nighttime O3 is also consistently
overestimated in July/October (Figure 3C-17).12
Bulk model performance statistics for Philadelphia (Table 3C-7) are again similar to
those for the Northeast as a whole, with more moderate performance compared to Boston during
both winter (not as poor) and summer/fall (not as good). The spring underestimate present in the
Boston comparisons is much smaller for Philadelphia (Figure 3C-18, Figure 3C-19), again
suggesting that the winter-spring underestimate is more pronounced at more northerly sites.
Philadelphia also exhibits the nighttime overestimates in the July/October hourly comparisons
seen in Boston, with slightly higher overestimates of peak July daytime concentrations.
11 Note that the MDA8 and hourly time series show average concentrations across all monitors within each urban
study area. The number of monitors included in this average sometimes changes by season since different
monitors within each study area take measurements over different periods of the year.
12 Note that the Y-axis scale for the various time series are not consistent.
3C-31
-------
Table 3C-5. CAMx model performance at monitoring sites in the Northeastern U.S.
Statistics shown are mean bias (MB), normalized mean bias (NMB), mean
error (ME), and normalized mean error (NME).
Season
MDA8 level
(PPb)
No. of obs
MB (ppb)
NMB (%)
ME (ppb)
NME (%)
Winter
Days < 60
7056
-6.4
-21.0
7.3
23.8
Days > 60
1
-26,7
-42.4
26.7
42.4
All Days
7057
-6,4
-21.0
7,3
23.8
Spring
Days < 60
7493
-6,2
-14,7
7,8
18.6
Days > 60
511
-5,1
-7,6
7,3
10.8
All Days
8004
-6.1
-14.0
7.7
17,8
Summer
Days < 60
7385
5,0
11,8
7.7
18.1
Days > 60
870
0,8
1.2
6.7
10,2
All Days
8255
4.5
10.1
7,6
16.9
Fall
Days < 60
7612
1.3
3.9
5,6
17,6
Days > 60
135
-0,9
-1,4
5,4
8.1
All Days
7747
1,2
3,7
5,6
17,3
Bias Summary: [ miri, 25th %, 50th %, 75th %, max ]
[-55, -26, -22. -16, 14]
Figure 3C-12. Normalized mean bias for MDA8 O3 in the Northeastern U.S., winter 2016.
3C-32
-------
P«1
Bias Summary: [ min, 25th %, 50th %, 75th %, max ]
[-21, 7.8, 17, 28, 85)
Figure 3C-14. Normalized mean bias for MDA8 O3 in the Northeastern U.S., summer
2016.
Bias Summary: [ min, 25th %, 50th %, 75th %, max ]
[-43, -20, -14, -7.1, 23]
Figure 3C-13. Normalized mean bias for MDA8 O3 in the Northeastern U.S., spring 2016.
3C-33
-------
® Q.
O
oo
Bias Summary: [ min, 25th %, 50th %. 75th %, max ]
[-37, 4.8, 15, 28, 93]
Figure 3C-15. Normalized mean bias for MDA8 O3 in the Northeastern U.S., fall 2016.
Table 3C-6. CAMx model performance at monitoring sites in the Boston study area.
Season
MDA8 level
(PPb)
No. of obs
MB (ppb)
NMB (%)
ME (ppb)
NME (%)
Winter
Days < 60
1346
-8.4
-25.6
8.9
27.2
Days > 60
0
NA
NA
NA
NA
All Days
1346
-8.4
-25.6
8.9
27.2
Spring
Days < 60
82
-9.1
-21.3
9.9
23.3
Days > 60
1476
-8,6
-12.6
10,4
15.2
All Days
1558
-9,0
-20.6
9.9
22,6
Summer
Days < 60
1484
3.6
9.0
6,2
15,7
Days > 60
146
1.2
1.8
5.9
8,9
All Days
1630
3.3
8.0
6.2
14,8
Fall
Days < 60
1482
-0.6
-1.8
5.4
17,4
Days > 60
8
0.3
0.43
5.4
8.4
All Days
1490
-0.6
-1.8
5.4
17.3
3C-34
-------
AOS MDA8 Comparison for Boston Monitors in 2016
eo
60 -
8 40 -
OQ
<
o
20 -
0 -
Figure 3C-16. Time series of monitored (black) and modeled (red) MDA8 O3 at Boston
monitoring sites in 2016.
AOS MDAfl
CAMx HDDM 2016fe
# sites:37
VrtiK ,
m w
01/01 02/02 03/05 04/06 05/08 06,¦'09 07/11 08/12 09/13 10/15 11/16 12/16
Date
3C-35
-------
AQS Comparison for Boston Monitors in January 2016
AOS Comparison for Boston Monitors in April 2016
AQS Hourly
CAMx HDDM 2016fe
60 -
3 40 '
3
6
I
20 -
03/31 04/03 04/06 04/08 04/11 04/13 04/16 04/19 04/21 04/24 04/27 04/29
40 -
AOS Hourly
CAMx HDDM 20l6le
30 -
i
8 20 -
•e
I
10 -
0101 01/03 01*06 01/09 01/11 01/14 01/17 01/19 01/22 01/25 01/27 01/30
AQS Comparison for Boston Monitors in July 2016 AQS Comparison for Boston Monitors in October 2016
AOS Hourly
CAMx HDDM 2016le
60 -
8
-
0
1
20 -
06/30 07AJ3 07/06 07/08 07/11 07/14 07/17 07/19 07/22 07/25 07/27 0730
AOS Hourly
CAMx HDDM 20161©
50 -
a
&
8
30 -
-
8
I
20 -
09/30 10/03 1006 1008 10/11 10/14 10/17 10/19 10/22 10/25 1027 1030
Figure 3C-17. Time series of monitored (black) and modeled (red) hourly O3
concentrations at Boston monitoring sites in January (top left), April (top
right), July (bottom left), and October (bottom right) 2016.
Table 3C-7. CAMx model performance at monitoring sites in the Philadelphia study area.
Season
MDA8 level
(PPb)
No. of obs
MB (ppb)
NMB (%)
ME (ppb)
NME (%)
Winter
Days < 60
2151
-5.0
-17.9
6.1
21.7
Days > 60
0
NA
NA
NA
NA
All Days
2151
-5.0
-17.9
6.1
21.7
Spring
Days < 60
2328
-4.5
-10.9
6.6
16.0
Days > 60
150
-3,0
-4.4
5.1
7.5
All Days
2478
-4,4
-10.3
6.5
15.2
Summer
Days < 60
2229
6.7
14.7
9.1
20.2
Days > 60
352
1.0
1.5
6.8
10,3
Ail Days
2581
5,9
12,3
OO
CO
18.3
Fall
Days < 60
2333
1,9
5,9
5.7
17.7
Days > 60
71
-1.0
-1.4
5.2
7.7
All Days
2404
1.8
5,5
5.7
17.1
3C-36
-------
AQS MDA8 Comparison for Philadelphia Monitors in 2016
80
a 60
Q.
Q.
8
3 40
20
0 J
AQS MDA8
CAMx HDDM 2016fe
# sites:47
01/01 02/02 03/05 04/06 05/08 06/09 07/11 08/12 09/13 10/15 11/16 12/18
Date
Figure 3C-18. Time series of monitored (black) and modeled (red) MDA8 Os at
Philadelphia monitoring sites in 2016.
3C-37
-------
AOS Comparison for Philadelphia Monitors in January 2016
AOS Comparison tor Philadelphia Monitors in April 2016
— AOS Hourly
CAMXHDDM 20161®
01/01 01/03 0106 01/09 01/11 01/14 01/17 01/19 01/22 01/25 01/27 01/30
Date
AOS Comparison for Philadelphia Monitors in July 2016
— AOS Hourly
CAMXHDDM 2016fe
03/31 04/03 0406 04/08 04/11 04/13 04/16 04/19 04.-21 04,24 04/27 04 29
Date
AQS Comparison for Philadelphia Monitors in October 2016
8
f 40
AOS Hourly
CAMx HDDM 2016fe
06*0 07,03 0706 0708 07/11 07/14 07/17 07/19 07/22 07/25 07/27 07/30
Date
70 -
60
50
I 40 H
8
Is 30
0
1
20
10 -
0 -
AOS Hourly
CAMXHDDM 2016te
09/30 1003 10O6 10/08 10/11 10/14 10/17 10/19 10/22 10/25 10/27 10-30
Date
Figure 3C-19. Time series of monitored (black) and modeled (red) hourly O3
concentrations at Philadelphia monitoring sites for January (top left), April
(top right), July (bottom left), and October (bottom right) 2016.
3C.4.2.2 Operational Evaluation in the Southeastern U.S.
In the Southeast region, mean bias for MDA8 O3 was generally less than ~5 ppb at most
sites in all seasons, as indicated in Table 3C-8. The exception is winter, where there were only
four days with measured MDA8 > 60 ppb and all were largely underpredicted. Spatial maps of
normalized mean bias are shown in Figure 3C-20 through Figure 3C-23. Performance was best
in the spring (slightly underestimated) and on high O3 days in the summer/fall. Atlanta was the
only one of the eight urban study areas located in the Southeast region.
Mean bias and normalized mean bias at Atlanta sites for the spring, summer, and fall
months were typical of performance throughout the Southeast region, with much better
performance in winter. The MDA8 O3 time series (Figure 3C-24) shows that the model
reasonably represents the variability occurring on high and low O3 concentration days. The
hourly time series plots (Figure 3C-25) also show reasonable model performance during daytime
hours but some persistent overestimates of both nighttime and peak daytime O3 occur, especially
in July.
3C-38
-------
Table 3C-8. CAMx model performance at monitoring sites in the Southeastern U.S.
Season
MDA8 level
(PPb)
No. of obs
MB (ppb)
NMB (%)
ME (ppb)
NME (%)
Winter
Days < 60
3775
-3.2
-9,2
5.3
15.4
Days > 60
4
-27.2
-40.6
27,2
40.6
All Days
3779
-3,2
-9.2
5.3
15.4
Spring
Days < 60
7193
-0,6
-1.4
5.2
11.7
Days > 60
468
-2.6
-4.0
5.0
7.8
All Days
7661
-0.7
-1.6
5,2
11,3
Summer
Days < 60
7825
5,2
13,9
7.6
20,2
Days > 60
396
0,4
0.6
6.2
9.5
All Days
8221
5.0
12.8
7.5
19.3
Fail
Days < 60
6456
3.4
8,7
6.0
15.5
Days > 60
139
0.6
0.9
4.8
7.6
All Days
6595
3.3
8.4
6.0
15,2
100
100
Bias Summary: [ min, 25th %, 50th %, 75th %, max ]
[-32. -9,6, -1.8, 7.6, 51]
Figure 3C-20. Normalized mean bias for MDA8 O3 in the Southeastern U.S., winter 2016.
3C-39
-------
100
- 20
-100
Bias Summary: [ min, 25th %, 50th %, 75th %, max )
[-18, -1.4, 5.1, 13, 44]
Figure 3C-21. Normalized mean bias for MDA8 O3 in the Southeastern U.S., spring 2016.
100
Bias Summary: [ min, 25th %, 50th %, 75th %, max ]
[-22, 16, 27, 39, 130]
Figure 3C-22. Normalized mean bias for MDA8 O3 in the Southeastern U.S., summer
2016.
3C-40
-------
Bias Summary: [ min, 25th %, 50th %, 75th %, max ]
[-12, 12, 29, 44, 120]
Figure 3C-23. Normalized mean bias for MDA8 O3 in the Southeastern U.S., fall 2016.
Table 3C-9. CAMx model performance at monitoring sites in the Atlanta study area.
Season
MDA8 level
(PPb)
No. of obs
MB (ppb)
NMB (%)
ME (ppb)
NME (%)
Winter
Days < 60
91
-0.9
-3,3
3,4
12.4
Days > 60
0
NA
NA
NA
NA
All Days
91
-0,9
-3.3
3.4
12,4
Spring
Days < 60
747
1,4
3,1
4.7
10,6
Days > 60
54
-1,4
-2.1
4.9
7,3
Ail Days
801
1,2
2.6
4.7
10,3
Summer
Days < 60
717
5,4
13.4
6.9
17.1
Days > 60
93
-1,1
-1.6
6.0
8.9
All Days
810
4,7
10,7
6.8
15.6
Fail
Days < 60
520
5,6
12,8
6.5
15.1
Days > 60
26
3,8
6,0
5.2
8.2
Ail Days
546
5,5
12,4
6.5
14.6
3C-41
-------
AQS MDA8 Comparison for Atlanta Monitors in 2016
80 1
_ 60 H
-O
CL
Q.
8
00 40
<
Q
20 -
0 -J
AQS MDA8
CAMx HDDM 2016fe
# sites:18
—lllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllilllllllllllllllllilllllllllllllllillllllllll
01/01 02/02 OS/OS 04/06 05/08 06/09 07/11 08/12 09/13 10/15 11/16 12/18
Date
Figure 3C-24. Time series of monitored (black) and modeled (red) MDA8 O3 at Atlanta
monitoring sites in 2016.
3C-42
-------
AOS Comparison for Atlanta Monitors in January 2016
AOS Comparison for Atlanta Monitors in April 2016
AOS Hourly
CAMx HODM 2016fe
01/01 01/03 01/06 01/09 01/11 01/14 01/17 01/19 01/22 01/25 01/27 01/30
Date
AOS Comparison for Atlanta Monitors in July 2016
8 40
AOS Hourly
CAMx HODM 2016te
03/31 04/03 04/06 04/08 04/11 04/13 04/16 04/19 04/21 04/24 04/27 04/29
Date
AOS Comparison for Atlanta Monitors in October 2016
AOS Hourly
CAMx HDDM 2016fe
ii
06/30 07/03 07/06 07/08 07/11 07/14 07/17 07/19 07/22 07/25 07/27 07/30
AOS Hourly
CAMx HDDM 2016fe
09/30 10/1)3 10/06 10/08 10/11 10/14 10/17 10/19 10/22 10/25 10/27 10/30
Date
Figure 3C-25. Time series of monitored (black) and modeled (red) hourly O3
concentrations at Atlanta monitoring sites in January (top left), April (top
right), July (bottom left), and October (bottom right) 2016.
3C.4.2.3 Operational Evaluation in the Midwest U.S.
Mean bias for MDA8 O3 in the Midwest region was around 6 ppb or less at most sites for
all seasons (Table 3C-10), except for high O3 days in spring. Normalized mean bias for MDA8
O3 was less than 15%, except in the winter when it was somewhat higher (~20%). Normalized
mean error was lowest on high O3 days in spring, summer, and fall, even though bias
performance was not notably better during these times. No distinct spatial patterns are apparent
from the maps of normalized mean bias (Figure 3C-26 through Figure 3C-29). Detroit was the
only one of the eight urban study areas located in the Midwest.
Detroit performance statistics for MDA8 O3 were similar to those from the rest of the
Midwest. However, under-estimates on high O3 days were more pronounced in Detroit than in
the rest of the region. The time series shows that the model accurately estimates both day and
nighttime hourly O3 in Detroit in April and July and generally captures the variations in MDA8
O3 throughout the year, although the persistent under-estimate in winter-spring is evident (Figure
3C-30, Figure 3C-31).
3C-43
-------
Table 3C-10. CAMx model performance at monitoring sites in the Midwest U.S.
Season
MDA8 level
(PPb)
No. of obs
MB (ppb)
NMB (%)
ME (ppb)
NME (%)
Winter
Days < 60
1775
-5.8
-20.2
6.4
22.4
Days > 60
0
NA
NA
NA
NA
All Days
1775
-5,8
-20,2
6,4
22.4
Spring
Days < 60
3635
-5,9
-14,1
7.6
18.1
Days > 60
370
-8.3
-12.5
9.2
14.0
All Days
4005
-6.1
-13.9
7.8
17,6
Summer
Days < 60
4680
3,3
7.8
7.4
17,8
Days > 60
556
-4.9
-7.3
8.6
12,8
All Days
5236
2.4
5.4
7.6
17.0
Fail
Days < 60
3439
2.2
6.7
5,1
15.3
Days > 60
51
3.3
5.1
5,6
8.6
All Days
3490
2.3
6.7
5,1
15,1
Bias Summary: [ min, 25th %, 50th %, 75th %, max ]
[-44, -27, -22, -16, -0.96]
Figure 3C-26. Normalized mean bias for MDA8 Oi in the Midwest U.S., winter 2016.
3C-44
-------
$°o
Bias Summary: [ min, 25th %, 50th %, 75th %, max ]
[-38, -18, -11, -4, 27]
Figure 3C-27. Normalized mean bias for MDA8 O3 in the Midwest U.S., spring 2016.
3?
100
Bias Summary: [ min, 25th %, 50th %, 75th %, max ]
[-20, 4,4, 11, 22, 87]
Figure 3C-28. Normalized mean bias for MI)A8 O3 in the Midwest U.S., summer 2016.
3C-45
-------
Bias Summary: [ min, 25th %, 50th %, 75th %, max ]
[-14, 10, 20, 31, 130]
Figure 3C-29. Normalized mean bias for MDA8 O3 in the Midwest U.S., fall 2016.
Table 3C-11. CAMx model performance at monitoring sites in the Detroit study area.
Season
MDA8 level
(PPb)
No. of obs
MB (ppb)
NMB (%)
ME (ppb)
NME (%)
Winter
Days < 60
29
-4,1
-19.5
5.9
26.3
Days > 60
0
NA
NA
NA
NA
All Days
29
-4,1
-19.5
5.9
26.3
Spring
Days < 60
337
-6.5
-15.8
8.3
20,0
Days > 60
28
-9.4
-13.5
10.0
14,4
All Days
365
-6.7
-15.5
8.4
19.3
Summer
Days < 60
485
2.0
4,7
6,8
16,1
Days > 60
59
-5,3
-8.1
7.9
12,1
All Days
544
1.2
2.7
6.9
15,5
Fall
Days < 60
245
3.1
9.7
5.6
17,2
Days > 60
3
-4.1
-6.7
4.1
6,7
All Days
248
3.0
9.3
5.5
17,0
3C-46
-------
AQS MDA8 Comparison for Detroit Monitors in 2016
80
~ 60
a
Q.
Q.
8
« 40
Q
20
0 -1
AQS MDA8
CAMx HDDM 2016fe
# sites21
01/01 02/02 03/05 04/06 05/08 06/09 07/11 08/12 09/13 10/15 11/16 12/18
Date
Figure 3C-30. Time series of monitored (black) and modeled (red) MDA8 O3 at Detroit
monitoring sites in 2016.
3C-47
-------
AQS Comparison for Detroit Monitors in January 2016
AOS Hourly
CAMxHDDM 20l6fe
014)1 01/03 01/06 01/09 01/11 01/14 01/17 01/19 01-22 01/25 01/27 01/30
Date
AQS Comparison for Detroit Monitors in July 2016
AOS Hourly
CAMxHDDM 2016le
= S.tfT. j'
06/30 07/03 07/06 07/08 07/11 07/14 07/17 07/19 07/22 07/25 07/27 07/30
Date
09/30 10/03 10/06 104)6 10/11 10/14 10/17 10/19 10/22 10/25 1027 10/30
Dale
Figure 3C-31. Time series of monitored (black) and modeled (red) hourly O3
concentrations at Detroit monitoring sites in January (top left), April (top
right), July (bottom left), and October (bottom right) 2016.
3C.4.2.4 Operational Evaluation in the Central U.S.
Mean bias for MDA8 O3 concentrations in the Central U.S. is within 4 ppb, except for
high days in winter (-6 ppb) and spring (-7 ppb) (Table 3C-12). Normalized mean error is within
15%, except for days < 60 ppb in winter and summer (-18%). Spatial maps of normalized mean
bias are shown in Figure 3C-32 through Figure 3C-35. Overall performance is best on lower O3
days in spring and high O3 days in summer and fall. St. Louis and Dallas were the only two of
the eight study areas which are located in the Central U.S. region.
St. Louis mean bias for MDA8 was within 5 ppb for all days and seasons. A north-south
gradient in NMB is apparent during both the winter and spring seasons in the maps shown in
Figure 3C-32 and Figure 3C-33, with larger underestimates visible at higher latitude/more
northerly monitors. Overall performance for St. Louis was best on high O3 days in summer. The
MDA8 time series shows reasonable agreement between CAMx and the monitor data for most of
the year (Figure 3C-36), with underestimates in January and overestimates in July also apparent
in the hourly time series (Figure 3C-37).
AQS Hourly
CAMxHDDM 2016le
# sites 20
AQS Comparison for Detroit Monitors in April 2016
03/31 04/03 04/06 04-06 04/11 04/13 04/16 04/19 04.21 04/24 04/27 04.29
Date
AQS Comparison for Detroit Monitors in October 2016
# sites 21
AOS Hourly
CAMxHDDM 2016(e
3C-48
-------
Performance statistics for MDA8 O3 in Dallas were better than those for the broader
region, with mean bias less than 5 ppb and normalized mean error just at or below 15% for all
days and seasons. The MDA8 and hourly time series also show excellent model performance,
with slightly underestimated peak day time O3 in January (Figure 3C-38, Figure 3C-39).
Overestimates of night-time O3 in April and October, although these overpredictions are less
pronounced in Dallas compared to many of the other urban study areas examined in the
assessment.
Table 3C-12. CAMx model performance at monitoring sites in the Central U.S.
Season
MDA8 level (ppb)
No. of obs
MB (ppb)
NMB (%)
ME (ppb)
NME (%)
Winter
Days < 60
4550
-4.0
-12.2
5.8
18.0
Days > 60
7
-5.7
-9.2
9.1
14.5
All Days
4557
-4.0
-12.2
5.8
18.0
Spring
Days < 60
7086
-1.7
-3.9
6.2
14,4
Days > 60
324
-7.0
-10.9
7.8
12.2
All Days
7410
-1.9
-4.3
6.2
14.3
Summer
Days < 60
8234
3.8
9.6
7.0
17.9
Days > 60
346
-2.7
-4.2
7.0
10.8
All Days
8580
3.5
8.7
7.0
17.4
Fail
Days < 60
7109
2.6
7.4
5.1
14.6
Days > 60
124
-1.8
-2.8
5.3
8.2
All Days
7233
2.5
7.1
5.1
14.4
- 40
-20
Bias Summary: [ min, 25th %, 50th %, 75th %, max ]
J-44, -13, -7.4, 1.6, 36]
Figure 3C-32. Normalized mean bias for MDA8 O3 in the Central U.S., winter 2016.
3C-49
-------
Bias Summary: [ min, 25th %, 50th %, 75th %, max ]
[-28, -7.6, -0.93, 6.6, 41 ]
Figure 3C-33. Normalized mean bias for MDA8 O3 in the Central U.S., spring 2016.
Bias Summary: [ min, 25th %, 50th %, 75th %, max ]
[-21, 6.9, 16, 24, 72]
Figure 3C-34. Normalized mean bias for MI)A8 O3 in the Central U.S., summer 2016.
3C-50
-------
Bias Summary: [ min, 25th %, 50th %, 75th %, max ]
[-9.1, 9.8, 19, 28, 75]
Figure 3C-35. Normalized mean bias for MDA8 O3 in the Central U.S., fall 2016.
Table 3C-13. CAMx model performance at monitoring sites in the Saint Louis study area.
Season
MDA8 level
(PPb)
No. of obs
MB (ppb)
NMB (%)
ME (ppb)
NME (%)
Winter
Days < 60
181
-5,9
-20,9
6,5
23.1
Days > 60
0
NA
NA
NA
NA
All Days
181
-5.9
-20.9
6.5
23.1
Spring
Days < 60
756
-3.5
-7,8
6.1
13,7
Days > 60
63
-7.2
-11,2
7.3
11.3
All Days
819
-3.7
-8.1
6.2
13,4
Summer
Days < 60
1061
5.8
13.7
8.4
19.6
Days > 60
121
-1.1
-1,6
8,1
12,1
All Days
1182
5.1
11.4
8.4
18.5
Fall
Days < 60
773
3,9
11.1
5,7
16,1
Days > 60
35
3,5
5.1
5,0
7,3
All Days
808
3,9
10.6
5.7
15,4
3C-51
-------
AQS MDA8 Comparison for SaintLouis Monitors in 2016
80 -
~ 60 -
» 40
20
0 J
AQS MDA8
CAMx HDDM 2016(e
# sites20
01/01 02/02 03/05 04/06 05/08 06/09 07/11 08/12 09/13 10/15 11/16 12/18
Date
Figure 3C-36. Time series of monitored (black) and modeled (red) MDA8 O3 at St. Louis
monitoring sites in 2016.
3C-52
-------
AOS Comparison for SaintLouis Monitors in January 2016
AQS Comparison for SaintLouis Monitors in April 2016
AQS Hourly
CAMx HDDM 2016le
01/01 01/03 01 >06 01/09 01/11 01/14 01/17 01/19 01/22 01/25 01/27 01/30
Date
AOS Comparison for SaintLouis Monitors in July 2016
70 -
60 -
50 -
40
30 -
20 -
10 -
0
AQS Hourly
CAMx HDDM 2016le
03/31 04/03 04/06 04/08 04/11 04/13 04/16 04/19 04/21 04/24 04/27 04/29
Date
AQS Comparison for SaintLouis Monitors in October 2016
AQS Hourly
CAMx HDDM 2016fe
I . II
06/30 07/03 07/06 07/08 07/11 07/14 07/17 07/19 07/22 07/25 07/27 07/30
Date
70 -
60 -
50 -
1
40
30 -
I
20 -
10 -
0 -
AQS Hourly
CAMx HDDM 2016fe
09/30 10/03 10/06 10.08 10/11 10/14 10/17 10/19 10/22 10/25 10/27 10/30
Date
Figure 3C-37. Time series of monitored (black) and modeled (red) hourly O3
concentrations at St. Louis monitoring sites in January (top left), April (top
right), July (bottom left), and October (bottom right) 2016.
Table 3C-14. CAMx model performance at monitoring sites in the Dallas study area.
Season
MDA8 level
(PPb)
No. of obs
MB (ppb)
NMB (%)
ME (ppb)
NME (%)
Winter
Days < 60
625
-3,2
-9.9
4.8
14.9
Days > 60
0
NA
NA
NA
NA
All Days
625
-3,2
-9.9
4.8
14.9
Spring
Days < 60
697
0,8
1.8
5.8
13.5
Days > 60
21
-4.9
-7,7
5.4
8.6
All Days
718
0,6
1.4
5,7
13.3
Summer
Days < 60
700
2,1
5,4
5.9
15.4
Days > 60
25
-2,8
-4.0
6.5
9.4
All Days
725
1,9
4,8
5.9
15.1
Fall
Days < 60
697
1.4
3.7
4.5
11.9
Days > 60
23
-3.6
-5.5
4.7
7,1
All Days
720
1,3
3.2
4.5
11.6
3C-53
-------
AQS MDA8 Comparison for Dallas Monitors in 2016
80 -i
60 -
40
20
0 J
AQS MDA8
CAMx HDDM 2016fe
# sites:24
l
01/01 02/02 03/05 04,06 05/08 06/09 07/11 08/12 09/13 10/15 11/16 12/18
Date
Figure 3C-38. Time series of monitored (black) and modeled (red) MDA8 O3 at Dallas
monitoring sites in 2016.
3C-54
-------
AQS Comparison for Dallas Monitors in January 2016
AQS Comparison for Dallas Monitors in April 2016
40
30
8
| 20
10
01/01 01/03 01 >06 01/09 01/11 01/14 01/17 01/19 01/22 01/25 01/27 01/30
Date
AQS Comparison for Dallas Monitors in July 2016
AQS Hourly
CAMxHDDM 2016fe
06*30 07/03 07/06 07/08 07/11 07/14 07/17 07/19 07/22 07/25 07/27 07/30
Date
AOS Hourly
CAMxHDDM 2016fe
03/31 04/03 04.06 0408 04/11 04/13 04/16 04/19 04/21 04/24 04/27 0429
Date
AQS Comparison for Dallas Monitors in October 2016
AQS Hourly
CAMxHDDM 2016te
09/30 10.03 10/06 10O8 10/11 10/14 10/17 10/19 10/22 10/25 10/27 10/30
Date
Figure 3C-39. Time series of monitored (black) and modeled (red) hourly O3
concentrations at Dallas monitoring sites in January (top left), April (top
right), July (bottom left), and October (bottom right) 2016.
3C.4.2.5 Operational Evaluation in the Western U.S.
Model statistics for MDA8 O3 in the Western U.S. are best on low O3 days in summer
and fall (Table 3C-15). High wintertime observations were substantially underestimated by the
model with an average MB of -26 but likely for different reasons. The high days in Riverside
California are probably due to traditionally understood O3 formation that occurs on warm sunny
days. The high O3 concentrations in Wyoming are an example of wintertime O3 formation that
occurs during cold pool meteorology events which have substantial snow cover and extreme
temperature inversions and are still an active area of research. Some spatial patterns in
normalized mean bias are apparent in the winter and in the summer (Figure 3C-40 through
Figure 3C-43), with overestimates on the West Coast and underestimates in the Intermountain
West. Two urban study areas are located in the Western U.S. and are evaluated in this section:
Sacramento and Phoenix.
The model performance for MDA8 O3 values in the Sacramento study area was best on
lower O3 days in summer and fall (Figure 3C-44). In Sacramento there were no days during the
AQS Hourly
CAMxHDDM 2016fe
3C-55
-------
winter with measured MDA8 O3 > 60 ppb. Normalized mean error is at or below 15% for all
seasons except winter. Hourly time series show good agreement in Sacramento, except for winter
when the model does not capture very much of the day to day variability in O3 concentrations
Figure 3C-45).
While normalized mean error was at or less than 15% in Phoenix on all days in all
seasons, the MDA8 time series shows frequent underestimates in winter-spring as well as
overestimates in summer-fall (Figure 3C-46). The hourly time series also show that though the
model captures some of the overnight O3 patterns in Phoenix, night time O3 is significantly
overestimated, particularly in January and October (Figure 3C-47).
Table 3C-15. CAMx model performance at monitoring sites in the Western U.S.
No. of obs
MB (ppb)
NMB (%)
ME (ppb)
NME (%)
Winter
Days < 60
15888
-2.8
-8.2
6.0
18.1
Days > 60
113
-25.8
-35.7
25.8
35.7
All Days
16001
-2.9
-8.7
6.2
18.4
Spring
Days < 60
15789
-4.6
-10.3
6.5
14.6
Days > 60
1471
-9.5
-14.7
10.0
15.4
All Days
17260
-5.0
-10.8
6.8
14.7
Summer
Days < 60
13254
1.2
2.6
6.7
14.9
Days > 60
4461
-6.6
-9.5
9.5
13.7
All Days
17715
-0.8
-1.6
7.4
14.5
Fall
Days < 60
15975
0.7
1.9
5.4
14.5
Days > 60
795
-9.2
-13.6
10.7
15.8
All Days
16770
0.2
0.6
5.6
14.6
3C-56
-------
Bias Summary: [ min, 25th %, 50th %, 75th %, max ]
[-47. -12, 8.3, 27, 110]
Figure 3C-40. Normalized mean bias for MDA8 O3 in the Western U.S., winter 2016.
c?
1
- 100
- 80
- 60
-40
- 20
- 0
- -20
- -40
- -60
- -80
1
- -100
Bias Summary: [ min, 25th %, 50th %, 75th %, max ]
[-60, -12, -5.7. 3.1, 82]
Figure 3C-41. Normalized mean bias for MDA8 O3 in the Western U.S., spring 2016.
3C-57
-------
n
-100
- 80
- 60
- 40
- 20
- 0
- -20
- -40
- -60
- -80
1
- -100
Bias Summary: [ min, 25th %, 50th %, 75th %, max ]
[-55, -5.2, 3.2, 16, 90]
Figure 3C-42. Normalized mean bias for MDA8 O3 in the Western U.S., summer 2016.
(?
<3?
¦
- 100
- 80
- 60
-40
- 20
- 0
- -20
- -40
- -60
- -80
- -100
Bias Summary: [ min, 25th %, 50th %, 75th %, max ]
[-57, 1.5, 16, 30, 120]
Figure 3C-43. Normalized mean bias for Ml) A 8 O3 in the Western U.S., fall 2016.
3C-58
-------
Table 3C-16. CAMx model performance at monitoring sites in the Sacramento study area.
Season
MDA8 level
(PPb)
No. of obs
MB (ppb)
NMB (%)
ME (ppb)
NME (%)
Winter
Days < 60
2359
-0.9
-3.2
5.5
18.9
Days > 60
0
NA
NA
NA
NA
All Days
2359
-0,9
-3.2
5.5
18.9
Spring
Days < 60
2474
-3,2
-7.9
5.6
13.6
Days > 60
116
-8.1
-12.6
9.4
14.6
All Days
2590
-3.5
-8,2
5,8
13,7
Summer
Days < 60
2157
0.6
1.3
5.8
13.7
Days > 60
628
-7.3
-10,8
00
CO
13.0
All Days
2785
-1.2
-2.5
6.5
13.5
Fail
Days < 60
2503
0.5
1,3
5.5
15.2
Days > 60
160
-7.7
-11.2
10.0
14.7
All Days
2663
0,0
0.0
5.7
15,1
AOS MDA8 Comparison for Sacramento Monitors in 2016
¦R
o.
CO
<
Ci
00
60
40
20 -
0 J
AQSMDA8
CAMx HDDM 201 fife
f sites:49
—! ti ¦ u 11;;! 11; 1111 Li 11, ; 111 -11! 'Jill! m 11 ¦ u i • ! 111:111 i'i 11;; i h < m i iin 111 uiiiii 11 n i (111 i i, 11;; 11.
01/01 02/02 03:05 04/06 05/00 06/09 07/11 08/12 09/13 10/15 11/16 12/16
Date
Figure 3C-44. Time series of monitored (black) and modeled (red) MDA8 O3 at
Sacramento monitoring sites in 2016.
3C-59
-------
AOS Comparison (or Sacramento Monitors in January 2016
AOS Comparison for Sacramento Monitors in April 2016
AQS Hourly
a
&
g 40 -
s
a
x
10 -
01/01 01 03 01/06 01/09 01/11 0114 0117 01.M9 01/22 01/25 0127 01,-30
AOS Comparison lor Sacramento Monitors in July 2016 AOS Comparison (or Sacramento Monitors in October 2016
B0
— AQS Hourly
CAM* HDDM 201 file
60
40
20
0
0630 07-03 07'06 07 08 07/11 07-14 07'16 07/19 07,22 07-25 07*7 0730
— AQS Hourly
CAM* HDDM 201618
20 -
09-30 10-03 10-06 10-08 10.11 10/14 10/16 1019 1022 10-25 10-27 1CKJ0
Figure 3C-45. Time series of monitored (black) and modeled (red) hourly O3
concentrations at Sacramento monitoring sites in January (top left), April
(top right), July (bottom left), and October (bottom right) 2016.
Table 3C-17. CAMx model performance at monitoring sites in the Phoenix study area.
Season
MDA8 level
(PPb)
No. of obs
MB (ppb)
NMB (%)
ME (ppb)
NME (%)
Winter
Days < 60
1292
-3,5
-9.8
5,3
15,0
Days > 60
3
-5.9
-9,7
5.9
9,7
Ail Days
1295
-3.5
-9,8
5.3
14,9
Spring
Days < 60
265
-5.6
-10.9
6.8
13,3
Days > 60
1082
-8,5
-13.3
9,6
14.9
All Days
1347
-6,2
-115
7,4
13.7
Summer
Days < 60
974
-2,1
-4.2
6,5
13,0
Days > 60
346
-4.7
-7,3
8.5
13,0
Ail Days
1320
-2.8
-5.2
7.1
13,0
Fall
Days < 60
1278
2.6
6.7
6.1
15,4
Days > 60
5
-3,8
-6.2
5.4
8.7
All Days
1283
2.6
6,6
6,1
15.4
3C-60
-------
AQS MDA8 Comparison for Phoenix Monitors in 2016
80 i
60 H
| 40
<
o
5
20 -
0 J
AQS MDA8
CAMx HDDM 2016fe
# sites:38
01/01 02/02 03/05 04.
-------
AOS Comparison for Phoenix Monitors in January 2016
AOS Comparison tor Phoenix Monitors in April 2016
50
40 -
s
8
30 -
3
20 -
10 -
0 -
— AOS Hourly
CAM* HDDM 201616
01/01 01.-03 01/06 01/09 01/11 01/14 01/17 01/19 01/22 0V25 Ot.27 01.30
Date
AOS Comparison for Phoenix Monitors in July 2016
— AOS Hourly
CAMx HDDI
0331 04.03 044)5 04-08 04/11 04/13 04/16 04/19 04/21 04,24 0426 0429
Date
AOS Comparison for Phoenix Monitors in October 2016
AOS Hourly
CAMx HDDM 2016IB
m
06^0 07/03 07/06 07*08 07/11 07/14 07/16 07/19 07*22 07/25 07/27 07/30
Date
8
8 40
AOS Hourly
CAMx HDDM 2016le
in
I
0930 10/03 10«8 10^08 10/11 10/14 10'16 10/19 1022 1025 10/27 10-30
Dale
Figure 3C-47. Time series of monitored (black) and modeled (red) hourly O3
concentrations at Phoenix monitoring sites in January (top left), April (top
right), July (bottom left), and October (bottom right) 2016.
3C.5 AIR QUALITY ADJUSTMENT TO MEET CURRENT AND
ALTERNATIVE AIR QUALITY SCENARIOS
3C.5.1 Overview of the Higher Order Direct Decoupled Method (HDDM)
In this section we present a model-based O3 adjustment methodology that allows for
adjustments to observed hourly O3 concentrations to reflect the expected impacts of changes in
NOx emissions. This methodology uses the CAMx model, described above in section 3C.4,
instrumented with the Higher order Decoupled Direct Method (HDDM) - a tool that generates
modeled sensitivities of O3 to emissions changes. The outputs of the HDDM are used to estimate
the distribution of O3 concentrations associated with just meeting three air quality scenarios (O3
monitor design values of 75 ppb, 70 ppb, and 65 ppb) within multiple urban study areas. The
HDDM sensitivities are applied to ambient air measurements of O3 to estimate how O3
concentrations would respond to changes in U.S. anthropogenic emissions. This approach, based
on Simon et al. (2013), was applied previously for the 2015 O3 NAAQS review.
3C-62
-------
The CAMx photochemical modeling incorporates emissions from non-anthropogenic
sources and anthropogenic emissions from sources in the U.S and in the portions of Canada and
Mexico within the regional modeling domain. Pollution from sources in other locations within
and outside of North America is included as transport into the boundary of the modeling domain.
3C.5.1.1 Capabilities
Chemical transport models, such as CAMx, simulates physical and chemical processes in
the atmosphere to predict 3-dimensional (3-D) gridded pollutant concentrations. These models
account for the impacts of emissions, transport, chemistry, and deposition on spatially and
temporally varying pollutant concentrations. Required model inputs include time-varying
emissions and meteorology fields, time varying concentrations of pollutants at the boundaries of
the model domain (i.e. boundary conditions), and a characterization of the 3-D field of chemical
concentrations to initialize the model (i.e. initial conditions).
Beyond modeling the ambient air concentrations of O3, chemical transport models can be
used to estimate the response of ambient air O3 concentrations to changes in emissions. One
technique to simulate the response of O3 to emissions changes, the brute force method, requires
the modeler to explicitly model this response by directly altering the emissions inputs in the
model simulation. This technique provides an estimate of the O3 concentration at the altered
emission level, but often does not provide accurate information regarding the response of O3 to
other levels of emissions since the chemistry for O3 formation is nonlinear. Therefore, when
using only the brute force method, a new model simulation would need to be performed for
every emissions scenario under consideration.
Other analytical techniques have been developed to estimate the O3 response to emission
perturbations without performing multiple simulations. One such method is termed the
Decoupled Direct Method (DDM) (Dunker, 1984). DDM, solves for sensitivity coefficients
which are defined as the partial derivative of the atmospheric diffusion equations that underly the
model calculations, Equations (3C-1) and (3C-2).
Equation (3C-1)
p aCttr? _ ac'tCf)
^ i£j
Equation (3C-2)
3C-63
-------
Here, Sij(t), the sensitivity, gives the change in model concentration, Ci, (for instance O3
concentration) with an incremental change in any input parameter, pj (in this case emissions).
Equation (3C-2) allows us to normalize the sensitivity coefficient, Sij(t), so that it shows response
in relative terms for the input rather than in absolute units. Therefore, Pj (x,t) is the normalized
input and Sj is a scaling variable (Yang et al., 1997). In general terms, the sensitivity coefficient
tells us how a model output (O3 concentration) will change if a model input (emissions of NOx
or VOC) is perturbed. This first order sensitivity coefficient, Sij(t) is quite suitable for small
perturbations, but gives a linear response which is unlikely to represent the results of large
perturbations in very nonlinear chemical environments. Second (and third) order derivatives can
be calculated to give higher order sensitivity coefficients (Hakami et al., 2003). Higher order
sensitivity coefficients give the curvature and inflection points for the response curve and can
capture the nonlinearities in the response of O3 to emissions changes. Using Higher order DDM
(HDDM) allows for the sensitivities to be more appropriately applied over larger emissions
perturbations. Hakami et al. (2003) report that for an application in California, HDDM gave
reasonable approximations of O3 changes compared to that generated using brute force emissions
reductions of up to 50% using the first three terms of the Taylor series expansion, Equation (3C-
3).
C(+Ae) = CCO) + AeS(0> + S2(0) + ... + -T5Bt0) + Ru+1
2 ti! WT A
Equation (3C-3)
Here Ae represents the relative change in emissions (for instance Ae = -0.2 would be equivalent
to reducing emissions by 20%), Sn(0) is the nth order sensitivity coefficient, C(0) is the
concentration under baseline conditions (no perturbation in emissions) and Rn+i is a remainder
term.
A variant of DDM called DDM-3D has been implemented into several chemical transport
models, including CAMx, for both O3 and particulate matter (PM) predictions (Cohan et al.,
2005, Hakami et al., 2003, Napelenok et al., 2011, Dunker, 1984,Yang et al., 1997, Koo et al.,
2007, Zhang et al., 2012). These implementations allow the modeler to define the parameters for
which first and higher order sensitivities will be calculated. For instance, the sensitivity can be
calculated for emissions from a specific source type, for emissions in a specific geographic
region, and for emissions of a single O3 precursor or for multiple O3 precursors. In addition,
sensitivities can be calculated to boundary conditions, initial conditions, and various other model
inputs. Sensitivities to different sets of parameters can be calculated in a single model simulation
3C-64
-------
but computation time increases as the number of sensitivities increases. Outputs from an HDDM
simulation consist of time varying 3-D fields of first and second order sensitivities.
3C.5.1.2 Limitations
For the purposes of the O3 NAAQS analysis, an HDDM-based approach is well-suited
given its ability to 1) capture the non-linearity of O3 response to emissions changes, 2)
characterize different O3 responses at different locations (downtown urban versus downwind
suburban) and at different times of day, allowing us to incorporate temporal and spatial
variations in response into the O3 adjustment methodology, and 3) explicitly account for physical
and chemical processes influencing predicted sensitivities such as background O3 sources.
However, in addition to the many potential benefits of using HDDM to understand and
characterize O3 response to emissions changes, there are several limitations.
First, HDDM encompasses all of the uncertainties of the base photochemical model
formulation and inputs. So, uncertainties in how the physical and chemical processes are treated
in the model and in the model inputs propagate to the HDDM results. Also, HDDM can capture
response to larger emissions perturbations than DDM but it is still most accurate for small
perturbations. The larger the relative change in emissions, the less likely that the HDDM
sensitivities will properly capture the change in O3 that would be predicted by using brute force
emission reductions. Several studies have reported reasonable performance of HDDM for O3 up
to 50% emissions perturbations (Hakami et al., 2004, Hakami et al., 2003, Cohan et al., 2005),
but the magnitude of perturbation over which HDDM will give accurate estimates will depend on
the specific modeling episode, size of the model domain, emissions and meteorological inputs,
and the size of the emissions source to which the sensitivity is being calculated. In this work, we
applied sensitivities derived from model simulations done under varying NOx levels (see section
3C.5.2.2) and found that using this technique we were able to replicate O3 concentrations
estimated using brute force emission reductions with HDDM sensitivities for up to 90% NOx
emissions reductions with a mean bias of less than 3 ppb and a mean error of less than 4 ppb.
3C.5.2 Using CAMx/HDDM to Adjust Monitored Ozone Concentrations
3C.5.2.1 Conceptual Framework
This section outlines the methodology in which we apply CAMx/HDDM to estimate
hourly O3 concentrations that might result from just meeting three air quality scenarios (75 ppb,
70 ppb, and 65 ppb). These methods closely follow those documented in Simon et al. (2013) and
the risk and exposure assessment performed in the 2015 O3 NAAQS review (U.S. EPA, 2014).
As part of the methodology, photochemical modeling results are not used in an absolute sense,
3C-65
-------
but instead are applied to modulate ambient air measurements, thus tying estimated O3
distributions to measured values. The basic steps are outlined below and in Figure 3C-48.
Step 1: Run CAMx simulation with HDDM to determine hourly O3 sensitivities to NOx
emissions changes for the grid cells containing monitoring sites in an urban study area.
Step 2: For each monitoring site, season, and hour of the day use linear regression to relate first
order sensitivities of NOx (5var)to modeled O3 and second order sensitivities of NOx (S'nox) to
the first order sensitivities.
Step 3: For each measured hourly O3 value, calculate the first and second order sensitivities
based on monitoring site-, season-, and hour-specific functions calculated in Step 2.
Step 4: Adjust measured hourly 2015-2017 O3 concentrations for incrementally increasing levels
of emissions reductions using assigned sensitivities, then recalculate 2015-2017 design values
until all monitors in the urban study area just meet the levels of the air quality scenario.
3C-66
-------
Anthropogenic
Natural
r
U.S.
Canada and
Mexico
03 and 03 Precursor Emissions
f
7)
Meteorology
¦-
•••
Initial and Boundary conditions
Recent Monitored 03
(2015-2017)
Other Model Inputs
Step 1a
CAMx
HDDM Modeling
(Jan-Dec 2016)
/ Gridded hourly 03
-> Concentrations and
\ Sensitivities
Step 3.
Use Regressions and
Observed Ozone to
Predict Sensitivities
Unique Linear
Relationships Between
Sensitivities and Hourly
Ozone at Each Monitor
Location for Each
Season and
Hour-of-the-Day
1
Step 2:
Create Regressions
. /
I
\
/ Hourly Ozone \
' Observations Paired with j_
i, Sensitivities for 2015-2017 J
V. At All Monitor Locations J
elect Emissions'
V
-#c
Reductions to
which Sensitivities /
\ Will Be Applied /'
\ /
X s
v
Step 4:
Adjust Hourly Ozone
>1
to Meet Air Quality
*\
Scenarios
Step 1b:
Extract Output
Ozone Concentrations
& Sensitivities at
Locations of
Monitoring Sites for
Each Modeled Hour
'Adjusted Hourly Ozone"
Values for 2015-2017 at
Each Monitor Location to
Show Attainment with
V Air Quality Scenarios
Figure 3C-4S. Flow diagram demonstrating HDDM model-based O3 adjustment approach.
3C-67
-------
3C.5.2.2 Application to Measured O3 Concentrations in Urban Study Areas
The model-based adjustment approach described above was applied to the eight urban
study areas (Atlanta, Boston, Dallas, Detroit, Philadelphia, Phoenix, Sacramento, and St. Louis)
for an air quality scenario adjusted to just meet the current standard of 70 ppb and two alternative
air quality scenarios having design values of 75 ppb and 65 ppb. The analysis used CAMx
photochemical modeling for January-December of 2016 and ambient air data for the years 2015-
2017. When running CAMx with HDDM, additional information is required to designate model
inputs for calculating sensitivities. In this analysis, HDDM was set up to calculate the sensitivity
of O3 concentrations to U.S. anthropogenic NOx emissions.13
U.S. anthropogenic emissions were defined as all emissions in the following sectors:
commercial marine, rail, residential wood combustion, agricultural fires, onroad mobile, offroad
mobile, EGU point sources, oil and natural gas point, non-EGU point, non-point oil and gas, and
non-point area. These anthropogenic sectors account for 10.5 million of the total CONUS-wide
11.8 million tons per year of NOx emissions in 2016 (the remaining 1.3 million tons are from
biogenics and wildland fires, which included prescribed burns). Sensitivities were not calculated
for biogenic, wildland fire, Canadian, or Mexican emissions. In addition, sensitivities were not
calculated for any emissions originating from outside the domain (i.e., entering through the use
of boundary concentrations).
3C.5.2.2.1 Multi-step Application of HDDM Sensitivities
As discussed in section 3C.5.1.2 of this appendix, HDDM has been reported to
reasonably replicate brute force emissions reductions up to a 50% change in emissions. For this
analysis, it was desirable to have confidence that the HDDM sensitivities could replicate the
entire range of emissions reductions. Evaluations of the HDDM estimated O3 concentrations
compared to that estimated from brute force emissions reduction model runs confirm that the
HDDM estimates of O3 response to NOx reductions are fairly comparable for a 50% change.
However, O3 concentrations estimated from the HDDM sensitivities and the brute force method
begin to diverge in comparisons under larger emissions changes (90%). Consequently, two
additional CAMx/HDDM runs were performed under different levels of NOx emissions
reductions in order to characterize O3 sensitivities to NOx reductions over a larger range of
emissions perturbations. One CAMx/HDDM simulation was performed with U.S. anthropogenic
13 Sensitivities were only assessed using U.S. emissions in the contiguous 48 states. We did not assess responses to
VOC emission reductions in this analysis as a means to reduce computational costs because none of the urban
study areas considered here required VOC emission reductions to achieve the lower design values in the air
quality scenarios simulated in the 2014 HREA.
3C-68
-------
NOx emissions reduced by 50%. A second additional simulation was performed with a 90%
NOx reduction. Emissions of other species were not modified from the base case in these two
additional simulations. These additional HDDM simulations provide O3 sensitivities to NOx
under chemical regimes with lower NOx emissions. The sensitivities are used in a multistep
adjustment approach, as described in the following sections.
Figure 3C-49 provides a conceptual picture of the multistep adjustment procedure using
first-order sensitivities. Sensitivities from the base run are used to adjust O3 concentrations for
NOx emissions reductions up to X%. Additional emission reductions beyond X% use
sensitivities from the 50% NOx cut run until reductions exceed (X+Y)%. Finally, sensitivities
from the 90% NOx emissions reduction run are applied for any emission reductions beyond
(X+Y)%. In order to more closely approximate the non-linear O3 response to any level of
emissions reductions, 2nd order terms are added to the multistep approximation method in
Equations (3C-4) through (3C-7). P represents the percentage NOx cut for which the AO3 values
are being calculated, Sand S2 are the first and second order O3 sensitivities to U.S. NOx
emissions, and X and Y are described above.
AOa — ax SW£Mfbajt
a x SW£Mfbajt + 2 x b x Sj¥OiK50%rat + x •5|r0X5Bfkrat
c* ,
— C X + ~X Jj5oa^a% X
Equation (3C-5)
2x (P-X)
2 XY
100
for P < X
for X < P < (X + Y)
for P > (X + Y)
Equation (3C-6)
0
c = ujxfp-Cx-f-y'))
10c
far P < (X + F)
for 1Q0>P>(X + Y)
Equation (3C-7)
3C-69
-------
SB NOx fir st-order approximation from base DDM simulation
HI NOx first-order approximation from 50% NOx cut DDM simulation
i—t NOx s first-order approximation from 90% NOx cut DDM simulation
Theoretical ozone response curve to NOx emission reductions
£ CMAQ DDM simulations
Estimated
concentration
after P%
reduction in
baseline NOx
emissions
Starting concentration
T
100 P 90 X+Y 50 x
% reduction in baseline NOx emissions
Figure 3C-49. Conceptual picture of 3-step application of HDDM sensitivities.
The ideal value for equation transition points, X and Y, are determined by minimizing the
least square mean error between the adjusted concentrations using the multistep approach and
modeled concentrations from brute force NOx emissions reduction runs. We first determined the
value of X which gave the lowest error compared to brute forces estimates at 50% NOx
emissions reductions. Then holding X constant, we determined the value of Y which gave the
lowest error compared to brute force method O3 concentration estimates using 90% NOx
emissions reductions. This process was performed independently for each of the eight urban
study areas in this analysis.
Error in HDDM estimates of hourly O3 is defined here as the difference between HDDM
estimated O3 and O3 estimated using the brute force method. Based on equations (3C-4) through
(3C-7), this can be calculated from Equations (3C-8) and (3C-9) for 50% NOx emissions
reductions:
3C-70
-------
£ — £tOsonBgDS^S0 AOzonegp^
Equation (3C-8)
-JT , X2 2(50 -JQwn _ (2X^0-fl)'i(<>!
S ~ 100 ^ ^KOsJiass 2 x 100' * ^WOx_®e'se Jqq ^ ^W VJ-te^:fu. \ _
® ~ \2 X1001 2 x 10C" \ 100 100 2 "> 100" '
¦*" ^-^.V®r„%ent+ ~~ J®5 &0®wt»*iy»)
Equation (3C-10)
j _ (Ssm..ma: . tessatKaau \
' - IxlOS1 Sxlt®1 J
Equation (3C-11)
~~ \ 190 193 2X1W1 )
Equation (3C-12)
c - (~~%a%0%eut+ — & OzoneSF)W j
Equation (3C-13)
Next, the error is squared, summed over all points (error can be calculated for each
hourly O3 value at each monitoring location), and the derivative is set to 0 to determine X which
gives the least squares error (Equations (3C-14), (3C-15), and (3C-16)).
«* = A*X* + 2ABX9 +(2AC+ B *)X* + 2BCX + C1
Equation (3C-14)
L«2 = (£ j42)Jf4 + CI 2AB')X"d + C2 2AC + B2)X1 + (S2BC)X + IC2
Equation (3C-15)
3C-71
-------
(£«*)' = (4EA2)Jf3 + (3£ 2AB)X* + C2Z2AC + B2')X + (£2.BC) = 0
Equation (3C-16)
The value of X that gives the least squares error will occur at one of the three roots of the
trinomial in Equation (3C-16) or at 0 or 50. All real roots, 0, and 50 were input into equation
(3C-15) and X was set to the value which resulted in the lowest error in each city. An analogous
procedure was followed to determine Y using the 90% NOx emissions reduction brute force
simulation and Equations (3C-17) through (3C-23).
—A' ( X1 2r,„„
^ 100 ^ 2 X 100* ^ 100* ^
z'-r1 10(90 — (x + F))
+ 2 X 100* X Smx=^'k:M 100 x
11D:< 5S-'A'+Y)}}! 2
; l; ;: ~ -x ¦5jvtaW)%cur ~~ &O'zmiBBTi90
Equation (3C-17)
£l _ A2y* + 2ABY3+ (2AC + + 2BCY + C2
Equation (3C-18)
' ~ { 2.xiK! : ic:1 J
Equation (3C-19)
0 _ , iBxya-Qsgsjcut _ Wx '
' _ \ 100 100 2X1.0©1 }
Equation (3C-20)
(-X „ _ A'2 ^ 10 X (90 — X') ^ _ lOOxCW-Jf)*^ *'\
^ ~ I ^qq ^ATGxJtase 2 x 100J ^NOx^icut 4- 2 X 100s ^NOx^rbcitt ~ &OzsmeBp^ 1
Equation (3C-21)
£ g2 = (£ j4z)y4 + (£ 24S)rs + (£ 2ac + b *)y2 + (£, 2 bcjy + £ c2
Equation (3C-22)
3C-72
-------
CI>23' = C4£48)F8 + (3E2j4S)ra+ (2j24C + .B2)F + CS.2SC) = 0
Equation (3C-23)
The X and Y outpoints which have the least square error in each urban study area are
shown in Table 3C-18. This 3-step adjustment methodology was shown to be a robust method
for minimizing error in the HDDM applications for larger percentage changes in emissions by
Simon et al. (2013). Figure 3C-50 through Figure 3C-65 are density scatter plots that compare
hourly O3 estimates from brute force with hourly O3 estimates from the 3-step HDDM
adjustments at all monitor locations in each of the eight urban study areas evaluated in this study.
The colors in these plots depict the percentage of points falling at any one location. Mean error
for the 50% and 90% 3-step HDDM adjustment NOx emissions reductions cases compared to O3
concentrations estimated using the brute force method are less than 0.5 ppb and 2 ppb
respectively in all eight urban study areas.
Table 3C-18. X and Y cut-points used in Equations (3C-4) through (3C-7).
Urban Study Area
X
Y
Atlanta
37
48
Boston
38
45
Dallas
37
47
Detroit
37
45
Philadelphia
37
45
Phoenix
37
45
Sacramento
38
48
St. Louis
37
47
3C-73
-------
Atlanta Comparison : NOx cuts
o
C\J
o
CVJ
£
E s
o.
CD
RMSE = 0 3 ppb
20 40 60 80 100 120
03 at 50% NOx cut - brute force
Figure 3C-50. Comparison of brute force and 3-step HDDIY1 O3 estimates for 50% NOx
cut conditions in Atlanta.
3C-74
-------
E
"O
"O
3
O
X
O
Z
-o
O
to
"cc
CO
O
Boston Comparison : NOx cuts
Y = 0,21 + 0,99 * X
RMSE = 0,3 ppb
10
e
I- 6
K 4
0
20 40 60 80 100 120 140
03 at 50% NOx cut - brute force
Figure 3C-51. Comparison of brute force and 3-step HDDM O3 estimates for 50% NOx
cut conditions in Boston.
3C-75
-------
3 g
f— o
O °-|
"cc
CO
O 2
Dallas Comparison : NOx outs
RMSE = 0,2 ppb
20 40 60 80 100 120
03 at 50% NOx cut - brute force
Figure 3C-52. Comparison of brute force and 3-step HDDM O3 estimates for 50% NOx
cut conditions in Dallas.
3C-76
-------
Detroit Comparison : NOx cuts
o
o
CVJ
3 g
f— o
3
o
o
Z
CO
& °
"cc
CO
O
Y = 0,15 + 1 *X
/
¦
/
/
RMSE = 0,3 ppb
20 40 60 80 100 120
03 at 50% NOx cut - brute force
140
ID
- 6
Figure 3C-53. Comparison of brute force and 3-step HDDM O3 estimates for 50% NOx
cut conditions in Detroit.
3C-77
-------
Philadelphia Comparison : NOx outs
E
"O
"O
3
O
X
o
-o
o
uo
as
CO
O
a
T -\
O
CvJ -I
o
o H
o
CO
o.
CD
O
DJ
Y = 0,32 + 0,99 * X
/
/
RMSE = 0,5 ppb
10
- B
0 ~20 40 60 80 100 120 140
03 at 50% NOx cut - brute force
Figure 3C-54. Comparison of brute force and 3-step HDDM O3 estimates for 50% NOx
cut conditions in Philadelphia.
3C-78
-------
Phoenix Comparison : NOx cuts
o
5?
o
c\j -4
3 g
r— O
3
O
O
Z
CO
O ° -|
to
"cC
CO
o
CD
o
C\J
Y = 0.23 + 1 *X
RMSE = 0 3 ppb
10
- 6
0 20 40 60 80 100 120 140
03 at 50% NOx cut - brute force
Figure 3C-55. Comparison of brute force and 3-step HDDM O3 estimates for 50% NOx
cut conditions in Phoenix.
3C-79
-------
Sacramento Comparison : NOx cuts
E
"O
"O
3
O
X
O
Z
-c-
O
uo
"cC
CO
O
o
t
o
C\J -J
o
o
o
CO
o.
CD
o
C\j
Y = 0,18 + 1 *X
RMSE = 0,2 ppb
10
6
h4
2
0
0 20 40 60 80 100 120 140
03 at 50% NOx cut - brute force
Figure 3C-56. Comparison of brute force and 3-step HDDM O3 estimates for 50% NOx
cut conditions in Sacramento.
3C-80
-------
StLouis Comparison : NOx cuts
T3
tj
3
o
o
-o
o
uo
"cC
CO
O
o
o
CVJ
o
o
o
CO
o
CD
Y = 0,19 + 0.99 * X
RMSE = 0,2 ppb
10
- 6
L 0
0 20 40 60 80 100 120 140
03 at 50% NOx cut - brute force
Figure 3C-57. Comparison of brute force and 3-step HDDM O3 estimates for 50% NOx
cut conditions in St. Louis.
3C-81
-------
RMSE= 1.09 ppb
Atlanta Comparison : NOx cuts
o
CVI
£
E s
o.
CD
G
C\J
Y = -0.32 + 0.99 * X
20 40 60 80 100 120 140
03 at 90% NOx cut - brute force
h 4
h 2
J- 0
Figure 3C-58. Comparison of brute force and 3-step HDDM O3 estimates for 90% NOx
cut conditions in Atlanta.
3C-82
-------
o
¦•t
o
CvJ -I
?g
3
o
O
Z
CO
o £-1
CD
CD
CC
CO
O 2
G
C\J
0
Boston Comparison : NOx outs
Y = 0,22 + 0,99 'X
y
J
jf
/
/
s
/
/
/
/
/
RMSE = 0.941 ppb
20 40 60 80 100 120 140
03 at 90% NOx cut - brute force
14
12
- 10
- e
L o
Figure 3C-59. Comparison of brute force and 3-step HDDM O3 estimates for 90% NOx
cut conditions in Boston.
3C-83
-------
Dallas Comparison : NOx outs
o
¦ct
o
CVJ —
£
E s
3
O
X
O
Z
O
Gh
03
CO
o
o
CO
Y = -0,25 + 0,99 * X
RMSE = 0,969 ppb
40 60 80
03 at 90% NOx cut
100 120
brute force
140
10
- e
- 6
L o
Figure 3C-60. Comparison of brute force and 3-step HDDM O3 estimates for 90% NOx
cut conditions in Dallas.
3C-84
-------
Detroit Comparison : NOx cuts
E
"O
"O
3
O
X
O
Z
O
o>
"cC
CO
O
o
¦•t
o
CVJ
o
o
o
CO
o.
CD
o
C\J
Y = 0.66 +0,97 #X
/
/
/
¦/'
/
/
RMSE = 1.07 ppb
12
10
- 6
L 0
0 20 40 60 80 100 120 140
03 at 90% NOx cut - brute force
Figure 3C-61. Comparison of brute force and 3-step HDDM O3 estimates for 90% NOx
cut conditions in Detroit.
3C-85
-------
o
CVJ -A
E
E °
o -
o
C\J
Y = 0.74 +0.95 *X
20 40 60 80 100 120 140
03 at 90% NOx cut - brute force
RMSE = 1.87 ppb
- 6
J- 0
Figure 3C-62. Comparison of brute force and 3-step HDDM O3 estimates for 90% NOx
cut conditions in Philadelphia.
3C-86
-------
RMSE = 1.05 ppb
Phoenix Comparison : NOx cuts
o
CiJ
£
E s
3
O
X
O
Z
o
CD
CO
CO
O
o _
CO
o
C\J
Y= 1.1 + 0,96 ' X
20 40 60 80 100 120 140
03 at 90% NOx cut - brute force
- 6
- 4
b 2
L 0
Figure 3C-63. Comparison of brute force and 3-step HDDM Os estimates for 90% NOx
cut conditions in Phoenix.
3C-87
-------
Sacramento Comparison : NOx cuts
o
¦f
o
CVI -J
3 g
f— o
3
o
o
Z
CO
o ° -|
m
cc
CO
O
CD
G
C\J
0
Y = 0,71 + 0.97 # X
/
RMSE - 0,706 ppb
20 40 60 80 100 120
03 at 90% NOx cut - brute force
140
- e
L o
Figure 3C-64. Comparison of brute force and 3-step HDDM O3 estimates for 90% NOx
cut conditions in Sacramento.
3C-88
-------
StLouis Comparison : NOx cuts
Y = 0,23 + 0,98 * X
/
z
RMSE = 0,918 ppb
20 40 60 80 100 120
03 at 90% NOx cut - brute force
140
ID
L 0
Figure 3C-65. Comparison of brute force and 3-step HDDM O3 estimates for 90% NOx
cut conditions in St. Louis.
3C.5.2.2.2 Relationships between HDDM Sensitivities and Modeled O3
Concentrations
First and second order hourly O3 sensitivities to NOx emissions reductions were extracted
from the HDDM simulation for model grid cells that contained the O3 monitors in the eight
urban study areas. Extracted data included modeled sensitivities at monitor locations for all
modeled hours in 2016. These sensitivities cannot be applied directly to observed values for two
reasons: 1) high modeled O3 days/hours do not always occur concurrently with high observed O3
days/hours and 2) the modeling time period includes only 2016 but the time period we are
3C-89
-------
analyzing in this assessment includes three full years of ambient air data, 2015-2017. As to the
first point, photochemical models are generally used in a relative sense for purposes of projecting
design values. In this manner, model predictions are "anchored" to ambient air measurements. In
general, the average response on high modeled days is used for this purpose. This allows for
more confidence in calculated results when "less than ideal model performance [occurs] on
individual days" (U.S. EPA, 2007). Similarly, for this analysis we believe it is appropriate to
account for the fact the model does not always perfectly agree with measurements and that
sensitivities from a low O3 modeled day would not be appropriate to apply to a high O3 measured
day (and vice-versa) even if they occur on the same calendar day. For this reason, a method was
developed to generalize the modeled site-, season-, and hour-specific sensitivities so that they
could be applied to ambient air data during 2015-2017.14
Simon et al. (2013) describe how first order sensitivities are generally well correlated
with hourly modeled O3 concentrations and second order sensitivities are well correlated to first
order sensitivities. Based on their analysis, we create a separate linear regression for Snox as a
function of hourly O3 (i.e. Snox = m> 100 ppb) while the brute force
emissions simulations for the 90% reduction show much lower O3 (< 40 ppb). In these isolated
cases, base modeled O3 is low due to NOx titration and increases occur with reductions of NOx.
The HDDM sensitivities for these few points appear to be too high to be applied over large
(>50%) emissions changes because of strongly nonlinear chemistry. However these extreme
cases are not relevant for this analysis, since the largest emissions reduction required for
Philadelphia was 53% to meet the air quality scenario for 65 ppb (Table 3C-19). The two urban
study areas requiring emission cuts larger than —50%, Phoenix and Sacramento, both show much
better agreement between the 90% brute force and HDDM predictions (Figure 3C-63 and Figure
3C-64 respectively).
14 The 12 months modeled covered a variety of conditions such that we can use the results from this modeled time
period in conjunction with the ambient data from the longer 3-year period for estimating responses and applying
adjustments
15 Seasons are defined as follows: Winter = December, January, February; Spring = March, April, May; Summer =
June, July, August; Fall = September, October, November.
3C-90
-------
For the 50% and 90% emissions reduction CAMx/HDDM simulation, regressions were
performed for first order NOx sensitivities with modeled O3 from the base HDDM simulation.
The regression technique was performed for the first and second order NOx sensitivities from the
base run and the 50% emissions reduction and 90% emissions reduction simulations. The
sensitivities from the emissions reduction runs were fitted to hourly O3 concentrations in the base
simulation. Simon et al. (2013) found that correlation coefficients using for sensitivities from
NOx reduction simulations to base case O3 concentrations were similar to those with O3
concentrations from the NOx reduction runs.
3C.5.2.2.3 Application of Sensitivity Regressions to Ambient Air Data
To apply the HDDM adjustments to observed data, sensitivities must be determined for
each hour from 2015-2017 at each site based on the linear relationship from the modeled data
and the observed O3 concentration. The linear regression model also allows us to quantify the
standard error of each predicted sensitivity value at each hour and site.
Observed hourly O3 from 2015-2017 at each monitor location was adjusted by applying
incrementally increasing emissions reductions using equations (3C-4) through (3C-8) and
recalculating MDA8 values for incrementally increasing emissions reductions until an emissions
level is reach for which all monitors in an urban study area achieved design values at the level of
the air quality scenario being evaluated (design values of 75, 70, or 65 ppb). Therefore, all
monitors within an urban study area were treated as responding to the same percentage reduction
in NOx emissions.
The precursor reductions used to estimate spatial and temporal patterns of O3
concentrations for the three air quality scenarios were NOx-only reductions. We focused on
NOx-only reductions in light of several key findings from analyses for the 2014 HREA that
explored the use of both NOx and VOC reductions versus NOx-only scenarios (2014 HREA,
Appendix 4D). There were several key findings from that comparison. First, in most of the urban
study areas, the NOx /VOC scenario did not affect O3 response at the monitor having the highest
design value in such a way to reduce the total required emissions cuts. Further, evidence in the
literature has shown that locations in the U.S. have gotten more NOx-limited since 2007 (the
year modeled in the 2014 HREA) (Jin et al., 2017, Laughner and Cohen, 2019) and thus VOC
reductions would be expected to have less impact on resulting O3 concentrations in our scenarios
for the 2016 modeling used here than they had in the previous analysis. Finally, the two areas
(Denver and Chicago) in which VOC emissions had the most impact in the 2014 HREA were not
included in the current analysis. For these reasons, NOx-only reductions were determined to be
3C-91
-------
the most appropriate scenarios for this analysis. The final emissions reductions that were applied
in each urban study area are given in Table 3C-19 below.16
Table 3C-19. Percent emissions changes used for each urban study area to just meet each
of the air quality scenarios evaluated.
Urban Study
Area
75 ppb
70 ppb
65 ppb
Atlanta
0%
25%
44%
Boston
+7%
14%
40%
Dallas
15%
32%
45%
Detroit
+18%
21%
47%
Philadelphia
23%
43%
53%
Phoenix
14%
49%
68%
Sacramento
45%
58%
72%
Saint Louis
+11%
13%
38%
The 2014 HREA included a thorough analysis of the standard error associated with the
predicted O3 concentrations produced using the HDDM adjustment approach. This analysis
found that while the error in predicted values varied by site and air quality scenario being
evaluated, the magnitudes were small (<1.5 ppb in most cases). We did not repeat such an
analysis here given the small magnitude of the standard errors found in this previous assessment.
3C.6 INTERPOLATION OF ADJUSTED AIR QUALITY USING
VORONOI NEIGHBOR AVERAGING
The APEX exposure model uses spatial fields of ambient air quality concentrations at
variable spatial scales (e.g., 500 m regular grid, census tract centroid) as inputs, but requires that
there be no missing values. The final air quality data used as inputs to the APEX model were the
hourly O3 concentrations at monitoring sites adjusted using CAMx/HDDM, then interpolated to
each census tract centroid in the eight urban study areas using the Voronoi Neighbor Averaging
(VNA; Gold et al., 1997; Chen et al., 2004) technique described below. A cross-validation
analysis supporting the use of the VNA technique for the creation of hourly O3 spatial fields was
conducted in the previous review (U.S. EPA, 2014; Appendix 4A).
16 Note that these emissions reductions and broad nationwide emission cuts are not intended to represent
recommended control scenarios since they would not be the most efficient method for achieving a particular
standard in many areas.
3C-92
-------
The following paragraphs provide a numerical example of VNA used to estimate an O3
concentration value for census tract "E" in Figure 3C-66 below.
The first step in the VNA technique is to identify the set of nearest monitors for each
census tract. The left-hand panel of Figure 3C-66 presents a numerical example with nine census
tracts (squares) and seven monitoring sites (stars), with the focus on identifying the set of nearest
neighboring sites to census tract "E" in the center of the panel. The Delaunay triangulation
algorithm identifies the set of nearest neighboring monitors by drawing a set of polygons called
the "Voronoi diagram" around the census tract "E" centroid and each of the monitoring sites.
Voronoi diagrams have the special property that each edge of each of the polygons are the same
distance from the two closest points, as shown in the right-hand panel of Figure 3C-66.
A
B
Monitor:
90 ppb +
15 miles
C
*
D
Monitor: *
80 ppb
10 miles
~7e
—„ # ..—
/
F
— ie
Monitor:
60 ppb
15 miles
G
*
/ H
*
Monitor:
100 ppb
20 miles
t
*
#= Census Tract "E" Centroid # = Census Tract "E" Centroid
ic
= Air Quality Monitor * .. _ „ , _ ..
1 = Air Quality Monitor
Figure 3C-66. Numerical example of the Voronoi Neighbor Averaging (VNA) technique.
The VNA technique then chooses the monitoring sites whose polygons share a boundary
with the census tract "E" centroid. These monitors are the "Voronoi neighbors", which are used
to estimate the concentration value for census tract "E". The VNA estimate of the concentration
value in census tract "E" is the inverse distance squared weighted average of the four monitored
concentrations. The further the monitor is from the center of census tract "E", the smaller the
weight. For example, the weight for the monitor in census tract "D" 10 miles from the census
tract "E" centroid is calculated as follows:
1/102
= 0 4675
1/102 + 1/152 + 1/152 + 1/202
3C-93
-------
Equation (3C-24)
The weights for the other monitors are calculated in a similar fashion. The final VNA
estimate for census tract "E" is calculated as follows:
VNA(E) = 0.4675 * 80 + 0.2078 * 90 + 0.2078 * 60 + 0.1169 * 100 = 80.3 ppb
Equation (3C-25)
The adjusted hourly O3 concentrations in the eight urban study areas were used to
calculate VNA estimates for approximately 9,725 census tracts * 26,304 hours * 3 air quality
scenarios ~ 767 million values. The computations were executed using the R statistical
computing program (R Core Team, 2018), with the Delaunay triangulation algorithm
implemented in the "deldir" package (Turner, 2018).
3C.7 RESULTS FOR URBAN STUDY AREAS
3C.7.1 Design Values
Table 3C-20 through Table 3C-27 provide the design values for ambient monitoring sites
in each of the eight urban study areas for 2015-2017 based on the observed data, and based on
the adjusted O3 concentrations for the three air quality scenarios (i.e., air quality meeting the
current standard of 70 ppb, and air quality meeting two alternative levels of 75 ppb and 65 ppb).
In each table, the highest design value for each scenario is displayed in bold text. The data in
these tables demonstrate that high O3 values at monitors within some urban study areas respond
differently to reductions in NOx emissions.
In five of the eight urban study areas, the monitor with the highest observed design value
remained the highest when the air quality was adjusted in each of the three air quality scenarios.
For example, Atlanta monitor 131210055 had the highest 2015-2017 design value of 75 ppb, as
well as design values of 70 ppb and 65 ppb for the 70 ppb and 65 ppb scenarios, respectively.
The other study areas where the same monitor had the highest design value in the observations as
well as the 75 ppb, 70 ppb, and 65 ppb scenarios were Dallas (481210034), Detroit (261630019),
Sacramento (060570005), and St. Louis (291831002).
Boston and Philadelphia saw shifts in the highest monitor as a result of the adjustments.
In Boston, monitor 250051004 in Fall River, MA was highest in the observations and following
the upward adjustment to meet 75 ppb. Monitor 250051004 and two other monitors (440090007
in Narragansett, RI and 440090007 east of Providence, RI) had design values of 70 ppb for the
adjustment to meet the current standard. After the final adjustment for the 65 ppb scenario, the
highest design value occurred at the Narragansett monitor. In Philadelphia, monitor 420170012
near Trenton, NJ was highest in the observations. However, following each of the adjustments to
3C-94
-------
75 ppb, 70 ppb and 65 ppb, the location of the highest monitor shifted slightly west to monitor
421010024 (east of downtown Philadelphia).
The pattern for Phoenix was unique among the eight urban study areas. One monitor
(040139997) was consistently high in the observations and for all adjusted levels. However, two
other monitors were equally as high in the observations (040132005; 040131003 - also high at
75 ppb) but responded more strongly to the applied NOx reductions. While monitors 040132005
and 040131003 are slightly removed from downtown Phoenix (near Pinnacle Peak to the
northeast and Mesa to the southeast, respectively), monitor 040139997 is closer the center of the
Phoenix metropolitan area. This location is likely near higher concentrations of urban NOx
sources, making this monitor slightly less responsive to the NOx emissions adjustments.
Table 3C-20. 2015-2017 design values for monitors in the Atlanta study area.
Monitor ID
Observed
75 ppb
70 ppb
65 ppb
130590002
64
64
59
54
130670003
67
67
62
57
130770002
63
63
59
54
130850001
65
65
61
56
130890002
71
71
66
59
130970004
69
69
64
58
131210055
75 A
75
70
65
131350002
71
71
66
60
131510002
71
71
65
59
132230003s
N/A
N/A
N/A
N/A
132319991
67
67
62
56
132470001
69
69
64
57
A Highest DV for each scenario is displayed in bold.
B Monitor used to develop AQ surfaces but DVs not calculated because data were incomplete.
3C-95
-------
Table 3C-21. 2015-2017 design values for monitors in the Boston study area.
Monitor ID
Observed
75 ppb
70 ppb
65 ppb
090159991
70
72
68
61
250010002s
N/A
N/A
N/A
N/A
250051004
<
CO
h-
75
70
63
250051006
69
71
68
62
250092006
66
68
65
61
250094005
65
67
64
59
250095005
62
64
61
56
250170009
64
66
62
57
250213003
70
72
68
62
250230005
68
70
65
60
250250042
61
62
61
58
250270015
65
67
64
59
250270024
66
68
64
59
330012004
59
61
57
53
330111011
62
64
61
57
330115001
67
65
65
60
330131007
63
64
61
56
330150014
63
65
61
57
330150016
66
68
65
59
330150018
65
67
64
59
440030002
72
74
70
63
440071010
70
72
68
62
440090007
71
73
70
65
A Highest DV for each scenario is displayed in bold.
B Monitor used to develop AQ surfaces but DVs not calculated because data were incomplete.
3C-96
-------
Table 3C-22. 2015-2017 design values for monitors in the Dallas study area.
Monitor ID
Observed
75 ppb
70 ppb
65 ppb
400130380s
N/A
N/A
N/A
N/A
480850005
74
72
67
63
481130069
74
72
68
63
481130075
74
72
68
63
481130087
64
62
58
54
481210034
<
o>
h-
75
70
65
481211032
74
71
66
62
481390016
65
63
60
56
481391044
64
61
58
55
482210001
67
65
61
58
482311006
62
60
56
53
482510003
73
70
65
60
482570005
61
59
56
53
483491051
63
61
58
56
483670081
70
67
63
59
483970001
66
63
60
57
484390075
71
69
65
60
484391002
72
70
67
62
484392003
73
71
67
62
484393009
75
73
69
64
484393011
67
65
61
57
A Highest DV for each scenario is displayed in bold.
B Monitor used to develop AQ surfaces but DVs not calculated because data were incomplete.
23. 2015-2017 design values for monitors in the Detroit study area.
Monitor ID
Observed
75 ppb
70 ppb
65 ppb
260490021
67
70
65
60
260492001
67
71
65
59
260910007
66
70
64
58
260990009
71
73
69
63
260991003
66
68
65
61
261250001
70
72
68
63
261470005
71
74
69
64
261610008
67
69
65
60
261619991
69
72
66
59
261630001
66
69
65
60
261630019
<
CO
h-
75
70
65
2616300936
N/A
N/A
N/A
N/A
2616300946
N/A
N/A
N/A
N/A
A Highest DV for each scenario is displayed in bold.
B Monitor used to develop AQ surfaces but DVs not calculated because data were incomplete.
3C-97
-------
Table 3C-24. 2015-2017 design values for monitors in the Philadelphia study area.
Monitor ID
Observed
75 ppb
70 ppb
65 ppb
100010002
66
62
57
53
100031007
67
64
59
55
100031010
74
70
65
60
100031013
71
67
63
58
100032004
72
68
63
58
240150003
74
70
64
59
340010006
64
60
55
51
340070002
77
74
68
63
340071001
68
64
60
56
340110007
66
62
56
53
340150002
74
70
68
60
420110006
66
63
57
53
420110011
70
67
61
58
420170012
80A
75
69
64
420290100
73
69
63
58
420450002
71
69
64
60
420910013
72
69
64
59
421010004B
N/A
N/A
N/A
N/A
421010024
78
75
70
65
421010048
76
72
67
63
A Highest DV for each scenario is displayed in bold.
B Monitor used to develop AQ surfaces but DVs not calculated because data were incomplete.
3C-98
-------
Table 3C-25. 2015-2017 design values for monitors in the Phoenix study area.
Monitor ID
Observed
75 ppb
70 ppb
65 ppb
040130019
74
74
68
62
040131003
76
75
69
63
040131004
75
74
69
63
040131010
74
74
69
62
040132001
68
67
64
59
040132005
76A
74
67
60
040133002
72
72
67
62
040133003
69
68
63
59
040134003
70
69
65
60
040134004
71
70
64
59
040134005s
N/A
N/A
N/A
N/A
040134008
70
69
64
58
040134010
68
68
63
59
040134011
63
62
58
54
040135100B
N/A
N/A
N/A
N/A
040137003
66
65
60
56
040137020
72
72
67
61
040137021
75
74
67
60
040137022
75
74
67
60
040137024
72
71
66
60
040139508
73
72
66
61
040139702
72
71
64
57
040139704
70
69
63
57
040139706
68
68
63
57
040139997
76
75
70
65
040213001
74
73
66
60
040213003
66
65
61
57
040213007
68
67
62
59
040217001
65
64
59
55
040218001
73
72
65
60
A Highest DV for each scenario is displayed in bold.
B Monitor used to develop AQ surfaces but DVs not calculated because data were incomplete.
3C-99
-------
Table 3C-26. 2015-2017 design values for monitors in the Sacramento study area.
Monitor ID
Observed
75 ppb
70 ppb
65 ppb
060170010
83
71
65
59
060170012s
N/A
N/A
N/A
N/A
060170020
80
69
63
56
060570005
86A
75
70
65
0605700076
N/A
N/A
N/A
N/A
060610003
84
72
66
58
060610004
77
67
62
56
060610006
79
71
65
58
060611004
64
61
60
58
060612002
75
67
61
54
060670002
78
70
65
58
060670006
77
71
66
59
060670010
69
63
59
54
060670011
68
61
56
50
060670012
82
72
66
59
060670014B
N/A
N/A
N/A
N/A
060675003
78
69
63
57
061010003
64
56
52
47
061010004s
N/A
N/A
N/A
N/A
061130004
63
55
52
47
061131003
69
60
55
50
A Highest DV for each scenario is displayed in bold.
B Monitor used to develop AQ surfaces but DVs not calculated because data were incomplete.
3C-100
-------
Table 3C-27. 2015-2017 design values for monitors in the St. Louis study area.
Monitor ID
Observed
75 ppb
70 ppb
65 ppb
170830117s
N/A
N/A
N/A
N/A
170831001B
N/A
N/A
N/A
N/A
171170002
65
68
63
57
171190008
69
72
67
62
171191009
68
71
66
61
171193007
70
73
68
62
171199991
67
70
65
58
171630010
68
71
67
61
290990019
68
71
66
59
291130003s
N/A
N/A
N/A
N/A
291130004s
N/A
N/A
N/A
N/A
291831002
72A
75
70
65
291831004
70
73
67
62
291890005
65
67
63
58
291890014
69
72
67
62
295100085
66
69
65
61
A Highest DV for each scenario is displayed in bold.
B Monitor used to develop AQ surfaces but DVs not calculated because data were incomplete.
3C.7.2 Distribution of Hourly O3 Concentrations
Figure 3C-67 through Figure 3C-74 display diurnal boxplots of hourly O3 concentrations
for 2015-2017 at monitor locations in each urban study area. For each hour of the day, the
rectangular box represents the 25th and 75th percentiles of the distribution, with a solid line
representing the median of the distribution through the center. Each box has "whiskers" which
extend up to 1.5 times the interquartile range (i.e., the 75th percentile minus the 25th percentile)
from the box, and dots which represent outlier values. Black boxplots represent observed hourly
O3 concentrations, while blue boxplots represent hourly O3 concentrations adjusted to meet the
current standard of 70 ppb. Red boxplots represent hourly O3 concentrations adjusted for the 75
ppb17 scenario, and green boxplots represent hourly O3 concentrations adjusted for the 65 ppb
scenario.
The boxplots include the observed O3 concentrations as well as the concentrations
adjusted to just meet the current standard and the two alternative air quality scenarios. Note that
these plots include data from all sites in the study area, and thus the plots provide the overall
distribution of O3 at both the urban core sites and the downwind suburban sites. The hourly plots
17 No adjusted values are shown for the 75 ppb scenario for Atlanta because the observed design value was 75 ppb,
and thus no adjustments were made to the hourly O3 concentrations for that scenario.
3C-101
-------
show similar patterns in most of the urban study areas. O3 concentrations during daytime hours
decrease from observed values (black) to values adjusted to meet the current standard of 70 ppb
(blue) and decrease further under the alternative scenario of 65 ppb (green). These daytime
decreases are mainly seen on high O3 days represented by outlier dots extending above the box
and whiskers. Some study areas had observed 2015-2017 design values already meeting the
alternative scenario of 75 ppb, therefore some plots show increases in O3 concentrations while
other study areas show decreases in O3 concentrations for the 75 ppb scenario.
In some urban study areas O3 concentrations on the mid-range days, represented by the
25th - 75th percentile boxes, remained fairly constant (e.g. Boston) while in other urban study
areas O3 on mid-range days decreased (e.g. Atlanta). Although daytime O3 decreased,
concentrations during morning rush-hour period generally increase. These increases are
associated with VOC-limited and NOx titration conditions near NOx sources during rush-hour
periods. Reducing NOx under these conditions results in less O3 titration and thus increases O3
concentrations. Nighttime increases in O3 as a results of NOx reductions are often seen to a lesser
extent than morning rush-hour period increases. Collectively these features generally lead to a
flattening of the diurnal O3 pattern with smaller differences between daytime and nighttime
concentrations as NOx emissions are reduced. Urban study areas that required more substantial
NOx reductions for the 65 ppb scenario generally had more pronounced patterns of decreases in
daytime O3 and increases in nighttime O3 leading to a flatter diurnal O3 pattern (e.g., Sacramento
in Figure 3C-73).
Figure 3C-75 through Figure 3C-82 display the same information as Figure 3C-67
through Figure 3C-74 but for monthly rather than diurnal distributions. Similar to the diurnal
plots, the seasonal distributions become flatter when adjusted to meet the 70 ppb and 65 ppb
scenarios, especially on the highest O3 days. This is due to more O3 decreases during summer
months and more O3 increases in winter months. The O3 increases in the winter are consistent
with the understanding that solar insolation rates are lower in the winter reducing total
photochemical activity and shifting the net effect of NOx emissions on O3 which can both create
O3 through photochemical pathways and destroy O3 through titration. In addition, the decreases
on the highest O3 days and increases on the lowest O3 days show a visible compression of the O3
distribution in these plots, similar to what was seen in the diurnal plots.
3C-102
-------
o
¦
o
CM -
O
O -
O
00
_Q
Q.
Q_
n°-
U CD
O
o.
CM
Atlanta sites: 2015-2017
observed
70 ppb
65 ppb
o
o
o
o
o
o
2 8
8
8 I
© a
§ g B ft
« S S Q
- e s
H
§
©
I :
O
6) 8
n
- e
u
w © y y
0
8
10 12
hour
14 16 18 20 22
Figure 3C-67. Diurnal distribution of hourly O3 concentrations at monitoring sites in the
Atlanta study area. Note: Observed concentrations in this area have a design
value of 75 ppb.
3C-103
-------
Boston sites: 2015-2017
o
1
¦ observed
¦ 75 ppb
o ¦ 70 ppb
¦ 65 ppb
o- o
I 1 1 I 1 I I I I I 1 I I I I I I 1 I 1 1 I I I
0 2 4 6 8 10 12 14 16 18 20 22
hour
Figure 3C-6S. Diurnal distribution of hourly O3 concentrations at monitoring sites in the
Boston study area.
3C-104
-------
o
o
CM
O
O
O
00
_Q
Q.
Cl
n°
U CD
O
O
CM
Dallas sites: 2015-2017
observed
75 ppb
70 ppb
65 ppb
o
o o
8 Q
y g
~w< 'S'
o °
© 8
lO, 0
I '
« -
o
o o
r\
ft a
e a o
,v>
§
o
oft?
a i " "
- ¦' P: § 8
§
>w<
©
o
o
G
a
o
8 R §
o - o
5 O
O
a
q a © :
¦ i|
UJLIlUjU
-1- —
—i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—
0 2 4 6 8 10 12 14 16 18 20 22
hour
Figure 3C-69. Diurnal distribution of hourly O3 concentrations at monitoring sites in
Dallas study area.
3C-105
-------
o
o
CM
O
O
O
00
_Q
Q.
Q.
n°
U CD
O
o
CM
Detroit sites: 2015-2017
observed
75 ppb
70 ppb
65 ppb
0 ~
V- £„»
n
u
r\, V / W r>
8 .. - © © r
^ * u< ~ ipi
0 " - ~
O -
Tl
m ¦
n
¦¦
-i 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 r
n 1 1 r~
0 2 4 6 8 10 12 14 16 18 20 22
hour
Figure 3C-70. Diurnal distribution of hourly O3 concentrations at monitoring sites in
Detroit study area.
3C-106
-------
Philadelphia sites: 2015-2017
o
C\J
o
o
o
00
_Q
Q.
ClO
"—CD
co
O
o
"d-
o
CM
observed
75 ppb
70 ppb
65 ppb
o
° ° i i i °
' 1 8 § I § a
0 u a d 6 -
o
§ m
1 *
1 8
s I
8 o
„ o
i °
8
: i p § § © I
Q
R
I I
I
I I
I I
I——I
1
u
n
J i_r_i i—i —i
i—i
W
i i
i i
i i
¦ ¦
i i
»
LJ
mi
0 @
—i 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1—
0 2 4 6 8 10 12 14 16 18 20 22
hour
Figure 3C-71. Diurnal distribution of hourly O3 concentrations at monitoring sites in
Philadelphia study area.
3C-107
-------
Phoenix sites: 2015-2017
o
CVJ
o
o
o
CO
_Q
Q.
Q.O
•—CD
CO
O
o
"3-
o
CM
observed
75 ppb
70 ppb
65 ppb
o
o
8
9
0
1
K S a s g
o «
° 8 Q Q q Q o °
n
n
UHl
I I i
WW
n
0
§ 0
O
©
Q
tH
F
©
'x
!
© O
? 0
w
9 6
O
n
hi
U
i i
LJ|
Jill
I I™
9 g « -
—1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
0 2 4 6 8 10 12 14 16 18 20 22
hour
Figure 3C-72. Diurnal distribution of hourly O3 concentrations at monitoring sites in
Phoenix study area.
3C-108
-------
o
o
CM
O
O
O
00
_Q
Q.
Q.
n°
U CD
O
o
CVJ
Sacramento sites: 2015-2017
observed
75 ppb
70 ppb
65 ppb
o o
§ Q
Q I
° o
° Q o
o - O
§ i ° §
I I I 1 I
§ 1
o
o
©
n ? a
§ I h
© @ i g % 0 | g \
I
- ® i"i q
*rt
1_«
cpmrnm
LJ
LJ
LJ
U
UJ
LJ
rn
u
M
n
¦¦
n
l I
rn
n
u
LJ
fCDr
nnffl
yyuy
—I 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1—
0 2 4 6 8 10 12 14 16 18 20 22
hour
Figure 3C-73. Diurnal distribution of hourly O3 concentrations at monitoring sites in
Sacramento study area.
3C-109
-------
SaintLouis sites: 2015-2017
o
1
¦ observed
¦ 75 ppb
o ¦ 70 ppb
¦ 65 ppb
4- y iii pi *" -r ~
o- - ~ - - u
1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
0 2 4 6 8 10 12 14 16 18 20 22
hour
Figure 3C-74. Diurnal distribution of hourly O3 concentrations at monitoring sites in
St. Louis study area.
3C-110
-------
Atlanta sites: 2015-2017
o
C\J -
O
o -
o.
CO
_Q
Q.
Q-O.
v—CD
CO
o
o.
o.
CM
observed
70 ppb
65 ppb
o
0
yj
o
'o1
I
s
I
o
o
o
o
I
u
—I—
11
6 7
month
10
12
Figure 3C-75. Monthly distribution of hourly O3 concentrations at monitoring sites in the
Atlanta study area. Note: Observed concentrations in this area have a design
value of 75 ppb.
3C-111
-------
Boston sites: 2015-2017
O
CVJ -
o
o -
observed
75 ppb
70 ppb
65 ppb
Q
o
o.
CO
_Q
Q.
Q.O.
' CD
CO
O
o.
Tl-
CD .
CM
n
u
n
mi
H
I I
6 7
month
10 11
12
Figure 3C-76. Monthly distribution of hourly O3 concentrations at monitoring sites in the
Boston study area.
3C-112
-------
Dallas sites: 2015-2017
o
c\J -
o
o -
o.
oo
observed
75 ppb
70 ppb
65 ppb
o
o
0
o
Q
Q
O
O
O
o
0
1
_Q
Q.
Q. O.
-—CO
CO
o
o
"tf-
o.
(M
I
o
o
6 7
month
10 11
12
Figure 3C-77. Monthly distribution of hourly O3 concentrations at monitoring sites in
Dallas study area.
3 C -113
-------
Detroit sites: 2015-2017
o
c\J -
o
o -
observed
75 ppb
70 ppb
65 ppb
O
o
o.
CO
_Q
Q_
Q.O.
' CD
CO
O
o.
Tl-
o.
CM
I
IhI
IhI
1
15
—r-
9
—i 1—
10 11
6 7
month
12
Figure 3C-7S. Monthly distribution of hourly O3 concentrations at monitoring sites in the
Detroit study area.
3C-114
-------
Philadelphia sites: 2015-2017
o
CM -
O
O ¦
observed
75 ppb
70 ppb
65 ppb
o
0
n
ft
1
o
o
i
§
o
o
o
o
5
o
ia
o.
CO
_Q
CL
Q. o.
-—co
CO
O
o
8
o
o
8
0
0
o
"3-
o.
CM
¥
g
a
—r~
3
—r—
4
—i—
5
—i—
9
—i—
10
6 7
month
11 12
Figure 3C-79. Monthly distribution of hourly O3 concentrations at monitoring sites in the
Philadelphia study area.
3 C-115
-------
Phoenix sites: 2015-2017
o
c\J -
o
o -
o.
CO
_Q
Q.
Q.O.
' CD
CO
O
o.
Tt"
o.
CM
observed
75 ppb
70 ppb
65 ppb
o
ft
o
ft
F
L
i
3
—i—
4
—i—
5
6 7
month
—i—
8
—i—
9
—i—
10
11 12
Figure 3C-80. Monthly distribution of hourly O3 concentrations at monitoring sites in
Phoenix study area.
3C-116
-------
Sacramento sites: 2015-2017
o
c\j -
o
o -
observed
75 ppb
70 ppb
65 ppb
o
o
o
o
@
o.
00
o
o
_Q
Q.
Q. O.
co
o
o
"tf-
o.
(M
¦
I
2
—r—
3
I
4
—I—
5
—r-
8
—i 1 1 1—
9 10 11 12
6 7
month
Figure 3C-81. Monthly distribution of hourly O3 concentrations at monitoring sites in
Sacramento study area.
3C-117
-------
SaintLouis sites: 2015-2017
o
CVJ -
o
o -
observed
75 ppb
70 ppb
65 ppb
O .
CO
_Q
Q_
Q.O.
' CD
CO
O
o.
o.
CM
1
J
n
LJ
—r-
2
—i—
3
—i—
8
—i—
11
4 5 6 7
month
10
12
Figure 3C-82. Monthly distribution of hourly O3 concentrations at monitoring sites in the
St. Louis study area.
3C.7.3 Air Quality Inputs for the Exposure and Risk Analyses
The air quality inputs for the exposure and risk analyses discussed in chapter 3 of this PA
include spatial surfaces of hourly O3 concentrations estimated for each census tract in the eight
urban study areas using the VNA technique described in section 3C.6. In this section, we present
three types of figures which summarize the data from the hourly VNA surfaces for observed air
3 C-118
-------
quality, and air quality adjusted to meet the current standard of 70 ppb, and air quality adjusted
to meet alternative scenarios of 75 ppb18 and 65 ppb.
The first set of figures (Figure 3C-83 through Figure 3C-90) shows density scatter plots
of the change in MDA8 O3 concentrations versus the observed concentrations based on the
hourly VNA estimates in each study area. In each of these figures, the left-hand panel shows the
observed MDA8 values (x-axis) versus the change in those values that occur when air quality is
adjusted for the 75 ppb scenario (y-axis). The middle panel shows the MDA8 values for air
quality adjusted to meet the 75 ppb scenario (x-axis) versus the additional change in those values
that occur when air quality is adjusted to meet the current standard of 70 ppb (y-axis). Finally,
the right-hand panels show the corresponding changes from the current standard to the 65 ppb
scenario. Within each panel, the x and y values are rounded to the nearest integer and colored to
show the relative frequency of each lxl ppb square within the plot region. Values falling
outside of the plot region were set to the nearest value within the plot region, and frequencies
above the range in the color bar were set to the highest value within the color bar.
The second set of figures (Figure 3C-91 through Figure 3C-106) provides maps of the
adjusted design values (3-year average of the annual 4th highest MDA8 values) and May-
September average MDA8 values based on the ambient air data and the hourly VNA surfaces, as
well as difference maps showing the changes between these surfaces. For the difference maps,
the panels on the left show the changes in these values that occur when air quality is adjusted for
the 75 ppb scenario, the panels in the middle show the additional changes in these values that
occur when air quality is further adjusted to meet the current standard of 70 ppb, and the right-
hand panels show the additional changes that occur then air quality is further adjusted for the 65
ppb scenario. Within each panel, squares show values based on observed data at ambient air
monitoring sites while circles show values based on VNA estimates at census tract centroids.
While each panel shows both monitors in the study area for each selected urban study area as
well as some additional monitors located outside of the study area, only the monitors located
within the study area were used when determining the emissions reductions necessary to meet
the various standards.
The third set of figures (Figure 3C-107 through Figure 3C-114) shows changes in design
values (3-year average of the annual 4th highest MDA8 values) and May-September average
MDA8 values in the eight urban case study areas versus population and population density. The
total population and population density information for each census tract were obtained from the
U.S. Census Bureau based on the 2010 U.S. Census. Each panel shows a histogram of the total
18 Atlanta was already just meeting the 75 ppb scenario for the 2015-2017 period. Boston, Detroit, and St. Louis
were below 75 ppb for 2015-2017; design values for these urban study areas were adjusted upward to just meet
75 ppb.
3C-119
-------
population stratified by the change in design value or seasonal average. The bars are also color-
coded by population density bin. Values falling outside of the plot region set to the nearest
values within the plot region.
In general, the density scatter plots show that the HDDM adjustment procedure predicts
increases in MDA8 O3 at low ambient air concentrations and decreases in MDA8 O3 at high
concentrations (Figure 3C-83 through Figure 3C-90). The vast majority of the increases in
MDA8 O3 occur at ambient air concentrations below 50 ppb. The relationship between the
starting concentrations and the changes in these values based on the HDDM adjustments is fairly
linear with strong negative correlation in all eight urban study areas.19 In some study areas, such
as Philadelphia and Detroit, there is a bimodal pattern near the center of the distribution, which
may be indicative of differing behavior near the urban population center versus the surrounding
suburban areas.
The maps reveal consistent spatial patterns of O3 changes across the urban study areas.
The design values generally decreased when air quality was adjusted to meet the current standard
of 70 ppb20 and continued to decrease when air quality was further adjusted for the 65 ppb
scenario (Figure 3C-91 through Figure 3C-106). The design values tend to decrease more
quickly in suburban and rural areas than in the urban population centers. The May-September
"seasonal" average MDA8 values also followed this trend to some extent, although the behavior
in the urban population centers varied slightly amongst the urban study areas (Figure 3C-107
through Figure 3C-114). In summary, these figures show that using the CAMx/HDDM
adjustment methodology, peak O3 concentrations are reduced in urban study areas with large
domain-wide reductions in U.S. anthropogenic NOx emissions.
19 Except for the "Observed - 75 ppb" changes for the three urban study areas where the design values were adjusted
upwards: Boston, Detroit, and St. Louis.
20 All design values from the VNA surfaces decreased when going from recent conditions to the 75 ppb adjustment
scenario, with the exceptions of study areas that required upward adjustments for the 75 ppb scenario: Boston,
Detroit, and St. Louis.
3C-120
-------
Change in MDA8 03 from Observed to 75 ppb
Change in MDA8 03 from 75 ppb to 70 ppb
Change in MDA8 03 from 70 ppb to 65 ppb
O o
O o
0 20 40 60 80 100
MDA8 03 (ppb) - Qbseived 2015 - 2017
20 40 60 80 100
MDA8 03 (ppb) - Meeting 75 ppb
20 40 60 80 100
MDA8 03 (ppb) - Meeting 70 ppb
0.0 0.2 0 4 Frequency (%) 0.6 0.8
Figure 3C-83. Changes in MDA8 O3 based on HDDM adjustments in the Atlanta study area.
1.0
3C-121
-------
Change in MDA8 03 from Observed to 75 ppb
Change in MDA8 03 from 75 ppb to 70 ppb
Change in MDA8 03 from 70 ppb to 65 ppb
O o
O o
O o
0 20 40 60 80 100
MDA8 03 (ppb) - Qbseived 2015 - 2017
20 40 60 80
MDA8 03 (ppb) - Meeting 75 ppb
20 40 60 80 100
MDA8 03 (ppb) - Meeting 70 ppb
0.0 0.2 0.4 Frequency (%) 0.6 0.8
Figure 3C-84. Changes in MDA8 O3 based on HDDM adjustments in the Boston study area.
1.0
3C-122
-------
Change in MDA8 03 from Observed to 75 ppb
Change in MDA8 03 from 75 ppb to 70 ppb
Change in MDA8 03 from 70 ppb to 65 ppb
O o
O o
20 40 60 80 100
MDA8 03 (ppb) - Qbseived 2015 - 2017
20 40 60 80
MDA8 03 (ppb) - Meeting 75 ppb
20 40 60 80
MDA8 03 (ppb) - Meeting 70 ppb
0.0 0.2 0.4 Frequency (%) 0.6
Figure 3C-85. Changes in MDA8 O3 based on HDDM adjustments in the Dallas study area.
0.8
1.0
3C-123
-------
Change in MDA8 03 from Observed to 75 ppb
Change in MDA8 03 from 75 ppb to 70 ppb
Change in MDA8 03 from 70 ppb to 65 ppb
O o
O o
0 20 40 60 80 100
MDA8 03 (ppb) - Qbseived 2015 - 2017
20 40 60 80 100
MDA8 03 (ppb) - Meeting 75 ppb
20 40 60 80 100
MDA8 03 (ppb) - Meeting 70 ppb
0.0 0.2 0.4 Frequency (%) 0.6 0.8
Figure 3C-86. Changes in MDA8 O3 based on HDDM adjustments in the Detroit study area.
1.0
3C-124
-------
Change in MDA8 03 from Observed to 75 ppb
Change in MDA8 03 from 75 ppb to 70 ppb
Change in MDA8 03 from 70 ppb to 65 ppb
0 20 40 60 80 100
MDA8 03 (ppb) - Qbseived 2015 - 2017
O o
O o
20 40 60 80 100
MDA8 03 (ppb) - Meeting 75 ppb
20 40 60 80 100
MDA8 03 (ppb) - Meeting 70 ppb
0.0 0.2 0.4 Frequency (%) 0.6 0.8
Figure 3C-87. Changes in MDA8 O3 based on HDDM adjustments in the Philadelphia study area.
1.0
3C-125
-------
Change in MDA8 03 from Observed to 75 ppb
Change in MDA8 03 from 75 ppb to 70 ppb
Change in MDA8 03 from 70 ppb to 65 ppb
O o
O o
0 20 40 60 80 100
MDA8 03 (ppb) - Qbsen/ed 2015 - 2017
20 40 60 80 100
MDA8 03 (ppb) - Meeting 75 ppb
20 40 60 80 100
MDA8 03 (ppb) - Meeting 70 ppb
0.0 0.2 0.4 Frequency (%) 0.6 0.8
Figure 3C-88. Changes in MDA8 O;? based on HDDM adjustments in the Phoenix study area.
1.0
3C-126
-------
Change in MDA8 03 from Observed to 75 ppb
Change in MDA8 03 from 75 ppb to 70 ppb
Change in MDA8 03 from 70 ppb to 65 ppb
O o
O o
O o
20 40 60 80 100
MDA8 03 (ppb) - Qbseived 2015 - 2017
20 40 60 80
MDA8 03 (ppb) - Meeting 75 ppb
20 40 60 80
MDA8 03 (ppb) - Meeting 70 ppb
0.0 0.2 0.4 Frequency (%) 0.6 0.8
Figure 3C-89. Changes in MDA8 O3 based on HDDM adjustments in the Sacramento study area.
1.0
3C-127
-------
Change in MDA8 03 from Observed to 75 ppb
Change in MDA8 03 from 75 ppb to 70 ppb
Change in MDA8 03 from 70 ppb to 65 ppb
o o
O o
0 20 40 60 80 100
MDA8 03 (ppb) - Qbseived 2015 - 2017
20 40 60 80 100
MDA8 03 (ppb) - Meeting 75 ppb
20 40 60 80 100
MDA8 03 (ppb) - Meeting 70 ppb
0.0 0.2 0.4 Frequency (%) 0.6 0.8
Figure 3C-90. Changes in MDA8 O3 based on HDDM adjustments in the St. Louis study area.
1.0
3C-128
-------
Observed 2015-2017
Meeting 75 ppb
Meeting 70 ppb
Meeting 65 ppb
Figure 3C-91. Annual 4th highest MDA8 O3 and May-September mean MDA8 O3 based on HDDM adjustments in the
Atlanta study area.
3C-129
-------
75 ppb - Observed 70 ppb - 75 ppb 65 ppb - 70 ppb
_Q
Q_
h ™ Q.
ro
O
CO
<
Q
Q.
0)
^C/)
>>
_ C3
Figure 3C-92. Changes in annual 4th highest MDA8 O3 and May-September mean MDA8 O3 based on HDDM adjustments
in the Atlanta study area.
3C-130
-------
Observed 2015-2017
Meeting 75 ppb
Meeting 70 ppb
Meeting 65 ppb
7^
u d>
Figure 3C-93.
Annual 4th highest MDA8 O3 and May-September mean MDA8 O3 based on HDDM adjustments in the
Boston study area.
3C-131
-------
75 ppb - Observed
70 ppb - 75 ppb
65 ppb - 70 ppb
ojro
Figure 3C-94. Changes in annual 4th highest MDA8 O3 and May-September mean MDA8 O3 based on HDDM adjustments
in the Boston study area.
3C-132
-------
Observed 2015-2017 Meeting 75 ppb Meeting 70 ppb Meeting 65 ppb
Figure 3C-95. Annual 4th highest MDA8 O3 and May-September mean MDA8 O3 based on HDDM adjustments in the Dallas
study area.
3C-133
-------
75 ppb - Observed 70 ppb - 75 ppb 65 ppb - 70 ppb
Figure 3C-96. Changes in annual 4th highest MDA8 O3 and May-September mean MDA8 O3 based on HDDM adjustments
in the Dallas study area.
3C-134
-------
Observed 2015-2017 Meeting 75 ppb Meeting 70 ppb Meeting 65 ppb
Figure 3C-97. Annual 4th highest MDA8 O3 and May-September mean MDA8 O3 based on HDDM adjustments in the
Detroit study area.
3C-135
-------
75 ppb - Observed 70 ppb - 75 ppb 65 ppb - 70 ppb
OlCO
Figure 3C-98. Changes in annual 4th highest MDA8 O3 and May-September mean MDA8 O3 based on HDDM adjustments
in the Detroit study area.
3C-136
-------
Observed 2015-2017 Meeting 75 ppb Meeting 70 ppb Meeting 65 ppb
Figure 3C-99. Annual 4th highest MDA8 O3 and May-September mean MDA8 O3 based on HDDM adjustments in the
Philadelphia study area.
3C-137
-------
75 ppb - Observed 70 ppb - 75 ppb 65 ppb - 70 ppb
CNro
Figure 3C-100. Changes in annual 4th highest MDA8 O3 and May-Septeinber mean MDA8 O3 based on HDDM adjustments in
the Philadelphia study area.
3C-138
-------
Observed 2015-2017 Meeting 75 ppb Meeting 70 ppb Meeting 65 ppb
Figure 3C-101. Annual 4th highest MDA8 O3 and May-September mean MDA8 O3 based on HDDM adjustments in the
Phoenix study area.
3C-139
-------
75 ppb - Observed 70 ppb - 75 ppb 65 ppb - 70 ppb
Figure 3C-102. Changes in annual 4th highest MDA8 O3 and May-Septeinber mean MDA8 O3 based on HDDM adjustments in
the Phoenix study area.
3C-140
-------
Observed 2015-2017 Meeting 75 ppb Meeting 70 ppb Meeting 65 ppb
Figure 3C-103. Annual 4th highest MDA8 O3 and May-September mean MDA8 O3 based on HDDM adjustments in the
Sacramento study area.
3C-141
-------
75 ppb - Observed 70 ppb - 75 ppb 65 ppb - 70 ppb
csiro
Figure 3C-104. Changes in annual 4th highest MDA8 O3 and May-Septeinber mean MDA8 O3 based on HDDM adjustments in
the Sacramento study area.
3C-142
-------
Observed 2015-2017 Meeting 75 ppb Meeting 70 ppb Meeting 65 ppb
Figure 3C-105. Annual 4th highest MDA8 O3 and May-September mean MDA8 O3 based on HDDM adjustments in the St.
Louis study area.
3C-143
-------
75 ppb - Observed 70 ppb - 75 ppb 65 ppb - 70 ppb
1
CL
Q_
- CM CO
O
CO
_ CSJ (I)
c
.Q
Q_
CM Q_
CO
o
^co
<
Q
CL
a>
^c/)
>>
-------
>,
ro
Observed 2015 - 2017
Meeting 75 ppb
Meeting 70 ppb
Meeting 65 ppb
50 60 70 80
Concentration (ppb)
50 60 70 80
Concentration (ppb)
50 60 70 80
Concentration (ppb)
60 70
Concentration (ppb)
J
30 40 50 60
Concentration (ppb)
30 40 50 60 30 40 50 60
Concentration (ppb) Concentration (ppb)
Population Density (people/kmA2)
Less than 500
30 40 50 60
Concentration (ppb)
500 to 2500
More than 2500
Figure 3C-107. Annual 4th highest MDA8 O3 and May-September mean MDA8 O3 by population based on HDDM
adjustments in the Atlanta study area.
3C-145
-------
Observed 2015 - 2017
Meeting 75 ppb
Meeting 70 ppb
Meeting 65 ppb
>,
50 60 70 80
Concentration (ppb)
0
co
to
r
CM
0
—
O
E
CM
cr
ID
(>
03
-5
O
£_i_
T—
O
LL
ID
O
O
O
50 60 70 80
Concentration (ppb)
O
CO
___
10
>
r
CM
0
—
O
F
CM
c
LD
< >
03
O
£_i_
—
( )
LL
U1
O
O
O
50 60 70 80
Concentration (ppb)
0
co
m
03
"J
O
LI
T—
O
f)
O
O
O
O
CO
__
m
if}
r
CM
0
—
O
F
CM
c
ID
< >
03
"J
O
Q.
T—
O
n
ID
60 70
Concentration (ppb)
30 40 50 60
Concentration (ppb)
40 50
Concentration (ppb)
30 40 50 60
Concentration (ppb)
30 40 50 60
Concentration (ppb)
Less than 500
Population Density (peop!e/kmA2)
500 to 2500
More than 2500
Figure 3C-108. Annual 4th highest MDA8 O3 and May-September mean MDA8 O3 by population based on HDDM
adjustments in the Boston study area.
3C-146
-------
Observed 2015 - 2017
Meeting 75 ppb
Meeting 70 ppb
Meeting 65 ppb
>»
CO
50 60 70 80
Concentration (ppb)
30 40 50 60
Concentration (ppb)
O
co
ir)
r
CM
0
—
0
E
CM
£T
m
(>
03
0
£_i_
T—
O
LL
U";
0
0
0
50 60 70 80
Concentration (ppb)
0
CO
10
)
r
CM
0
—
O
F
CM
c
< >
03
O
O.
—
( >
n
u1
0
0
0
50 60 70 80
Concentration (ppb)
30 40 50 60 ~ 30 40 50 60
Concentration (ppb) Concentration (ppb)
Population Density (peopie/kmA2)
Less than 500
0
CO
LD
tjy
r
CM
0
—
O
F
CM
c
ID
< >
03
~i
O
U-
—
0
LL
ID
O
O
O
O
CO
__
ID
<13
r
CM
0
—
O
F
CM
c
LD
(>
03
~J
O
Q.
T—
O
n
ID
O
O
O
50 60 70 80
Concentration (ppb)
30 40 50 60
Concentration (ppb)
500 to 2500
More than 2500
Figure 3C-109. Annual 4th highest MDA8 O3 and May-September mean MDA8 O3 by population based on HDDM
adjustments in the Dallas study area.
3C-147
-------
Observed 2015 - 2017
Meeting 75 ppb
Meeting 70 ppb
Meeting 65 ppb
>»
CO
i.
50 60 70 80
Concentration (ppb)
0
co
ID
Ui
r
CM
0
—
O
E
CM
ST
ID
( >
CO
~5
O
UL
T—
O
Q_
LD
O
O
O
60 70
Concentration (ppb)
O
CO
___
10
(/)
r
CM
0
—
O
F
CM
c
utj
< >
03
1
0
£_i_
—
( >
LL
ID
O
O
O
50 60 70
Concentration (ppb)
0
co
___
LO
U)
r
CM
0
—
O
F
CM
CI
LO
< >
TO
~5
O
£-1.
^—
O
LL
iO
O
O
O
O
CO
__
m
if}
r
CM
0
—
O
E
CM
c
LO
(>
ro
O
Q.
T—
O
n
10
50 60 70 80
Concentration (ppb)
30 40 50 60
Concentration (ppb)
40 50 60 30 40 50 60
Concentration (ppb) Concentration (ppb)
Population Density (peopie/kmA2)
Less than 500
30 40 50 60
Concentration (ppb)
500 to 2500
More than 2500
Figure 3C-110. Annual 4th highest MDA8 O3 and May-September mean MDA8 O3 by population based on HDDM
adjustments in the Detroit study area.
3C-148
-------
Observed 2015 - 2017
Meeting 75 ppb
Meeting 70 ppb
Meeting 65 ppb
>,
ro
.It
St]
Si
50 60 70 80
Concentration (ppb)
0
co
UT)
IJ>
r
CSJ
0
—
O
F
CM
£T
IT)
(>
03
~5
O
Li.
T—
O
Q_
LO
O
O
O
50 60 70 80
Concentration (ppb)
0
CO
10
>
r
CM
0
—
O
F
OJ
c
uo
< >
03
0
£_i_
—
( )
n
0
0
0
50 60 70 80
Concentration (ppb)
50 60 70 80
Concentration (ppb)
30 40 50 60
Concentration (ppb)
40 50
Concentration (ppb)
40 50
Concentration (ppb)
30 40 50 60
Concentration (ppb)
Less than 500
Population Density (people/kmA2)
500 to 2500
More than 2500
Figure 3C-111. Annual 4th highest MDA8 O3 and May-September mean MDA8 O3 by population based on HDDM
adjustments in the Philadelphia study area.
3C-149
-------
>,
Observed 2015 - 2017
Meeting 75 ppb
Meeting 70 ppb
Meeting 65 ppb
k
50 60 70 80
Concentration (ppb)
60 70
Concentration (ppb)
60 70
Concentration (ppb)
50 60 70 80
Concentration (ppb)
S= LO
30 40 50 60
Concentration (ppb)
30 40 50 60 30 40 50 60
Concentration (ppb) Concentration (ppb)
Population Density (peop!e/kmA2)
Less than 500
30 40 50 60
Concentration (ppb)
500 to 2500
More than 2500
Figure 3C-112. Annual 4th highest MDA8 O3 and May-September mean MDA8 O3 by population based on HDDM
adjustments in the Phoenix study area.
3C-150
-------
>,
03
30
Observed 2015 - 2017
CO
0
O
O
co
co
00
<
^10
^ to
Q
c ™ "
0
— 0
O
in O
-t—¦
(/)
"
E™ '
CD
sz
c to
c= 10
O)
0 -
0 "
X
"ro
=5 O
to
=5 O
sz
CLt-
Q- T—
~rci
0
0
3
n_i t_r—i_|
c
c
<
0
0
0
50 60 70 80
Concentration (ppb)
0
CO
0
0
0
CO
CO
OO
<
0
to
, to
g CSJ '
0
£ CM "
O
IE 0
= O
c
03
=.
O
Meeting 75 ppb
Meeting 70 ppb
Meeting 65 ppb
40 50 60
Concentration (ppb)
60 70
Concentration (ppb)
60 70
Concentration (ppb)
60 70
Concentration (ppb)
40 50
Concentration (ppb)
30 40 50 60
Concentration (ppb)
30 40 50 60
Concentration (ppb)
Population Density (peopie/kmA2)
500 to 2500
More than 2500
1 I Less than 500
Figure 3C-113. Annual 4th highest MDA8 O3 and May-September mean MDA8 O3 by population based on HDDM
adjustments in the Sacramento study area.
3C-151
-------
>,
Observed 2015 - 2017
Meeting 75 ppb
Meeting 70 ppb
Meeting 65 ppb
50 60 70 80
Concentration (ppb)
50 60 70
Concentration (ppb)
50 60 70 80
Concentration (ppb)
50 60 70 80
Concentration (ppb)
30 40 50 60
Concentration (ppb)
40 50 60 30 40 50 60
Concentration (ppb) Concentration (ppb)
Population Density (people/kmA2)
Less than 500
30 40 50 60
Concentration (ppb)
500 to 2500
More than 2500
Figure 3C-114. Annual 4th highest MDA8 O3 and May-September mean MDA8 O3 by population based on HDDM
adjustments in the St. Louis study area.
3C-152
-------
REFERENCES
Carlton, AG, Bhave, PV, Napelenok, SL, Edney, EO, Sarwar, G, Pinder, RW, Pouliot, GA and
Houyoux, M (2010). Model Representation of Secondary Organic Aerosol in
CMAQv4.7. Environ Sci Technol 44(22): 8553-8560.
Chen, J, Zhao, R and Li, Z (2004). Voronoi-based k-order neighbour relations for spatial
analysis. J Photogramm Remote Sens 59(1): 60-72.
Cohan, DS, Hakami, A, Hu, Y and Russell, AG (2005). Nonlinear Response of Ozone to
Emissions: Source Apportionment and Sensitivity Analysis. Environ Sci Technol 39(17):
6739-6748.
Dunker, AM (1984). The decoupled direct method for calculating sensitivity coefficients in
chemical kinetics. J Chem Phys 81(5): 2385-2393.
ERG (2017). Technical Report: Development of Mexico Emission Inventories for the 2014
Modeling Platform. Research Triangle Park, NC. Available at:
ftp://newftp.epa.gov/Air/emismod/2014/v2/2014fd/emissions/EPA%205-
18 %20Report_ Clean %20Final_01042017. pdf
Gery, MW, Whitten, GZ, Killus, JP and Dodge, MC (1989). A photochemical kinetics
mechanism for urban and regional scale computer modeling. J. Geophys. Res 94(12):
925-912.
Gilliam, RC and Pleim, JE (2009). Performance Assessment of New Land Surface and Planetary
Boundary Layer Physics in the WRF-ARW. J Appl Meteorol Climatol 49(4): 760-774.
Gold, CM, Remmele, PR and Roos, T (1997). Voronoi methods in GIS. Algorithmic
Foundations of Geographic Information Systems. Springer Berlin Heidelberg. Berlin,
Heidelberg 1340: 21-35.
Hakami, A, Bergin, M and Russell, A (2004). Ozone formation potential of organic compounds
in the eastern United States: a comparison of episodes, inventories, and domains.
Environ. Sci. Technol.
Hakami, A, Odman, MT and Russell, AG (2003). High-Order, Direct Sensitivity Analysis of
Multidimensional Air Quality Models. Environ Sci Technol 37(11): 2442-2452.
Heath, NK, Pleim, JE, Gilliam, RC and Kang, D (2016). A simple lightning assimilation
technique for improving retrospective WRF simulations. J Adv Model Earth Syst 8(4):
1806-1824.
Henderson, B, Dolwick, P, Jang, C, Misenis, C, Possiel, N, Timin, B, Eyth, A, Vukovich, J,
Mathur, R, Hogrefe, C, Pouliot, G, Appel, W and Brehme, K (2018). Hemispheric
CMAQ Application and Evaluation for 2016. 2018 CMAQ Annual Conference Chapel
Hill, NC.
3C-153
-------
Houyoux, MR, Vukovich, JM, Coats Jr., CJ, Wheeler, NJM and Kasibhatla, PS (2000). Emission
inventory development and processing for the Seasonal Model for Regional Air Quality
(SMRAQ) project. Journal of Geophysical Research: Atmospheres 105(D7): 9079-9090.
Koo, B, Dunker, AM and Yarwood, G (2007). Implementing the decoupled direct method for
sensitivity analysis in a particulate matter air quality model. Environ Sci Technol 41(8):
2847-2854.
Napelenok, SL, Foley, KM, Kang, D, Mathur, R, Pierce, T and Rao, ST (2011). Dynamic
evaluation of regional air quality model's response to emission reductions in the presence
of uncertain emission inventories. Atmos Environ 45(24): 4091-4098.
Nenes, A, Pandis, SN and Pilinis, C (1998). ISORROPIA: A New Thermodynamic Equilibrium
Model for Multiphase Multicomponent Inorganic Aerosols. Aquat Geochem 4(1): 123-
152.
Pouliot, G and Bash, J (2015). Updates to Version 3.61 of the Biogenic Emission Inventory
System (BEIS). Presented at Air and Waste Management Association conference
Raleigh, NC
R Core Team (2018). R: A Language and Environment for Statistical Computing. R Foundation
for Statistical Computing Vienna, Austria. Available at https://www.R-project.org.
Ramboll Environ (2014). wrfcamx version 4.3 Release Notes. December 17, 2014 Ramboll
Environ International Corporation. Novato, CA. Available at: http://www.camx.com/
Simon, H, Baker, KR, Akhtar, F, Napelenok, SL, Possiel, N, Wells, B and Timin, B (2013). A
direct sensitivity approach to predict hourly ozone resulting from compliance with the
National Ambient Air Quality Standard. Environ Sci Technol 47(5): 2304-2313.
Skamarock, C, Klemp, J, Dudhia, J, Gill, D, Barker, D, Duda, M, Huang, X-Y, Wang, W and
Powers, G (2008). A Description of the Advanced Research WRF Version 3 (No.
NCAR/TN-475+STR). Mesoscale and Microscale Meteorology Division, National
Center for Atmospheric Research. Boulder, Colorado, . Available at:
https://opensky. ucar. edu/islandora/object/technotes:500.
Turner, R (2018). ddeldir: Delaunay Triangulation and Dirichlet (Voronoi) Tesselation. Version
0.1-15. Available at https://cran.R-project.org/package=deldir/.
U.S. EPA (2007). Guidance on the Use of Models and Other Analyses for Demonstrating
Attainment of Air Quality Goals for Ozone, PM2.5, and Regional Haze. U.S.
Environmental Protection Agency. Research Triangle Park, North Carolina. U.S. EPA.
EPA -454/B-07-002.
https://nepis. epa.go v/Exe/ZyPDF. cgi/Pl009OLl.PDF?Dockey=Pl009OLl.PDF.
U.S. EPA (2014). Health Risk and Exposure Assessment for Ozone. (Final Report). Office of Air
Quality Planning and Standards. Research Triangle Park, NC. U.S. EPA. EPA-452/R-14-
3C-154
-------
004a. August 2014. Available at:
https://nepis. epa.go v/Exe/ZyPURL. cgi ?Dockey=Pl 00KB UF. txt.
U.S. EPA (2017). Meteorological Model Performance for Annual 2016 Simulation WRF v3.8.
Office of Air Quality Planning and Standards. Research Triangle Park, NC. Available at:
https://www3. epa.gov/ttn/scram/guidance/met/MET_ TSD_2016.pdf.
U.S. EPA (2019). Techical Support Document: Preparation of Emissions Inventories for the
Version 7.1 2016 North American Emissions Modeling Platform. Office of Air Quality
Planning and Standards, U.S. Environmental Protection Agency. Research Triangle Park,
NC. Available at: https://www.epa.gov/sites/production/files/2019-
08/documents/2016v7. l_northamerican_emismod_tsd. pdf.
Yang, Y-J, Wilkinson, JG and Russell, AG (1997). Fast, Direct Sensitivity Analysis of
Multidimensional Photochemical Models. Environ Sci Technol 31(10): 2859-2868.
Yarwood, G, Whitten, GZ and Jung, J (2010). Development, Evaluation and Testing of Version
6 of the Carbon Bond Chemical Mechanism (CB6). Texas Commission on
Environmental Quality, ENVIRON International Corporation Austin, Texas
https://www.tceq.texas.gov/assets/public/implementation/air/am/contracts/reports/pm/58
20784005FY1026-20100922-environ-cb6.pdf.
Zhang, W, Capps, SL, Hu, Y, Nenes, A, Napelenok, SL and Russell, AG (2012). Development
of the high-order decoupled direct method in three dimensions for particulate matter:
enabling advanced sensitivity analysis in air quality models. Geosci Model Dev 5(2):
355-368.
3C-155
-------
APPENDIX 3D
EXPOSURE AND RISK ANALYSIS FOR THE OZONE NAAQS REVIEW
TABLE OF CONTENTS
3D. 1 INTRODUCTION 3D-9
3D. 1.1 Planning and Scientific/Public Review of the Current Analysis 3D-10
3D. 1.2 Overview 3D-11
3D.1.3 2014 Ozone Exposure and Risk Assessment 3D-12
3D. 1.4 Current Analysis 3D-14
3D.2 POPULATION EXPOSURE AND RISK APPROACH 3D-15
3D.2.1 Urban Study Areas 3D-16
3D.2.2 Simulated Populations 3D-20
3D.2.3 Ambient Air Concentrations 3D-34
3D.2.4 Meteorological Data 3D-48
3D.2.5 Construction of Human Activity Pattern Sequences 3D-50
3D.2.6 Microenvironmental Concentrations 3D-61
3D.2.7 Estimating Exposure 3D-72
3D.2.8 Estimating Risk 3D-72
3D.2.9 Assessing Variability/Co-Variability and Characterizing Uncertainty 3D-87
3D.3 POPULATION EXPOSURE AND RISK RESULTS 3D-93
3D.3.1 Characteristics of the Simulated Population and Study Areas 3D-94
3D.3.2 Exposures at or above Benchmark Concentrations 3D-96
3D.3.3 Lung Function Risk 3D-116
3D.3.4 Uncertainty Characterization 3D-144
3D.4 REFERENCES 3D-178
ATTACHMENTS
1. Estimating U.S. Census Tract-level Asthma Prevalence (2013-2017)
2. ICF Technical Memo: Identification of Simulated Individuals at Moderate Exertion
3. ICF Technical Memo: Updates to the Meteorology Data and Activity Locations within CHAD
4. Detailed Exposure and Risk Results
3D-1
-------
TABLE OF TABLES
Table 3D-1. Criteria used to identify and select urban study areas for inclusion in the O3
exposure and risk analyses 3D-19
Table 3D-2. General description of ambient air quality domains for the eight study
areas 3D-20
Table 3D-3. Descriptive statistics for children and adult asthma prevalence, using all census
tracts within eight consolidated statistical areas (CSAs) in the APEX asthma
prevalence file 3D-24
Table 3D-4. Regression parameters used to estimate RMR by sex and age groups 3D-29
Table 3D-5. List of states, counties, and O3 seasons that define the air quality and exposure
spatial and temporal modeling domain in each study area 3D-36
Table 3D-6. List of ambient air monitor IDs, range of O3 design values, and number of
monitors in each study area 3D-42
Table 3D-7. Range of the percent NOx emission changes needed to adjust air quality in the
eight study areas for the three air quality scenarios 3D-44
Table 3D-8. Study area meteorological stations, locations, and hours of missing data 3D-49
Table 3D-9. Overview of Studies Included in the APEX Activity Data Files 3D-51
Table 3D-10. Comparison of time spent outdoors and exertion level by asthma status for
children and adult diaries used by APEX 57
Table 3D-11. Number of diary days in CHAD for children and adults, grouped by temperature
and day-type categories 3D-58
Table 3D-12. Microenvironments modeled and calculation method used 3D-64
Table 3D-13. Air exchange rates (AER, hr1) for indoor residential microenvironments with A/C
by study area and temperature 3D-66
Table 3D-14. Air exchange rates (AER, hr1) for indoor residential microenvironments without
A/C by study area and temperature 3D-67
Table 3D-15. Individual air exchange rate data (hr1) obtained from three studies used to
develop an AER distribution used for schools in all study areas 3D-68
Table 3D-16. A/C prevalence from US Census American Housing Survey (AHS) data by study
area 3D-69
Table 3D-17. Parameter values for distributions of penetration and proximity factors used for
estimating in-vehicle ME concentrations 3D-71
Table 3D-18. VMT (2015-2017) derived conditional probabilities for interstate, urban, and local
roads used to select inside-vehicle proximity factor distributions in each study
area 3D-71
Table 3D-19. Responses reported in 6.6-hr controlled human exposure studies at a given
benchmark concentration 3D-75
3D-2
-------
Table 3D-20. Summary of controlled human exposure study data stratified by concentration
level and lung function decrements, corrected for individual response that
occurred while exercising in clean air, ages 18-35 3D-77
Table 3D-21. Estimated coefficients for the MSS lung function model 3D-85
Table 3D-22. Age term parameters for application of the MSS model to all ages 3D-87
Table 3D-23. Summary of how variability was incorporated into the exposure and risk analysis.
3D-89
Table 3D-24. Important components of co-variability in exposure modeling 3D-92
Table 3D-25. Summary of study area features and the simulated population 3D-96
Table 3D-26. Percent of people estimated to experience at least one exposure at or above
benchmarks while at moderate or greater exertion, for air quality adjusted to just
meet the current standard 3D-100
Table 3D-27. Number of people estimated to experience at least one exposure at or above
benchmarks while at moderate or greater exertion, for air quality adjusted to just
meet the current standard 3D-101
Table 3D-28. Percent of people estimated to experience at least two exposures at or above
benchmarks while at moderate or greater exertion, for air quality adjusted to just
meet the current standard 3D-102
Table 3D-29. Number of people estimated to experience at least two exposures at or above
benchmarks while at moderate or greater exertion, for air quality adjusted to just
meet the current standard 3D-103
Table 3D-30. Percent of people estimated to experience at least four exposures at or above
benchmarks while at moderate or greater exertion, for air quality adjusted to just
meet the current standard 3D-104
Table 3D-31. Number of people estimated to experience at least four exposures at or above
benchmarks while at moderate or greater exertion, for air quality adjusted to just
meet the current standard 3D-105
Table 3D-32. Percent of people estimated to experience at least one exposure at or above
benchmarks while at moderate or greater exertion, for the 75 ppb air quality
scenario 3D-107
Table 3D-33. Percent of people estimated to experience at least two exposures at or above
benchmarks while at moderate or greater exertion, for the 75 ppb air quality
scenario 3D-108
Table 3D-34. Percent of people estimated to experience at least four exposures at or above
benchmarks while at moderate or greater exertion, for the 75 ppb air quality
scenario 3D-109
Table 3D-35. Percent of people estimated to experience at least one exposure at or above
benchmarks while at moderate or greater exertion, for the 65 ppb air quality
scenario 3D-111
3D-3
-------
Table 3D-36. Percent of people estimated to experience at least two exposures at or above
benchmarks while at moderate or greater exertion, for the 65 ppb air quality
scenario 3D-112
Table 3D-37. Percent of people estimated to experience at least four exposures at or above
benchmarks while at moderate or greater exertion, for the 65 ppb air quality
scenario 3D-113
Table 3D-38. Comparison of current assessment to 2014 HREA for percent of children
estimated to experience at least one exposure at or above benchmarks while at
moderate or greater exertion 3D-115
Table 3D-39. Comparison of current assessment to 2014 HREA for percent of children
estimated to experience at least two exposure at or above benchmarks while at
moderate or greater exertion 3D-116
Table 3D-40. Percent of people estimated to experience at least one lung function decrement at
or above the indicated level, for air quality adjusted to just meet the current
standard, using the population-based (E-R function) risk approach 3D-119
Table 3D-41. Number of people estimated to experience at least one lung function decrement at
or above the indicated level, for air quality adjusted to just meet the current
standard, using the population-based (E-R function) risk approach 3D-120
Table 3D-42. Percent of people estimated to experience at least two lung function decrements at
or above the indicated level, for air quality adjusted to just meet the current
standard, using the population-based (E-R function) risk approach 3D-121
Table 3D-43. Number of people estimated to experience at least two lung function decrements
at or above the indicated level, for air quality adjusted to just meet the current
standard, using the population-based (E-R function) risk approach 3D-122
Table 3D-44. Percent of people estimated to experience at least four lung function decrements
at or above the indicated level, for air quality adjusted to just meet the current
standard, using the population-based (E-R function) risk approach 3D-123
Table 3D-45. Number of people estimated to experience at least four lung function decrements
at or above the indicated level, for air quality adjusted to just meet the current
standard, using the population-based (E-R function) risk approach 3D-124
Table 3D-46. Percent of people estimated to experience at least one lung function decrement at
or above the indicated level, for the 75 ppb air quality scenario, using the
population-based (E-R function) risk approach 3D-125
Table 3D-47. Percent of people estimated to experience at least two lung function decrements at
or above the indicated level, for the 75 ppb air quality scenario, using the
population-based (E-R function) risk approach 3D-126
Table 3D-48. Percent of people estimated to experience at least four lung function decrements
at or above the indicated level, for the 75 ppb air quality scenario, using the
population-based (E-R function) risk approach 3D-127
3D-4
-------
Table 3D-49. Percent of people estimated to experience at least one lung function decrement at
or above the indicated level, for the 65 ppb air quality scenario, using the
population-based (E-R function) risk approach 3D-128
Table 3D-50. Percent of people estimated to experience at least two lung function decrements at
or above the indicated level, for the 65 ppb air quality scenario, using the
population-based (E-R function) risk approach 3D-129
Table 3D-51. Percent of people estimated to experience at least four lung function decrements
at or above the indicated level, for the 65 ppb air quality scenario, using the
population-based (E-R function) risk approach 3D-130
Table 3D-52. Percent of people estimated to experience at least one lung function decrement at
or above the indicated level, for air quality adjusted to just meet the current
standard, using the individual-based (MSS model) risk approach 3D-132
Table 3D-53. Number of people estimated to experience at least one lung function decrement at
or above the indicated level, for air quality adjusted to just meet the current
standard, using the individual-based (MSS model) risk approach 3D-133
Table 3D-54. Percent of people estimated to experience at least two lung function decrements at
or above the indicated level, for air quality adjusted to just meet the current
standard, using the individual-based (MSS model) risk approach 3D-134
Table 3D-55. Number of people estimated to experience at least two lung function decrements
at or above the indicated level, for air quality adjusted to just meet the current
standard, using the individual-based (MSS model) risk approach 3D-135
Table 3D-56. Percent of people estimated to experience at least four lung function decrements
at or above the indicated level, for air quality adjusted to just meet the current
standard, using the individual-based (MSS model) risk approach 3D-136
Table 3D-57. Number of people estimated to experience at least four lung function decrements
at or above the indicated level, for air quality adjusted to just meet the current
standard, using the individual-based (MSS model) risk approach 3D-137
Table 3D-58. Percent of people estimated to experience at least one lung function decrement at
or above the indicated level, for the 75 ppb air quality scenario, using the
individual-based (MSS model) risk approach 3D-138
Table 3D-59. Percent of people estimated to experience at least two lung function decrements at
or above the indicated level, for the 75 ppb air quality scenario, using the
individual-based (MSS model) risk approach 3D-139
Table 3D-60. Percent of people estimated to experience at least four lung function decrements
at or above the indicated level, for the 75 ppb air quality scenario, using the
individual-based (MSS model) risk approach 3D-140
Table 3D-61. Percent of people estimated to experience at least one lung function decrement at
or above the indicated level, for the 65 ppb air quality scenario, using the
individual-based (MSS model) risk approach 3D-141
3D-5
-------
Table 3D-62. Percent of people estimated to experience at least two lung function decrements at
or above the indicated level, for the 65 ppb air quality scenario, using the
individual-based (MSS model) risk approach 3D-142
Table 3D-63. Percent of people estimated to experience at least four lung function decrements
at or above the indicated level, for the 65 ppb air quality scenario, using the
individual-based (MSS model) risk approach 3D-143
Table 3D-64. Characterization of key uncertainties in exposure and risk analyses using
APEX 3D-146
Table 3D-65. Percent of children estimated to experience at least one lung function decrement
at or above the indicated level, for air quality adjusted to just meet the current
standard, using the population-based (E-R function) risk approach 3D-163
Table 3D-66. Estimated lung function risk contribution resulting from selected 7-hr average O3
exposures in children, using the E-R function risk approach, 2016 3D-166
Table 3D-67. MSS model risk estimates from varying the number of simulated children. 3D-168
Table 3D-68. Estimated lung function risk contribution resulting from selected 7-hr average O3
exposures in children, using the MSS model risk approach, 2016 3D-169
Table 3D-69. Percent of children experiencing one or more FEVi decrements >10, 15, 20%,
2016 air quality adjusted to just meet the current standard, considering influence
of moderate or greater exertion level in the MSS model and E-R function risk
approaches 3D-174
Table 3D-70. Percent of children experiencing one or more FEVi decrements >10, 15, 20%,
2016 air quality adjusted to just meet the current standard, considering the setting
of variability parameter, vi, in the MSS model 3D-178
3D-6
-------
TABLE OF FIGURES
Figure 3D-1. Locations of the eight study areas selected for the current O3 exposure and risk
analysis 3D-18
Figure 3D-2. County boundaries, census tract population densities, and meteorological stations
in the Atlanta (top) and Boston (bottom) study areas 3D-37
Figure 3D-3. County boundaries, census tract population densities, and meteorological stations
in the Dallas (top) and Detroit (bottom) study areas 3D-38
Figure 3D-4. County boundaries, census tract population densities, and meteorological stations
in the Philadelphia (top) and Phoenix (bottom) study areas 3D-39
Figure 3D-5. County boundaries, census tract population densities, and meteorological stations
in the Sacramento (top) and St. Louis (bottom) study areas 3D-40
Figure 3D-6. Hourly O3 distributions by hour-of-day (left panel) and month (right panel) at
ambient air monitoring sites in Philadelphia for observed air quality (black), air
quality adjusted to meet the current standard (70 ppb, blue) and two other design
values (75 ppb, red; and 65 ppb, green). From PA, Appendix 3C, Figures 3C-71
and 3C-79, respectively 3D-45
Figure 3D-7. Histograms of hourly O3 concentrations (ppb, x-axis) for the air quality scenario
just meeting the current O3 standard in the eight study areas. The x-axis midpoint
concentrations range from 0 to 70 ppb, in 2 ppb increments (rightmost, maximum
histogram bar for all study areas represents the frequency of all hourly
concentrations >70 ppb) 3D-47
Figure 3D-8. Calculated design values for census tracts in the Philadelphia study area, derived
from a VNA interpolation of CAMx/HDDM adjusted O3 concentrations. Figure
modified from PA, Appendix 3C, Figure 3C-99 3D-48
Figure 3D-9. Percent of children (5-18 years) and adults (19-90 years) having afternoon time
outdoors while at moderate or greater exertion, categorized by daily maximum
temperature (°F) and time (hours/day) groups 3D-58
Figure 3D-10. Illustration of the mass balance model used by APEX to estimate concentrations
within indoor microenvironments 3D-62
Figure 3D-11. Controlled human exposure data for FEVi responses in individual study
subjects 3D-78
Figure 3D-12. Median value of Bayesian fit population-based E-R function data (left panel) and
illustrative curves (right panel) for FEVi decrements >10% (top panel), >15
(middle panel), >20% (bottom panel). Drawn from the 2014 HREA, Table 6A-1
with processing and model development described by Abt (2013) 3D-81
Figure 3D-13. Conceptual representation of the two-compartment model used by the MSS
model. C is exposure concentration, Vis ventilation rate, t is time, X\s an
intermediate quantity, a is a decay constant. Adapted from Figure 1 in McDonnell
etal. (1999) 3D-83
Figure 3D-14. Comparison of a probit function curve (blue line) with the Bayesian logistic/linear
function curve (red) in estimating the probability of lung function decrements
3D-7
-------
>15% (based on data in Table 3D-20). Confidence intervals for the probit model
reflect variability in the regression model coefficients 3D-161
Figure 3D-15.
Figure 3D-16.
Figure 3D-17.
Figure 3D-18.
3D-8
Estimated lung function risk contribution resulting from selected 7-hr average O3
exposures in children, using the E-R function risk approach and air quality
adjusted to just meet the current standard, for one decrement (top panel) and two
decrements (bottom panel), 2016 3D-167
Lung function risk contribution resulting from selected 7-hr average O3 exposures
in children, using the MSS model risk approach and air quality adjusted to just
meet the current standard, for one decrement (top panel) and two decrements
(bottom panel), 2016 3D-170
Example time-series of O3 exposures, EVR, and FEVi reductions estimated using
MSS model for a simulated child in the Atlanta study area, based on a day in a
year (2016) of the current standard air quality scenario 3D-172
Time-series of O3 exposures, EVR, and FEVi reductions of 10% (left panel), 15%
(middle panel), and 20% (right panel) estimated using MSS model for two
simulated children (interpersonal variability parameter U= 0.963, top panel; II =
1.78, bottom panel) in the Atlanta study area on three days in a year (2016) of the
current air quality scenario 3D-177
-------
3D.1 INTRODUCTION
This appendix to the O3 Policy Assessment (PA) summarizes the quantitative exposure
and risk analysis performed for the current O3 NAAQS review. The analysis builds upon the
methodology and lessons learned from the human exposure and risk analyses conducted in the
prior O3 review (2014 HREA; U.S. EPA, 2014), analysis plans outlined in the Integrated Review
Plan (IRP; U.S. EPA, 2019d), and information provided in the 2020 O3 Integrated Science
Assessment (ISA; U.S. EPA, 2020), which builds on the 2013 ISA (U.S. EPA, 2013).
For the current O3 NAAQS review, exposures and risks were modeled for people residing
in eight U.S. urban study areas,1 considering three hypothetical air quality scenarios developed
from ambient air O3 monitoring data adjusted based on a photochemical model-based approach
for a single 3-year period (2015 to 2017), and based on health effects observed in controlled
human exposure studies. The three air quality scenarios were for O3 concentrations across the
study area such that the location with the highest design value2 just meets: (1) the current
standard (i.e., a design value of 70 ppb), (2) a design value of 75 ppb, and (3) a design value of
65 ppb. The exposures and risks were estimated for (1) all school-age children (ages 5-18), (2)
school-age children with asthma (ages 5-18), (3) all adults (ages 19-90),3 and (4) adults with
asthma (ages 19-90),4 each while at moderate or greater exertion level at the time of exposure.
The strong emphasis on children and people with asthma reflects the conclusion based on the
currently available evidence that these are important at-risk groups, as summarized in section
3.3.2 of this PA and described in the ISA (ISA, section IS.6.1).
Health risk is characterized in two ways in these analyses, producing two types of risk
metrics: one involving comparison of population exposures, while at elevated exertion, to
benchmark concentrations, and the second involving estimated population occurrences of
1 For the 2014 HREA, controlled human exposure-based health risk was estimated in 15 urban study areas
considering five air quality scenarios and two 3-year periods (2006-2008 and 2008-2010). In addition, an
epidemiologic-based health risk approach was applied in 12 urban study areas also considering the same five air
quality scenarios and for two single-year periods (2007 and 2009). Further, an epidemiologic-based health risk
approach was applied to the continental U.S. considering a single air quality scenario (unadjusted, as is ambient
air concentrations).
2 The design value for these scenarios is the 3-year average of the annual 4th highest daily maximum 8-hr average O3
concentration. For example, a monitoring site meets the current standard if the design value, derived from the data
for that site, is less than or equal to 70 ppb.
3 For the 2014 HREA, older adults (ages 65-95) were simulated as a separate group. In the current assessment, older
adults within this age group are included in the simulation of all adults. Additionally, the upper age limit in the
current assessment is 90 years given data limitations since recognized in CHAD for older age entries.
4 For the 2014 HREA, adults with asthma (ages 19-95) were simulated, similar to the group simulated for the current
assessment. Additionally, the upper age limit in the current assessment is 90 years given data limitations since
recognized in CHAD for older age entries.
3D-9
-------
ambient air Cte-related lung function decrements (PA, Figure 3-3). The first risk metric is based
on comparison of estimated daily maximum 7-hour (7-hr) average exposures for individuals
breathing at elevated rates to concentrations of potential concern (benchmark concentrations),5
and the second uses exposure-response (E-R) information for study subjects experiencing FEVi
decrements (specifically Cb-related decrement of 10% or more) 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%.
A description of the exposure and risk modeling performed, including a summary of (1)
the ways in which scientific and public review of the current analysis occurred, and (2) the 2014
HREA and important updates in modeling tools and approaches that contributed to planning and
completion of the analyses presented in this document is provided in sections 3D. 1.1 through
3D. 1.4. The detailed description of the modeling tools, algorithms, input data and output metrics,
along with an assessment of how variability is addressed in the analysis is provided in section
3D.2. Finally, the exposure and risk results, including a characterization of uncertainties, are
found in section 3D.3.
3D. 1.1 Planning and Scientific/Public Review of the Current Analysis
As described in section 1.4 of the PA, a consultation with the Clean Air Scientific
Advisory Committee (CASAC) was held in November 2018 on the draft IRP to receive their
input and comments from the public were also solicited on the draft IRP. Both comments from
the CASAC and the public were considered in shaping the analysis plans, which were
summarized in the final IRP.
The draft PA, with a draft version of this appendix was provided to the CASAC for its
review and to the public for public comment, as summarized in section 1.4 of this (final) PA. In
consideration of comments from the CASAC (Cox, 2020) and the public a number of additional
analyses and presentations have been added. In consideration of CASAC recommendations and
public comments, this document includes presentations reflecting further analyses, investigations
and/or clarifications of the available data with regard to a number of areas.
• 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 (section 3D.2.5.3);
• Estimates for the comparison-to-benchmarks analysis additionally summarized in light
of the estimates from the last review (section 3D.3.2.4);
5 The exposure duration and approach for identifying simulated individuals at moderate or greater exertion have
been updated from what was used in the 2014 HREA to more closely match the circumstances of the controlled
human exposure studies, as described in section 3D.2.2.3.3 and 3D.2.8.1.
3D-10
-------
• 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 (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 (section 3D.3.4.1);
• Evaluation of uncertainty with the E-R function and risk estimates (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 (section 3D.3.4.1).
3D. 1.2 Overview
Estimates of human exposure to O3 can provide meaningful answers to policy-relevant
questions regarding exposures of concern and resulting risk estimates. This is particularly true
when the important elements of O3 exposure, i.e., the frequency, magnitude, duration, and
pattern, are accounted for and when the exposures are estimated using policy-relevant ambient
air quality scenarios, i.e., ambient air conditions that either just meet the current O3 standard or
other air quality scenarios. Further, the policy-relevance of these estimated O3 exposures can be
extended when they are linked with adverse health outcome data obtained from controlled
human exposure studies to quantitatively estimate health risk. As a result, via the quantitative
relationships that exist between ambient air concentrations, exposures, and health effects, one
can estimate the impact varying air quality conditions have on public health.
Exposure to O3 can be directly estimated by monitoring the concentration of O3 in a
person's breathing zone (close to the nose/mouth) using a personal exposure monitor. Studies
employing this measurement approach have been reviewed in the current and 2013 O3 ISAs and
in past O3 Air Quality Criteria Documents (AQCDs; U.S. EPA, 1986, 1996, U.S. EPA, 2006).
Personal exposure measurements from these studies can be useful in describing a general range
of exposure concentrations (among other reported measurement data) and in identifying factors
that may influence varying exposure levels. However, these measurement studies of personal
exposure to O3 are largely limited by the disparity between measurement sample durations and
durations of interest, and in appropriately capturing variability in population exposure occurring
over large geographic areas, particularly when considering both O3 concentrations in ambient air
(e.g., spatial variability) and population (e.g., age, sex) attributes that greatly influence exposure.
Because of these limitations in personal exposure measurement data, more commonly
human exposure is estimated using sophisticated models that better account for physical (e.g.,
meteorology) or personal (e.g., age) attributes that may strongly influence variability in
exposures. These exposure models can combine information on ambient air O3 concentrations in
various microenvironments, e.g., near roads, in schools, etc., with information on activity
3D-11
-------
patterns for individuals sampled from the general population or specific subpopulations, e.g.,
children with asthma. When integrating these varied data (among many others such as population
demographics and disease prevalence) and understanding the key factors affecting exposure,
exposure models can be more informative than the limited information given by measurement
data alone.
Ozone exposure is highly dependent on the ambient air concentrations in an urban area,
which vary spatially and temporally. An exposure model can reasonably estimate exposures for
any perceivable at-risk population (e.g., people with asthma living in a large urban area) and
considering any number of defined hypothetical air quality conditions (e.g., those in which
concentrations just meet a particular air quality standard) provided underlying data exist to
generate such estimates. Further, exposure models that account for variability in human
physiology can also realistically estimate pollutant intake dose by using activity-specific
ventilation rates. Each of these important features of O3 exposure cannot realistically be
measured for a study group or population of interest over wide ranging temporal and spatial
scales, particularly when considering time, cost, and other constraints, and serve as the
justification for using a modeling approach to estimate exposure and health risks.
3D.1.3 2014 Ozone Exposure and Risk Assessment
The 2014 HREA included two types of risk analyses. The first type of risk analysis,
exposure-based risk, used health effect information obtained from controlled human exposure
studies (summarized in the IRP, section 5.1.1.1). The second type, epidemiologic-based risk,
used concentration-response functions derived from epidemiologic studies (IRP, section 5.1.1.2).
Because we used only the exposure-based risk analysis approach for this review (see section
3D. 1.4 below; IRP, section 5.1.2), it is only these results that are succinctly summarized in this
section.6'7
6 Details regarding all of the risk analyses performed for the prior review can be found in chapters 5 (exposure-based
health benchmark risk), 6 (exposure-based lung function risk), and 7 (epidemiologic-based risk) of the 2014
HREA.
7 We note that the CASAC comments on the draft PA included several related to development of risk estimates from
epidemiological study results (Cox, 2020). Because an epidemiologic-based risk analysis was not performed for
this review, the issues raised by those comments are not considered here.
3D-12
-------
For the 2014 HREA, two exposure-based risk analyses8 were performed in a set of 15
urban study areas9 and for five different air quality scenarios: unadjusted ambient air O3
conditions, air quality adjusted to just meet the then-existing standard (75 ppb, annual 4th highest
daily maximum 8-hr average concentration, averaged over a 3-year period), and air quality
adjusted to just meet potential alternative O3 standards having the same form and averaging
times, with levels of 70, 65 and 60 ppb.10 The scenarios were based on air quality from two 3-
year periods: 2006-2008 and 2008-2010. The first exposure-based risk analysis involved
comparison of population exposures, while at elevated exertion, to benchmark concentrations.
The exposure-to-benchmark comparison characterizes the extent to which individuals in at-risk
populations could experience exposures of concern (i.e., average exposure concentrations at or
above specific benchmarks while at moderate or greater exertion levels) while engaging in their
daily activities in study areas with air quality adjusted to just meet the then-existing standard and
other O3 air quality conditions. Results were characterized using three benchmark concentrations
(60, 70, and 80 ppb O3), exposures to which in controlled human exposure studies yielded
different occurrences and severity of respiratory effects in the human subjects (2014 HREA,
section 5.2.8). The second exposure-based risk analysis involves estimated population
occurrences of ambient air 03-related lung function decrements. The lung function risk analysis
provides estimates of the extent to which populations in such areas could experience decrements
in lung function. Based on the range of health effects considered clinically relevant and the
potential for varied responses in healthy individuals versus people with asthma, the lung function
risk analysis reported estimates for risk of lung function decrement at or above three different
magnitudes, i.e., forced expiratory volume in one second (FEVi) reductions of at least 10%,
15%, and 20% (2014 HREA, section 6.2.1).
Key observations and insights from the O3 exposure-to-benchmark comparison and lung
function risks, in addition to important caveats and limitations, were addressed in Section II.B of
the Final Rule notice (80 FR 65312 to 65315, October 26, 2015). The exposure-based analyses in
8 For the primary analysis results in the 2014 HREA, population exposures were used to estimate health benchmark
and lung function risks using an individual-based approach. In addition, a population-based E-R function
approach was used to estimate lung function risk but done mainly for comparison with the individual-based
approach and with prior review assessment results.
9 The 15 urban study areas assessed were Atlanta, Baltimore, Boston, Chicago, Cleveland, Dallas, Denver, Detroit,
Houston, Los Angeles, New York, Philadelphia, Sacramento, St. Louis, and Washington, DC.
10 These scenarios reflect air quality with design values that equal the level of the now-current standard and two
others having levels just above and below the current standard. The air quality data were generated using a
combined ambient monitor data and modeling approach similar to that used for the current assessment. These
simulations were intended to be illustrative and do not reflect any consideration of specific control programs
designed to meet the specified standards. Further, these simulations were not intended to represent predictions of
when, whether, or how areas might meet a specified standard.
3D-13
-------
the 2014 HREA, and most particularly the exposure to benchmarks analysis were important
considerations in the 2015 decision on revisions to the primary O3 standard (80 FR 65362-65365,
October 26, 2015).
3D.1.4 Current Analysis
As described in the IRP (section 5.1.2.2), the quantitative analyses for this review focus
on the comparison to benchmark exposure-based risk analysis approach, based on the controlled
human exposure studies. In part, this is because substantial updates to data, information, models,
and tools are available, ensuring that the new exposure and risk estimates are both improved and
appropriately targeted. Additionally, estimates from the exposure-based analyses, particularly the
comparison of daily maximum exposures to benchmark concentrations, were most informative to
the Administrator's decision in the last review (IRP, section 3.1.2). This largely reflected the
EPA conclusion that "controlled human exposure studies provide the most certain evidence
indicating the occurrence of health effects in humans following specific O3 exposures," and
recognition that "effects reported in controlled human exposure studies are due solely to O3
exposures, and interpretation of study results is not complicated by the presence of co-occurring
pollutants or pollutant mixtures (as is the case in epidemiologic studies)" (80 FR 65343, October
26, 2015). In the last review, the Administrator placed relatively less weight on the air quality
epidemiologic-based risk estimates, in recognition of an array of uncertainties, including, for
example, those related to exposure measurement error (80 FR 65346, October 26, 2015).
3D.1.4.1 Aspects updated since 2014
A number of aspects of the exposure-based risk analyses were updated since the 2014
HREA. The updates were based on important uncertainties characterized in the last review and
having newly available data, information, models, and tools that could provide risk estimates in
which we have greater confidence that was the case for the risks estimated in the last review, as
summarized in Appendix 5A of the IRP. These updates include:
• Air quality
- More recent (2015-2017) ambient air monitoring data from US EPA's Air Quality
System (AQS) having unadjusted concentrations at or near the current standard
(section 3D.2.3.2);
- Updated photochemical model (CAMx version 6.5)11 to adjust ambient air
concentrations to just meet the air quality scenarios to be assessed (section
3D.2.3.3).
• Exposure and risk model
11 CAMx is the Comprehensive Air Quality Model with Extensions. This model is briefly described in Appendix 3C.
Additional information and model download can be found at http://www.camx.com/.
3D-14
-------
- More recent (2010) U.S. Census demographics and commuting data (section
3D.2.2.1);
- More recent (2013-2017) asthma prevalence for census tracts in all study areas
(section 3D.2.2.2);
- Updated equations to estimate resting metabolic rate (RMR) (section 3D.2.2.3.2)
and associated ventilation rate (Ve) (section 3D.2.2.3.3);
- Improved matching of controlled human exposure study duration (6.6-hr) and
target ventilation rate to that estimated for simulated individuals (7-hr duration,
distribution accounting for resting ventilation) and used for benchmark
comparisons and population-based E-R lung function risk (section 3D.2.2.3.3 and
3D.2.8.1);
- More recent (2015-2017) meteorological data to reflect the assessment years
(section 3D.2.4)
- Increased number of diary-days and added new activity descriptions to activity
pattern data base (section 3D.2.5.1);
- Most recent MSS-FEVi model (McDonnell et al., 2013) to estimate individual
lung function risk (section 3D.2.8.2.2);
- New evaluations of important uncertainties (section 3D.3.4.2): form of E-R
function, E-R function risk confidence intervals, low exposure concentration
contribution to lung function risk, influence of ventilation rate on lung function
risk, influence of variability parameter settings in MSS-FEVi model.
3D.2 POPULATION EXPOSURE AND RISK APPROACH
This section describes the data, information, models, and tools used to characterize
exposure and health risk associated with O3 in ambient air for three air quality scenarios. As
summarized above in section 3D. 1.4, the overall analysis approach is based on linking the health
effects information observed in controlled human exposure studies to estimated population-based
exposures that reflect our current understanding of concentrations of O3 in the ambient air.
Population exposures and risks were estimated using the EPA's Air Pollution Exposure
Model (APEX), version 5. APEX is a multipollutant, population-based, stochastic,
microenvironmental model that can be used to estimate human exposure via inhalation for
criteria and toxic air pollutants. APEX is designed to estimate human exposure to these
pollutants at the local, urban, and consolidated metropolitan level. In this analysis, we used
APEX to estimate exposure and risk in eight study areas, the details of which are provided in the
following subsections. Additional information not provided here regarding all of APEX modules,
algorithms, and modeling options can be found in the APEX User's Guide (U.S. EPA, 2019a;
U.S. EPA, 2019b).
Briefly, APEX calculates the exposure time-series for a user-specified duration and
number of individuals. Collectively and by design, these simulated individuals are intended to be
3D-15
-------
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 the exposure
and risk analyses performed here, the core demographic geographical units for estimating
exposure are census tracts. For each simulated person, the following general steps are performed:
• Select personal attribute variables and choose values to characterize the simulated
individual (e.g., age, sex, body weight, disease status);
• 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);
• Calculate the pollutant concentrations in the microenvironments (MEs) that simulated
individuals visit;
• Calculate the simulated individual's exposure, and simultaneously, their breathing
rate for each exposure event and summarize for the selected exposure metric.
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. As described
above regarding air quality and in the sections that follow describing APEX model inputs and
approaches to estimating exposure, 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. The population distributions of exposures are then combined
with the health effects information to characterize associated risk via two types of metrics: a
comparison to benchmark concentrations and lung function risk. The details of the model input
data and general approaches used for estimating exposure and risk are described in the sections
that follow.
3D.2.1 Urban Study Areas
To identify a list of urban areas for the current analysis, we first considered the list of 15
urban study areas evaluated in the 2014 HREA, which represented a range of geographic areas,
encompassing variability in air quality, climate, and population demographics. We also
considered other candidate study areas (e.g., Phoenix). As was done for the 2014 HREA, we
developed criteria to select urban study areas for the current exposure and risk analysis. Those
criteria are as follows:
3D-16
-------
• Have at least 10 ambient air monitors having complete year data for the 2015-2017
period;
• Combined statistical area (CSA)/metropolitan statistical area (MSA) ambient air
monitor design values are between 60-80 ppb, thus having minimal adjustment
needed to just meet the current 8-hr O3 NAAQS;
• CSA/MSA population between 2 to 10 million;
• Anticipated reasonable air quality model performance12; and
• Reasonable geographic distribution across continental U.S.
Based on these selection criteria, we chose the eight study areas listed in Table 3D-1 (and
shown in Figure 3D-1) to develop our population exposure estimates. Included also are the nine
other study areas considered but not selected for the current exposure and risk analysis. We
recognize the Sacramento study area does not meet the design value criterion (i.e., 86 ppb is
outside the range of values considered), however we relaxed this criterion to include a study area
in the Pacific/West region of the U.S and because exposure and risk was evaluated in the 2014
HREA (as opposed to using Los Angeles which was also evaluated in the 2014 HREA but has a
2015 -17 design value of 112 ppb).
We broadly defined the study areas using geographic coordinates to center the overall
exposure modeling domain for the APEX modeling (Table 3D-2). A wide city radius (i.e., 30
km) along with standard political/statistical county aggregations (e.g., whether in a CSA/MSA)
were then used to identify the specific counties that comprise each study area. As a result, 131
counties containing 9,725 census tracts were used to define the air quality domain in the eight
study areas.13 As done for prior exposure-based assessments, ambient air O3 concentrations were
estimated to census tracts to capture spatial heterogeneity that may exist within each study area
(PA, Appendix 3C) and to link with the population input data sets (section 3D.2.2).
12 While we expect air quality models to effectively capture relationships between ozone and its chemical precursors
in most areas, there are known situations (e.g. documented influence of stratospheric ozone intrusions) that may
be more challenging for air quality models to represent. We therefore excluded some of these more challenging
areas from this analysis (see Table 3D-1).
13 The identification of specific counties and census tracts are provided in the APEX ambient air concentration input
files for each study area. The approach used to estimate O3 concentrations is summarized in section 3D.2.3 below
and is described fully in the Appendix 3C of this PA.
3D-17
-------
Detroit*
Boston*
{*#}
~Sacramento
St. Louis*
Philadelphia
Phoenix
Dallas
Atlanta
Figure 3D-1. Locations of the eight study areas selected for the current O3 exposure and
risk analysis.
3D-18
-------
Table 3D-1. Criteria used to identify and select urban study areas for inclusion in the O3
exposure and risk analyses.
Selected
for
Analysis?
Study Area
Census
Division A
U.S. Climate
RegionB
CSA/MSA
Population
c (millions)
CSA/MSA
Land Area
D (Km2)
Ambient
Air
Monitors
(n)
Design ValuesE
(ppb)
2017
2008, 2010
Yes
Atlanta
South Atlantic
Southeast
6.6
26,873
11
75
95, 80
Boston
New England
Northeast
8.3
22,780
22
73
82, 76
Dallas
West S Central
South
8.0
36,411
20
79
91,86
Detroit
East N Central
Upper Midwest
5.4
14,972
11
73
82, 75
Philadelphia
Mid Atlantic
Northeast
7.2
15,391
19
80
92, 83
Phoenix
Mountain
Southwest
4.9
37,725
28
76
81, 77
Sacramento
Pacific
West
2.6
20,709
18
86
99, 99
St. Louis
West N Central
Ohio Valley
2.9
23,504
12
72
82, 77
No
Baltimore
South Atlantic
Northeast
2.8
6,738
5
75
91,89
ChicagoF
East N Central
Ohio Valley
9.9
21,941
21
78
78, 74
Cleveland
East N Central
Ohio Valley
3.5
9,322
15
74
84, 77
DenverF
Mountain
Southwest
3.6
33,824
10
79
86, 78
Houston
West S Central
South
7.2
27,744
19
81
91,84
Los AngelesF
Pacific
West
18.8
87,943
41
112
119,112
New YorkF
Mid Atlantic
Northeast
23.5
30,544
36
83
89, 82
Salt Lake CityF
Mountain
Southwest
2.6
46,517
10
78
82, 74
Washington DC
South Atlantic
Southeast
6.2
14,341
15
71
87, 81
A U.S Census Division data are found at: https://www.ncdc.noaa.gov/monitoring-references/maps/us-census-divisions.php.
B U.S. Climate Region data are found at: https://www.ncdc.noaa.gov/monitoring-references/maps/us-ciimate-regions.php.
c U.S. Census CSA/MSA population data are found at: https://www.census.gov/data/tabies/time-series/demo/popest/2010s-totai-metro-and-micro-
statisticai-areas.html.
D U.S. Census land area data taken from "G001 Geographic Identifiers, 2010 SF1 100% data file" available at:
https://factfinder.census.gov/faces/nav/jsf/pages/searchresuits.xhtmi?refresh=t.
E Ozone ambient air monitor design values (see .xlsx sheet 'Table6. Monitor Trends') are found at: https://www.epa.gov/air-trends/air-quaiity-design-
vaiues.
F Potential air quality modeling/adjustment issues: VOC-limited (Chicago, Denver), stratospheric O3 issues (Denver), low monitor density (Salt Lake
City), monitor issues (New York), and high DVs (Los Angeles).
3D-19
-------
Table 3D-2. General description of ambient air quality domains for the eight study areas.
CSA/MSA
Coordinates
Counties A
(n)
Tracts
(n)
Longitude
(degrees)
Latitude
(degrees)
Name
ID#
Abbrev.
Atlanta-Athens-Clarke County-
Sandy Springs, GA-AL
122
ATL
-84.3880
33.7490
39
1,077
Boston-Worcester-Providence,
MA-RI-NH-CT
148
BOS
-71.0589
42.3601
19
1,753
Dallas-Fort Worth, TX-OK
206
DAL
-96.7970
32.7767
21
1,422
Detroit-Warren-Ann Arbor, Ml
220
DET
-83.0458
42.3314
10
1,583
Philadelphia-Reading-Camden,
PA-NJ-DE-MD
428
PHI
-75.1652
39.9526
16
1,725
Phoenix-Mesa, AZ
429
PHX
-112.0740
33.4484
2
988
Sacramento-Roseville, CA
472
SAC
-121.4944
38.5816
7
539
St. Louis-St. Charles-Farmington,
MO-IL
476
STL
-90.2003
38.6303
17
638
A Delineations promulgated by the Office of Management and Budget (0MB) in February of 2013 (PA, Appendix 3C, section 3C.2).
3D.2.2 Simulated Populations
APEX 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). Each simulated person is represented by a personal
profile. The personal profile includes specific attributes such as an age, a home tract, a work tract
(or is not employed), housing characteristics, physiological parameters, and so on. The profile
does not correspond to any particular individual that resides in the study area, but rather
represents a simulated person. Accordingly, 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.
APEX generates population-based exposures using several population databases. Based
on the geographic boundaries defining the study areas and the study groups of interest, APEX
simulates representative individuals using appropriate geographic, demographic, and health
status information provided by existing population-based surveys. For the current exposure and
risk analysis, population input data sets are organized by U.S. census tracts.
3D-20
-------
Several updates were made to the APEX model inputs and algorithms for use in
simulating the populations of interest in this exposure and risk analysis and are described in the
following sections: population demographic data that are based on the 2010 census (section
3D.2.2.1), asthma prevalence rates based on the 2013-2017 National Health Interview Survey
(NHIS) that vary by age, sex and geographic location (section 3D.2.2.2), and data and equations
used to approximate personal attributes such as body weight, resting metabolic rate, and
breathing rate (section 3D.2.2.3).
3D.2.2.1 Demographics
As briefly described in section 3D.2.1 (and more fully in section 3D.2.3 below and the
PA, Appendix 3C), ambient air concentrations were modeled to census tracts in each study area
to capture spatial heterogeneity in ambient air O3 concentrations. Population data were generated
using the same spatial scale to also account for variability in population demographics. Tract-
level population counts were obtained from the 2010 Census of Population and Housing
Summary File l.14 Summary File 1 contains what the Census program calls "the 100-percent
data," which is the compiled information from the questions asked of all (100% of) people and
housing units in the U.S. Three national-based APEX input files15 are used for the current
exposure and risk analysis as follows.
• Population sectors US 2010.txt. census tract identifiers (IDs), latitudes and
longitudes in degrees.
• Population female All 2010.1x1: census tract IDs, tract-level population counts for
females, stratified by 23 age groups.16
• Population male All 2010.txt: census tract IDs, tract-level population counts for
males, stratified by the same 23 age groups as done for females.
3D.2.2.2 Asthma Prevalence
The four population study groups included in this exposure assessment are adults (19 to
90 years old),17 children (5 to 18 years old),18 and those within each of the two groups having
14 Technical documentation - 2010 Census Summary File 1—Technical Documentation/prepared by the U.S. Census
Bureau, Revised 2012 - available at: http://www.census.gov/prod/cen2010/doc/sfl.pdf.
15 The names of all APEX files are provided here to link the brief description with the appropriate APEX input file.
16 The age groups in this file are: 0-4, 5-9, 10-14, 15-17, 18-19, 20-20, 21-21, 22-24, 25-29, 30-34, 35-39, 40-44, 45-
49, 50-54, 55-59, 60-61, 62-64, 65-66, 67-69, 70-74, 75-79, 80-84, >84.
17 The upper limit for adults was set to age 90 due to the limited information available in CHAD for modeling
activity patterns and physiological processes for adults >90.
18 As in other NAAQS reviews, we do not estimate exposures and risk for children younger than 5 years old due to
the more limited information contributing relatively greater uncertainty in modeling their activity patterns and
physiological processes than children between the ages of 5 to 18.
3D-21
-------
asthma, based on their identification as an at-risk population (PA, section 3.3.2; ISA, section
IS.4.4.2). To best approximate the number (and percent) of individuals comprising the latter two
population groups in each study area, we considered several influential variables that could affect
asthma prevalence. It is widely recognized that there are significant differences in asthma
prevalence based on age, sex, U.S. region, and family income level, among other factors.19 There
is spatial heterogeneity in family income level across census geographic areas (and also across
age groups)20 and spatial variability in local scale ambient air concentrations of O3 (e.g., PA,
Appendix 3C, Figures 3C-91 through 3C-106). Thus, we accounted for these particular attributes
of this study group and their spatial distribution across each of the study areas to better estimate
the variability in population-based O3 exposures and risks for these at-risk population groups.
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. Thus, our primary
objective with these data was to generate census tract-level prevalence that reflect variability in
asthma prevalence contributed by several known influential attributes (i.e., age, sex, family
income level, geographic location). Two data sets were identified and linked together to estimate
asthma prevalence used for this exposure and risk analysis: asthma prevalence and population
data.
First, asthma prevalence data were obtained from the 2013-2017 National Health
Interview Survey (NHIS) and are stratified by NHIS defined regions (Midwest, Northeast, South,
and West), age, and sex.21 These asthma prevalence data are particularly useful given that age is
expressed as a continuous variable, a feature not found in other asthma prevalence data that are
available (e.g., state or county level data). We explored variables that were available in the NHIS
data set that contributed to variability in asthma prevalence and that could be used to extrapolate
the asthma prevalence to a finer geographic scale than the NHIS-provided four regions. The
linking variable had to be common with variables available in the population demographic data.
Based on this criterion, we selected family income level to poverty thresholds (i.e., whether the
family income was considered at/below or above a factor of 1.5 of the U.S. Census estimate of
poverty level for the given year) and used that as an additional variable to stratify the NHIS
asthma prevalence.
19 For example, see the Center for Disease Control report "National Surveillance of Asthma: United States, 2001-
2010", available at: https://www.cdc.gov/nchs/data/series/sr_03/sr03_035.pdf.
20 For example, see the U.S. Census report "Income and Poverty in the United States: 2016", available at:
https://www.census.gov/content/dam/Census/library/publications/2017/demo/P60-259.pdf.
21 Information about the NHIS is available at: http://www.cdc.gov/nchs/nhis.htm.
3D-22
-------
Then, we obtained population data from the 2017 Census American Community Survey
(ACS) to estimate family income level to poverty thresholds at the census tract level and
stratified by several ages and age groups.22 By combining the NHIS and U.S. Census population
data sets, we developed census tract level asthma prevalence for children (by age in years) and
adults (by age groups), also stratified by sex (male, female) that were weighted by the individual
census tract population and family income level proportions. Finally, we adjusted the census
tract-level asthma prevalence data based on individual state-level prevalence data from the 2013-
2016 Behavioral Risk Factor Surveillance System (BRFSS).23 This was done because overall, the
asthma prevalence data reported from BRFSS were consistently higher than that derived from
the NHIS data, particularly when considering adults, and thus resulted in an upward adjustment
to the initially derived NHIS census tract level data set. A detailed description of how the NHIS,
U.S. Census, and BRFSS data were processed and combined to create the data set used for input
to APEX is provided in Attachment 1. The national-based APEX input file is used for the current
exposure and risk analysis as follows:
• asthma_prev 1317 tract 053119 adjusted.txt\ census tract IDs, tract-level asthma
prevalence (in fractional form) stratified by sex, 18 single year ages (for ages < 18),24
and 7 age groups (for ages > 17).
The asthma prevalence varies for the different ages and sexes of children and adults25 that
reside in each census tract of each study area. We evaluated the spatial distribution of the asthma
prevalence using the tracts that comprise the air quality domain in each study area. We first
separated the estimates for children from those for adults and calculated the distribution of
asthma prevalence for the tracts, stratified by sex (Table 3D-3). These summary statistics
represent the range of age- and sex-specific probabilities for the census tracts comprising each
study area that are used by APEX to estimate the number of individuals that have asthma.
22 Census tract level data is the finest scale geographical unit having family income information. The family
income/poverty ratio threshold used was 1.5, that is the surveyed person's family income was considered either <
or > than a factor of 1.5 of the U. S. Census estimate of poverty level for the given year.
23 Table C2.1 (for each adults and children) was downloaded to obtain the 2013-2016 BRFSS current asthma
prevalence by state and sex, available at: https://www.cdc.gov/asthma/brfss/default.htm. Table CI was also
downloaded to obtain the asthma prevalence for the two age groups not stratified by sex. Accessed 5/3/19.
24 The census data only had children for single years up to and including age 17, after that age they are provided in
groups. The upper portion of this age range differs from those considered as children in estimating exposures (i.e.,
in our exposure assessment children are considered upwards to 18 years old). To simulate the number of children
with asthma age 18, estimated prevalence from the first adult group were used (i.e., individuals age 18-24).
25 While prevalence was estimated for all ages of children (in single years 5-17), for adults they were estimated for
seven age groups: 18-24 years, 25-34 years, 35-44 years, 45-54 years, 55-64 years, 65-74 years, and >75 years old
(see Attachment 1 for more information).
3D-23
-------
Table 3D-3. Descriptive statistics for children and adult asthma prevalence, using all
census tracts within eight consolidated statistical areas (CSAs) in the APEX
asthma prevalence file.
CSA Name - ID#
(# tracts)
and
Population group
Sex
Asthma Prevalence across all ages (or age groups) and census tracts A
Mean
Standard
Deviation
Minimum
Median
95th
percentile
99th
percentile
Maximum
Atlanta-122
(1,077)
adult
female
11.1%
1.8%
7.7%
11.1%
14.0%
15.9%
20.9%
male
5.5%
0.8%
4.3%
5.4%
7.1%
7.5%
7.9%
child
female
9.7%
1.7%
6.5%
9.6%
12.9%
13.9%
15.0%
male
14.1%
1.7%
10.6%
14.0%
16.8%
17.6%
18.3%
Boston-148
(1,753)
adult
female
13.8%
1.8%
10.5%
13.5%
17.3%
20.5%
28.9%
male
7.6%
0.9%
5.4%
7.5%
9.1%
10.0%
12.9%
child
female
9.4%
2.0%
5.6%
9.5%
12.4%
13.5%
17.1%
male
15.4%
2.5%
8.7%
15.1%
19.5%
20.8%
23.4%
Dallas-206
(1,422)
adult
female
9.3%
1.5%
6.5%
9.3%
11.8%
13.5%
16.5%
male
4.9%
0.7%
3.8%
4.9%
6.4%
6.8%
9.7%
child
female
7.6%
1.3%
5.0%
7.4%
10.0%
10.9%
13.5%
male
11.0%
1.4%
8.3%
11.0%
13.2%
13.8%
18.1%
Detroit-220
(1,583)
adult
female
13.3%
2.5%
7.8%
13.4%
17.8%
20.6%
25.6%
male
7.9%
2.2%
1.0%
7.6%
12.4%
14.7%
19.0%
child
female
8.6%
1.5%
6.4%
8.2%
11.6%
12.5%
13.2%
male
13.3%
3.0%
7.7%
12.7%
19.9%
23.6%
25.5%
Philadelphia-
428
(1,725)
adult
female
12.1%
2.3%
8.2%
12.0%
16.4%
19.8%
26.5%
male
6.5%
0.9%
4.6%
6.4%
8.1%
9.0%
11.4%
child
female
9.1%
1.9%
5.6%
9.2%
12.0%
13.1%
15.3%
male
13.6%
2.4%
8.2%
13.3%
17.8%
19.2%
21.1%
Phoenix-429
(988)
adult
female
11.6%
1.6%
8.6%
11.7%
14.4%
16.0%
19.7%
male
7.0%
1.5%
5.1%
7.1%
9.1%
11.7%
16.7%
child
female
7.6%
1.5%
4.6%
8.0%
9.5%
9.6%
9.6%
male
11.5%
1.8%
8.5%
11.6%
14.8%
15.9%
17.1%
Sacramento-
472
(539)
adult
female
10.4%
1.4%
7.7%
10.5%
12.7%
14.0%
16.5%
male
5.7%
1.1%
4.2%
5.9%
7.3%
9.0%
13.6%
child
female
8.5%
1.7%
5.2%
9.0%
10.7%
10.9%
10.9%
male
10.8%
1.7%
8.1%
10.9%
13.7%
14.8%
16.2%
St. Louis-476
(638)
adult
female
11.8%
2.1%
6.8%
11.9%
15.0%
17.4%
21.5%
male
6.5%
1.8%
0.9%
6.5%
9.9%
11.8%
14.5%
child
female
9.2%
2.0%
5.3%
9.1%
12.9%
14.2%
15.6%
male
11.1%
2.4%
6.5%
10.7%
15.9%
19.3%
21.9%
A Prevalence is based on single year ages (children) or age groups (adults) and sex derived from 2013-2017 CDC NHIS asthma prevalence
and considering U.S. census tract level family income/poverty ratio data. Data presented are not population-weighted and represent the
distribution of applied probabilities used by APEX for tracts having a non-zero population. Note, upper and lower percentiles could represent
prevalence for a single year age/sex residing in a single tract within a study area.
3D-24
-------
In general and consistent with broadly defined national asthma prevalence (e.g., Table 3-
1 of the PA), male children have higher rates than female children26 and adult females have
higher rates than adult males.27 The overall asthma prevalence for children was similar to that
estimated for adults, largely the result of having a greater BRFSS adjustment applied to adult
females compared to that applied to children of either sex.28 As described above, and by design
(i.e., in using age, sex, and family income variables) there is wide ranging spatial variability in
the estimated asthma prevalence. For instance, the Boston, Detroit, and Philadelphia study areas
have some of the highest asthma prevalence for boys and adult women considering most of the
descriptive statistics, with rates of 25% or higher in one or more census tracts for a given year of
age (Table 3D-3). In contrast, the Dallas study area exhibits some of the lowest asthma
prevalence (and low variability) for any of the four age/sex groups compared to the other study
areas.
There are other personal attributes shown to influence asthma prevalence, such as race,
ethnicity, obesity, smoking, health insurance, and activity level (e.g., Zahran and Bailey, 2013).
The set of variables chosen to stratify asthma prevalence for use in this 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 Census population demographic data sets. Many of the additional influential factors
identified here are not available in the census population data and/or have limited representation
in the asthma prevalence data (e.g., the survey participant does/does not have health insurance, or
they did/did not provide a response to a question regarding their body weight). 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
26 Population weighted asthma prevalence, when not categorized by the eight study areas, is greater in boys (mean of
11.1%) than that of girls (mean of 7.3%). Nationally, asthma prevalence for boys is 9.5%, for girls is 7.3% (Table
3-1 of the PA).
27 Population weighted asthma prevalence, when not categorized by the eight study areas, is greater in women (mean
of 12.0%) than that of men (mean of 6.5%). Nationally, asthma prevalence for women is 9.8%, for men is 5.4%
(Table 3-1 of the PA).
28 Population weighted asthma prevalence, when not categorized by the eight study areas and sex, is similar for
children (mean of 9.2%) and adults (mean of 9.3%). Nationally, asthma prevalence for children is 8.4% and for
adults is 7.7% (Table 3-1 of the PA).
3D-25
-------
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.
3D.2.2.3 Personal Attributes
In addition to using the above demographic information to construct the simulated
individuals, each modeled person is assigned anthropometric and physiological attributes by
APEX. All of these variables are treated probabilistically, accounting for interdependencies
where possible, and reflecting variability in the population. It is not the intention of this
document to provide detailed description of all the model inputs in each of the files and the data
used in their derivation, and where additional details exist, appropriate reference materials are
provided. We describe further a few APEX model inputs that have been recently updated and
that are available for use in this exposure and risk analysis. These are new statistical distributions
for estimating body weight, equations for estimating resting metabolic rate, and equations for
estimating activity-specific ventilation rate. Each of these data and algorithms are important,
particularly the ventilation rate (section 3D.2.2.3.3), because the health response observed in the
controlled human exposure studies is concomitant with elevated breathing rate. Brief
descriptions of the data used to develop these generalized (i.e., non-03 specific) input files are
provided in the sections below. For additional detail, see U.S. EPA (2018) Appendices G and H,
and the data within the APEX input files.
3D.2.2.3.1 Body Weight and Surface Area
Anthropometric attributes utilized by APEX in various assessments for estimating
exposures or doses can include height, body weight (BW), and body surface area (BSA). Two
key personal attributes determined for each individual in this assessment are BW and BSA, both
of which are used in the calculation of a number of other variables associated with estimating
exposures (e.g., ventilation rate).
Regarding the estimation of body weight, a new APEX input file was recently generated
using 2009-2014 National Health and Nutrition Examination Survey (NHANES) data.29 Briefly,
body weight and height data for surveyed individuals were obtained and stratified by sex and
single years for ages 0 - 79; all ages above 80 were combined as a single age group. Statistical
form of the age- and sex-specific body weight and height distributions were evaluated using a
29 NHANES questionnaire datasets for 2009-2010, 2011-2012, 2013-2014 are available at
https://wwwn.cdc.gov/nchs/nhanes/Default.aspx. Details regarding the data used and the derivation of the APEX
input file data distributions is found in U.S. EPA (2018), Appendix G.
3D-26
-------
log-likelihood statistic. Body weight was found to best fit a lognormal distribution; height was
found to best fit a normal distribution. Because height and body weight are not independent, the
joint distributions of height and logarithm of body weight were fit assuming a bivariate normal
distribution. Then, parameters defining the joint distributions30 were smoothed using a natural
cubic spline to have them represent continuous functions of age rather than vary discontinuously.
In addition, having the smoothed parameters could be used to extrapolate information obtained
from the single age year distributions (ages 0 - 79) to approximate statistical distributions of
body weight for ages >80. To do so, a linear function was fit to ages 70 and above to extrapolate
the parameter values (and hence the statistical distributions of body weight) up to age 100.
These body weight distributions are randomly sampled by APEX to estimate an age and
sex-specific body weight for each simulated individual. Comparison of the new distributions to
the body weight distributions previously used by APEX and developed from the 1999-2004
NHIS indicate, for both sexes and across all ages, simulated body weight is about two percent
greater using the updated distributions. This difference is expected given the consistent trend of
increasing body weight that has occurred in the U.S. population over the past few decades.
Age- and sex-specific body surface area, a variable used in conjunction with breathing
rate to approximate moderate or greater exertion (section 3D.2.2.3.3) is estimated for each
simulated individual (Equation 3D-1) and is based on an equation provided in Burmaster (1998):
BSA = e"2-2781 x BW0-6821 Equation 3D-1
One standard APEX input file is used for the current O3 exposure and risk analysis:
• Physiology051619 Ufixedtxt: Provides parameters for estimating body weight (log BW,
standard deviation of BW, lower and upper bounds of BW, by single age years 0-100 and
by two sexes) and regression coefficients used in estimating BSA for all sexes and ages.
3D.2.2.3.2 Energy Expenditure and Oxygen Consumption Rates
Energy expended by different individuals engaged in different activities can have an
important role in pollutant-specific exposure and/or dose. For example, energy expenditure is
related to ventilation rate, which is an important variable in estimating exposure and risk given
that the 03-induced lung function response has been documented to occur under conditions of
elevated ventilation (PA, section 3.3.1.1). In addition, because we are also interested in
exposures that occur over relatively short durations (i.e., < 8 hours), estimating activity-specific
ventilation rate (Ve) has always been an important motivation behind the development of the
algorithm used by APEX. The fundamental basis for Ve algorithm is founded in energy
expenditure which, for our modeling purposes here, can be related to an individual's resting
30 Five parameters were used for each age and sex: mean log(BW), standard deviation of log(B W), mean height,
standard deviation of height, and body weight-height correlation coefficients.
3D-27
-------
metabolic rate (RMR) or the energy expended while an individual is at complete rest, along with
the energy expended while an individual performs activities involving greater exertion, termed
here as metabolic equivalents of work (METs) (McCurdy, 2000). The approaches used by APEX
for estimating RMR and METs are described below, beginning first with the update to the
equations used for estimating a simulated individual's RMR.
Since the 2014 HREA,31 we have reviewed recent RMR literature and other published
sources containing individual data and have compiled the associated individual RMR
measurements, along with associated influential attributes such as age, sex, and body weight,
where available. Data from these individual studies were then combined with RMR data reported
in the Oxford-Brookes database (Henry, 2005; IOM, 2005) and screened for duplicate entries. In
addition, observations missing values for RMR, BW, age, or sex were deleted, resulting in a
dataset containing 16,254 observations (9,377 males and 6,877 females). Using this new RMR
dataset and having a goal of updating the previous RMR equations and reducing discontinuities
in RMR between age groups, new equations were developed.
Details regarding the data, the derivation, and performance evaluation of the new
equation that APEX uses to estimate RMR are provided in U.S. EPA (2018), Appendix H.
Briefly, the equations follow the general format of a multiple linear regression (MLR) model,
using age and body weight as independent variables to estimate each simulated individual's
RMR, along with a residual error term (f).32 It is known that RMR and BW, as well as RMR and
age, are not exactly linearly related; the algorithms developed here use BW (in kg), age (in
years), and the natural logarithms of BW and (age+1)33 as follows in Equation 3D-2, with their
parameter estimates provided in Table 3D-4.
RMR = /?0 + /?XBW + /?2 log(BW) + (33Age + /?3log (Age) + et Equation 3D-2
When comparing observed versus predicted values, the new RMR equations have a bias
of less than 0.5%, compared to the previously used APEX equations which had a bias of between
1-2%. Further, the discontinuities in RMR seen across particular age group boundaries using the
31 The algorithm used to estimate RMR for the 2014 HREA was based on analyses by Schofield (1985) who used
clinical subject data from studies conducted as far back as 60 years prior to that publication. In addition, the
Schofield (1985) RMR equations contained abrupt discontinuities at some of the equation boundaries (e.g.,
between age 59 and 60). As a result, we felt it was important to obtain newly available study data to develop
RMR equations that better represent a more recent population and having fewer discontinuities.
32 The residual error term largely accounts for the estimation of inter-personal variability in RMR for individuals
having the same body weight and age. There are other potentially influential sources of variability that are not
explicitly accounted for by the equation (e.g., seasonal influences on RMR) and thus remain as an uncertainty.
33 The "+1" modifier allows APEX to round age upwards instead of downwards to whole years, which is necessary
to avoid undefined log(0) values.
3D-28
-------
previous equations have been reduced when using these updated equations in APEX. One
standard APEX input file is used for the O3 exposure and risk analysis:
• Physiology051619 Ufixedtxt: Regression coefficients used to estimate RMR (kcal day"1)
for two sexes and six age groups.
Table 3D-4. Regression parameters used to estimate RMR by sex and age groups.
Sex
Age
Group
Subjects
(n)
BW
log(BW)
Age
log(Age)
Intercept
Standard
Deviation
male
0-5
625
13.19
270.2
-18.34
131.3
-208.5
69.10
6-13
1355
10.21
260.2
13.04
-205.7
333.4
115.3
14-24
4123
0.207
1078.0
115.1
-2794.0
3360.6
161.1
25-54
2531
2.845
729.6
3.181
-191.6
-1067
178.2
55-99
743
9.291
264.8
-5.288
181.5
-705.9
163.6
female
0-5
625
11.94
261.5
-22.31
120.9
-183.6
64.16
6-13
1618
5.296
409.1
40.37
-524.9
392.7
99.43
14-29
2657
0.968
676.9
40.89
-1002
772.7
143.1
30-53
1346
4.935
355.4
16.28
-896.0
2225
145.3
54-99
631
2.254
445.9
5.464
-489.9
944.2
124.5
Units: RMR = kilocalories/day; BW = kilograms; Age = years
Following the estimation of an age- and sex-specific RMR for simulated individuals, the
next variable used for estimating ventilation rate involved an approximation of the energy
expended for activities an individual performs throughout their day. As mentioned above,
activity-specific energy expenditure is highly variable and can be estimated using metabolic
equivalents of work (METs), or the ratios of the rate of energy consumption for non-rest
activities to the resting metabolic rate of energy consumption, as follows in Equation 3D-3:
EE = MET X RMR Equation 3D-3
where,
EE = Energy expenditure (kcal/minute)
MET = Metabolic equivalent of work (unitless)
RMR = Resting metabolic rate (kcal/minute)
Statistical distributions of METs were developed for simulated activities using the
physical-activity compendium (Ainsworth et al., 2011; hereafter "the compendium"). The
compendium contains a point value for the MET associated with each of several hundred
different activities. Activity-specific MET distributions were developed by cross-walking the
3D-29
-------
activities described in the compendium with the descriptions of activities in the activity pattern
data base used by APEX (section 3D.2.5). The shape of the statistical distribution (e.g., normal,
lognormal, triangular, point) for each activity was assigned based on the number of
corresponding activities in the compendium and goodness-of-fit statistics. When simulating
individuals, APEX randomly samples from the activity-specific METs distributions to obtain
values for every activity performed. Two standard APEX input files are used for the current O3
exposure and risk analysis:
• METdistributions 092915.txt: MET distribution number, statistical form, distribution
parameters, lower and upper bounds, activity description
• MET mapping 071018.txt: activity codes, age group (where applicable), occupation
group, MET distribution number, and activity description used to link of MET
distributions to activities performed
The rate of oxygen consumption (VO2, Liters min"1) for each activity is then calculated
from the energy expended (kcal min"1) using an energy conversion factor (ECF, Liters O2 kcal"1)
as follows in Equation 3D-4:
V02 = EE X ECF Equation 3D-4
The value of the ECF is randomly selected from a uniform distribution for each person,
U[0.20, 0.21] (Johnson, 2002, adapted from Esmail et al., 1995). One standard APEX input file
is used for the current O3 exposure and risk analysis:
• Physiology051619 Ufixed.txt\ Parameters of the uniform distribution representing the
ECF used for all ages and both sexes.
3D.2.2.3.3 Ventilation Rate
Human activities are variable over time, with a wide range of activities possible within
only a single hour of the day. The type of activity an individual performs, such as sleeping or
jogging (as well as individual-specific factors such as age, weight, RMR) will influence their
ventilation rate. APEX estimates minute-by-minute ventilation rates that account for the
expected variability in the activities performed by simulated individuals. Ventilation rate is
important in this assessment because the lung function responses associated with short-term O3
exposures coincide with moderate or greater exertion (2013 ISA, Table 6-1). In our exposure
modeling approach, APEX generates the complete time-series of activity-specific ventilation
rates and the corresponding time-series of estimated O3 exposures and is directly used for the
individual-based lung function risk (section 3D.2.8.2.2). APEX can then aggregate both the
ventilation rate and exposure concentration for the duration of interest (e.g., 7-hr average), and
they can be used for the benchmark comparison (section 3D.2.8.1) and estimating the
population-based lung function risk (section 3D.2.8.2.1). Thus, the model provides O3 exposure
3D-30
-------
estimates for the simulated individuals that pertain to specific target levels for both ventilation
rate and exposure concentration. The approach to estimating activity-specific energy expenditure
and associated ventilation rate involves several algorithms and physiological variables, with
details found in the APEX User's Guide (U.S. EPA, 2019a, U.S. EPA, 2019b).
Using the existing measurement Ve dataset from Graham and McCurdy (2009), new Ve
algorithms were developed for predicting activity specific Ve in the individuals simulated by
APEX (Appendix H of U.S. EPA (2018)). The new Ve algorithms do not directly employ
previously used variables to stratify the data (age groups, sex) and explain variability (age, body
weight, height) in ventilation rate, effectively simplifying and reducing the number of equations.
The new algorithms utilize a new variable, the maximum volume of oxygen consumed (VChm)
as an input.34 Body weight, height, and sex - as well as fitness level (which is often represented
by VChm) - influence oxygen consumption for a particular activity. However, variability for
each of these influential variables are already captured in the algorithm used to estimate each
simulated individual's RMR, and subsequently, the estimation of their activity-specific VO2.35
Thus, the only input variables needed for the new Ve algorithm are VO2 and VChm,36 both of
which are estimated by APEX.
Details for the derivation of and performance evaluation of the new equation that APEX
uses to estimate ventilation rate are provided in U.S. EPA (2018) Appendix H. Briefly, the Ve
dataset contains 6,636 observations, with 4,565 males and 2,071 females. Similar to the earlier
ventilation equation by Graham and McCurdy (2009), a mixed-effects regression (MER) model
was fit because the MER separates residuals into within-person (ew) and between-person (eh)
effects, known as intrapersonal and interpersonal effects, respectively.37 It was found that the
actual values of VO2 and VChm are less relevant than the fraction of maximum capacity,
represented by fi = VCte/VCtem. The variable fi may operate non-linearly (for example, fi = 0.9 is
likely more than twice as encumbering as fi = 0.45). A transformation regression approach
34 Use of VO2111 as an explanatory variable in separate related research on metabolic equivalents of task (MET)
values for persons with unusual maximum capacity for work suggests that their MET distributions are modified in
a predictable way by their maximum MET (or, equivalently, by VChm), thus providing support for use of this
variable in the new Ve algorithms Details are provided in Appendix H of U.S. EPA (2018).
35 Oxygen consumption associated with activities performed is based on the activity specific metabolic equivalents
for work (METs), an individual's estimated RMR, and an energy to oxygen conversion factor (Equations 3D-3
and 3D-4 above).
36 Distributions of V02m used by APEX were derived from 20 published studies reporting individual data and
grouped mean (and standard deviation) data obtained from 136 published studies. Details are provided in Isaacs
and Smith, 2005 (and found in Appendix B of U.S. EPA (2009).
37 N(0, t'h) is a normal distribution with mean zero and standard deviation e/, = 0.09866 meant to capture
/'wterpersonal variability, which is sampled once per person. N(0, ew) is an;/?/re/personal residual with standard
deviation of c„ = 0.07852, which is sampled daily due to natural /'n/rapersonal fluctuations in VE that occur daily.
3D-31
-------
(using PROC TRANSREG; SAS, 2017) was used to determine the most appropriate variable
transformation, indicating a power of 4 to 5 be used when only the log transformed VO2 was
used as the independent variable and described in Equation 3D-5.
Y _ e(3.300 + 0.8i28xin(7oz)+ 0.5126 x (7o2^o2m)4+w(o,eb)+w(o,ew)) Equation 3D-5
In comparing the statistical fit of the new equation with the equations used by APEX
previously to estimate ventilation rate, the resulting coefficient of determination (R2 values) for
the new equation (R2 = 0.94) indicates an improved fit compared to that of the previous
equations (R2 = 0.89 - 0.92). Further, because the data were not stratified by age groups (or any
other groupings), there are no discontinuities in predictions made across age boundaries as was
observed when employing the previous equations. Information used in estimating ventilation rate
is found in the following APEX two input files:
• Physiology051619 Ufixedtxt: parameters describing statistical distributions of
normalized maximum oxygen consumption rate (NVChm) for two sexes by single age
years (0-100) (see, Isaacs and Smith, 2005).
• Ventilation 062117.txt: minimum and maximum age ranges, regression coefficients,
between and within error terms used to estimate individual activity-specific
ventilation.
To use this information to estimate health risks for children, the ventilation rates observed
for the adult controlled human exposure study subjects need to be converted into rates that best
reflect the different physiology of children. Consistent with prior REAs (U.S. EPA, 2009, 2014,
2018; Whitfield et al., 1996) we used an equivalent ventilation rate (EVR, L/min-m2), which is
essentially an allometrically normalized ventilation rate (Equation 3D-6), to estimate instances
when any simulated individual reaches a ventilation rate relatively as high as that of the study
subjects (i.e., termed here as moderate or greater exertion).
EVR = Equation 3D-6
BSA M
Before discussing the value used to determine whether a simulated individual is at
moderate or greater exertion, a brief description of the controlled human exposure study protocol
is warranted. Most of the controlled human exposure studies evaluating O3 health effects of
interest for our exposure benchmark analysis (e.g., Adams, 2006; Folinsbee et al., 1988) were
conducted over a 6.6-hr exposure period, thus, the most relevant exposures and associated
breathing rates for the exposure benchmark comparisons would be those occurring on average
over a 6.6-hr period (not an 8-hr period as was used in previous REAs). The typical protocol for
the 6.6-hr controlled human exposure studies employed a mixture of exercise and rest periods
varied across the duration of the study, with an expectation that the study subject achieves, on
3D-32
-------
average, a target EVR of 20 L/min-m2 (i.e., a ventilation rate of -35 L/min in females and -40
L/min in males) while exercising using a treadmill or cycle ergometer (e.g., Schelegle et al.,
2009). Most researchers collected the ventilation data during periods of exertion and therefore
reported the exercise-only conditions (e.g., Horstman et al., 1990; Folinsbee et al., 1988).
More specifically, during the 6.6-hr study experiments, 5 hours were used for exercise
(i.e., six 50-minute (min) periods on a treadmill or cycle ergometer), with the remaining 1.6
hours comprised of a series of 10-min rest periods occurring immediately after the exercise along
with a 35-min lunch break before the fourth exercise period. As a result of these rest/lunch
periods, the study subject's actual ventilation rates (and hence EVRs) are expected to be less than
the target/observed exercise levels reported in the controlled human exposure studies. Note, the
simulated individuals used to estimate exposure and risk perform numerous activities throughout
the day, each having varied durations and exertion levels (e.g., jogging, sleeping, eating). As
such, when time-averaging across a simulated exposure period of interest, the period likely
would contain ventilation rates of varying duration and intensity. For this review, to better match
the ventilation information obtained from the controlled human exposure studies with that of the
simulated individuals, we accounted for the impact from the rest/lunch time ventilation rate
along with that attained during exercise to estimate an appropriate EVR for the study subjects.
Attachment 2 provides details regarding the data and approach used to estimate the EVR,
an APEX model variable used to identify when a simulated individual is at moderate or greater
exertion. Briefly, the controlled human exposure study data set available used to calculate EVR
was comprised of 177 study subjects, each evaluated for 2 or more exposure levels (i.e., totaling
485 experiments), and having multiple measurements for each exercise period, yielding 4,024
individual EVR data points. Of these six studies providing raw data,38 only Schelegle et al.
(2009) mentioned resting Ve (and hence a resting EVR), with an average value for males and
females estimated as 7.61 and 8.05 L/min-m2, respectively and based on regression equations
provided by Aitken et al. (1986). We calculated total (exercise and rest) EVR for each person
across the 6.6-hr study period as a weighted average based on the observed EVR for the 5 hours
of exercise and the estimated EVR for 1.6 hours of rest/lunch. Descriptive statistics were
calculated and indicated the person-level EVR data were normally distributed, having a mean
value of 17.32 (L/min-m2) and a standard deviation of 1.25 (L/min-m2). To reflect variability
across simulated individuals, an EVR is probabilistically selected from this distribution once per
person and used for the duration of their simulation period. This new approach for assigning a
unique EVR to every simulated individual, one that accounts for rest and exercise periods and
38 The six studies include Folinsbee et al. (1988), Folinsbee et al. (1994), Horstman et al. (1990), Kim et al. (2011),
McDonnell et al. (1991), and Schelegle et al. (2009).
3D-33
-------
based on the distribution of ventilation rates achieved by all controlled human exposure study
subjects, more appropriately reflects the EVR variability expected to exist in the simulated
population compared to the approach used in the last two reviews (e.g., U.S. EPA, 2007a; U.S.
EPA, 2007b; 2014 HREA) that assigned a single lower bound EVR value to all individuals.39
For practical and tractable modeling reasons, this individual-level EVR threshold is
applied to APEX simulated individuals using a 7-hr averaging time (representing the 6.6-hr
period rounded to whole numbers) in order to better represent the exposure study design than the
previously used 8-hr average. Then, once a simulated individual is identified as having surpassed
their personal 7-hr average EVR threshold in a given day, the level of their simultaneously
occurring 7-hr average O3 exposure is recorded by APEX. Retained for each simulated
individual are the daily maximum 7-hr average exposure concentration(s) that occurred while at
moderate or greater exertion over the assessment period.
3D.2.3 Ambient Air Concentrations
Ambient air concentrations serve as a fundamental input used by APEX to estimate
exposure. There are two important attributes of ambient air concentrations to consider when
estimating population exposure and risk using APEX: spatial and temporal variability. This is
because there can be significant spatial and temporal heterogeneity in O3 concentrations across
each of the study areas and there is substantial flexibility by APEX in handling ambient air
concentrations at varying scales, both temporally (e.g., hourly, daily) and spatially (e.g. 500-
meter grid, census tract).
For this exposure and risk analysis (as done for the last review), we were interested in
having hourly O3 concentrations at the census tract level. Having these temporally and spatially
resolved ambient air concentrations in each study area allows for better utilization of APEX
temporal and spatial capabilities in estimating exposure and risk (e.g., the population data
described in section 3D.2.2 are at a census tract level). Because APEX simulates where
individuals are located and what they are doing at specific times of the day, more realistic
exposure estimates are obtained in simulating the contact of individuals with these temporally
and spatially diverse concentrations.
Ambient air monitors for O3 capture the temporal scale of interest (i.e., hourly) and can
provide general information regarding O3 levels across an urban area. However, given their
39 The EVR used in prior REAs (e.g., U.S. EPA, 2007b; U.S. EPA, 2007a; 2014 HREA) was based on a single lower
bound EVR value of 13 L/min-m2 selected from a range provided by Whitfield et al. (1996). For the current
assessment approach, assigning randomly sampled values from an EVR distribution of N{17.32,1.25} still allows
for some simulated individuals to be considered at elevated exertion when exceeding an EVR of-13-14 L/min-
m2 (Appendix 3D, Attachment 2, Table 3) but overall, leads to fewer individuals achieving a moderate or greater
exertion level when compared to simulations employing a single lower bound EVR value of 13 L/min-m2.
3D-34
-------
limited spatial representativeness, i.e., tens of monitors extending across areas >10,000 km2, the
monitors may not fully inform concentration variability that may exist at a finer spatial scale. In
addition, of interest in this review are concentrations that represent a specific air quality scenario
(e.g., ambient air quality that just meets the current standard). In general, due to varying levels of
precursor emissions and meteorological conditions, most monitored 3-year periods do not have
O3 concentrations that just meet a specific air quality scenario of interest. Therefore, due to these
two realities, modeling methods are used to achieve the desired temporal and spatial scale along
with estimating ambient air O3 concentrations that represent a specific air quality scenario.
The sections that follow briefly summarize the data and approaches used to estimate the
air quality concentrations used by APEX. A detailed description on the air quality data
collection, processing, adjustment, and evaluation is provided in Appendix 3C. First, section
3D.2.3.1 below provides information for the overall bounding of the modeling domains. The
identification of ambient air monitoring data used as a foundation for representing fine-scale
temporal and broad-scale spatial concentration variability is provided in section 3D.2.3.2. The
approach used to adjust concentrations to just meet air quality scenarios of interest is described
in section 3D.2.3.3. And finally, Section 3D.2.3.4 describes the technique used to interpolate the
concentrations from the monitor locations to the desired spatial scale (i.e., census tracts). It is
these estimated hourly census tract O3 concentrations representing air quality scenarios that serve
as the basic ambient air concentrations from which each simulated individual's
microenvironmental concentrations and exposures are estimated (sections 3D.2.6 and 3D.2.7,
respectively). Multiple unique APEX input files are used for the current exposure and risk
analyses, one for each year and study area, and in the following two formats:
• concsCSA[number]S[air quality scenario]Y[year].txt: Tract IDs, hourly
concentrations (ppm), calendar date, by study area and year
• districtsCSA[number]Y[year].txt\ Tract IDs, latitude, longitude, begin and end date
3D.2.3.1 Spatial and Temporal Boundaries of Modeling Domains
APEX has several options to select air quality data to use for estimating exposure and
risk. For this exposure and risk analysis, we used the list of counties that comprise each
CSA/MSA and their geographic boundaries to define the broad spatial characteristics of each
study area (Table 3D-5). As a result, simulated individuals residing within these counties would
be part of the exposure modeling domain and any ambient air concentrations estimated within
these counties would be used by APEX. Figure 3D-2 to Figure 3D-5 depict the spatial extent of
the exposure and risk modeling domain in each study area, along with a visualization of tract-
level population density and location of meteorological stations (see section 3D.2.4). The air
3D-35
-------
radius for APEX, a variable used to define the modeling domain, was set at 30 km to include all
air quality receptors (i.e., census tracts) within each county to model exposures and risks.
For each study area, three years of recent air quality were selected to estimate exposures.
The exposure periods are the O3 seasons40 for which routine hourly O3 monitoring data were
available, and defined by 40 CFR part 58, Appendix D, Table D-3. These periods are designed to
reasonably capture variability in ambient air O3 concentrations and meteorology and include the
high concentration events occurring in each area. Having this range of air quality data across
multiple years allows us to realistically estimate a range of exposures, rather than using a single
year of air quality. The number of O3 monitors in operation did not vary from year to year, thus,
the overall spatial representation of each study area by the ambient air monitors (and that using
the statistically interpolated data) remained constant for each year over the simulation period.
Table 3D-5. List of states, counties, and O3 seasons that define the air quality and
exposure spatial and temporal modeling domain in each study area.
Study Area
State Abbreviation: County ListA
O3 seasonB
Atlanta
GA: Barrow, Bartow, Butts, Carroll, Cherokee, Clarke, Clayton, Cobb, Coweta,
Dawson, DeKalb, Douglas, Fayette, Forsyth, Fulton, Gordon, Gwinnett, Flail,
Flaralson, Fleard, Flenry, Jackson, Jasper, Lamar, Madison, Meriwether, Morgan,
Newton, Oconee, Oglethorpe, Paulding, Pickens, Pike, Polk, Rockdale, Spalding,
Troup, Upson, Walton.
March to
October
Boston
CT: Windham. MA: Barnstable, Bristol, Essex, Middlesex, Norfolk, Plymouth,
Suffolk, Worcester. NH: Belknap, Flillsborough, Merrimack, Rockingham, Strafford.
Rl: Bristol, Kent, Newport, Providence, Washington.
March to
September
Dallas
TX: Bryan, Collin, Cooke, Dallas, Denton, Ellis, Fannin, Grayson, Flenderson,
Flood, Flopkins, Flunt, Johnson, Kaufman, Navarro, Palo Pinto, Parker, Rockwall,
Somervell, Tarrant, Wise.
January to
December
Detroit
Ml: Genesee, Lapeer, Lenawee, Livingston, Macomb, Monroe, Oakland, St. Clair,
Washtenaw, Wayne.
March to
October
Philadelphia
DE: Kent, New Castle. MD: Cecil. NJ: Atlantic, Burlington, Camden, Cape May,
Cumberland, Gloucester, Salem PA: Berks, Bucks, Chester, Delaware,
Montgomery, Philadelphia.
March to
October
Phoenix
AZ: Maricopa, Pinal.
January to
December
Sacramento
CA: El Dorado, Nevada, Placer, Sacramento, Sutter, Yolo, Yuba.
January to
December
St. Louis
IL: Bond, Calhoun, Clinton, Jersey, Macoupin, Madison, Marion, Monroe, St. Clair,
MO: Franklin, Jefferson, Lincoln, St. Charles, St. Francois, St. Louis, Warren, St.
Louis City.
March to
October
A Delineations promulgated by the Office of Management and Budget (0MB) in February of 2013 (PA Appendix 3C, section 3C.2).
B These are the regulatorily required monitoring seasons (see section 2.3.1 of the main PA).
40 In this current analysis and for practical purposes, even though there are different durations of monitoring data
available across the study areas (i.e., some areas perform a full year of monitoring, others less than a full year), an
O3 season is considered to be synonymous with a year and exposure results are reported on a per year basis.
3D-36
-------
Gilmer
White
Lumpkin
Gordon
Pickens
Dawson
Cherokee
Bartow
Forsyth
Barrow
Gwinnett
Paulding
Oglethorpe
DeKalb
Haralson
Walton
Douglas / Fulron
tockdal
Carroll
Newton
Fayette
Butts
Spalding
Chambers
Upson
Talbot
Kilometers
Population Density - Year 2010
Atlanta Study Area
Population Density per Square Mile
0 - 600 EBB 2 000 " 3'700 I
600 - 2,000 3,700- 6,200
6,200 - 116,000
Legend:
Study Area Center ~ Meteorological Stations I I Modeled Counties
Walker
DeKalb
IU
Chattooga
Cherokee
Calhoun,
Clebume
Clay Randolph \ Heard
Lamar
Crav
Oconee
Stephens
Anderson
, Banks Franklin Hart V
Hall^^^B
Jackson ) Madison - Elbert
Morgan \ Greene Taliaferro
Warren
Jasper B Putnam
Baldwin
Monroe Jones
Hancock
Washington
Wilkinson
Source; US Census Bureau, Esri, DigitalGlobe, GeoEye. Earthstar Geographies,
CNES/Airbus DS, USDA, USGS, AeroGRID. IGN, and the GIS User Community
Carroll
Belknap
Merrimack
Rockingham
Hillsborough
Middlesex
Worcester
Norfolk
Plymouti
Bristol
Providei
Windham
Barnstable.
fashingtoi
Population Density - Year 2010
Boston Study Area
Grafton
Cumberland
Rutland Windsor
Sullivan
Bennington
Windham
Cheshire
Franklin
Hampshire
Hampden
Litchfield Hartford Tolland
New London
Dukes
New Haven
Nantucket
Legend:
Study Area Center Meteorological Stations Modeled Counties
Source: US Census Bureau, Esri, DigitalGlobe, GeoEye, Earthstar Geographies,
jcNES/Airbus DS, USDA, USGS, AeroGRID, IGN, and the GIS User Community
Population Density per Square Mile >'
0-600 2,000-3,700 6,200 - 116,000
600 - 2,000 3,700- 6,200
Figure 3D-2. County boundaries, census tract population densities, and meteorological
stations in the Atlanta (top) and Boston (bottom) study areas.
3D-37
-------
Denton
ockwa
Hood
Johnson
Henderson
Navarro
Carter
Marshall
Population Density - Year 2010
Dallas Study Area
Legend:
yf1 Study Area Center Meteorological Stations ~ Modeled Counties
Population Density per Square Mile
0 - 600 !9| 2.°00 ' 3J°° I
600 - 2,000 3,700- 6,200
6,200 - 116,000
Wichita
Archer
Cooke
Montague
McCurtain
Grayson
Fannin
Young
Wise
Rains
Palo Pinto
Parker
Smith
Comanche
Jefferson
Hunt
Hopkins
Titus
Camp
Upshur
Kaufman
Van Zandt
Eastland
Freestone
Bosque
Erath
Source: US Census Bureau, Esri, DigitalGlobe,
GeoEye, Earthstar Geographies, CNES/Airbus DS,
USDA, USGS, AeroGRID, IGN, and the GIS User
Community
0^1 and
Macomb
Washtei
Monroe .
Hillsdale
0 15
30
60
Lenawee
Lucas
ulton
Source: US Census Bureau, Esri, DigitalGlobe,
GeoEye, Earthstar Geographies, CNES/Airbus DS,
USDA, USGS, AeroGRID, IGN, and the GIS User
Community
Legend:
jjf1 Study Area Center Meteorological Stations ~ Modeled Counties
Population Density per Square Mile
0 - 600 2,000 - 3,700 I
600 - 2,000 3,700- 6,200
6,200- 116,000
Population Density - Year 2010
Detroit Study Area
Gratiot
Ionia Clinton
Eaton
Saginaw
Shiawassee
Tuscola
Ingham
Genesee
Livingston
Lapeer
Sanilac
St. Clair
Calhoun
I
Jackson
Figure 3D-3. County boundaries, census tract population densities, and meteorological
stations in the Dallas (top) and Detroit (bottom) study areas.
3D-38
-------
Union
(Hunterdon Somerset
Middlesex
Bucks
Monmouth
Mercer
lontgomery
0ti ester
lelawa
Burlington
Ocean
[oucester
Salem
Atlantic
imberland
;Carolini
Kilometers
Schuylkill
Dauphin
Lebanon
-
Lancaster
Baltimore
N
Harford \
—f
k
Anne Arundel
V f
SussexP
¦ Source: US Census Bureau, Esri, DigitalGlobe, GeoEye, Earthstar Geographies,
¦CNES/Airbus DS, USDA, USGS, AeroGRID, IGN, and the GIS User Community
Population Density - Year 2010
Philadelphia Study Area
Legend:
[Jf1 Study Area Center ~ Meteorological Stations Modeled Counties
Population Density per Square Mile
0 - 600 2,000 - 3,700 ^
600 - 2,000 3.700- 6,200
6,200 - 116.000
Population Density - Year 2010
Phoenix Study Area
Navajo
Yavapa
Gila
Maricopa
Pinal
0 15 30 60
Pima
Source: US Census Bureau, Esri, DigitalGlobe. GeoEye, Earthstar Geographies,
CNES/Airbus DS, USDA, USGS, AeroGRID, IGN, and the GIS User Community
Leqend:
Study Area Center Meteorological Stations |__J Modeled Counties
Population Densitv per Sauare Mile
0 - 600 2,000 - 3,700 6,200 - 116,000
600 - 2,000 3,700- 6,200
4-
Figure 3D-4. County boundaries, census tract population densities, and meteorological
stations in the Philadelphia (top) and Phoenix (bottom) study areas.
3D-39
-------
Washoe
Nevada
Placer
Sutter
El Dorado
ramento
l Kilometers
Population Density - Year 2010
Sacramento Study Area
Population Density per Square Mile
0 - 600 mi 2,000 - 3,700 |
600 - 2,000 3,700- 6,200
6,200- 116,000
Legend:
¦fj}!1 Study Area Center it Meteorological Stations ~ Modeled Counties
Glenn
Sonoma
Solano
Sierra
Colusa
^¦Source: US Census Bureau, Esri, DigitalGlobe, GeoEye, Earthstar Geographies,
¦CNES/Aittus DS, USDA, USGS, AeroGRID, IGN, and the GIS User Community
Amador
San Joaquin
Calaveras
Tuolumne
Population Density - Year 2010
St. Louis Study Area
Christian j
Shelby
Macoupin
Audrain
Montgomery
Effingham
Jersey
Lmco n
Fayette
Bond
Madison
St. Charles *j ^
St. Louis
Warren
Marion
Clinton
** •»«! Twr
St. Clair
Franklin
Jefferson
Monroe
Jefferson
Randolph
Franklin
Crawford Washington
El
Ste. Genevieve
Jackson
Saline
St. Francois
Kilometers
Source: US Census Bureau, Esri, DigitalGlobe. GeoEye, Earthstar Geographies,
CNES/Airbus DS, USDA, USGS. AeroGRID, IGN, and the GIS User Community
Legend:
Study Area Center Meteorological Stations Modeled Counties
Population Density per Square Mile
0-600 ^^¦ 2,000-3,700 6.200 - 116,000
600 - 2,000 3.700- 6.200
Figure 3D-5. County boundaries, census tract population densities, and meteorological
stations in the Sacramento (top) and St. Louis (bottom) study areas.
3D-40
-------
3D.2.3.2 Ambient Air Monitoring Data
We used hourly O3 concentrations from ambient air monitors in each study area for the
2015-2017 period to develop the air quality surface used for estimating exposure and risk (Table
3D-6; details in PA, Appendix 3C, section 3C.3).41 Design values for monitors in each study area
were used to determine the direction and magnitude of adjustments needed to just meet the
current standard and the other two air quality scenarios (section 3D.2.3.3). The two other air
quality scenarios are O3 concentrations for which the highest design value in the area is just
above or just below the current standard level: 75 ppb and 65 ppb. Ambient air monitors outside
each study area, but within 50 km, were also used to improve spatial interpolation of air quality
near the edges of the study areas (section 3D.2.3.4). All available ambient air O3 monitor data
were used to develop the adjusted air quality surfaces, however design values were not
calculated for monitors having incomplete data.
41 Briefly, hourly O3 concentration data for all U.S. monitoring sites for 2015-2017 were retrieved from the EPA's
Air Quality System (AQS) database. Monitors within the CSA boundary for each urban study area were identified
and used to determine the NOx emissions changes necessary to meet the air quality scenarios of interest (section
3D.2.3.3). Monitors within 50 km of the CSA boundary were identified to provide additional data for spatial
interpolation (section 3D.2.3.4).
3D-41
-------
Table 3D-6. List of ambient air monitor IDs, range of O3 design values, and number of
monitors in each study area.
Study Area
State: Ambient Air Monitor IDs A
O3 Design
Values (ppb)
(# of
monitors)
Atlanta
GA:130590002,130670003,130770002,130850001,130890002,130970004,
131210055,131350002,131510002, 132230003,132319991,132470001
63-75
(12)
Boston
CT: 090159991 MA: 250010002, 250051004, 250051006, 250092006,
250094005, 250095005, 250170009, 250213003, 250230005, 250250042,
250270015, 250270024 NH:330012004, 330111011, 330115001,330131007,
330150014, 330150016, 330150018 Rl: 440030002, 440071010, 440090007
59-73
(23)
Dallas
OK: 400130380TX: 480850005, 481130069, 481130075, 481130087,
481210034, 481211032, 481390016, 481391044, 482210001, 482311006,
482510003, 482570005, 483491051, 483670081, 483970001, 484390075,
484391002, 484392003, 484393009, 484393011
61 -79
(21)
Detroit
Ml: 260490021, 260492001, 260910007, 260990009, 260991003, 261250001,
261470005,261610008, 261619991, 261630001, 261630019, 261630093,
261630094
66-73
(13)
Philadelphia
DE: 100010002,100031007,100031010,100031013,100032004 MD:
240150003 NJ: 340010006, 340070002, 340071001, 340110007, 340150002
PA: 420110006, 420110011, 421010004, 420170012, 420290100, 420450002,
420910013, 421010024, 421010048
64-80
(20)
Phoenix
AZ:040130019, 040131003, 040131004, 040131010, 040132001, 040132005,
040133002, 040133003, 040134003, 040134004, 040134005, 040134008,
040134010, 040134011, 040135100, 040137003, 040137020, 040137021,
040137022, 040137024, 040139508, 040139702, 040139704, 040139706,
040139997, 040213001, 040213003, 040213007, 040217001, 040218001
63-76
(30)
Sacramento
CA: 060170010, 060170012, 060170020, 060570005, 060570007, 060610003,
060610004, 060610006, 060611004, 060612002, 060670002, 060670006,
060670010, 060670011, 060670012, 060670014, 060675003, 061010003,
061010004, 061130004, 061131003
63-86
(21)
St. Louis
IL: 170830117, 170831001,171170002,171190008,171191009,171193007,
171199991,171630010 MO: 290990019, 291130003, 291130004, 291831002,
291831004, 291890005, 291890014, 295100085
65-72
(16)
A Bold font indicates monitor(s) design value used to adjust ambient air concentrations to just meet selected air quality scenarios. From PA,
Appendix 3C, Tables 3C-20 to 3C-27. Italic/b/tf indicates monitor did not meet completeness criteria to calculate a design value.
3D.2.3.3 Model Adjusted Concentrations at Monitor Locations to Represent Air
Quality Scenarios
Details of the approach used to develop the three air quality scenarios (design values of
70, 65 and 75 ppb) are provided in the PA, Appendix 3C, sections 3C.4 and 3C.5. Briefly, the
ambient air concentrations described above in section 3D.2.3.2 were adjusted to just meet the
current standard (70 ppb, annual 4th highest daily maximum 8-hr average concentration,
3D-42
-------
averaged over a 3-year period) and two other air quality scenarios (75 and 65 ppb, annual 4th
highest daily maximum 8-hr average concentration, averaged over a 3-year period)42 using a
model-based O3 methodology that adjusts the observed hourly O3 concentrations to reflect the
expected spatially and temporally varying impacts of changes in NOx emissions. The
methodology is similar to that used for the 2014 HREA and employs a photochemical air quality
model combined with a tool that calculates modeled sensitivities of O3 to precursor emission
changes.
For the current analysis, the Comprehensive Air Quality Model with Extensions
(CAMx)43 served as the chemical transport model,44 with 2016 selected as the base year for
determining the adjustments needed for the 2015-2017 ambient air monitoring data. Model
inputs include meteorological data,45 emissions,46 and initial and boundary conditions.47 The
evaluation of modeled versus observed O3 concentrations for 2016 indicated CAMx generally
reproduced the observed spatial and temporal patterns, with the exception of concentration
underestimates occurring in winter across almost all regions (PA, Appendix 3C, section 3C.4.2).
The CAMx model was instrumented with the Higher order Decoupled Direct Method
(HDDM) to calculate modeled nonlinear sensitivities of O3 to emission changes (PA, Appendix
3C, section 3C.5). The photochemical modeling outputs included both modeled O3
concentrations and sensitivities of O3 concentrations to changes in NOx emissions for each hour
in a single year at all ambient air monitor locations (Appendix 3C, sections 3C.4 and 3C.5).
Linear regression was used with these single-year 2106 model outputs to create relationships
between the sensitivities and O3 concentrations for each hour of each of the four seasons at each
monitoring location. The relationships between hourly sensitivities and hourly O3 for each
season were then used with three years of ambient air monitoring data at each location to predict
42 In these scenarios, the air quality conditions were adjusted such that the monitor location with the highest
concentrations in each area had a design value just equal to either 75 ppb or 65 ppb.
43 The Comprehensive Air Quality Model with Extensions and associated documentation is found at
www.camx.com.
44 The 2014 HREA used the Community Multiscale Air Quality Modeling System (CMAQ) to model air quality.
45 Horizontal wind components (i.e., speed and direction), temperature, moisture, vertical diffusion rates, and rainfall
rates for each 12 Km grid cell in each vertical layer was derived from version 3.8 of the Weather Research and
Forecasting Model (WRF; http://wrf-model.org). For details, see PA, Appendix 3C, section 3C.4.1.4.
46 Emissions from electric generating units, other point sources, area sources, agricultural sources (ammonia only),
anthropogenic fugitive dust sources, nonroad mobile sources, onroad mobile sources, and biogenic sources are
based on the alpha version of the Inventory Collaborative 2016 emissions modeling platform
(,http://views.cira.colostate.edu/wiki/wiki/9169). For details, see PA, Appendix 3C, section 3C.4.1.5.
47 Initial and lateral boundary concentrations for the 12 km domain are provided by the hemispheric version of the
Community Multi-scale Air Quality model (H-CMAQ) v5.2.1. The H-CMAQ model was run for 2016 with a
horizontal grid resolution of 108 km and 44 vertical layers up to 50 hPa. For details, see PA, Appendix 3C,
section 3C.4.1.6.
3D-43
-------
hourly sensitivities for the complete 3-year record at each monitoring location. From these, we
calculated hourly O3 concentrations at each monitor location based on iteratively increasing NOx
reductions to determine the adjustments necessary for the monitor location with the highest
design value in each study area to just meet the target value, e.g., 70 ppb for the current standard
scenario (Appendix 3C, section 3C.5). For the 75 ppb air quality scenario, we note that three
areas required an increase in NOx emissions as their highest O3 design values were below 75
ppb. For the other five study areas and that same air quality scenario and for all study areas with
the other two air quality scenarios (i.e., 65 and 70 ppb), emission reductions were required
(Table 3D-7).
Table 3D-7. Range of the percent NOx emission changes needed to adjust air quality in the
eight study areas for the three air quality scenarios.
Design Value for
each Air Quality
Scenario
Range of NOx Emission Changes
Applied Across the Eight Study Areas
75 ppb
+18% to-45%
70 ppb
-13% to -58%
65 ppb
-38% to -72%
From PA, Appendix 3C, Table 3C-19.
3D.2.3.4 Interpolation of Adjusted Monitor Concentrations to the Census Tracts
Comprising Each Study Area
As described above, model-based relationships between O3 and NOx emissions were
used to adjust hourly O3 concentrations at the ambient air monitor locations (section 3D.2.3.2) to
represent conditions in which the study area just meets the selected air quality scenario (section
3D.2.3.3). Simulated O3 concentrations were then needed at a finer spatial scale than that given
by the monitor sites to better represent the spatial heterogeneity in O3 concentrations across
locations frequented by the simulated population (and during the times frequented) across the
study area. To accomplish this in each of the eight study areas, the adjusted hourly O3
concentrations at monitoring sites were interpolated to census tract centroids using the Voronoi
Neighbor Averaging (VNA; PA, Appendix 3C, section 3C.6). Nearby monitoring concentrations,
for each hour, inform the estimation of O3 for a given census tract using inverse distance
weighting. In so doing, both spatial and temporal gaps in the desired air quality surface are filled
simultaneously, resulting in a final dataset of ambient air O3 concentration estimates with high
temporal and spatial resolution (hourly concentrations in 500 to 1700 census tracts) for each of
the eight study areas and for years 2015 to 2017 (Appendix 3C, section 3C.7).
3D-44
-------
3D.2.3.5 Evaluation of Temporal and Spatial Characteristics of the Simulated Air
Quality Surfaces
We applied the above described approaches to simulate air quality surfaces that represent
fine-scale temporal (i.e., hourly) and spatial (i.e., census tract) variability in O3 concentrations
for the three air quality scenarios in each study area. Then, characteristics of the simulated air
quality surfaces were evaluated for trends and patterns that would be informative for interpreting
the simulated exposure and risk results. For example, Figure 3D-6 illustrates the temporal
variability across the three years of monitoring data, stratified by hour-of-day (left panel) and
month (right panel), in Philadelphia for the ambient air measurements, and for the three
simulated air quality scenarios (following the model-based adjustment at each monitor location).
Philadelphia sites: 2015-2017
• observed
¦ 75 ppb standard
• 70 ppb standard
¦ 65 ppb standard
0 e
0 2 4 6 8 10 12 14 16 18 20 22
hour
Philadelphia sites: 2015-2017
a
Q
Sg
8
| ot»erv«d
1 75 opb sttmdRrd
' 70 ppb standard
65 ppb standard
« 8
I »
2
i
Billlll
2 3 4 5 6 7
month
10 11 12
Figure 3D-6. Hourly O3 distributions by hour-of-day (left panel) and month (right panel) at
ambient air monitoring sites in Philadelphia for observed air quality (black),
air quality adjusted to meet the current standard (70 ppb, blue) and two other
design values (75 ppb, red; and 65 ppb, green). From PA, Appendix 3C, Figures
3C-71 and 3C-79, respectively.
The diurnal and seasonal temporal patterns for the three air quality scenarios are similar
to the monitor observations, with highest O3 concentrations during the during late
morning/afternoon hours and during spring/summer months. In addition, the upper end of the O3
concentration distributions decrease from observed values (black) to values adjusted to meet the
current standard of 70 ppb (blue) and decrease further when adjusted to meet a design value of
65 ppb (green). These decreases can be seen when evaluating the highest O3 hours-of-the day
and represented by the data points that extend beyond the whiskers of the boxplots. Further, the
3D-45
-------
overall pattern flattens when decreasing the level of the O3 standard, considering both the diurnal
and monthly distributions. Regarding the diurnal pattern, O3 increases during early morning
hours are associated with VOC-limited and NOx titration conditions near NOx sources during
rush-hour periods. Lower O3 concentrations in the winter months result from lower solar
insolation rates and a reduction in total photochemical activity. See PA, Appendix 3C (Section
3C.7.2 and Figures 3C-67 through 3C-82) for details for temporal characteristics of all eight
study areas.
We also evaluated the hourly O3 concentrations by considering the overall shape of the
concentration distribution using the census-tract resolution interpolated data. Even though both
the temporal and spatial attributes may be conflated in such a presentation, a histogram can be
useful in illustrating important features of the distribution (e.g., skewness, kurtosis, upper
percentile tails) that may be influential in estimated exposures and risks. For example, Figure
3D-7 illustrates the overall shape48 of the hourly concentration distribution in each of the eight
study areas for the air quality scenario just meeting the current standard. The distribution for all
study areas are skewed to the right, generally representing a lognormal form.
There are notable differences across the collection of study areas. For example, the
distributions for Boston, Dallas, Philadelphia, and Sacramento are slender (i.e., leptokurtic),
showing much higher peaks around the mean value, relative to the other four study areas,
Atlanta, Detroit, Phoenix, and St. Louis which exhibit relatively flatter (i.e., platykurtic)
distributions, and the latter three of which, show an increased frequency of upper percentile
concentrations. Phoenix, in particular, exhibits the greatest right-most shift in the hourly O3
concentration distribution and would reflect other areas of the U.S. having a similar distribution
of ambient air O3 concentrations. Also, there are only limited instances of hourly O3
concentrations >70 ppb in all study areas for the air quality scenario just meeting the current
standard (Figure 3D-7). This is consistent with recent (unadjusted) ambient air monitoring data,
whereas hourly O3 concentrations are rarely at or above 100 ppb when design values are <70 ppb
(i.e., <0.02% frequency; see Appendix 2A, Table 2A-4). This is important to note because these
distinct features of the O3 concentration distribution, along with the spatial and temporal
intersection of concentrations with population demographics and activity patterns, play an
important role in contributing to variation in the estimated population exposures and risks
presented in section 3D.3 below.
48 Figure 3D-7 is intended to illustrate the differences in the shape of the distributions. All histograms have the exact
same range of values for the x-axis, i.e., the midpoint concentrations range from 0 to 70 ppb, in 2 ppb increments
(maximum value represents frequency of all hourly concentrations >70 ppb. Because there are varied distribution
shapes, the range of values for the y-axis differ across the study areas. The actual value of the y-axis is
unimportant in this context because of interest here are the relative differences that exist across the concentration
distributions (e.g., frequency of high O3 concentrations relative to the occurrence of low O3 concentrations).
3D-46
-------
ll
inn Atlanta
jll IL.
1 Boston
as 11! 1! 11! ], i i
n Dallas
!!11111 ill III.-
ilfll il lnrinn~„n
rat lLimil.JLl.-LL
1 Detroit
i
Philadelphia
»11!II1!II111
¦In Phoenix
UnnOrinn^n
| 1 J 1 J 1 I ! . 1 lillSnran.
-Dm
_j
Sacramento
A
L St. Louis
Frequency
~
[tlOflo^
l llllflfillflfllflilOnnonn^
Hourly Ambient O3 (0 to 70+ ppb)
Figure 3D-7. Histograms of hourly O3 concentrations (ppb, x-axis) for the air quality scenario just meeting the current O3
standard in the eight study areas. The x-axis midpoint concentrations range from 0 to 70 ppb, in 2 ppb increments
(rightmost, maximum histogram bar for all study areas represents the frequency of all hourly concentrations >70 ppb).
3D-47
-------
Regarding spatial variability, Figure 3D-8 displays census tract design values for each of
the three air quality scenarios in Philadelphia. A decline in the highest ambient air O3
concentrations is predicted across the study area when considering air quality scenarios at lower
design values.
Meeting 75 ppb
Meeting 70 ppb
Meeting 65 ppb
Figure 3D-8, Calculated design values for census tracts in the Philadelphia study area,
derived from a VNA interpolation of CAMx/HDDM adjusted O3
concentrations. Figure modified from PA, Appendix 3C, Figure 3C-99.
3D.2.4 Meteorological Data
Temperature data are used by APEX in selecting human activity data and in estimating
air exchange rates (AERs) for indoor residential microenvironments (MEs). When developing
profiles, APEX uses temperature data from the closest weather station to each Census tract.
Hourly surface temperature measurements were obtained from the National Oceanic and
Atmospheric Administration (NOAA) Integrated Surface Hourly (ISH) data files.49 The weather
stations used for each study area are given in Table 3D-8, along with general locations provided
in Figure 3D-2 to Figure 3D-5.
In general, the occurrence of missing temperature data was limited to a few hours per
year. Missing hourly temperature data were estimated by the following procedure. Where there
were consecutive strings of missing values (data gaps) of 9 or fewer hours, missing values were
estimated by linear interpolation between the observed values at the ends of the gap. Remaining
missing values at a meteorological station were estimated by fitting linear regression models for
each hour of the day, with each of the other monitors, and choosing the model which maximizes
R2, for each hour of the day, subject to the constraints that R2 be greater than 0.40 and the
number of regression data values (days) is at least 100. If there no suitable regression models to
fill the missing values, for gaps of 12 or fewer hours, missing values were estimated by linear
49 See: ftp://ftp.ncdc.noaa.gov/pub/data/noaci/isd-lite/
3D-48
-------
interpolation between the valid values at the ends of the gap. Any remaining missing values were
replaced with the value at the closest station for that hour. Because there were limited instances
of missing data, there were negligible differences between the statistically filled and the original
temperature data with missing values.
Table 3D-8. Study area meteorological stations, locations, and hours of missing data.
Study Area
Station Name
WBAN A
Latitude
Longitude
Number of hours with
missing temperature
2015
2016
2017
Atlanta
HARTSFIELD-JACKSON ATLANTA
13874
33.630
-84.442
6
4
5
FULTON CO-BROWN FLD ARPT
03888
33.779
-84.521
34
84
220
DEKALB-PEACHTREE AIRPORT
53863
33.875
-84.302
13
6
47
DOBBINS AIR RESERVE BASE
13864
33.917
-84.517
171
142
58
Boston
LAURENCE G HANSCOM FLD
14702
42.470
-71.289
55
164
19
BEVERLY MUNICIPAL AIRPORT
54733
42.584
-70.918
56
8
7
GEN E L LOGAN INTERNATIONAL
14739
42.361
-71.010
5
4
5
NORWOOD MEMORIAL AIRPORT
54704
42.191
-71.174
17
38
17
Dallas
DALLAS LOVE FIELD AIRPORT
13960
32.852
-96.856
5
5
5
DALLAS/FT WORTH INTERN AT
03927
32.898
-97.019
5
5
5
DALLAS EXECUTIVE AIRPORT
03971
32.681
-96.868
27
14
36
Detroit
DETROIT METRO WAYNE COUNTY
94847
42.231
-83.331
462
547
619
GROSSE ILE MUNICIPAL AIRPORT
54819
42.099
-83.161
484
397
44
DETROIT CITY AIRPORT
14822
42.409
-83.010
25
22
69
OAKLAND CO. INTNL AIRPORT
94817
42.665
-83.418
16
11
17
Philadelphia
WINGS FIELD AIRPORT
64752
40.100
-75.267
150
241
324
SOUTH JERSEY REGIONAL ARPT
93780
39.941
-74.841
na
90
69
PHILADELPHIA INTERNATIONAL
13739
39.873
-75.227
5
6
5
NE PHILADELPHIA AIRPORT
94732
40.079
-75.013
28
13
51
Phoenix
PHOENIX SKY HARBOR INTL
23183
33.428
-112.004
13
8
6
SCOTTSDALE AIRPORT
03192
33.623
-111.911
9
19
10
Sacramento
SACRAMENTO EXECUTIVE
23232
38.507
-121.495
10
21
87
SACRAMENTO MCCLELLAN AFB
23208
38.667
-121.400
366
368
89
SACRAMENTO INTL AIRPORT
93225
38.696
-121.590
28
53
41
St. Louis
SCOTT AIR FORCE BASE/MIDAMER
13802
38.550
-89.850
110
49
45
LAMBERT-ST LOUIS INTERNAT
13994
38.753
-90.374
11
7
7
ST LOUIS DOWNTOWN AIRPORT
03960
38.571
-90.157
12
49
7
A Weather Bureau Army Navy (WBAN) number of the meteorological stations,
"na" is no data available
3D-49
-------
Multiple unique APEX input files are used for the current exposure and risk analyses, one
for each year and study area, and in the following two formats:
• METdataCSA[number]Y[year].txt\ meteorological station IDs, hour of day, hourly
temperature (°F) for each meteorological station, by study area and year
• METlocsCSA [number]Y[year].txt: meteorological station IDs, latitudes and
longitudes, start and stop dates of temperature data
3D.2.5 Construction of Human Activity Pattern Sequences
Exposure models use human activity pattern data to estimate exposure to pollutants.
Different human activities, such as outdoor exercise, indoor reading, or driving a motor vehicle
can lead to different pollutant exposures, intakes and doses. This may be due to differences in the
pollutant concentration in the varied locations where different activities are performed as well as
to differences in the energy expended in performing the activities (because energy expended
influences inhalation and thus may influence pollutant intake). To model exposures to ambient
air pollutants, it is critical to have information on the locations where people spend time and the
activities performed in such locations. The following subsections describe the activity pattern
data, population commuting data, and the approaches used to simulate where individuals might
be and what they might be doing.
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 and the activities they
perform. The following subsections describe the population data we used in the assessment to
assign key features of the simulated individuals, and approaches used to simulate the basic
physiological functions important to the exposure estimates for this exposure and risk analysis.
3D.2.5.1 Consolidated Human Activity Database
The Consolidated Human Activity Database (CHAD) provides time series data on human
activities through a database system of collected human diaries, or daily time location activity
logs (U.S. EPA, 2019c). The purpose of CHAD is to provide a basis for conducting multi-route,
multi-media exposure assessments (McCurdy, 2000). The data contained within CHAD come
from multiple surveys with variable, study-specific structure (e.g., real time minute-by-minute
recording of diary events versus a recall method using time-block-averaging). Common to all of
the peer-reviewed studies, individuals provided information on their locations visited and
activities performed for each surveyed day. Personal attribute data for the surveyed individuals,
such as age and sex, are included in CHAD and are used as variables to link to the population
data. The latest version of CHAD contains data for nearly 180,000 individual diary days. Most of
the CHAD data are from studies conducted since 2000, several of which are newly included or
3D-50
-------
updated since the 2014 HREA.50 Table 3D-9 provides the survey study information including the
geographic coverage, year, and the number of diaries available for use by APEX.51
Table 3D-9. Overview of Studies Included in the APEX Activity Data Files.
Study Name
(abbreviation)
Geographic
Coverage
Study Year
Number of Diary Days A
Age Range
Reference
Ages 5-18
Any Age
min
max
American Time
Use Survey,
Bureau of Labor
Statistics (BLS)
Entire US
2003-11
7,559
123,932
15
85
US Bureau of Labor
Statistics (2014)
Baltimore
Retirement Home
Study (BAL)
Baltimore
County, MD
1997-98
0
390
72
93
Williams et al. (2000)
California Activity
Pattern Studies
(CAA, CAC, CAY)
California
CAA: 1987-88
36
1,570
18
94
Wiley et al. (1991a),
Wiley et al. (1991b)
CAC: 1989-90
680
1,197
0
11
CAY: 1987-88
182
182
12
17
Cincinnati Activity
Patterns Study
(CIN)
Cincinnati,
OH
1985
736
2,595
0
86
Johnson (1989)
Detroit Exposure
and Aerosol
Research Study
(DEA)
Detroit, Ml
2004-2007
5
336
18
74
Williams et al. (2009)
Denver, Colorado
Personal Exposure
Study (DEN)
Denver, CO
1982-1983
7
784
18
70
Johnson (1984):
Johnson et al. (1986)
EPA Longitudinal
Studies (EPA)
Central NC
1999-2000,
2002, 2006-08,
2012-2013
0
1,780
0
72
Isaacs et al. (2013)
Los Angeles
Ozone Exposure
Study: Elementary
School/High
School (LAE, LAH)
Los
Angeles, CA
1989-1990
49
49
10
12
Roth Associates
(1988): Spier et al.
(1992)
43
43
13
17
National Human
Activity Pattern
Study (NHAPS):
Air/Water (NHA,
NHW)
48 states
1992-1994
659
4,723
0
93
Klepeis et al. (1995):
Tsang and Klepeis
(1996)
713
4,663
0
93
50 CHAD updates since the 2014 HREA include expansion of activity codes, revision to the METs distributions,
filling missing temperatures, characterizing ambiguous location entries, etc. See U.S. EPA, 2019c and
Attachment 3.
51 Following stated updates to improve the CHAD diary information, some diaries in the CHAD master database
remain unusable for exposure and risk modeling. Most commonly this is from having excessive missing or
unknown location or activity data (e.g., >3 hours/day).
3D-51
-------
Study Name
(abbreviation)
Geographic
Coverage
Study Year
Number of Diary Days A
Age Range
Reference
Ages 5-18
Any Age
min
max
Population Study
of Income
Dynamics PSID 1,
II, III (ISR)
Whole US
1:1997
3,302
5,327
0
13
University of
Michigan, 2016
II: 2002-2003
4,816
4,825
5
19
111:2007-2008
2,633
2,690
10
19
National-scale
Activity Study
(NSA)
7 US metro
areas
2009
0
6,820
35
92
Knowledge Networks
(2009)
RTI Ozone
Averting Behavior
Study (OAB)
35 US metro
areas
2002-2003
1,941
2,872
2
12
Mansfield et al.
(2009)
RTP Particulate
Matter Panel Study
(RTP)
Wake and
Orange
Counties,
NC
2000-2001
0
874
55
85
(Williams et al.,
2003a, 2003b),
Williams et al., 2001
Study of Use of
Products and
Exposure-related
Behaviors (SUP)
California
2006-2010
1,293
8,831
1
88
Bennett et al. (2012)
Seattle Study
(SEA)
Seattle, WA
1999-2001
317
1,645
6
91
Liu et al. (2003)
Valdez Air Health
Study (VAL)
Valdez, AK
1990-1991
72
387
11
71
Goldstein et al.
(1992)
Washington, DC
Study (WAS)
Washington,
DC
1982-1983
11
695
18
98
Hartwell et al. (1984):
Johnson et al.
(1986): Settergren et
al. (1984)
All Studies, Areas, and Years (TOTAL):
25,054
177,210
0
98
A The APEX activity data file differs from that of the CHAD master database by removing what are considered as unusable diaries for our
exposure and risk analyses (-2,000 diary days). The four criteria used to screen the CHAD master database are as follows: 1) Daily
maximum temperature is missing, 2) daily average temperature is missing, 3) the day-of-week is missing, and 4) at least 3 hours of events
have activity or location codes of "unknown" and/or "missing".
Three standard APEX input files are used for the current exposure and risk analyses to
create the activity pattern profiles for all simulated individuals.
• CHADEvents 060419A.txt: CHAD ID, clock hour (hhmm), duration of event
(minutes), CHAD activity code, and CHAD location code, serving as a daily
sequence of locations visited, activities performed, and their duration
• CHADQuest 060419A.txt. CHAD ID, day-of-week, sex, race, employment status,
age, maximum daily temperature, average temperature, occupation, missing time
(minutes), record count, commute time (see also section 3D.2.5.2)
• CHADSTATSOutdoor 060419A.txt: CHAD ID, total daily time spent outdoors
(minutes) (see also section 3D.2.5.4)
3D-52
-------
3D.2.5.2 Commuting and Employment Data
Exposures can vary across a study area based on spatial heterogeneity in ambient air
concentrations and how that corresponds with a simulated individual's activity pattern and
geographic location. APEX approximates home-to-work commuting flows between census
designated areas for each employed individual, and thus accounts for differing ambient air
concentrations that may occur in these geographic locations. APEX has a national commuting
database originally derived from 2010 Census tract level data collected as part of the U.S. DOT
Census Transportation Planning Package. The data used to generate the APEX commuting file
are from the "Part 3-The Journey to Work" files. The Census files contain counts of individuals
commuting from home to work locations at a number of geographic scales. These data have been
processed to calculate fractions (and hence commute probabilities) for each tract-to-tract flow to
create the national commuting data distributed with APEX. This database contains commuting
data for each of the 50 states and Washington, D.C. This dataset does not differentiate people
that work at home from those that commute within their home tract. A companion file to the
commuting flow file is the commuting times file, i.e., an estimate of the usual amount of time in
minutes it takes for commuters to get from home to work each day and tract-to-tract commuting
distances. The commuting times file information is used to select CHAD activity pattern data
from individuals having time spent inside vehicles similar to the census commute times and
associated distances travelled. Two standard APEX input files are used for the current exposure
and risk analysis, as listed here.
• Commuting times US 2010.txt: census block IDs, count of all employed
individuals, count of employed individuals that do not work at home, 7 groups of
block-level one-way commuting times (in minutes)
• Commuting flow US 20I0.txt: census tract IDs, tract-to-tract commute cumulative
probabilities (in fractional form), commute distance (km)
Another population-based file associated with commuting is the employment file. This
APEX input file contains the probability of employment separately for males and females by age
group (starting at age 16) and by census tract (the only census unit available for this type of
data). The 2010 Census collected basic population counts and other data using the short form but
collected more detailed socioeconomic data (including employed persons) from a relatively
small subset of people using the 5-year American Community Survey (ACS).52 The ACS dataset
52 2010 U.S. Census American FactFinder: http://factfinder2.census.gov/. For instance, to obtain the table ID
B23001 "Sex by age by employment status for the population 16 years and over", the following steps were
performed. First, select the "guided search option", choose "information about people" and select "employment
3D-53
-------
provides the number of people in the labor force, which were stratified by sex/age/tract,
considering both civilian workers and workers in the Armed Forces. The data were stratified by
sex and age group and were processed so that each sex-age group combination is given an
employment probability fraction (ranging from 0 to 1) within each census tract. Children under
16 years of age were assumed to be unemployed. One national-based APEX input file is used for
the current exposure and risk analyses as follows:
• Employment US 2010.txt: census tract IDs, employment probabilities (in
fractional form), stratified by 13 age groups.53
3D.2.5.3 Assignment of Activity Pattern Data to Individuals
Once APEX identifies the basic personal attributes of a simulated individual (section
3D.2.2) and daily air temperatures (section 3D.2.4), activity pattern data obtained from CHAD
(section 3D.2.5.1) are then selected based on age, sex, temperature category, and day of the
week. These attributes are considered first-order attributes in selecting CHAD diaries when
modeling human exposures (Graham and McCurdy, 2004). The particular locations people visit,
amount of time spent there, and frequency of these visits can also be influenced by local weather
conditions. When considering seasonal temperature ranges (i.e., cold/not cold during cool
months; hot/not hot during warm months), (Graham and McCurdy, 2004) found daily maximum
temperature (DMT) influences time spent outdoors. Participation rate and amount of time
outdoors was found lower on cold DMT days compared to the other three temperature
categories, while the participation rate on hot days was less than that on not hot days. Because of
these findings, we use a similar DMT range (<55, 55-83, >84 °F) to select activity pattern data
that best match each study area's meteorological data for every day of the simulated individual's
exposure profile. This information for the selecting of activity pattern data is found in the
following APEX input file, varying by study area and simulation year:
• Functions O3 CSA[number] 040219.txt: probabilities and interval definitions
associated with a few input variables. For activity diary selection - day of week
intervals (weekend or weekday) by three temperature ranges.
While there may be other important attributes that may influence activity patterns (e.g.,
obesity, disease status), there are limits to our ability to link to all the possible personal attributes
(labor force) status", "sex" and "age". For geography type select "census tract -140" for each state. Tables
containing the employment numbers were downloaded and used to calculate the employment probabilities for
each age group.
53 The age groups in this file are: 16-19, 20-21, 22-24, 25-29, 30-34, 35-44, 45-54, 55-59, 60-61, 62-64, 65-69, 70-
74, and >75.
3D-54
-------
that may be of interest in modeling an individual's activities to the CHAD data. This is largely
because CHAD is a compilation of data collected from numerous individual activity pattern
studies conducted over several decades, many of which had a unique survey design. As a result,
there is a varying amount of missing personal attribute data for the surveyed individuals in
CHAD. For instance, there are only a limited number of CHAD diaries with survey-requested
health information (e.g., the health status of respondents). Specifically regarding whether or not a
survey participant had asthma, very few of the available diaries have either a 'yes' or 'no'
response to this health condition. When considering the 177,210 diary days used by APEX, there
are only 4,935 diary days from individuals having asthma (of which 3,133 are children ages 5-
18),54 representing a small fraction of the CHAD data. On its own, having approximately 5,000
diaries may appear to be a large number of diaries, however, following a grouping of the diaries
by their first-order attributes when developing simulated profiles (e.g., age, sex, day-of-week,
etc., daily temperature), would likely result in fewer than 100 diaries available for simulating a
single day for a particular individual. Accordingly, the selection of diaries to use for APEX-
simulated individuals does not consider health status (i.e., any diary is used, regardless of
whether the individual indicated they did or did not have asthma, or that information was
unknown).
This restriction in the number of diaries from individuals having asthma is not considered
to be a significant limitation for estimating exposures for simulated individuals with asthma. In
general, modeling people with asthma similarly to healthy individuals (i.e., using the same time-
location-activity profiles) is supported by the activity analyses reported by van Gent et al. (2007)
and Santuz et al. (1997). Other researchers, for example, Ford et al. (2003), have shown
significantly lower leisure time activity levels in asthmatics when compared with individuals
who have never had asthma. Based on these inconsistent findings, we evaluated this issue in the
2014 HREA and, using the available activity pattern data in the CHAD database, we compared
participation in afternoon outdoor activities at elevated exertion levels among people having
asthma, people not having asthma, and unknown health status (2014 HREA, Appendix G,
section 5G-1.4). The 2014 HREA analysis indicated health status had little to no impact on the
participation in afternoon activities at elevated exertion levels. A similar analysis was repeated
here to include the diary data currently used by APEX, not just those that would be included in
the simulations for the 2014 HREA (i.e. -50,000 diaries).
Of interest in this current risk and exposure analysis are instances when individuals
experience their highest O3 exposures. As shown in 2014 HREA, the highest exposures occur
54 The American Time Use Survey, a study contributing the largest number of diaries (n=124,517) to CHAD, did not
include a question for whether a surveyed individual has asthma.
3D-55
-------
when individuals spend time outdoors, particularly during the afternoon hours (2014 HREA,
Appendix 5G section 5G-2). To prepare the APEX activity dataset for analysis here, afternoon
hours were characterized as the time between 12 PM and 8 PM and only those persons that spent
some time outdoors were retained. As is done by APEX in simulating individuals, level of
exertion was estimated by sampling from the specific METs distributions assigned for each
person's activity performed. Then, we identified activities having a METs value of greater than
three as instances where a person was at moderate or greater exertion levels (U.S. DHHS, 1999).
Afternoon outdoor time was then stratified by exertion level, summed for two study groups of
interest (i.e., children and adults), and presented in percent form within Table 3D-10.
Regarding the diaries for children of interest for these exposure and risk analyses (ages 5-
18), about 13% are from an individual having asthma, 48% are from those who do not have
asthma, and the remaining portion of children's diaries have unknown health status. About 1% of
CHAD diaries for adults are from individuals with asthma and about 11% are from those who do
not have asthma. Far fewer children's diaries are from persons whose asthma status is unknown
(40%) compared to adults (88%), and the proportions are smaller still in terms of the total
available person-days. On average, about 42% of all children having known asthma status spent
some afternoon time outdoors, and the percent is actually higher for children with asthma
(48.4%>) than for children not having asthma (40.5%). About half of the adults whose asthma
status was known spent afternoon time outdoors with a participation rate generally similar for
adults having asthma and adults not having asthma. Participation in outdoor events for children
having unknown asthma status varied little from that of persons with known asthma status.
Contrary to this, there were fewer adults with unknown asthma status that participated in outdoor
events (29%) when compared to those having known asthma status.
The amount of afternoon time spent outdoors by the persons that did so varied little
across the two study groups and two asthma classifications. Children, on average, spend
approximately 2XA hours of afternoon time outdoors, 80% of which is at a moderate or greater
exertion level, regardless of their asthma status. For children whose asthma status is unknown,
slightly more afternoon time is spent outdoors (about 150 minutes) but the percent of afternoon
time at moderate or greater exertion levels is slightly lower (about 69%). As seen with children,
adults spend approximately 7}A hours of afternoon time outdoors regardless of their asthma
status. However, the percent of afternoon time at moderate or greater exertion levels for adults
(about 55%>) is lower than that observed for children.
Based on this updated analysis and additional comparisons of CHAD diary days with
literature reported values of outdoor time participation at varying activity levels (see 2014
HREA), there are strong similarities in outdoor time, outdoor event participation, and activity
levels achieved among the two study groups and with those reported in independent studies of
3D-56
-------
people with asthma. Thus, we conclude the use of any CHAD diary, regardless of
known/unknown asthma status, is reasonable for purposes of simulating people with asthma in
this exposure and risk analysis.
Table 3D-10. Comparison of time spent outdoors and exertion level by asthma status for
children and adult diaries used by APEX.
CHAD: Children (5 to 18)A
CHAD: Adults (>18) B
Has Asthma?
Yes
No
Unknown
Yes
No
Unknown
Total Person Days (n)
3,133
11,948
9,973
1,279
16,323
127,377
Number of Person Days with Time
Spent Outdoors (% participation)
1,517
(48.4%)
4,840
(40.5%)
4,054
(40.6%)
569
(44.5%)
7,900
(48.4%)
36,949
(29.0%)
Overall Percent of Afternoon Hours
Spent Outdoors (%)
29.0%
27.3%
31.8%
28.3%
28.9%
27.2%
Overall Percent of Afternoon Time
Outdoors at Moderate or Greater
Exertion (%)
81.6%
81.1%
69.1%
55.4%
55.1%
62.3%
A CHAD studies for where a survey questionnaire response of whether or not child had asthma include CIN, ISR, NHA, NHW, OAB, and
SEA (see Table 3D-9 for study names).
B CHAD studies for where survey a questionnaire response of whether or not adult had asthma include CIN, EPA, ISR, NHA, NHW, NSA,
and SEA.
We also evaluated how temperature influences the amount of afternoon time spent
outdoors while at moderate or greater exertion by children (5-18 years) and adults (19-90 years).
This differs from analyses in Graham and McCurdy (2004) in which all outdoor time at any
exertion level was evaluated and the number of diary days available in CHAD was much less at
that time (-23,000 diary days). Also, in this current analysis, each CHAD/APEX diary day was
grouped by both DMT (<55, 55-83, or >84 °F) and day-type (weekday or weekend). Total
available diary days for each of these groups is provided in Table 3D-11. Then, afternoon time
outdoors (12:00 PM to 8:00 PM) was summed and place into one of five hourly groupings (0, 0-
4 hours per day) and the percent of diary days in each group was
calculated, the results of which are provided in Figure 3D-9 for children and adults.
Overall, the greatest proportion of diary days would be characterized as not having any
afternoon time spent outdoors at moderate or greater exertion (46 - 76%), with adults
consistently having a greater frequency of not spending afternoon time outdoors than children
(Figure 3D-9). Afternoon time outdoors at moderate or greater exertion for both children and
adults is less likely to occur on cold days (DMT <55 °F), with progressively increased frequency
of outdoor time with increasing temperatures for both day-types. Children are more frequently
spending afternoon time outdoors at elevated exertion levels, particularly when considering the
largest duration assessed (e.g., for durations of time outdoors >2 hours, the percent of child diary
days is greater than adults by a factor of 1.3 to 2.7).
3D-57
-------
Table 3D-11. Number of diary days in CHAD for children and adults, grouped by
temperature and day-type categories.
Daily
Maximum
Temperature
(°F)
Children (5-18 years)
Diary Days (n)
Adult (19-90 years)
Diary Days (n)
Weekday
Weekend
Weekday
Weekend
<55
3,888
3,504
19,316
17,136
55-83
6,823
5,800
36,034
32,982
>84
3,460
1,584
23,865
15,646
The number of diary days here can be used along with Figure 3D-9 to estimate the number of
diaries available in each time/hour group. The total number of diary days for this analysis is
170,033 and differs from CHAD/APEX (n=177,210) because of the age range selected.
Children (ages 5-18), Weekdays Adults (ages 19+), Weekdays
0 £ 1/2 1/2-2 2-4 >4
Afternoon Time Outdoors at Moderate or Greater Exertion (hours)
Adults (ages 19+), Weekends
0 £ 1/2 1/2-2 2-4 >4
Afternoon Time Outdoors at Moderate or Greater Exertion (hours)
Figure 3D-9. Percent of children (5-18 years) and adults (19-90 years) having afternoon
time outdoors while at moderate or greater exertion, categorized by daily
maximum temperature (°F) and time (hours/day) groups.
3D-58
<55
0 £1/2 1/2-2 2-4 >4
Afternoon Time Outdoors at Moderate or Greater Exertion (hours)
Children (ages 5-18), Weekends
o 40%
<55
0 £ 1/2 1/2-2 2-4 >4
Afternoon Time Outdoors at Moderate or Greater Exertion (hours)
-------
3D.2.5.4 Method for Longitudinal Activity Pattern Sequence
In order to estimate population exposure over a full year, a year-long activity sequence
needed to be created for each simulated individual based on CHAD, which is largely a cross-
sectional activity database of 24-hr records. On average, the typical surveyed subject provided in
CHAD has about two days of diary data. For this reason, the construction of a season-long
activity sequence for each individual requires some combination of repeating the same data from
one subject and using data from multiple subjects. The best approach would reasonably account
for the day-to-day and week-to-week repetition of activities common to individuals and
recognizing even these diary sequences are not entirely correlated, while maintaining realistic
variability among individuals comprising each study group.
APEX provides three methods of assembling composite diaries: a basic method, a
diversity and autocorrelation (D&A) method, and a Markov-chain clustering (MCC) approach.
We have selected the diversity and autocorrelation (D&A) method for this assessment based on
our consideration of the assessment objectives, an evaluation of differences in results produced
by the three methods, and consideration of flexibility provided by each approach with regard to
specifying key variable values, as discussed below. First a brief description of each method is
provided below.
The basic method involves randomly selecting an activity diary for the simulated
individual from a user-defined diary pool (e.g., age, sex). While the method is adequate for
estimating a mean short-term exposure for a population as a whole, it is less useful for estimating
how often individuals in a population may experience peak O3 exposures over a year.
The D&A method is a complex algorithm for assembling longitudinal diaries that
attempts to realistically simulate day-to-day (within-person correlations) and between-person
variation in activity patterns (and thus their exposures to the extent they are influenced by spatial
and temporal variability in ambient air and microenvironmental O3 concentrations). This method
was designed to capture the tendency of individuals to repeat activities, based on reproducing
realistic variation in a key diary variable, which is a user selected function of diary variables. The
method targets two statistics: a population diversity statistic (D) and a within-person
autocorrelation statistic (A). The D statistic reflects the relative importance of within and
between-person variance in the key variable. The A statistic quantifies the lag-one (day-to-day)
key variable autocorrelation. Values of D and A for the key variable are selected by the model
user and set in the APEX parameters file, and the method algorithm constructs longitudinal
diaries that preserve these parameters. Further details regarding this methodology can be found
in Glen et al. (2008).
The Markov-chain clustering (MCC) approach is similarly complex in attempting to
recreate realistic patterns of day-to-day variability. First, cluster analysis is employed to divide
3D-59
-------
the daily activity pattern records into three groups based on time spent in, for example, five
microenvironments: indoor-residence, other indoors, outdoor-near roads, other outdoors, and
inside vehicles. For each simulated individual, a single time-activity record is randomly selected
from each cluster. Then the Markov process determines the probability of a given time-activity
pattern occurring on a given day based on the time-activity pattern of the previous day and
cluster-to-cluster transition probabilities (and are estimated from the available multi-day time-
activity records), thus constructing a long-term sequence for a simulated individual. Details
regarding the MCC method and supporting evaluations are provided in U.S. EPA (U.S. EPA,
2019a, U.S. EPA, 2019b).
Che et al. (2014) performed an evaluation of the impact of the three APEX methods on
PM2.5 exposure estimates. As expected, little difference was observed across the methods with
regard to estimates of the mean exposures of simulated individuals. Differences were observed,
however, in the number of multiday exposures exceeding a selected benchmark concentration.
With regard to the number of simulated individuals experiencing 3 or more days above
benchmark concentrations, the MCC method estimates were approximately 12-14% greater than
either the random or D&A methods. For the number of persons experiencing at least one
exposure of concern, however, the MCC method estimates were approximately 4% lower than
those of the other two methods. For additional context, we note that, using all methods, there is
an order of magnitude difference in the number of persons exposed at least once versus three or
more times, indicating that, overall, the occurrence of simulated multiday exposures are rare
events regardless of method selection.
Che et al. (2014) concludes that while the MCC method produces a higher number of
multiday exposures, there remains a question whether the MCC method has greater accuracy
relative to the other two methods. We note this conclusion applies to both the estimations of
single day and multiday exposures, as there is an inverse relationship between the two when
simulating exposures using APEX and a finite set of activity pattern data. Thus, the MCC
method produces a smaller number of single day exposures above benchmarks relative to the
other two methods, estimations also subject to a degree of uncertainty.
In the absence of having a robust data set (e.g., multiday/week diary data from a random
population) to better evaluate the accuracy of any of the methods, we considered selection of the
longitudinal approach for this assessment from a practical perspective, guided by a balancing of
the single day and multiday exposures that can be estimated by each method. In so doing, we
selected the D&A approach, recognizing that the D&A method allows for flexibility in the
selection of the key influential variable and its setting values, and also the ability to directly
observe the impact of changes to these values on model outputs.
3D-60
-------
The key variable selected for this exposure and risk analysis is the amount of time an
individual spends each day outdoors, as that is the most important determinants of exposure to
high levels of O3 (2014 HREA, Appendix 5G, section 5G-2). In their evaluation, Che et al.
(2014) varied the values of D and A for this variable to determine the impact to estimated
exposures. Compared to their base level simulation (i.e., D=0.19 and A=0.22), increasing both D
and A by 100% increased the number of persons having at least three exposures above the
selected benchmark by about 4%, while also reducing the percent of persons experiencing at
least one day above benchmarks by less than 1% (Che et al., 2014). In recognizing uncertainty in
the parameterization of D and A (i.e., based on Xue et al., 2004) a limited field study of a small
subset of the population, children 7-12) and that the Che et al., 2014 base level simulation D&A
values produced a lower estimate of repeated exposures compared with the MCC method, we
have used values of 0.5 for D and 0.2 for A for all ages to potentially increase representation of
multiday exposures without significantly reducing the percent of the population experiencing at
least one day at or above benchmark concentrations.
3D.2.6 Microenvironmental Concentrations
In APEX, exposure of simulated individuals occurs in microenvironments (MEs) rather
than assuming people are exposed continuously and consistently to ambient air. To best estimate
personal exposures, it is important to maintain the spatial and temporal sequence of MEs people
inhabit and to appropriately represent the time series of concentrations that occur within them.
Two methods are available in APEX for calculating pollutant concentrations within MEs: a mass
balance model and a transfer factor approach. In both approaches, ME concentrations depend on
the ambient (outdoor) air O3 concentrations and ambient air temperatures, as well as statistical
distributions to parameterize the variables used by each approach. Further, the statistical
distributions of some of the key variables depend on values of other variables in the model. For
example, the distribution of air exchange rates inside an individual's residence depends on the
type of heating and air conditioning present, which are also probabilistic inputs to the model. The
value of a variable can be set as a constant for the entire simulation (e.g., house volume remains
identical throughout the exposure period), or APEX can sample a new value hourly, daily, or
seasonally from user-specified statistical distributions. APEX also allows the user to specify
diurnal, weekly, or seasonal patterns for certain ME parameters. Details regarding the two
methods can be found in (U.S. EPA, 2019a, U.S. EPA, 2019b) and are briefly described below.
The mass balance method, used for the indoor MEs, assumes that an enclosed
microenvironment (e.g., a room within a home) is a single well-mixed volume in which the air
concentration is approximately spatially uniform (Figure 3D-10). The concentration of an air
pollutant in such a microenvironment is estimated using (1) inflow of air into the
3D-61
-------
microenvironment, (2) outflow of air from the microenvironment, (3) removal of a pollutant
from the microenvironment due to deposition, filtration, and chemical degradation, and (4)
emissions from sources of a pollutant inside the microenvironment (not used for this exposure
and risk analysis). Considering the microenvironment as a well-mixed fixed volume of air, the
mass balance equation for a pollutant in the microenvironment can be written in terms of
concentration as follows in Equation 3D-7:
££W= c - C -C
dt m out removal
Equation 3D-7
where,
C(t) = Concentration in the microenvironment at time t
C i„ = Rate of change in C(t) due to air entering the microenvironment
C out = Rate of change in C(t) due to air leaving the microenvironment
C removal = Rate of change in C(t) due to all internal removal processes
Microenvironment
Air
outflow
Indoorsources
Air
inflow
Removal due to:
•Chemical reactions
•Deposition
•Filtration
Figure 3D-10. Illustration of the mass balance model used by APEX to estimate
concentrations within indoor microenvironments.
3D-62
-------
The factors model (used for the outdoor and inside vehicle MEs) is simpler than the mass
balance model. In this method, the value of the ME concentration is not dependent on the ME
concentration during the previous time step. Rather, this model uses Equation 3D-8 to calculate
the concentration in an ME from the ambient air quality data:
Cmean ~ Cambient ^ fproximity ^ fpollutant Equation 3D-8
where,
Cmean = Mean concentration over the time step in a microenvironment (ppm)
Cambient= The concentration in the ambient (outdoor) air (ppm)
fproximity = Proximity factor (unitless)
fpollutant = fraction of ambient air pollutant entering microenvironment (unitless)
Based on findings from the 2014 HREA, we have specified seven MEs to simulate in this
assessment, largely based on two factors: the expectation of a particular ME leading to exposures
of interest and the availability of factors needed to reasonably model the ME. The 2014 HREA
indicated that high (>50 ppb) 8-hr daily maximum O3 exposures occurred while individuals spent
much larger amounts of afternoon time outdoors compared with those experiencing low (< 50
ppb) exposure levels (2014 HREA, Appendix 5G, Figure 5G-5). Given that finding and the
objective for the exposure assessment (i.e., understanding how often and where maximum O3
exposures occur), we recognized the added efficiency of minimizing the number of MEs
compared to that done in the 2014 HREA (i.e., 28 microenvironments), particularly reducing the
number of lower-exposure indoor MEs that were parameterized and included at that time.
Accordingly, we aggregated the number of MEs to seven and estimate exposures of
ambient air origin that occur within a core group of indoor, outdoor, and inside vehicle MEs.
Four indoor MEs (indoor-residence, indoor-restaurant, indoor-school, and indoor-other55) were
modeled based on having specific air exchange rate data available for each (section 3D.2.6.1).
All outdoor locations were assumed to have O3 concentrations equivalent to ambient air,
however there were two MEs used to do so, distinguished by whether or not they occurred near
roads. The outdoor near road ME was modeled separately due to the expected decrease in
concentrations occurring in that ME relative to that of ambient air concentrations. And finally, an
inside-vehicle ME was modeled based on the expectation that it would lead to some instances of
relatively lower exposures compared with ambient air concentrations. Table 3D-12 lists the
seven microenvironments selected for this analysis and the exposure calculation method used for
55 The indoor-other ME is comprised of all non-residential MEs, thus could include office buildings, stores, etc.
3D-63
-------
each. The variables and their associated parameters used to calculate ME concentrations are
summarized in subsequent sections below.
Table 3D-12. Microenvironments modeled and calculation method used.
Microenvironment (ME)
APEX ME
Calculation
Variables A
Indoor - Residence
1
Mass balance
AER&RM
Indoor - Restaurant
2
Mass balance
AER&RM
Indoor - School
3
Mass balance
AER&RM
Indoor - Other
4
Mass balance
AER&RM
Outdoor - General
5
Factors
None
Outdoor - Near road
6
Factors
PR
Inside - Vehicle
7
Factors
PE
A AER = air exchange rate, RM = removal rate, PR = proximity factor, PE = fraction of pollutant
entering microenvironment, None = ME concentration is equal to ambient air concentration.
The seven microenvironments were mapped to the 115 CHAD locations56 because using
such a large number of MEs would go well beyond the practical scale needed for the exposure
and risk analyses. Note that the ambient air concentration used in calculating ME concentration
for each exposure event varies temporally and spatially. For example, commuters (i.e., employed
individuals who do not work at home) are assigned to either their home tract or work tract
concentration, depending on whether the population probabilities and commuting data base
produce either a home or work event. Additionally, depending on the particular ME (i.e., other
than home or work), the mapping of CHAD locations to the seven MEs also uses an identifier
that designates the relative location in the air quality surface from which the ambient air
concentration (used to calculate the ME concentration) is selected. For this assessment, such
locations would include the Census tract for a simulated individual's home (H), work (W), near
work (NW), near home (NH), last (L, either NH or NW), other (O, average of all), or unknown
(U, last ME determined) location. Specific designations are provided in the APEX ME mapping
file, with selection based on known factors and professional judgement. For example, when an
individual is in their home, the ambient air concentration in the home tract is used to calculate
their ME concentration. When the individual is at work, the tract the individual commuted to is
used to calculate their ME concentration. Travel inside vehicles used the ambient air
concentration data from the tract used to calculate the prior ME concentration. Most other MEs
(both indoor and outdoor) use ambient air concentration data selected from near home tracts.
56 The location codes indicate specific MEs that extend beyond simple aggregations of indoor, in-vehicle, and
outdoor locations where people spend time. For example, CHAD has a location code for when individuals spent
time inside their residence while in the kitchen.
3D-64
-------
Status attribute variables are also important in estimating ME concentrations, and can
include, but are not limited to, housing type, whether the house has air conditioning, and whether
the car has air conditioning. Because outdoor MEs are expected to contribute the most to an
individuals' highest O3 exposure (and potential health risk) and the status attribute variables
pertain to indoor MEs, the setting of these particular variables will have limited impact to the
exposure and risk results generated here. In this assessment, a number of temperature ranges are
used in selecting the particular distribution for estimating air exchange rates (AERs). Maximum
daily temperature is also used in diary selection to best match the study area meteorological data
for the simulated individual (Graham and McCurdy, 2004) and air conditioning use.
Multiple APEX input files (the first and third in the list below), of the same general
format, are used for estimating ME concentrations in each study area. A single APEX ME
mapping file is used for all study areas. These ME input files contain the parameter settings for
all variables described in the subsections that follow.
• ME descriptions 03 7MEsCSA [number].txt: defines ME calculation method,
conditional variables used (e.g., temperature categories - see functions file), distribution
type, distribution parameters (mean, standard deviation, minimum, maximum) for AERs,
decay rates, proximity factors, and PE fractions used to estimate O3 in 7 MEs.
• Microenvironmenl mappings 07 MI is. txt: maps 115 CHAD locations to the 7 APEX
MEs and assigns the tract-level ambient air concentrations to use for each location.
Contains CHAD location code, CHAD description, APEX ME number, and ambient air
concentration location identifier
• Functions O3 CSA[number]040219.txt: variables used for selecting AER - air
conditioning (A/C) prevalence (home has A/C, does not have A/C) by five temperature
ranges for air exchange rate (<50, 50-67, 68-76, 77-85, or >85 °F). (see section 3D.2.6.1)
3D.2.6.1 Indoor Microenvironments
As described above, all four indoor MEs (indoor-residential, indoor-restaurant, indoor-
school, and indoor-other) were modeled using a mass balance model. The three variables used to
calculate ME concentrations, air exchange rates (section 3D.2.6.1.1), air conditioning prevalence
(section 3D.2.6.1.2), and ozone removal rate (section 3D.2.6.1.3) are described below.
3D.2.6.1.1 Air Exchange Rates
Distributions of air exchange rates (AERs, hr1) for the indoor residential ME were
developed using data from several studies. The analysis of these data and the development of
most of the distributions used in the modeling were originally described in detail in the 2007
exposure analysis (U.S. EPA (2007a), Appendix A) and updated in the 2014 HREA (see
Appendix 5E). Briefly, AER distributions for the residential microenvironments depend on the
type of air conditioning (A/C) and on the outdoor temperature, among other variables for which
we do not have sufficient data to estimate. AER distributions were found vary greatly across
3D-65
-------
cities, A/C types, and temperatures, so that the selected AER distributions for the modeled cities
should also depend on these attributes. For example, the mean AER for residences with A/C
ranges from 0.38 in Research Triangle Park, NC at temperatures > 25 °C upwards to 1.244 in
New York, NY considering the same temperature range (2014 HREA, Appendix 5E). For each
combination of A/C type, city, and temperature with a minimum of 11 AER values, exponential,
lognormal, normal, and Weibull distributions were fit to the AER values and compared.
Generally, the lognormal distribution was the best-fitting of the four distributions, and so, for
consistency, the fitted lognormal distributions are used for all the cases.
There were a number of limitations in generating study-area specific AER stratified by
temperature and A/C type. For example, AER data and derived distributions were available only
for selected cities, and yet the summary statistics and comparisons demonstrate that the AER
distributions depend upon the city as well as the temperature range and A/C type. As a result,
city-specific AER distributions were used where possible; otherwise staff selected AER data
from a similar city. Another important limitation of the analysis was that distributions were not
able to be fitted to all of the temperature ranges due to limited number of available measurement
data in these ranges. A description of how these limitations were addressed can be found in the
2014 HREA, Appendix 5E. The AER distributions used for the exposure modeling are given in
Table 3D-13 (Residences with A/C) and Table 3D-14 (Residences without A/C).
Table 3D-13. Air exchange rates (AER, hr1) for indoor residential microenvironments with
A/C by study area and temperature.
Study Area
Daily Mean
Temperature
(°C)
Lognormal Distribution
GM, GSD, min, max
(hr1)
Original AER Study Data
Used
Atlanta
<10
0.962,1.809,0.1,10
Research Triangle Park, NC
10-20
0.562,1.906,0.1,10
20-25
0.397,1.889,0.1,10
>25
0.380,1.709,0.1,10
Boston, Philadelphia
<10
0.711,2.108,0.1,10
New York, NY
10-25
1.139,2.677,0.1,10
>25
1.244,2.177,0.1,10
Dallas, Phoenix
<20
0.407,2.113,0.1,10
Houston, TX
20-25
0.467,1.938,0.1,10
25-30
0.422,2.258,0.1,10
>30
0.499,1.717,0.1,10
Detroit
<10
0.744,1.982,0.1,10
Detroit, Ml or New York, NY
10-20
0.811,2.653,0.1,10
20-25
0.785,2.817,0.1,10
>25
0.916,2.671,0.1,10
3D-66
-------
Study Area
Daily Mean
Temperature
(°C)
Lognormal Distribution
GM, GSD, min, max
(hr1)
Original AER Study Data
Used
Sacramento
<25
0.503,1.921,0.1,10
Sacramento
>25
0.830,2.353,0.1,10
St. Louis
<10
0.921,1.854,0.1,10
St. Louis
10-20
0.573,1.990,0.1,10
20-25
0.530,2.427,0.1,10
25-30
0.527,2.381,0.1,10
>30
0.609,2.369,0.1,10
Table 3D-14. Air exchange rates (AER, hr1) for indoor residential microenvironments
without A/C by study area and temperature.
Study Area
Daily Mean
Temperature
(°C)
Lognormal Distribution
GM, GSD, min, max
(hr-1)
Original AER Study Data
Used
Atlanta, St. Louis
<10
0.923,1.843,0.1,10
St. Louis
10-20
0.951,2.708,0.1,10
>20
1.575,2.454,0.1,10
Boston, Philadelphia
<10
1.016,2.138,0.1,10
New York, NY
10-20
0.791,2.042,0.1,10
>20
1.606,2.119,0.1,10
Dallas, Phoenix
<10
0.656,1.679,0.1,10
Houston, TX
10-20
0.625,2.916,0.1,10
>20
0.916,2.451,0.1,10
Detroit
<10
0.791,1.802,0.1,10
Detroit, Ml or New York, NY
10-20
1.056,2.595,0.1,10
20-25
1.545,2.431,0.1,10
>25
1.860,2.437,0.1,10
Sacramento
<10
0.526,3.192,0.1,10
Sacramento
10-20
0.665,2.174,0.1,10
20-25
1.054,1.711,0.1,10
>25
0.827,2.265,0.1,10
The AER distribution (hr1) used for indoor restaurants in all study areas is a fitted
lognormal distribution, having a geometric mean = 3.712, geometric standard deviation = 1.855
and bounded by the lower and upper values of the sample data set {1.46, 9.07}. This distribution
was developed using data from Bennett et al. (2012) who measured AER in restaurants (details
on derivation provided in the 2014 HREA, Appendix 5E). The AER distribution (hr1) used for
3D-67
-------
indoor schools in all study areas is a fitted Weibull distribution,57 having a threshold (t) = 0,
shape (C) = 1.26, and scale (a) = 1.75, bounded by a lower and upper range {0, 10}. This
distribution was developed from Lagus Applied Technology, 1995, Shendell et al., 2004, and
Turk et al., 1989 who measured AER in schools (raw data provided in Table 3D-15).
Table 3D-15. Individual air exchange rate data (hr1) obtained from three studies used to
develop an AER distribution used for schools in all study areas.
Individual Air
Exchange Rate Data (hr1)
Lagus Applied Technology (1995)
Shendell etal. (2004)
Turk etal. (1989)
0.56
1.34
1.92
2.71
0.1
0.3
0.6
0.8
0.74
1.46
2.26
2.76
0.1
0.4
0.6
1.3
0.76
1.48
2.26
2.81
0.1
0.4
0.6
1.8
0.8
1.58
2.27
2.82
0.1
0.4
0.9
2
0.98
1.61
2.29
2.83
0.2
0.4
0.9
2.2
1.15
1.61
2.33
2.87
0.2
0.4
1.2
2.2
1.19
1.67
2.38
2.93
0.2
0.4
1.3
3
1.21
1.67
2.4
3.03
0.2
0.5
1.3
1.22
1.73
2.53
3.23
0.2
0.5
1.4
1.23
1.8
2.53
3.7
0.3
0.6
1.8
1.23
1.84
2.57
4.38
0.3
0.6
2.9
1.27
1.9
2.68
5.03
0.3
0.6
5.4
1.33
1.91
2.71
8.72
The AER distribution (hr1) used for indoor other in all study areas is a fitted lognormal
distribution, having a geometric mean = 0.949, geometric standard deviation = 1.857 and
bounded by the lower and upper values of the sample data set {0.30, 4.02}. This distribution was
developed using data from Bennett et al. (2012) who measured AER in non-residential buildings
(details on derivation provided in the 2014 HREA, Appendix 5E).
3D.2.6.1.2 Air Conditioning Prevalence
The selection of an AER distribution for the indoor residence ME is conditioned on the
presence or absence of A/C. We assigned this housing attribute to indoor residential
microenvironments using A/C prevalence data from the American Housing Survey (AHS).58 The
57 Of the three statistical distributions evaluated (lognormal, gamma, Weibull), results of a Cramer-von Mises
goodness of fit test indicated the data distribution was not statistically different than a Weibull distribution.
58 2015 and 2017.xlsx files were downloaded from https://www.census.gov/programs-
surveys/ahs/data/interactive/ahstablecreator.html for Atlanta, Boston, Dallas, Detroit, Philadelphia, and Phoenix
(accessed on 3/4/2019). The most recent data available for Sacramento and St. Louis was 2011 and available at
3D-68
-------
A/C prevalence data were assigned to our study areas where the AHS data best matched our
exposure simulation years and or study area. In all study areas and for each year, housing units
containing either central or 3 or more room AC were summed, followed by the calculation of the
A/C prevalence. If multiple years were available, these data were averaged to generate the final
A/C prevalence (unitless) for each study area (Table 3D-16). For the other three indoor MEs
(indoor-restaurant, indoor-school, and indoor-other) mechanical ventilation was assumed to be
present in all buildings (i.e., A/C prevalence = 1.0).
Table 3D-16. A/C prevalence from US Census American Housing Survey (AHS) data by
study area.
Study Area
Total Housing
Units
(x1,000)
Central AC
(x1,000)
Room AC
3 or more
(x1,000)
Year
AC
Prevalence
(unitless)
Mean AC
Prevalence
(unitless)
No AC
Prevalence
(unitless)
Atlanta
1982.8
1875.2
27.3
2015
0.96
0.96
0.04
2109
2001
22.7
2017
0.96
Boston
1838.4
649
311.9
2015
0.523
0.531
0.469
1854
674.6
322.1
2017
0.538
Dallas
2471.2
2323.1
49.9
2015
0.96
0.966
0.034
2565
2444
46.7
2017
0.971
Detroit
1709
1267.1
34
2015
0.761
0.761
0.239
1723
1280
31.1
2017
0.761
Philadelphia
2216.1
1395.4
295.9
2015
0.763
0.776
0.224
2308
1516
303.1
2017
0.788
Phoenix
1644
1591.3
7.4
2015
0.972
0.968
0.032
1686
1619
6.7
2017
0.964
Sacramento
783.7
677.5
4.6
2011
0.87
0.87
0.13
St. Louis
1115.2
1013.1
23.2
2011
0.929
0.929
0.071
3D.2.6.1.3 Ozone Decay and Deposition Rates
As done for the 2014 HREA, a distribution for combined O3 decay and deposition rates
was obtained from the analysis of measurements from a study by Lee et al. (1999). This study
measured decay rates in the living rooms of 43 residences in Southern California. Measurements
of decay rates in a second room were made in 24 of these residences. The 67 decay rates range
from 0.95 to 8.05 hr"1. A lognormal distribution was fit to the measurements from this study,
https://www.census.gov/programs-surveys/ahs/data/201 l/ahs-2011 -summary-tab les/ahs-metropolitan-summary-
tables.html (accessed on 4/2/2019).
3D-69
-------
yielding a geometric mean of 2.51 hr"1 and a geometric standard deviation of 1.53 hr"1. These
values are constrained to lie between 0.95 and 8.05 hr"1. This combined O3 decay and deposition
rate distribution was used for all four indoor microenvironments.
3D.2.6.2 Outdoor Microenvironments
As mentioned above, the two outdoor MEs (outdoor-general and outdoor-near road) used
the factors approach to estimate ME concentrations. The factors approach uses two variables in
combination with ambient air O3 concentrations: a proximity factor and a factor expressing the
fraction of a pollutant entering (PE factor) an ME, and these are discussed below.
Proximity factors are used to adjust ambient air O3 concentrations, based on the ME
location relative to that of the ambient air concentration. For the outdoor-general ME, there is no
adjustment used (proximity = 1.0); it is assumed that wherever an individual is outdoors, the
individual experiences the ambient air O3 concentrations for the tract they are present in at that
time (e.g., at home, at work, or nearby census tract). For the outdoor-near road ME, a proximity
factor is used, recognizing that ambient air concentrations measured away from roadways tend to
increase with distance. As done for the 2014 HREA, we employed the distribution for local roads
(i.e., a normal distribution {0.755, 0.203}, bounded by 0.422 and 1.0) derived from the
Cincinnati Ozone Study (American Petroleum Institute, 1997, Appendix B; Johnson et al., 1995),
based on the assumption that most of the outdoors-near-road ozone exposures will occur
proximal to local roads (see Table 3D-17 and details below in section 3D.2.6.3).
PE factors are used to adjust for the percent of a pollutant entering a ME. PE factors for
the outdoor-general and outdoor-near road MEs, because they are effectively aligned with the
ambient air O3 concentrations, are set equivalent to 1.
3D.2.6.3 Inside-Vehicle Microenvironments
As done for the 2014 HREA, for the in-vehicle ME, proximity and PE factor distributions
were obtained from the Cincinnati Ozone Study (American Petroleum Institute, 1997, Appendix
B; Johnson et al., 1995). This field study was conducted in the greater Cincinnati metropolitan
area in August and September 1994. Vehicle tests were conducted according to an experimental
design specifying the vehicle type, road type, vehicle speed, and ventilation mode. Vehicle types
were defined by the three study vehicles: a minivan, a full-size car, and a compact car. Road
types were interstate highways (interstate), principal urban arterial roads (urban), and local roads
(local). Nominal vehicle speeds (typically met over 1-min intervals within 5 mph) were at 35
mph, 45 mph, or 55 mph. Ozone concentrations were measured inside the vehicle, outside the
vehicle, and at six fixed-site monitors in the Cincinnati area. Table 3D-17 lists the parameters of
the normal distributions developed for proximity and PE factors (both are unitless) for in-vehicle
microenvironments used in this exposure and risk analysis.
3D-70
-------
A daily conditional variable was used to select the three proximity factor distributions to
use in estimating the inside-vehicle ME concentrations. The 2015-2017 Vehicle Miles of Travel
(VMT) data available from the U.S. Department of Transportation (DOT) were used to generate
these daily conditional variables.59 For local and interstate road types, the VMT for the same
DOT categories were used. For urban roads, the VMT for all other DOT road types were
summed (i.e., other freeways/expressways, other principal arterial, minor arterial, and collector).
Table 3D-18 summarizes the conditional variables used for each study area to select for the
proximity factor distribution used to estimate inside-vehicle ME concentrations.
Table 3D-17. Parameter values for distributions of penetration and proximity factors used
for estimating in-vehicle ME concentrations.
ME Factor
Road Type
Arithmetic
Mean
(unitless)
Standard
Deviation
(unitless)
Lower Bound A
(unitless)
Upper Bound
(unitless)
PE
All
0.300
0.232
0.100
1.0
Proximity
Local
0.755
0.203
0.422
1.0
Urban
0.754
0.243
0.355
1.0
Interstate
0.364
0.165
0.093
1.0
A A 5th percentile value estimated using a normal approximation as Mean -1.64 x standard deviation.
Table 3D-18. VMT (2015-2017) derived conditional probabilities for interstate, urban, and
local roads used to select inside-vehicle proximity factor distributions in each
study area.
Study Area
Conditional Probabilities for Vehicle Proximity Factors (unitless)
Interstate
Urban
Local
Atlanta
0.339
0.392
0.269
Boston
0.416
0.455
0.129
Dallas
0.496
0.453
0.051
Detroit
0.357
0.531
0.112
Philadelphia
0.361
0.523
0.116
Phoenix
0.364
0.542
0.094
Sacramento
0.456
0.433
0.111
St. Louis
0.460
0.363
0.177
59 Data were downloaded (accessed on 3/13/2019) from U.S. Department of Transportation (DOT) Federal Highway
Administration (FHA) Highway Statistics Series Publications. The three individual years (2015-2017) of data
were downloaded from dropdown menu available at: https://www.jhwa.dot.gov/policyinformation/statistics.cfm.
3D-71
-------
3D.2.7 Estimating Exposure
APEX estimates the complete time series of exposure and breathing rate for every
simulated individual. This is because APEX accounts for important factors that influence
exposure and include the magnitude, duration, frequency of exposures, and the breathing rate of
individuals at the time of exposure. APEX can summarize exposure data using standardized time
metrics (e.g., hourly or daily average, daily maximum 7-hr average), as is needed for comparison
to benchmark concentrations (section 3D.2.8.1) or can output the minute-by-minute exposure
concentrations and simultaneous breathing rate, as is needed for the lung function risk modeling
(section 3D.2.8.2.2). As a reminder, calculated exposures are distinct from that of ambient air
concentrations by accounting for simulated individual's time-location-activity patterns and O3
concentration decay/variation occurring within the occupied microenvironments. Further,
exposures (and hence health risks) are estimated for four groups of individuals residing in each
study area: children (individuals aged 5 to 18 years), children with asthma, adults (individuals
older than 18 years), and adults with asthma.
3D.2.8 Estimating Risk
We derived two types of metrics to characterize potential population health risk: a
comparison of simulated exposures to benchmark concentrations (section 3D.2.8.1) and by using
simulated exposures to estimate lung function risk (section 3D.2.8.2). As done in the last review,
these two approaches are based on the body of evidence from the controlled human exposure
studies reporting lung function decrements (as measured by changes in FEV1)60 along with
supporting health evidence from Cb-related epidemiologic studies. As discussed in Appendix 3
of the ISA, there is a significant body of controlled human exposure studies reporting lung
function decrements and respiratory symptoms in adults associated with 1- to 6.6-hr exposures to
O3, all but a few of which were available in the last review and no new studies that included 6.6-
hour exposures were available (ISA, Appendix 3, section 3.1.4.1.1; 2013 ISA, section 6.2.1.1).
The exposure studies of greatest interest are those that have exposed subjects during exercise
(ISA, Appendix 3; 2013 ISA, section 6.2.1.1). In general, the 1- to 2-hr exposure studies utilize
an intermittent exercise protocol in which subjects rotate between periods of exercise and rest,
though a limited number of these studies use a continuous exercise regime. A quasi-continuous
exercise protocol is common to the 6.6-hr exposure studies where subjects complete six 50-min
60 There are other respiratory responses resulting from O3 exposures that were measured in these studies, including
increased lung inflammation, increased respiratory symptoms, increased airway responsiveness, and impaired
host defenses. While the available quantitative information is inadequate to reasonably model these other health
endpoints, nevertheless the observed responses remain informative in characterizing overall risks.
3D-72
-------
periods of exercise followed by 10-min rest periods (along with a 35-min lunch/rest period)
(ISA, Appendix 3, section 3.1.4.1.1).
For lung function risk, we estimate risk of an Cb-related decrement at or above 10%, 15%
and 20%. These sizes of decrements have been used in the risk assessments for the past three
reviews, i.e., those completed in 2015, 2008 and 1997 (2014 HREA; U.S. EPA, 2007a, U.S.
EPA, 2007b; Whitfield et al., 1996). In the last review, the CASAC concurred with the EPA's
use in the 2014 HREA of estimated FEVi decrements of >15% as a scientifically relevant
surrogate for adverse health outcomes in active healthy adults, and an FEVi decrement of >10%
as a scientifically relevant surrogate for adverse health outcomes for people with asthma and
lung disease (Frey, 2014, p. 3).
3D.2.8.1 Comparison to Benchmark Concentrations
For the comparison of simulated exposures to benchmark concentrations that reflect
observations from the 6.6-hr controlled human exposure studies, APEX estimates the daily
maximum 7-hr average O3 exposure61 for every simulated individual, stratified by exertion level
at the time of exposure. This indicator was selected based on controlled human exposure studies
where reported adverse health responses were associated with exposure to O3 and while the study
subject was exercising.62 A 7-hr average exposure concentration is more appropriate than using
an 8-hr average (as was done for the prior REAs) because it aligns more closely to the 6.6-hr
durations of the controlled human exposure studies on which the benchmark concentrations are
based.63 The 7-hr average exposure concentrations experienced by simulated individuals while at
moderate or greater exertion (EVR >17.32 ± 1.25 L/min-m2 body surface area; see above section
3D.2.2.3.3) are then compared to the benchmark concentrations.
Benchmark concentrations used in this assessment include O3 exposure concentrations of
60, 70 and 80 ppb; the same benchmarks used for the 2014 HREA (based on there being no new
6.6-hr controlled human exposure studies that might inform consideration of alternatives).
Estimating the occurrence of ambient air-related 7-hr average O3 exposures at and above these
61 Only the maximum 7-hr average O3 exposure concentration is retained by APEX for each day simulated, per
person.
62 Health responses observed in the controlled human exposure studies are from 6.6-hr exposures to O3, that
involved quasi-continuous exercise. Therefore, it is possible that the effects observed at benchmark levels
identified using a 6.6-hr exposure could occur at slightly lower concentrations for a comparable 7-hr exposure and
occur at still lower concentrations for a comparable 8-hr exposure. From a practical perspective, there would be a
greater number of individuals estimated at or above a particular benchmark when averaging exposures across a
6.6-hr period than when compared to simulations using 7-hr or 8-hr averaging (the latter of which was used in the
prior assessments and recognized specifically in the 2014 HREA, section 5.2.8, footnote 18).
63 Note that the 8-hr averaging time for ambient air O3 concentrations associated with the current standard remains
the same as used in prior assessments. The only difference is that for the current exposure and risk analysis, 8-hr
ambient air O3 concentrations are now evaluated with a more appropriate exposure and risk metric (i.e., a 7-hr
average exposure benchmark).
3D-73
-------
benchmark levels is intended to provide perspective on the potential for public health impacts of
03-related health effects observed in human clinical and toxicological studies, but for which
available data do not support development of E-R functions, precluding their evaluation in
quantitative risk assessments (e.g., lung inflammation, increased airway responsiveness, and
decreased resistance to infection), as well as lung function decrements which are currently
evaluated in quantitative risk assessments. The 80 ppb benchmark concentration represents an
exposure where multiple controlled human exposure studies (of the 6.6-hr, with exercise design)
demonstrate a range of 03-related respiratory effects including lung inflammation and airway
responsiveness, as well as respiratory symptoms, in healthy adults. The 70 ppb benchmark
concentration reflects a study that found statistically significant decrements in lung function as
well as increased respiratory symptoms. The 60 ppb benchmark level represents the lowest
exposure level at which statistically significant decrements in lung function, but not respiratory
symptoms, have been observed in studies of healthy individuals (see Table 3-2 of PA).64 This is
summarized in Table 3D-19 below. Further details on the body of evidence supporting the
selection of these benchmark levels is described in the ISA, Appendix 3 and summarized in the
PA, section 3.3 and Appendix 3A.
APEX then calculates two general types of exposure estimates for the population of
interest: the estimated number of people exposed to a specified O3 concentration level and, the
number of days per year that they are so exposed, while at moderate or greater exertion. The
former highlights the number of individuals exposed one or more times per year (i.e., at least
once) at or above a selected benchmark level. The latter is expressed as multiday exposures, that
is, the number of times per year each simulated individual experiences a daily maximum
exposure at or above a benchmark. These same exposure results are also used in estimating
population-based lung function risk (section 3D.2.8.2.1).
64 Prolonged exposure to 40 ppb O3 results in a small decrease in group mean FEVi that is not statistically different
from responses following exposure to filtered air (Adams, 2002; Adams, 2006).
3D-74
-------
Table 3D-19. Responses reported in 6.6-hr controlled human exposure studies at a given
benchmark concentration.
Benchmark
Concentration
(ppb)
Responses Reported in Controlled Human Exposure Studies A
Decrements in Lung Function, and Other Effects
Respiratory Symptoms
>80
Prolonged exposure to an average O3 concentration of 80 ppb, 100
ppb, or 120 ppb O3 results in statistically significant group mean
decrements in FEVi ranging from 6B to 8%, 8 to 14%, and 10 to
16%, respectively.0 Statistically significant increases in multiple
inflammatory response indicators and in airway responsiveness.
Statistically significant
increases in respiratory
symptoms (ISA, section
3.1.4.2.1).
>70
Prolonged exposure to an average O3 concentration of 70 ppb
results in a statistically significant group mean decrement in FEVi
of about 6%.D
>60
Prolonged exposure to an average O3 concentration of 60 ppb
results in group mean FEVi decrements ranging from 1.7% to
3.5%.E Based on data from multiple studies, the weighted average
group mean decrement was 2.5%. In some analyses, these group
mean decrements in lung function were statistically significantF
while in other analyses they were not.G Statistically significant
increases in sputum neutrophils, an indicator of inflammatory
response.
None of studies at this
exposure concentration
have observed a
statistically significant
increase in symptom
scores (ISA, section
3.1.4.2.1).
A Information is drawn from Table 3A-1 of Appendix 3A of the PA for 6.6-hr exposure protocol with exercise EVR of 20 L/min/m2 (see also
ISA, Figure 3-3). These studies have been performed with healthy adult subjects.
B Measurements collected at 80 ppb exposure for 30 subjects as part of the Kim et al. (2011) study that were presented only in Figure 5 of
McDonnell et al. (2012) indicate a group mean decrement of 3.5%.
c Folinsbee et al. (1994), Florstman et al. (1990), McDonnell et al. (1991), Adams (2002), Adams (2006), Adams (2000), Adams and Ollison
(1997), Schelegle etal. (2009).
D Schelegle et al. (2009).
E Adams (2002), Adams (2006), Schelegle etal. (2009), and Kim etal. (2011).
F Brown etal. (2008), Kim etal. (2011). In an analysis of the Adams (2006) data, Brown etal. (2008) addressed the more fundamental
question of whether there were statistically significant differences in responses before and after the 6.6-hr exposure period and found the
study group average effect on FEV1 at 60 ppb to be small, but statistically significant using several common statistical tests, even after
removal of potential outliers.
G Adams (2006), Schelegle et al. (2009).
3D.2.8.2 Lung Function Risk
We used two approaches to estimate health risk. As done for the lung function risk
assessments conducted during the prior O3 NAAQS reviews, the first approach used a Bayesian
Markov Chain Monte Carlo technique to develop probabilistic population-based Exposure-
Response (E-R) functions. These population-based E-R functions were then combined with the
APEX estimated population distribution of 7-hr maximum exposures for people at or above
moderate exertion (EVR >17.32 ± 1.25 L/min-m2 body surface area) to estimate the number of
people expected to experience lung function decrements. The second approach is based on the
McDonnell-Stewart-Smith (MSS) FEVi model (McDonnell et al., 2013). The MSS model uses
the time-series of O3 exposure and corresponding ventilation rates for each APEX simulated
3D-75
-------
individual to estimate their personal time-series of FEVi reductions, selecting the daily
maximum reduction for each person. As done for the exposure benchmark analysis, APEX
calculates, for the population of interest, the estimated number of simulated individuals expected
to experience an FEVi response at or above a selected level and the number of days per year that
may occur per person. A key difference between these approaches is that the population-based E-
R method directly approximates a population distribution of FEVi reductions while the MSS
model estimates FEVi reductions at the individual level (which are then aggregated to a
population level). Each of these approaches is discussed in detail below.
3D.2.8.2.1 Population-based E-R function
For developing the population-based E-R function, we used the exact same E-R function
as used for the 2014 HREA given CAS AC advice on the approach used for the prior O3 review
(Henderson, 2006) and that there were no new controlled human exposure study data to justify
the generating of a new E-R function for this current analysis. Briefly, data from several
controlled human exposure studies that evaluated 6.6-hr exposures at moderate exertion were
combined and used to estimate E-R functions. Considering the above discussion and as done in
the 2014 HREA, we separated the controlled human exposure study data into three lung function
decrement categories. The mid- to upper-end of the range of moderate levels of functional
responses and higher (i.e., FEVi decrements >15% and >20%) are included to generally
represent potentially adverse lung function decrements in active healthy adults, while for people
with asthma or lung disease, a focus on moderate functional responses (FEVi decrements down
to 10%) may be appropriate (Table 3D-20 and Figure 3D-11). The controlled human exposure
study data in this table were first corrected on an individual basis for study effects in clean
filtered air to remove any systemic bias that might be present in the data attributable to the
effects of the experimental procedures and extraneous responses (e.g., exercise, diurnal
variability, etc.) (2013 ISA, pp. 6-4 and 6-5). This is done by subtracting the FEVi decrement in
filtered air from the FEVi decrement (at the same time point) during exposure to O3. An example
of this calculation is given in the 2014 HREA, Appendix 6D.
3D-76
-------
Table 3D-20. Summary of controlled human exposure study data stratified by
concentration level and lung function decrements, corrected for individual
response that occurred while exercising in clean air, ages 18-35.
Study, Grouped by
Average O3 Exposure
Protocol
Study
Subjects
(n)
Subjects Responc
inq (n)A
AFEV1
>10%
AFEV1
>15%
AFEVt
>20%
0.040ppmOs
Adams (2002)
Square-wave (constant level), face mask
30
2
0
0
Adams (2006)
Variable levels (exercise avg = 0.040 ppm)
30
0
0
0
0.060ppmO3
Adams (2006)
Square-wave
30
2
0
0
Variable levels (exercise avg = 0.060 ppm)
30
2
2
0
Kim et al. (2011)
Square-wave
59
3
1
0
Schelegle et al. (2009)
Variable levels (exercise avg =0.060 ppm)
31
4
2
1
0.070ppmO3
Schelegle et al. (2009)
Variable levels (exercise avg= 0.070 ppm)
31
6
3
2
0.080ppmO3
Adams (2002)
Square-wave, face mask
30
6
5
2
Adams (2003)
Square-wave, chamber
30
6
2
1
Square-wave, face mask
30
5
2
2
Variable levels (exercise avg=0.080 ppm),
chamber
30
6
1
1
Variable levels (exercise avg=0.080 ppm),
face mask
30
5
1
1
Adams (2006)
Square-wave
30
7
2
1
Variable levels (exercise avg=0.080 ppm)
30
9
3
1
F-H-M1
Square-wave
60
17
11
8
Kim et al. (2011)
Square-wave
30
4
1
0
Schelegle et al. (2009)
Variable levels (exercise avg=0.080 ppm)
31
10
5
4
0.0870ppm O3
Schelegle et al. (2009)
Variable levels (exercise avg=0.087 ppm)
31
14
10
7
0.100 ppm O3
F-H-M1
Square-wave
32
13
11
6
0.120 ppm O3
Adams, 2002
Square-wave, chamber
30
17
12
10
Square-wave, face mask
30
21
13
7
F-H-MB
Square-wave
30
18
15
10
A Data from 2014 HREA, Table 6-3 and were originally compiled by Abt (2013). Individual subject responses were corrected using pre- and
post-exposure observations.
B F-H-M combines data from Folinsbee etal. (1988), Horstman etal. (1990), and McDonnell etal. (1991).
3D-77
-------
70
60
£50
LU
CD
CO 40
>-
Q
=>
2 30
20
10
A Adams (2006)
X Adams (2003)
+ Adams (2002)
= Kim etal. (2011)
^ Schelegle (2009)
D Folinsbee; Horstman; McDonnell
>10% a15%>20%
o
o
0.030
0.040
~
~
+
+
~
+
+
ffl
+
0.050 0.060 0.070 0.080 0.090
OZONE (PPM)
0.100
0.110
0.120
0.130
Figure 3D-11. Controlled human exposure data for FEVi. responses in individual study
subjects.
A Bayesian Markov Chain Monte Carlo (BMCMC) approach (Lunn et aL, 2012)
developed as part of an earlier O? exposure and risk analysis (U.S. EPA, 2007a, U.S. EPA,
2007b, section 3 .1.2) was modified for the 2014 HREA and used to generate the population-
based E-R functions using the updated controlled human exposure study data (Abt, 2013).65
Briefly, we considered both linear and logistic functional forms in estimating the E-R function
and chose a 90 percent logistic/10 percent piecewise-linear split using a BMCMC approach. For
each of the three measures of lung function decrement, we first assumed a 90 percent probability
that the E-R function has the following 3-parameter logistic form indicated by Equation 3D-9:66
y(x; a,J3,y) =
a er{l-O
(1 + er)(l + e^T)
Equation 3D-9
65 In some of the controlled human exposure studies, subjects were exposed to a given O3 concentration more than
once - for example, using a constant (square-wave) exposure pattern in one protocol and a variable (triangular)
exposure pattern in another protocol. However, because there were insufficient data to estimate subject-specific
response probabilities, we assumed a single response probability (for a given definition of response) for all
individuals and treated the repeated exposures for a single subject as independent exposures in the binomial
distribution.
66 The 3-parameter logistic function is a special case of the 4-parameter logistic, in which the function is forced to go
through the origin, so that the probability of response to 0.0 ppm is 0.
3D-78
-------
where x denotes the O3 concentration (in ppm) to which the individual is exposed, y
denotes the corresponding response (decrement in FEVi > 10%, > 15% or > 20%), and a, /?, and
y are the three parameters whose values are estimated.
We then assumed a 10 percent probability that the E-R function has the following 2-piece
linear with threshold (hockey stick) form67 indicated by Equation 3D-10:
The selection of the 90 percent logistic/10 percent piecewise-linear split was based
largely on the results of sensitivity analyses in the 2007 O3 risk assessment combined with
CASAC advice on the model form (U.S. EPA, 2007b),68 and from model fit determined in the
2014 HREA.69 Therefore, as done for the 2014 HREA, we are using only the 90/10 E-R function
in the current analysis to estimate risk. Further, because there were no newly available controlled
human exposure study data for 6.6-hr duration exposures since the 2014 HREA, we used the
exact same 90/10 E-R function derived at that time, the overall approach of which is briefly
described below.
To generate the E-R functions, prior distributions needed to be specified to estimate the
posterior distribution for each of the unknown parameters (Box and Tiao, 1973). For the logistic
functional form, we assumed lognormal priors and used Max likelihood estimates (MLE) of the
means and variances for the 3 parameters. For the linear functional form, we assumed normal
priors using ordinary least square (OLS) estimates for the means and variances for the
parameters.
For each of the two functional forms (logistic and linear), we derived the posterior
distributions using the binomial likelihood function and prior distributions for each of the
unknown parameters. Specifically, we used three Markov chains (each chain corresponds to a set
67 The 2-piece linear models estimate no occurrences below about 40 ppb for the 10% lung function decrement and
below about 60 ppb for the 15% and 20% lung function decrements based on the limited available data at those
exposure levels. Note that as these two-piece linear model forms are combined with a second model form
(logistic) for the final model, their contribution to estimated responses is low.
68 The 1997 risk assessment used a linear form consistent with the advice from the CASAC 03 panel at the time that
a linear model reasonably fit the available data at exposures of 0.08, 0.10, and 0.12 ppm. Following the addition
of exposures data at 0.06 and 0.04 ppm in the 2007 assessment, a logistic model was found to provide a good fit
to the data. The CASAC O3 panel for that review noted that there are only limited data at the two lowest exposure
levels and, as a result, a linear model could not entirely be ruled out, resulting in the combined model based on
both the logistic and linear forms (U.S. EPA, 2007b).
69 Analyses using the updated data available for the 2014 HREA determined that for each of the three E-R curves,
the 90/10 logistic/linear mix has smaller error in fit (weighted RMSE) relative to the other two E-R curves
evaluated: one having a 80/20 logistic/linear mix and the other having a 50/50 mix.
Equation 3D-10
3D-79
-------
of initial parameter values) and for each chain we used 4,000 iterations as the "burn-in" period70
followed by 96,000 iterations for the estimation. Each iteration corresponds to a set of estimates
for the parameters of the (logistic or linear) exposure-response function. We then examined the
outputs using the options WinBUGS provides to check convergence and auto-correlation (e.g.,
trace plot, auto correlation). Finally, we combined 8,100 sets of values from the logistic model
runs (the last 2,700 iterations from each chain) with 900 sets of values from the linear model runs
(the last 300 iterations from each chain) to obtain a single combined distribution for each
predicted value, reflecting the 90 percent/10 percent assumptions stated above (WinBUGS v
1.4.3; Lunn et al., 2012).
We selected the median (50th percentile) E-R function from the 9,000 sets of functions to
estimate the risk for changes in FEVi >10%, >15%, and >20% (Figure 3D-12). The original E-R
data to which the curves were fit are also provided in the figure, along with the derived E-R
function data used to combine with the daily maximum 7-hr exposures for the simulated
population, while at moderate exertion (section 3D.2.8.1). The population at-risk is estimated by
multiplying the expected response rate by the number of people exposed in the relevant
population (and stratified by 7-hr average exposures, in 0.01 ppm increments), as shown in
Equation 3D-11:
N
Rk = ^ PjX(RRk | ej) Equation 3D-11
j=i
where:
ej = (the midpoint of) the /lh interval of personal exposure to O3
Pj = fraction of the population with O3 exposures of ej ppm
RRk | ej = kth response rate at O3 exposure concentration e,
N = number of intervals (categories) of O3 personal exposure concentration.
The number of 0.01 ppm intervals was maximally set to 16 (Figure 3D-12), however,
given the adjusted air quality scenarios, the midpoint values used in the risk calculation typically
ranged from 0.05 to 0.095 ppm. Conventional rounding was applied to the sum of the calculated
risk value.71
70 Markov chain Monte Carlo (MCMC) simulations require an initial adaptive "burn-in" set of iterations, which are
not used as part of the E-R curve output but allow the BMCMC sampling to stabilize.
71 For calculated risks (i.e., the summed number of people at each daily maximum 7-hr average exposure interval)
where the tenths value was less than 0.5, data were rounded down to the next lowest integer. For calculated risks
where the tenths value was greater than or equal to 0.5, data were rounded up to the next highest integer.
3D-80
-------
03
>
LU
FEV,
FEV,
(ppm)
>10%
>15%
>20%
0
0
0
0
0 005
00008
0.0001
0
0.010
0.0019
0.0002
0
0.015
00035
0.0004
0.0001
0.020
0.0056
0.0007
0.0001
0 025
0.0084
0.0Q11
0.0002
0.030
0.0123
0.0018
0.0003
0.035
0.0176
0 0029
00006
0 040
0.0249
0.0045
0.0Q11
0 045
0.0362
0 0070
0.0019
0.050
0.0495
0.0109
0.0033
0 055
0.0665
0.0167
0.0060
0.060
0.0883
0 0260
0.0108
0.065
0.1160
0 0404
0.0180
0 070
01497
0.0595
0.0296
0 075
0.1905
0.0860
00476
0 080
0.2378
0.1212
0.0738
0 085
0.2894
0.1642
0 1083
0 090
0.3415
0.2115
0 1482
0.095
0.3948
0.2614
0.1879
0.100
0.4474
0.3116
0.2219
0.105
04961
0.3560
0.2493
0.110
0.5393
0.3922
0.2704
0.115
0.5756
0.4199
0.2853
0.120
0.6055
0.4408
0.2952
0.125
0.6292
0.4567
0.3012
0.130
0.6477
0.4695
0.3047
0.135
0.6639
0.4789
0.3068
0 140
0.6774
04867
0.3082
0.145
0.6893
0.4912
0.3089
0.150
0.6999
0.4941
03093
0.155
0.7084
0.4959
0.3096
0.160
0 7133
0,4968
0.3097
80% T
—-Median Fit Function
Study Data (n subjects)
90
32
« 30%
331
c 20%
V 10%
ISO
60
0%
0.16
0
0.04
0.08
0.12
6.6-hr Ozone Exposure (ppm)
80%
—Median Fit Function
ft 70%
Study Data (n subjects)
60%
50%
« 30%
c 20%
«j 10%
331
ISO
60
0%
0
0.04
0.08
0.12
0.16
6.6-hr Ozone Exposure (ppm)
80% t
—Median Fit Function
§ 70%
fM
Al
.*60%
LU
X 50%
Study Data (n subjects)
90
« 30%
c 20%
10% (top panel),
>15 (middle panel), >20% (bottom panel). Drawn from the 2014 HREA, Table
6A-1 with processing and model development described by Abt (2013).
3D-81
-------
From a practical perspective, the population-based E-R function risk approach takes into
account that there is a fraction of the population that could experience a lung function decrement
at any daily maximum 7-hr average exposure level (i.e., from the minimum to the maximum,
including the level of the exposure benchmarks), having a low probability of decrements
resulting from low exposures and higher probability at the highest exposures. That said, the
approach allows for decrements to occur at exposures below those tested/observed in the
controlled human exposure studies, albeit a small population fraction (e.g., see the response
frequency for exposures below 60 ppb in Figure 3D-12), recognizing there is potential for
variability in the degree of sensitivity between the controlled human exposure study subjects and
the simulated population. Note also that because there is a strict limit on attaining a particular
ventilation rate for the simulated individuals (i.e., 7-hr average exposures for individuals must
simultaneously occur at moderate or greater exertion, section 3D.2.2.3.3), there may be some
potential to underestimate lung function responses if they were to occur at the higher end of the
exposure distribution (i.e., where exposures are >60 ppb) that coincide with breathing rates just
below those specified by the moderate or greater exertion requirement.
3D.2.8.2.2 The McDonnell-Stewart-Smith (MSS) Model
The McDonnell-Stewart-Smith (MSS) model, a statistical model to estimate FEVi
responses for individuals associated with short-term exposures to O3, was developed using
controlled human exposure data72 from studies using varying exposure durations and varying
exertion levels and breathing rates (McDonnell et al., 2007). Following the development of the
model by McDonnell et al., 2007), Schelegle et al. (2009) found a delay in response when
modeling FEVi decrements as a function of accumulated dose and estimated a threshold
associated with the delay. McDonnell et al. (2012) refit a 2010 version of the model that included
a body mass index (BMI) variable (McDonnell et al., 2010), adding data from eight additional
studies73 and incorporating a threshold parameter into the model, which allows for modeling a
delay in response until accumulated dose (i.e., accounting for decreases over time according to
first order reaction kinetics) reaches a threshold value. The threshold is not a concentration
threshold and does not preclude responses at low concentration exposures.
The MSS model was first used for estimating lung function risk in the 2014 HREA and
was based on the revised version of the model available at that time (McDonnell et al., 2012).
Another version of the MSS model has become available since the last review, which differs
72 Data were from 15 controlled human exposure studies that included 531 volunteers (ages 18 to 35), exposed to O3
on a total of 864 occasions (McDonnell et al., 2007).
73 Data from these eight additional studies included 201 individuals.
3D-82
-------
from the prior model in that it assumes that the intra-subject variance term (Var(s)) increases
with the response (McDonnell et al., 2013).74 Therefore, with a fixed ventilation rate, Var(s) in
this most recent version of the MSS model will be larger for higher exposure concentrations and
smaller for lower exposure concentrations. The most recent version of the MSS model is the
model described here and is the model used in this risk analysis.
The lung function model is conceptually a two-compartment model (Figure 3D-13). The
accumulated amount O3 (exposure concentration x ventilation rate, used to represent dose) is
modeled in the first compartment and modified by an exponential decay factor to yield an
intermediate quantity X. The response (FEVi reduction) of the individual to X is modeled in the
second compartment as a sigmoid-shaped function of the net accumulated dose. A threshold
parameter imposes the constraint that there is no response while the value of X is below the
threshold value.
compartment 1
compartment 2
logistic
response
X(t)
C(t)v(t)
(dose)
» AFEV
dX/dt=C(t)V(t)-aX
aX (metabolism)
Figure 3D-13. Conceptual representation of the two-compartment model used by the MSS
model. C is exposure concentration, V is ventilation rate, t is time, X is an
intermediate quantity, a is a decay constant. Adapted from Figure 1 in
McDonnell et al. (1999).
X is given by the solution of the differential Equation 3D-12:
^ - (3sX(t) Equation 3D-12
X(t) increases with "normalized dose" (C-l^6) over time for an individual and allows for
removal of "normalized dose" with a half-life of 1 Ifis through the 2nd term in Equation 3D-12.
74 The MSS model used for the 2014 HREA (McDonnell et al., 2012) assumed intra-subject variability was constant
for all exposures and responses. It had been shown previously that FEVi response varies within individuals
experiencing the same exposure and that the range of variation in response increases with higher exposure and
response (McDonnell et al., 1983). Evaluations based on a goodness-of-fit test and visual inspection of observed
versus predicted values indicate the most recent MSS model that better accounts for intra-subject variation is
improved in its estimation capabilities when compared to the previous MSS model (McDonnell et al., 2013).
3D-83
-------
The response function M is described in Equation 3D-13:
Mijk = (ft + P2Aik + (38Bik) [1+pJ-p3Tijk ~ 77^;} Equation 3D-13
where,
Tijk = max{0, Xijk - ft9} Equation 3D-14
P9 is a threshold parameter which allows X to increase up to the threshold before the
median response is allowed to exceed zero. By construction, when X = 0, then M=0. Because P3
and P4 are positive, when X > 0 then M > 0. Because X is never negative, neither is M. This
model calculates the percent FEVi decrement due to O3 exposure (compartment 2) as:
%AFEVlijk = eu< Mijk + eijk Equation 3D-15
Var^Eijk) = v1 + v2 eUi Mi]k Equation 3D-16
Note that a positive value of %AFEV1 means a decrease in effective lung volume or a
decrement in lung function. The above variance structure also allows for negative %AFEV1
values or an increase in lung volume, i.e., an improvement in lung function. The indices ij,k in
Equations 3D-12 to 3D-16 refer to the ith subject at the jth time for the kth exposure protocol for
that subject, while the variables are defined as:
t = time (minutes)
to = time at the start of the event
ti = time at the end of the event
C(t) = O3 exposure (ppm) at time t during the event
Ve(!) = expired minute volume (L min"1) at time l
BSA = body surface area (m2),
V(t) = VE(t)/BSA (L/min-m2) at time t
Aik = age (years) of the ith subject in the klh exposure protocol minus 23.8, the mean
age of the subjects
Bik = the body mass index (BMI, kg/m2) of the i'h subject in the klh exposure protocol
minus 23.1, the mean BMI of the subjects
Ui = subject-level zero-mean Gaussian random effect error/variability term (between-
individual variability not otherwise captured by the model)
eijk = Gaussian error/variability term, which includes measurement error and within-
individual variability not otherwise captured by the model
3D-84
-------
vi, V2 = constants used to parameterize the variance of syk. vi captures the intra-
individual noise in FEVi that is not due to ozone exposure, while V2 captures the
remaining intra-individual variability in FEVi.
Pi to fio unitless fitted model parameters (constant for all simulated individuals)
In general, this model would be considered a non-linear random-effects model (Davidian
and Giltinan, 2003). The best fit values (based on maximum likelihood) of the Ps and the
variances {sijk} were estimated from fits of the model to the clinical data (see McDonnell et al.,
2013) and are provided in Table 3D-21.
Table 3D-21. Estimated coefficients for the MSS lung function model.
Values for MSS Model Coefficients Used in Equations 3D-12 to 3D-17 A
Pi
02
03
04
05
06
08
09
Vi
V2
var(U)B
9.763
-0.4315
0.01281
30.92
0.002921
0.9525
0.4890
32.94
9.112
2.166
1.123
A Based on "Model 3" from McDonnell et al. (2013).
B The random sampling from the var(U) distribution was limited to ± 2 standard deviations.
As described above in estimating exposure, APEX uses activity pattern data to represent
a sequence of events that simulate the movement of a modeled person through geographical
locations and microenvironments during the simulation period. Each of these events are defined
by a geographic location, start time, duration, microenvironment visited, and activity performed.
Events in APEX are intervals of constant activity and exposure concentration, where an
individual is in one microenvironment and can range in duration from 1 to 60 minutes. In APEX,
because the exposure concentration C(t), exertion level, and normalized ventilation rate V(t) are
constant over an event, Equation 3D-17 provides an analytic solution for each event:
X(t±) = +^7(^)^(1 - e-Ps(ti-to)) Equation 3D-17
P 5
Note that C(t\) and V(t\) denote the (constant) values of C(t) and V(t) during the event75
from time to to time t\. In APEX, values of Ui and Sijk are drawn from Gaussian distributions with
mean zero and variances var(C7) and var(e), constrained to be within ±2 standard deviations from
the means (when sampled values fall outside of this range, they are discarded and resampled).
75 Events in APEX are intervals of constant activity and concentration, where an individual is in one
microenvironment. Events range in duration from one to 60 minutes. C(t\) and V(t\) denote the (constant) values
of C(t) and V(t) during the event from time to to time it.
3D-85
-------
The values of Ui are chosen once for each individual and remain constant for individuals
throughout the simulation. Values for Syk are sampled daily for each individual.
We are using this model to estimate lung function decrements for people ages 5 and
older. However, this model was developed using only data from individuals aged 18 to 35 and
the age adjustment term [(3i + (32 (Age ik - 23.8)] in the numerator of Equation 3D-13 is not
appropriate for all ages.76 Clinical studies data for children which could be used to fit the model
for children are not available at this time. In the absence of data, we are extending the model to
ages 5 to 18 by holding the age term constant at the age 18 level. Since the response increases as
age decreases in the range 18 to 35, this trend may extend into ages of children, in which case the
responses of children could be underestimated. However, the slope of the age term in the MSS
model is estimated based on data for ages 18 to 35 and does not capture differences in age trend
within this range; in particular, we do not know at what age the response peaks, which could be
above or below age 18. The evidence from clinical studies indicates that the responsiveness of
children to O3 is about the same as for young adults (ISA, Appendix 3, section 3.1.4.1.1) This
suggests that the age term for children should not be higher than the age term for young adults
(2014 HREA).77
Because the responses to O3 continuously declines from age 18 to 55 and for ages >55 the
response is generally considered minimal,78 here we assume the MSS model age term for ages 35
to 55 linearly decreases to zero and set it to zero for ages >55.79 To extend the age term to ages
outside the range of ages the MSS model is based on (ages 18-35), we re-parameterized the age
term in the numerator of Equation 3D-13 by [Pi + (32(ai Age + 012)], for different ranges of ages
(ai and 012 depend on age), requiring that these terms match at each boundary to form a piecewise
linear continuous function of age. As a result, the values of ai and 012 for four age ranges are
provided in Table 3D-22.
76 Note that the effect of age is also accounted for by using age-specific ventilation rate and body surface area. In
addition, APEX lung function risk for different age groups is also influenced by the time spent outdoors and the
activities engaged in by those groups, which vary by age.
77 See 2014 HREA Chapter 6 (sections 6.4.2 and 6.5.3) and Appendices 6D and 6E for details.
78 There is a recent 3-hr controlled human exposure study (EVR = 15-17 L/min-m2 during six 15-min exercise
periods) performed on healthy adults (ages 59.9 ± 4.5) that found 3-hr O3 exposures of 120 ppb yielded a
statistically significant reduction FEVi when compared to the filtered air response (Aijomandi et al., 2018). How
this relates to the magnitude and duration of exposures and ventilation rates of interest in this exposure and risk
analysis remain uncertain at this time.
79 "In healthy individuals, the fastest rate of decline in O3 responsiveness appears between the ages of 18 and
35 years ... During the middle age period (35-55 years), O3 sensitivity continues to decline, but at a much lower
rate. Beyond this age (>55 years), acute O3 exposure elicits minimal spirometric changes." (2013 ISA, p. 6-22)
3D-86
-------
Table 3D-22. Age term parameters for application of the MSS model to all ages.
Age Range
Pi
P2
ai
0(2
5-17
9.763
-0.4315
0
-5.8
18-35
9.763
-0.4315
1
-23.8
36-55
9.763
-0.4315
0.5714
CO
CO
I
>55
0
-0.4315
0
0
See Table 3D-21 for related MSS model coefficients.
As described above for the population-based E-R function risk approach (section
3D.2.8.2.1), the individual-based MSS model risk approach also allows for decrements to occur
at exposures below those tested/observed in the controlled human exposure studies, however, for
this approach there is not a strict limit on the ventilation per se. Indeed, FEVi decrements are
more likely to occur with high breathing rates (and concomitant with high exposures), but it is
not necessary that an individual's 7-hr average EVR reach their particular threshold (EVR
>17.32 ± 1.25 L/min-m2) for an individual to experience an adverse response as is used for both
the exposure benchmarks and the E-R function risk approach. The time-series of exposures,
breathing rate, and FEVi will vary with each diary event, with FEVi non-linearly dependent on
exposure levels/breathing rate from both the prior and current exposure/breathing events. That
said in doing so, the MSS approach could overstate risk when including instances where both the
exposures and ventilation rates are less than that tested/observed in the controlled human
exposure studies.
3D.2.9 Assessing Variability/Co-Variability and Characterizing Uncertainty
An important issue associated with any population exposure and risk assessment is the
assessment of variability and characterization of uncertainty. Variability refers to the inherent
heterogeneity in a population or variable of interest (e.g., residential air exchange rates). The
degree of variability cannot be reduced through further research, only better characterized with
additional measurement. Uncertainty refers to the lack of knowledge regarding the values of
model input variables (i.e., parameter uncertainty), the physical systems or relationships used
(i.e., use of input variables to estimate exposure or risk or model uncertainty), and in specifying
the scenario that is consistent with purpose of the assessment (i.e., scenario uncertainty).
Uncertainty is, ideally, reduced to the maximum extent possible through improved measurement
of key parameters and iterative model refinement.
Section 3D.2.9.1 summarizes how variability and co-variability are addressed in the
current exposure and risk analysis and is based on the above described input data and model
algorithms used. Section 3D.2.9.2 summarizes the overall approach used for the uncertainty
3D-87
-------
characterization. The outcome of the updated uncertainty characterization, which builds upon the
important uncertainties identified in the IRP (Appendix 5A) and addressed in this current
exposure and risk analyses, is discussed below in section 3D.3.4.
3D.2.9.1 Variability and Co-variability Assessment
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. The details regarding many of the variability distributions used as model inputs are
described above, while details regarding the variability addressed within its algorithms and
processes are found in the APEX User Guides (U.S. EPA, 2019a, U.S. EPA, 2019b).
APEX 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. APEX simulates individuals and then calculates O3 exposure and lung function
risk for each of these simulated individuals. This 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 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.
We note also that correlations and non-linear relationships between variables input to the
model can result in the model producing inaccurate results if the inherent relationships between
these variables are not preserved. APEX is designed to account for co-variability, or linear and
nonlinear correlation among the model inputs, provided that enough is known about these
relationships to specify them. This is accomplished by providing inputs that enable the
correlation to be modeled explicitly within APEX. For example, there is a non-linear relationship
between the outdoor temperature and air exchange rate in homes. One factor that contributes to
this non-linear relationship is that windows tend to be closed more often when temperatures are
at either low or high extremes than when temperatures are moderate. This relationship is
explicitly modeled in APEX by specifying different probability distributions of air exchange
rates for different ambient air temperatures. Note that where possible, we identified and
3D-88
-------
incorporated the observed variability in input data sets rather than employing standard default
assumptions and/or using point estimates to describe model inputs. In any event, APEX models
variability and co-variability in two ways:
• Stochastically. The user provides APEX with probability distributions characterizing the
variability of many input parameters. These are treated stochastically in the model and
the estimated exposure distributions reflect this variability. For example, the rate of O3
decay in houses can depend on a number of factors which we are not able to explicitly
model at this time, due to a lack of data. However, we can specify a distribution of
removal rates that reflects observed variations in O3 decay. APEX randomly samples
from this distribution to obtain values that are used in the mass balance model. Further,
co-variability can be modeled stochastically through the use of conditional distributions.
If two or more parameters are related, conditional distributions that depend on the values
of the related parameters are input to APEX. For example, the distribution of air
exchange rates (AERs) in a house depends on the outdoor temperature and whether or not
air conditioning (A/C) is in use. In this case, a set of AER distributions is provided to
APEX for different ranges of temperatures and A/C use, and the selection of the
distribution in APEX is driven by the temperature and A/C status at that time.
• Explicitly. For some variables used in modeling exposure, APEX models variability and
co-variability explicitly and not stochastically. For example, the complete series of hourly
ambient air O3 concentrations and hourly temperatures are used in the exposure and risk
calculations. These are input to the model continuously in the time period modeled at
different spatial locations, and in this way the variability and co-variability of hourly O3
concentrations and hourly temperatures are modeled explicitly.
Important sources of the variability and co-variability accounted for by APEX and used
for this exposure and risk analysis are provided in Table 3D-23 and Table 3D-24, respectively.
Table 3D-23. Summary of how variability was incorporated into the exposure and risk
analysis.
Component
Variability Source
Summary
Ambient Air
Concentration Input
(Appendix 3C)
CAMx Air Quality
Modeling
Spatial: model results are output at 12 km spatial resolution for the full
CONUS domain.
Temporal: model results are calculated and archived at hourly resolution
for the full 2016 calendar year.
CAMx/HDDM
estimates of 1-hr
ambient air O3
concentrations
Spatial: simulations of O3 response to changes in emissions predicted to
multiple monitors in eight geographically representative study areas.
Temporal: hourly O3for each of three years (2015-2017).
Ambient air monitor
hourly
concentrations
Spatial: local ambient air monitor sites used to interpolate adjusted 03
concentrations to census tracts, including monitors outside of the study
area.
Temporal: pattern of hourly O3 concentrations at census tracts also
informed by local ambient air monitors.
3D-89
-------
Component
Variability Source
Summary
Population data
Individuals are randomly sampled from U.S. census tracts used in each
study area, stratified by age (single years) and sex probabilities (U.S.
Census Bureau, 2012).
Employment
Work status is randomly generated from U.S. census tracts, stratified by
aqe and sex employment probabilities (U.S. Census Bureau, 2012).
Simulated
Individuals
Activity pattern data
Data diaries used to represent locations visited and activities performed
by simulated individuals are randomly selected from CFIAD (nearly
180,000 diaries) using six diary pools stratified by two day-types
(weekday, weekend) and three temperature ranges (< 55.0 °F, between
55.0 and 83.9 °F, and >84.0 °F). CFIAD diaries capture real locations that
people visit and the activities they perform, ranging from 1 -min to 1 -hr in
duration (U.S. EPA, 2019c).
Commuting data
Employed individuals are probabilistically assigned ambient air
concentrations originating from either their home or work block based on
U.S. Census derived tract-level commuter data (U.S. DOT, 2012; U.S.
Census Bureau, 2012).
Longitudinal
profiles
A sequence of diaries is linked together for each individual that preserves
both the inter- and intra-personal variability in human activities (Glen et
al„ 2008).
Asthma prevalence
Asthma prevalence is stratified by sex, single age years for children (5-
17), seven adult age groups, (18-24, 25-34, 35-44, 45-54, 55-64, 65-74,
and, >75), three regions (Midwest, Northeast, and South), and U.S.
Census tract level poverty ratios (Attachment 1).
Resting metabolic
rate
Five age-group and two sex-specific regression equations, use body mass
and aqe as independent variables (U.S. EPA (2018), Appendix Fl).
Metabolic
equivalents by
activity
Randomly sampled from distributions developed for specific activities
(some age-specific) (U.S. EPA, 2019c)
Oxygen uptake per
unit of energy
expended
Randomly sampled from a uniform distribution to convert energy
expenditure to oxygen consumption (U.S. EPA, 2019a, U.S. EPA, 2019b).
Physiological
Factors Relevant to
Ventilation Rate
Body mass
Randomly selected from population-weighted lognormal distributions with
age- and sex-specific geometric mean (GM) and geometric standard
deviation (GSD) derived from the National Flealth and Nutrition
Examination Survey (NFIANES) for the years 2009-2014 (U.S. EPA
(2018), Appendix G).
Body surface area
Sex-specific exponential equations using body mass as an independent
variable (Burmaster, 1998).
Height
Randomly sampled from population-weighted normal distributions
stratified by single age years and two sexes developed from 2009-2014
NFIANES data (U.S. EPA (2018), Appendix G).
Ventilation rate
Event-level activity-specific regression equation using oxygen
consumption rate (VO2) and maximum VO2 as independent variables, and
accounting for intra- and inter-personal variability (U.S. EPA (2018),
Appendix Fl).
Fatigue and EPOC
APEX approximates the onset of fatigue, controlling for unrealistic or
excessive exercise events in an individual's activity time-series while also
estimating excess post-exercise oxygen consumption (EPOC) that may
occur following vigorous exertion activities using several equations and
3D-90
-------
Component
Variability Source
Summary
input variable distributions (Isaacs et al„ 2007; U.S. EPA, 2019a; U.S.
EPA, 2019b).
Equivalent
ventilation rate
A randomly sampled value is selected for each simulated individual from a
normal distribution derived from the controlled human exposure study
data. This approach accounts for interpersonal variability in exertion level
that occur during exposure events that include exercise and rest periods
(Attachment 2).
Microenvironmental
Approach
General
Seven total microenvironments are represented, including those expected
to be associated with high exposure concentrations (i.e., outdoors and
outdoor near-road). There is variability in particular microenvironmental
algorithm inputs. This results in differential exposures for each individual
(and event) because people spend varying amounts of time within each
microenvironment and ambient air concentrations vary within and among
study areas.
Spatial Variability
Ambient air concentrations used in microenvironmental algorithms vary
spatially within (i.e., census tracts) and among study areas (U.S.
qeoqraphic reqions).
Temporal
Variability
All exposure calculations are performed at the event-level when using
either factors or mass balance approach (durations can be as short as
one minute). For the indoor microenvironments, using a mass balance
model accounts for 03 concentrations occurring during a previous hour
(and of ambient air origin) to calculate a current event's indoor O3
concentrations.
Air exchange rates
For residences, several lognormal distributions are sampled for up to five
daily mean temperature ranges, study area region (2014 HREA Appendix
5E) and using study-area specific A/C prevalence rates from AHS survey
data (U.S. Census Bureau, 2019). For restaurants, a lognormal
distribution is sampled based on Bennett etal. (2012). For schools, a
Weibull distribution is sampled based on data from Lagus Applied
Technology (1995), Shendell etal. (2004), and Turk etal. (1989).
Removal rates
Values randomly selected from a lognormal distribution for the three
indoor microenvironments modeled (Lee et al„ 1999).
PE and PROX
factors
Penetration and proximity factors randomly sampled from probability
distributions for inside-vehicle and near-road microenvironments
(American Petroleum Institute (1997), Appendix B; Johnson et al„ 1995).
Lung Function Risk
Population-based
Exposure
Response Function
A continuous E-R function was derived using data from several controlled
human exposure studies and a logit-linear modeling approach. The full
distribution of population exposures was stratified by fine-scale bins (10
ppb) and linked to the continuous E-R function to estimate lung function
risk.
Individual-based
MSS model
Calculation accounts for variability in age, body mass, and the continuous
time-series of exposures and breathing rates. Residual terms (U and )
addresses intra- and inter-variability in responses across the simulated
population.
3D-91
-------
Table 3D-24. Important components of co-variability in exposure modeling.
Type of Co-variability
Modeled
by APEX?
Treatment in APEX / Comments
Within-person correlationsA
Yes
Sequence of activities performed, microenvironments
visited, and general physiological parameters (body
mass, height, ventilation rates).
Between-person correlations
No
Perhaps not important, assuming the same likelihood of
the population of individuals either avoiding or
experiencing an exposure event based on a social
(group) activity.
Correlations between profile variables and
microenvironment parameters
Yes
Profiles are assigned microenvironment parameters.
Correlations between demographic
variables and activities
Yes
Census tract demographic variables, appropriately
weighted and stratified by age and sex, are used in
activity diary selection.
Correlations between activities and
microenvironment parameters
No
Perhaps important, but do not have data. For example,
frequency of opening windows when cooking or smoking
tobacco products.
Correlations among microenvironment
parameters in the same microenvironment
Yes
Modeled with joint conditional variables.
Correlations between demographic
variables and air quality
Yes
Modeled with the spatially varying census tract
demographic variables (age and sex) and census tract air
quality data input to APEX.
Correlations between meteorological
variables and activities
Yes
Daily varying temperatures are used in activity diary
selection.
Correlations between meteorological
variables and microenvironment parameters
Yes
The distributions of microenvironment parameters can be
functions of temperature.
Correlations between drive times in CHAD
and commute distances traveled
Yes
CHAD diary selection is weighted by commute times for
employed persons during weekdays.
Consistency of occupation/school
microenvironmental time and time spent
commuting/busing for individuals from one
working/school day to the next.
No
Simulated individuals are assigned activity diaries
longitudinally without regard to occupation or school
schedule (note though, longitudinal variable used to
develop annual profile is time spent outdoors).
A The term correlation is used to represent linear and nonlinear relationships.
3D.2.9.2 Approach for Uncertainty Characterization
While it may be possible to capture a range of exposure or risk values by accounting for
variability inherent to influential factors, the true exposure or risk for any given individual within
a study area may be unknown, although it can be estimated. To characterize health risks,
exposure and risk assessors commonly use an iterative process of gathering data, developing
models, estimating exposures and risks, evaluating results for correctness and identifying areas
for potential improvement, given the goals of the assessment, scale and complexity of the
assessment performed, and limitations of the input data available. However, important
3D-92
-------
uncertainties often remain in any one of the data sets, tools, and approaches used and emphasis is
then placed on characterizing the nature of that uncertainty and its impact on exposure and risk
estimates.
The overall approach used for this exposure and risk generally follows that described by
WHO (2008) but varies in that a greater focus has been placed on evaluating the direction and
the magnitude of the uncertainty. This refers to qualitatively rating how the source of
uncertainty, in the presence of alternative information, may affect the estimated exposures and
health risk results. Following the identification of key uncertainties, we subjectively scale the
overall impact of the identified uncertainty by considering the relationship between the source of
uncertainty and the exposure concentrations (e.g., low, medium, or high potential impact). Also
to the extent possible, we include an assessment of the direction of influence, indicating how the
source of uncertainty may be affecting exposure or risk estimates (e.g., the uncertainty could lead
to over-estimates, under-estimates, or both directions). Further, and consistent with the WHO
(2008) guidance, we discuss the uncertainty in the knowledge-base (e.g., the accuracy of the data
used, recognition of data gaps) and, where possible, particular assessment design decisions (e.g.,
selection of particular model forms). 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.
We further recognize that uncertainties associated with APEX exposure modeling 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, 2008, 2010, 2014, 2018), all complementary to quantitative
uncertainty characterizations conducted for the 2007 O3 exposure assessment by Langstaff
(2007). Conclusions drawn from each of these characterizations are also considered here in light
of new information, data, tools, and approaches used in this exposure and risk analysis.
3D.3 POPULATION EXPOSURE AND RISK RESULTS
Exposure and risk results are presented here for simulated populations residing in the
eight study areas - Atlanta, Boston, Dallas, Detroit, Philadelphia, Phoenix, Sacramento, and St.
Louis - for a three-year air quality scenario in which air quality conditions just meet the current
primary 8-hr O3 standard (70 ppb, annual 4th highest daily maximum 8-hr average concentration,
averaged across 3-years) and two other air quality scenarios (i.e., design values of 75 and 65
ppb). Hourly concentrations of O3 in ambient air for the three hypothetical air quality scenarios
are estimated at census tracts in each study area as described in section 3D.2 above. Population
exposure and risk associated with these concentrations is estimated using the APEX model
simulations (section 3D.2) and is briefly described with the following.
3D-93
-------
APEX uses the hourly air quality surface in each study area, along with U.S. census tract
population demographics, to estimate the number of days per year each simulated individual in a
particular study area experiences a daily maximum 7-hr average O3 exposure at or above
benchmark levels of 60, 70, and 80 ppb (section 3D.2.8.1). These short-term exposures were
evaluated for children (5-18 years old), adults (>18 years old), and those with asthma within each
of these two study groups when the exposure corresponded with moderate or greater exertion
(i.e., the individual's EVR >17.32 ± 1.25 L/minute-m2).
Then, individuals expected to experience a lung function decrement (i.e., reduction in
FEVi >10%, >15%, >20%) were estimated using two approaches. The first approach linked the
population-based daily maximum 7-hr exposures while at moderate or greater exertion with an
exposure-response function derived from controlled human exposure study data (section
3D.2.8.2.1). The second lung function risk approach, considered an individual-based approach
here, used the McDonnell-Stewart-Smith (MSS) FEVi model (McDonnell et al., 2013) (section
3D.2.8.2.2). The MSS uses the time-series of O3 exposure and corresponding ventilation rates for
each APEX simulated individual to estimate their personal time-series of FEVi reductions,
selecting the daily maximum reduction for each person. The number of individuals estimated to
experience decrements are then aggregated to the population level. Again, of interest for both of
these lung function risk approaches is the number of days per year each simulated individual in a
particular study area experiences a lung function decrement.
Study area characteristics and the composition of the simulated population are provided
in section 3D.3.1. Exposure results are presented in a series of tables that allow for simultaneous
comparison of the exposure and risk metrics across the eight study areas and three simulation
years. Two types of results are provided for each study area: the percent (and number) of the
simulated population exposed at or above selected benchmarks, stratified by the number of
occurrences (i.e., days) in a year (section 3D.3.2) and the percent (and number) of the simulated
population experiencing a reduction in FEVi >10%, >15%, >20%, also stratified by the number
of days in a year (section 3D.3.3). Tables summarizing all of the exposure and risk results for
each study area are provided in Attachment 4.
3D.3.1 Characteristics of the Simulated Population and Study Areas
The eight study areas differ in population, geographic size, and demographic features
(Table 3D-25). In each of the eight study areas, APEX simulated O3 exposures and risks for
60,000 individuals,80 the demographic features of which were based on the information
80 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.
3D-94
-------
associated with the hundreds to thousands of census tracts within each area (as described in
section 3D.2.1 above).
Asthma prevalence in each modeling domain was estimated based on the 2013-2017
NHIS asthma prevalence data and the demographic characteristics for each study area (e.g., age,
sex and family income) using the methodology summarized in section 3D.2.2.2. Accordingly,
the percent of the simulated populations with asthma within the exposure modeling domain
varied by study area (Table 3D-25). The Dallas, Phoenix, and Sacramento study areas had the
lowest percent of children with asthma (9.2 to 9.6%), while Atlanta and Boston had the highest
percent of children with asthma (11.8 to 12.3%). The Dallas study area had the lowest percent of
adults with asthma (7.2%), while Boston and Detroit had the highest percent of adults with
asthma (both 10.9%). The statistics presented here are the aggregate of the study area as a whole,
within which asthma prevalence varied widely as the modeling approach fully accounted for the
variation in asthma prevalence across census tracts with demographic factors such as family
income to poverty ratios, age, and sex (and as described in section 3D.2.2.2).81 Nationally,
asthma prevalence is 7.9%; for children it is 8.4% and for adults it is 7.7% (Chapter 3, Table 3-
1). The asthma prevalence for children, adults, and the total population estimated for each of the
eight study areas are all greater than that of the national asthma prevalence, except for adults in
Dallas which has a slightly lower asthma prevalence. This suggests that overall, the at-risk
population simulated in the eight study areas could represent at-risk populations in other U.S.
urban areas that have a similarly above average asthma prevalence.
81 Representing the variation in asthma prevalence that occurs at the census tract level provides a level of resolution
for identification of at-risk individuals that is directly compatible with the resolution of the spatially varying
ambient air concentrations. In this way, the population in census tracts with higher concentrations is represented
appropriately with regard to asthma prevalence and exposures of the at-risk individuals with asthma are not
under-represented.
3D-95
-------
Table 3D-25. Summary of study area features and the simulated population.
Study Area
(Land Area - km2)A
Population
Group
(age range)
Simulated
Population
Simulated
Population with
Asthma
% of Simulated
Population with
Asthma
Atlanta
(30,655)
Children (5-18)
1,210,594
142,400
11.8
Adults (19-90)
4,226,009
359,375
8.5
All (5-90)
5,436,603
501,775
9.2
Boston
(25,117)
Children (5-18)
1,365,267
167,617
12.3
Adults (19-90)
5,870,125
642,224
10.9
All (5-90)
7,235,392
809,841
11.2
Dallas
(42,664)
Children (5-18)
1,418,728
130,421
9.2
Adults (19-90)
4,688,180
336,898
7.2
All (5-90)
6,106,908
467,319
7.7
Detroit
(16,884)
Children (5-18)
1,040,588
116,899
11.2
Adults (19-90)
3,932,484
427,221
10.9
All (5-90)
4,973,072
544,119
10.9
Philadelphia
(18,959)
Children (5-18)
1,309,547
146,982
11.2
Adults (19-90)
5,228,541
503,305
9.6
All (5-90)
6,538,088
650,287
9.9
Phoenix
(34,799)
Children (5-18)
849,200
81,396
9.6
Adults (19-90)
2,980,062
269,845
9.1
All (5-90)
3,829,262
351,240
9.2
Sacramento
(18,871)
Children (5-18)
465,845
45,208
9.7
Adults (19-90)
1,715,065
138,253
8.1
All (5-90)
2,180,910
183,461
8.4
St. Louis
(23,019)
Children (5-18)
546,393
56,039
10.3
Adults (19-90)
2,146,037
203,039
9.5
All (5-90)
2,692,430
259,078
9.6
All Study Areas
Combined
Children (5-18)
8,206,162
886,960
10.8
Adults (19-90)
30,786,503
2,880,160
9.4
All (5-90)
38,992,665
3,767,120
9.7
A From Appendix 3C, Table 3C-1.
3D.3.2 Exposures at or above Benchmark Concentrations
The exposure to benchmark comparisons are presented in a series of tables focusing on
the benchmark levels (i.e., people experiencing daily maximum 7-hr average O3 exposures >60,
70, and 80 ppb while at moderate or greater exertion). The full range of ambient air O3
concentrations for a 3-year O3 season (2015-2017) were used by APEX, providing a range of
estimated exposures. Adjusted air quality surfaces used to represent three air quality scenarios
were developed using 2015-2017 design values modeled sensitivities to changes in precursor
emissions (section 3D.2.3.3), and then interpolated to census tract centroids (section 3D.2.3.4).
3D-96
-------
Exposures were estimated for four study groups of interest (i.e., school-age children (5-18),
school-age children with asthma, adults (19-90), and adults with asthma).
In this exposure and risk analysis, we are primarily interested in O3 exposures associated
with the ambient air quality adjusted to just meet the current standard (70 ppb, annual 4th highest
daily maximum 8-hr average concentration, averaged over a 3-year period). Provided are the
percent and number of people in each study group estimated to experience 7-hr exposures at or
above the benchmarks, while at moderate or greater exertion (section 3D.3.2.1). For each
exposure metric and study group, the occurrence of single-day (at least one day per year) and
multi-day (at least 2, 4, or 6 days per year) exposures are presented. Exposure results for the two
other adjusted air quality scenarios (the 75 ppb and 65 ppb scenarios) are presented in sections
3D.3.2.2 and 3D.3.2.3, respectively. These two sections present only the percent of each study
group estimated to experience exposures at or above benchmarks while at moderate or greater
exertion, for single-day and multiday exposures during a year, and not also the number of
simulated individuals in each study group. The complete exposure results associated with all
simulated years, air quality scenarios, the four study groups, and eight study areas are found in
Attachment 4.
In general, and for all air quality scenarios, the percent of children estimated to
experience exposures at or above any of the benchmarks is consistently higher than that
estimated for adults. This is expected because children spend a greater amount of time outdoors,
and at a greater frequency, while at moderate or greater exertion when compared to adults (2014
HREA, sections 5.4.1 and 5.4.2). Estimated exposures for healthy people are similar to people
with asthma when considered on a percent of population basis. This is because similar diary data
are used to simulate the activity patterns of each study group, justified by evaluations that
indicated similarities in time spent outdoors, participation rate, and exertion level for people with
asthma when compared to healthy individuals (section 3D.2.5.3). When considering the
estimated exposures in terms of population counts, while children comprise about 20% of the
simulated population (Table 3D-25), the number of children experiencing exposures at or above
the benchmarks is greater than that of adults. Again, this a direct result of the differences in time
spent outdoors performing activities at elevated exertion. And finally, Detroit, Phoenix, and St.
Louis have a higher percent of individuals at or above benchmark levels relative to the other
study areas, likely influenced by their having an hourly O3 concentration distribution shape that,
overall, is more skewed to the right and/or has heavy tails at the uppermost percentiles (Figure
3D-7).
3D-97
-------
3D.3.2.1 Air Quality Just Meeting the Current Standard
With air quality adjusted to just meet the current standard, 0 to <0.1% of people in all
study groups were estimated to experience at least one daily maximum 7-hr exposure per year at
or above the 80-ppb benchmark (Table 3D-26). The occurrence of 7-hr O3 exposures at or above
70-ppb are also limited, even considering the worst year air quality in the three-year period, with
1% or fewer children (and children with asthma) in all study areas estimated to experience at
least one daily maximum 7-hr exposure per year at or above the 70-ppb benchmark. For the same
benchmark, 0.2% or fewer adults (and adults with asthma) were estimated to experience similar
exposures when considering the worst air quality year. When considering the 60-ppb benchmark,
on average, between about 3 to 9% of children (and children with asthma) experienced at least
one daily maximum 7-hr exposure at or above that benchmark, while during the worst air quality
year, the range in percent of children exposed extends slightly upwards (about 4 to 11%),
indicating limited variability in ambient air concentrations across the three-year period. Again,
there were fewer adults (and adults with asthma) exposed considering this same benchmark, on
average ranging from 0.2 to 1.5% of this study group and the worst air quality year ranging from
0.2 to 1.8%.
The number of simulated people in each study group estimated to experience at least one
7-hr exposure per year at or above the benchmarks is provided in Table 3D-27. As noted above,
there are few simulated people expected to experience a 7-hr exposure at or above the 80-ppb
benchmark, at most about 1,200 children and 500 adults when considering the worst year in a
single study area. Regarding the 70-ppb benchmark, on average, between about 700 to 8,300
children are estimated to experience at least one 7-hr exposure at or above that benchmark, while
the range for adults is about half that of children (400 to about 3,700), the range of which
considers the eight study areas. When considering the worst year, fewer than 12,000 children and
7,700 adults are estimated to experience at least one 7-hr exposure at or above the 70-ppb
benchmark in each study area. On average, the number of children estimated to experience at
least one 7-hr O3 exposure at or above the 60-ppb benchmark could be as high as nearly 70,000
in a few study areas, while for adults the number is just below 45,000. During the worst air
quality year, the estimated number of people experiencing at least one exposure at or above this
same benchmark could be as high as about 100,000 for children and 63,000 for adults. As a
whole, the patterns for people with asthma are similar though having smaller counts, the value of
which is dictated by the asthma prevalence in each area (Table 3D-25). In general, the number of
children with asthma at or above a benchmark would be about 10.8% of that estimated for all
children, while the number adults with asthma at or above a benchmark is about 9.4% of that
estimated for all adults.
3D-98
-------
Multiday exposures are limited when considering air quality adjusted to just meet the
current standard. For example, there are no children estimated to experience at least two days
with 7-hr O3 exposures at or above the 80-ppb benchmark and <0.1% at or above the 70-ppb
benchmark (Table 3D-28 and Table 3D-29). When considering the worst air quality year, <5% of
children (and <0.4% of adults) are estimated to experience at least two days with 7-hr O3
exposures at or above the 60-ppb benchmark. There are no people estimated to experience at
least four days with 7-hr O3 exposures at or above the 70-ppb benchmark except in one study
area (Table 3D-30 and Table 3D-31), and <0.5% experience at least six days with 7-hr O3
exposures at or above the 60-ppb benchmark (Attachment 4).
3D-99
-------
Table 3D-26. Percent of people estimated to experience at least one exposure at or above
benchmarks while at moderate or greater exertion, for air quality adjusted to
just meet the current standard.
Study
Group
Study Area
60 ppb Benchmark (7-hr)A
(% per Year)
70 ppb Benchmark (7-hr)A
(% per Year)
80 ppb Benchmark (7-hr)A
(% per Year)
Avg
Min
Max
Avg
Min
Max
Avg
Min
Max
Atlanta
3.3
1.4
5.2
0.4
0.1
0.8
<0.1
0
0.1
Boston
4.4
3.4
6.0
0.6
0.4
0.9
<0.1
<0.1
<0.1
Dallas
4.9
2.4
6.8
0.4
0.2
0.7
<0.1
0
<0.1
Children
Detroit
6.7
5.0
9.2
0.5
0.1
0.9
<0.1
0
<0.1
Philadelphia
4.1
3.9
4.2
0.4
0.3
0.4
<0.1
0
<0.1
Phoenix
8.2
6.0
10.6
0.2
<0.1
0.6
0
0
0
Sacramento
3.2
2.3
3.9
0.2
0.1
0.3
0
0
0
St. Louis
6.0
4.1
8.7
0.4
0.1
0.9
<0.1
0
<0.1
Atlanta
3.6
1.5
5.8
0.5
0.1
0.9
<0.1
0
0.1
Boston
5.1
3.7
7.0
0.7
0.5
1.0
<0.1
0
0.1
Children
with
Asthma
Dallas
5.3
2.2
7.4
0.4
0.3
0.7
<0.1
0
<0.1
Detroit
7.3
5.4
10.0
0.5
0.1
0.9
0
0
0
Philadelphia
4.3
4.1
4.4
0.4
0.4
0.4
0
0
0
Phoenix
8.8
6.6
11.2
0.3
0
0.7
0
0
0
Sacramento
3.3
2.6
4.0
0.2
0.1
0.3
0
0
0
St. Louis
6.0
3.9
9.0
0.3
<0.1
0.8
<0.1
0
<0.1
Atlanta
0.5
0.2
0.8
0.1
<0.1
0.1
<0.1
0
<0.1
Boston
0.5
0.3
0.8
0.1
<0.1
0.1
<0.1
<0.1
<0.1
Dallas
0.8
0.3
1.2
<0.1
<0.1
0.1
<0.1
0
<0.1
Adults
Detroit
1.0
0.8
1.6
0.1
<0.1
0.2
0
0
0
Philadelphia
0.5
0.5
0.5
<0.1
<0.1
<0.1
<0.1
0
<0.1
Phoenix
1.5
1.1
1.8
<0.1
<0.1
0.1
0
0
0
Sacramento
0.4
0.3
0.5
<0.1
<0.1
<0.1
0
0
0
St. Louis
0.9
0.5
1.3
<0.1
<0.1
0.1
0
0
0
Atlanta
0.4
0.2
0.6
<0.1
0
<0.1
0
0
0
Boston
0.4
0.2
0.7
<0.1
0
0.1
<0.1
0
<0.1
Dallas
0.6
0.2
0.9
<0.1
0
0.1
0
0
0
Adults with
Detroit
0.8
0.6
1.2
0.1
<0.1
0.2
0
0
0
Asthma
Philadelphia
0.4
0.3
0.5
<0.1
0
0.1
0
0
0
Phoenix
1.3
1.0
1.5
<0.1
0
<0.1
0
0
0
Sacramento
0.2
0.2
0.2
<0.1
<0.1
<0.1
0
0
0
St. Louis
0.7
0.4
1.2
<0.1
0
0.1
0
0
0
^ Calculated percent is rounded to the nearest tenth decimal using conventional rounding. Values equal to zero are designated by "0" (there
are no individuals exposed at that level). Small, non-zero values that do not round upwards to 0.1 (i.e., <0.05) are given a value of "<0.1
3D-100
-------
Table 3D-27. Number of people estimated to experience at least one exposure at or above
benchmarks while at moderate or greater exertion, for air quality adjusted to
just meet the current standard.
Study
Group
Study Area
60 ppb Benchmark (7-hr)A
(# per Year)
70 ppb Benchmark (7-hr)A
(# per Year)
80 ppb Benchmark (7-hr)A
(# per Year)
Avg
Min
Max
Avg
Min
Max
Avg
Min
Max
Children
Atlanta
39909
17291
63455
5199
1069
9947
464
0
1211
Boston
59549
46465
81939
8305
5438
11923
372
91
592
Dallas
69794
34499
96261
5864
3168
9718
173
0
284
Detroit
69627
52203
95509
5093
1492
9487
29
0
52
Philadelphia
53117
51116
54674
4656
4191
5151
44
0
87
Phoenix
69569
50754
89775
1953
269
4784
0
0
0
Sacramento
14928
10645
18378
727
272
1203
0
0
0
St. Louis
32841
22320
47609
2331
446
4863
12
0
36
Children
with
Asthma
Atlanta
5152
2078
8333
666
141
1271
67
0
202
Boston
8518
6166
11605
1145
796
1616
61
0
114
Dallas
6952
2908
9813
576
355
946
8
0
24
Detroit
8544
6209
11776
578
121
1110
0
0
0
Philadelphia
6264
6024
6504
597
524
655
0
0
0
Phoenix
7171
5336
9143
226
0
552
0
0
0
Sacramento
1517
1157
1871
93
54
155
0
0
0
St. Louis
3364
2195
4927
191
18
437
3
0
9
Adults
Atlanta
21318
9790
34160
2512
282
5001
117
0
352
Boston
30362
19274
48429
3391
2152
5283
294
98
489
Dallas
36646
14611
54461
2318
1328
4141
26
0
78
Detroit
40920
30215
62264
3692
1049
7668
0
0
0
Philadelphia
26375
25184
27973
1597
1481
1830
29
0
87
Phoenix
44552
33178
54585
745
149
1788
0
0
0
Sacramento
7318
4688
9176
400
229
600
0
0
0
St. Louis
18981
11016
28185
942
72
2075
0
0
0
Adults
with
Asthma
Atlanta
1385
775
2113
70
0
141
0
0
0
Boston
2544
1370
4207
294
0
685
65
0
98
Dallas
2109
781
3047
104
0
234
0
0
0
Detroit
3299
2425
5047
306
66
655
0
0
0
Philadelphia
2179
1569
2614
87
0
261
0
0
0
Phoenix
3377
2831
3973
50
0
99
0
0
0
Sacramento
295
257
343
38
29
57
0
0
0
St. Louis
1395
787
2325
72
0
179
0
0
0
A These values represent the population of individuals exposed in each study area. Values equal to zero are indicated by "0" (there are no
individuals exposed at the level).
3D-101
-------
Table 3D-28. Percent of people estimated to experience at least two exposures at or above
benchmarks while at moderate or greater exertion, for air quality adjusted to
just meet the current standard.
Study
Group
Study Area
60 ppb Benchmark (7-hr)A
(% per Year)
70 ppb Benchmark (7-hr)A
(% per Year)
80 ppb Benchmark (7-hr)A
(% per Year)
Avg
Min
Max
Avg
Min
Max
Avg
Min
Max
Atlanta
0.6
0.1
1.1
<0.1
0
<0.1
0
0
0
Boston
0.8
0.5
1.4
<0.1
<0.1
<0.1
0
0
0
Dallas
1.2
0.4
2.1
<0.1
<0.1
0.1
0
0
0
Children
Detroit
1.7
1.0
2.8
<0.1
<0.1
0.1
0
0
0
Philadelphia
0.8
0.7
0.9
<0.1
<0.1
<0.1
0
0
0
Phoenix
2.9
1.7
4.3
<0.1
0
<0.1
0
0
0
Sacramento
0.6
0.3
0.9
<0.1
0
<0.1
0
0
0
St. Louis
1.5
0.7
2.6
<0.1
<0.1
<0.1
0
0
0
Atlanta
0.7
0.1
1.2
<0.1
0
<0.1
0
0
0
Boston
1.0
0.6
1.6
<0.1
0
<0.1
0
0
0
Children
with
Asthma
Dallas
1.2
0.3
2.2
<0.1
0
0.1
0
0
0
Detroit
1.9
1.1
2.9
<0.1
0
<0.1
0
0
0
Philadelphia
0.9
0.8
0.9
<0.1
0
<0.1
0
0
0
Phoenix
3.2
1.8
4.9
<0.1
0
0.1
0
0
0
Sacramento
0.6
0.4
0.9
<0.1
0
<0.1
0
0
0
St. Louis
1.3
0.6
2.2
<0.1
0
<0.1
0
0
0
Atlanta
<0.1
<0.1
0.1
0
0
0
0
0
0
Boston
<0.1
<0.1
0.1
<0.1
0
<0.1
0
0
0
Dallas
0.1
<0.1
0.1
<0.1
0
<0.1
0
0
0
Adults
Detroit
0.1
0.1
0.2
<0.1
0
<0.1
0
0
0
Philadelphia
<0.1
<0.1
<0.1
<0.1
0
<0.1
0
0
0
Phoenix
0.3
0.2
0.4
<0.1
0
<0.1
0
0
0
Sacramento
<0.1
<0.1
0.1
<0.1
0
<0.1
0
0
0
St. Louis
0.1
<0.1
0.2
0
0
0
0
0
0
Atlanta
<0.1
0
0.1
0
0
0
0
0
0
Boston
<0.1
0
0.1
<0.1
0
<0.1
0
0
0
Adults
with
Asthma
Dallas
<0.1
<0.1
<0.1
0
0
0
0
0
0
Detroit
0.1
<0.1
0.1
0
0
0
0
0
0
Philadelphia
<0.1
0
<0.1
0
0
0
0
0
0
Phoenix
0.3
0.1
0.4
<0.1
0
<0.1
0
0
0
Sacramento
<0.1
0
0.1
0
0
0
0
0
0
St. Louis
0.1
<0.1
0.2
0
0
0
0
0
0
A Calculated percent is rounded to the nearest tenth decimal using conventional rounding. Values equal to zero are designated by "0" (there
are no individuals exposed at that level). Small, non-zero values that do not round upwards to 0.1 (i.e., <0.05) are given a value of "<0.1".
3D-102
-------
Table 3D-29. Number of people estimated to experience at least two exposures at or above
benchmarks while at moderate or greater exertion, for air quality adjusted to
just meet the current standard.
Study
Group
Study Area
60 ppb Benchmark (7-hr)A
(# per Year)
70 ppb Benchmark (7-hr)A
(# per Year)
80 ppb Benchmark (7-hr)A
(# per Year)
Avg
Min
Max
Avg
Min
Max
Avg
Min
Max
Atlanta
7365
1675
13801
155
0
282
0
0
0
Boston
11317
6690
18477
341
91
660
0
0
0
Dallas
17135
5226
29273
276
24
757
0
0
0
Children
Detroit
17829
10805
28894
243
69
520
0
0
0
Philadelphia
10142
9210
11764
124
65
175
0
0
0
Phoenix
24952
14153
36643
94
0
269
0
0
0
Sacramento
2601
1281
4278
16
0
31
0
0
0
St. Louis
8305
4071
14325
67
9
155
0
0
0
Atlanta
1002
202
1715
20
0
40
0
0
0
Boston
1669
1047
2617
30
0
68
0
0
0
Children
with
Asthma
Dallas
1600
378
2861
39
0
118
0
0
0
Detroit
2180
1301
3469
11
0
17
0
0
0
Philadelphia
1288
1113
1375
15
0
44
0
0
0
Phoenix
2609
1444
3977
24
0
71
0
0
0
Sacramento
282
179
396
5
0
8
0
0
0
St. Louis
713
337
1211
3
0
9
0
0
0
Atlanta
1925
211
3592
0
0
0
0
0
0
Boston
2446
1076
4794
98
0
196
0
0
0
Dallas
3724
1250
6798
26
0
78
0
0
0
Adults
Detroit
5178
2884
9438
44
0
131
0
0
0
Philadelphia
1917
1656
2266
29
0
87
0
0
0
Phoenix
8361
4718
11324
33
0
50
0
0
0
Sacramento
572
257
972
10
0
29
0
0
0
St. Louis
2587
858
4435
0
0
0
0
0
0
Atlanta
94
0
211
0
0
0
0
0
0
Boston
261
0
489
33
0
98
0
0
0
Adults
with
Asthma
Dallas
104
78
156
0
0
0
0
0
0
Detroit
328
197
590
0
0
0
0
0
0
Philadelphia
58
0
174
0
0
0
0
0
0
Phoenix
745
397
1142
17
0
50
0
0
0
Sacramento
38
0
86
0
0
0
0
0
0
St. Louis
191
72
358
0
0
0
0
0
0
A These values represent the population of individuals exposed in each study area. Values equal to zero are indicated by "0" (there are no
individuals exposed at the level).
3D-103
-------
Table 3D-30. Percent of people estimated to experience at least four exposures at or above
benchmarks while at moderate or greater exertion, for air quality adjusted to
just meet the current standard.
Study
Group
Study Area
60 ppb Benchmark (7-hr)A
(% per Year)
70 ppb Benchmark (7-hr)A
(% per Year)
80 ppb Benchmark (7-hr)A
(% per Year)
Avg
Min
Max
Avg
Min
Max
Avg
Min
Max
Atlanta
<0.1
<0.1
0.1
0
0
0
0
0
0
Boston
<0.1
<0.1
0.1
0
0
0
0
0
0
Dallas
0.1
<0.1
0.3
0
0
0
0
0
0
Children
Detroit
0.2
0.1
0.3
0
0
0
0
0
0
Philadelphia
0.1
<0.1
0.1
0
0
0
0
0
0
Phoenix
0.7
0.3
1.1
<0.1
0
<0.1
0
0
0
Sacramento
<0.1
<0.1
0.1
0
0
0
0
0
0
St. Louis
0.2
<0.1
0.3
0
0
0
0
0
0
Atlanta
<0.1
0
0.1
0
0
0
0
0
0
Boston
<0.1
0
0.1
0
0
0
0
0
0
Children
with
Asthma
Dallas
0.2
<0.1
0.4
0
0
0
0
0
0
Detroit
0.1
<0.1
0.2
0
0
0
0
0
0
Philadelphia
<0.1
<0.1
0.1
0
0
0
0
0
0
Phoenix
0.8
0.3
1.3
0
0
0
0
0
0
Sacramento
0.1
0
0.2
0
0
0
0
0
0
St. Louis
0.1
0
0.3
0
0
0
0
0
0
Atlanta
<0.1
0
<0.1
0
0
0
0
0
0
Boston
<0.1
0
<0.1
0
0
0
0
0
0
Dallas
<0.1
0
<0.1
0
0
0
0
0
0
Adults
Detroit
<0.1
0
<0.1
0
0
0
0
0
0
Philadelphia
<0.1
0
<0.1
0
0
0
0
0
0
Phoenix
<0.1
<0.1
<0.1
0
0
0
0
0
0
Sacramento
0
0
0
0
0
0
0
0
0
St. Louis
<0.1
0
<0.1
0
0
0
0
0
0
Atlanta
0
0
0
0
0
0
0
0
0
Boston
0
0
0
0
0
0
0
0
0
Adults
with
Asthma
Dallas
0
0
0
0
0
0
0
0
0
Detroit
0
0
0
0
0
0
0
0
0
Philadelphia
0
0
0
0
0
0
0
0
0
Phoenix
<0.1
<0.1
0.1
0
0
0
0
0
0
Sacramento
0
0
0
0
0
0
0
0
0
St. Louis
0
0
0
0
0
0
0
0
0
A Calculated percent is rounded to the nearest tenth decimal using conventional rounding. Values equal to zero are designated by "0" (there
are no individuals exposed at that level). Small, non-zero values that do not round upwards to 0.1 (i.e., <0.05) are given a value of "<0.1".
3D-104
-------
Table 3D-31. Number of people estimated to experience at least four exposures at or above
benchmarks while at moderate or greater exertion, for air quality adjusted to
just meet the current standard.
Study
Group
Study Area
60 ppb Benchmark (7-hr)A
(# per Year)
70 ppb Benchmark (7-hr)A
(# per Year)
80 ppb Benchmark (7-hr)A
(# per Year)
Avg
Min
Max
Avg
Min
Max
Avg
Min
Max
Atlanta
538
61
1190
0
0
0
0
0
0
Boston
471
137
865
0
0
0
0
0
0
Dallas
1986
260
4422
0
0
0
0
0
0
Children
Detroit
1665
746
3035
0
0
0
0
0
0
Philadelphia
662
349
1157
0
0
0
0
0
0
Phoenix
5997
2633
9554
5
0
14
0
0
0
Sacramento
158
8
411
0
0
0
0
0
0
St. Louis
862
209
1803
0
0
0
0
0
0
Atlanta
67
0
101
0
0
0
0
0
0
Boston
76
0
137
0
0
0
0
0
0
Children
with
Asthma
Dallas
213
24
473
0
0
0
0
0
0
Detroit
162
52
243
0
0
0
0
0
0
Philadelphia
58
22
109
0
0
0
0
0
0
Phoenix
637
212
1033
0
0
0
0
0
0
Sacramento
23
0
70
0
0
0
0
0
0
St. Louis
73
0
155
0
0
0
0
0
0
Atlanta
47
0
141
0
0
0
0
0
0
Boston
33
0
98
0
0
0
0
0
0
Dallas
104
0
234
0
0
0
0
0
0
Adults
Detroit
109
0
262
0
0
0
0
0
0
Philadelphia
29
0
87
0
0
0
0
0
0
Phoenix
646
199
894
0
0
0
0
0
0
Sacramento
0
0
0
0
0
0
0
0
0
St. Louis
60
0
143
0
0
0
0
0
0
Atlanta
0
0
0
0
0
0
0
0
0
Boston
0
0
0
0
0
0
0
0
0
Adults
with
Asthma
Dallas
0
0
0
0
0
0
0
0
0
Detroit
0
0
0
0
0
0
0
0
0
Philadelphia
0
0
0
0
0
0
0
0
0
Phoenix
83
50
149
0
0
0
0
0
0
Sacramento
0
0
0
0
0
0
0
0
0
St. Louis
0
0
0
0
0
0
0
0
0
A These values represent the population of individuals exposed in each study area. Values equal to zero are indicated by "0" (there are no
individuals exposed at the level).
3D-105
-------
3D.3.2.2 Additional Air Quality Scenario: 75 ppb
When considering air quality adjusted so that the design value at the highest monitor
location in each urban study area is equal to 75 ppb, there will be a greater percent and number
of people estimated to experience 7-hr O3 exposures at or above each of the benchmarks. For
example, estimated exposures to O3 concentrations at or above the 80-ppb benchmark are
limited, but not insignificant. When considering the worst air quality year, upwards to 0.6% of
children (and similarly for children with asthma) are estimated to experience at least one day
with a 7-hr exposure at or above the 80-ppb benchmark, while on average, most study areas had
at least 0.1% of children experiencing such an exposure (Table 3D-32). On average, between
about 1 to 2% of children (and similarly for children with asthma) would experience at least one
day with a 7-hr exposure at or above the 70-ppb benchmark, while for the worst air quality year
upwards to 3.4% of children (and 3.9% children with asthma) would experience such an
exposure. On average, between about 7 to 17% of children (and similarly for children with
asthma) would experience at least one day with a 7-hr exposure at or above the 60-ppb
benchmark, while for the worst year upwards to about 18% of children (and about 19% of
children with asthma) would experience such an exposure.
Under the 75 ppb air quality scenario, multiday exposures to the 80 ppb benchmark are
few, but not entirely eliminated as was shown with the exposure results considering air quality
adjusted to just meet the current standard. A small percent (<0.1 %) of children are estimated to
experience at least two days with 7-hr exposures at or above the 80-ppb (Table 3D-33). On
average, between 0.1 to 0.3% of children (and 0.1 to 0.4% of children with asthma) would
experience at least two days with 7-hr exposures at or above the 70-ppb benchmark, while for the
worst year upwards to 0.7% of children (and 0.8% of children with asthma) would experience
such an exposure. When considering the worst air quality year, between about 3 to 10% of
children (and 3 to 11% of children with asthma) and 0.2 to 1.2% of adults (and 0.1 to 1.1% of
adults with asthma) are estimated to experience at least two days with 7-hr O3 exposures at or
above the 60-ppb benchmark. On average, all study areas (and study groups) have a small
percent (<0.1 %) estimated to experience at least four days with 7-hr O3 exposures at or above the
70-ppb benchmark (Table 3D-34), and at most 2% of children (and 2.3% of children with
asthma) are estimated experience at least six days with 7-hr O3 exposures at or above the 60-ppb
benchmark for the worst air quality year (Attachment 4).
3D-106
-------
Table 3D-32. Percent of people estimated to experience at least one exposure at or above
benchmarks while at moderate or greater exertion, for the 75 ppb air quality
scenario.
Study
Group
Study Area
60 ppb Benchmark (7-hr)A
(% per Year)
70 ppb Benchmark (7-hr)A
(% per Year)
80 ppb Benchmark (7-hr)A
(% per Year)
Avg
Min
Max
Avg
Min
Max
Avg
Min
Max
Children
Atlanta
7.7
4.8
10.7
1.5
0.4
2.8
0.3
<0.1
0.6
Boston
6.6
5.0
8.8
1.3
0.9
1.9
0.1
0.1
0.1
Dallas
8.3
4.7
11.5
1.3
0.7
2.1
0.1
<0.1
0.1
Detroit
11.0
8.6
13.9
1.9
0.9
3.4
0.1
<0.1
0.1
Philadelphia
8.6
8.2
8.8
1.4
1.2
1.5
0.1
<0.1
0.1
Phoenix
15.7
13.2
17.9
2.0
0.9
3.4
<0.1
0
0.1
Sacramento
7.5
6.3
8.9
1.1
0.8
1.4
<0.1
<0.1
<0.1
St. Louis
10.6
8.5
13.0
1.7
0.8
3.2
0.1
0
0.1
Children
with
Asthma
Atlanta
8.5
5.2
11.8
1.7
0.4
3.1
0.3
<0.1
0.6
Boston
7.6
5.7
9.8
1.4
1.0
2.2
0.1
0.1
0.2
Dallas
8.9
4.6
11.9
1.4
0.9
2.2
0.1
<0.1
0.1
Detroit
12.0
9.6
15.0
2.1
1.1
3.9
<0.1
0
0.1
Philadelphia
9.4
9.1
9.6
1.5
1.3
1.6
0.1
<0.1
0.1
Phoenix
17.1
14.4
19.2
2.1
1.0
3.8
0.1
0
0.2
Sacramento
7.8
6.9
9.3
1.1
0.9
1.5
0.1
<0.1
0.1
St. Louis
10.6
8.4
13.2
1.6
0.6
3.2
0.1
0
0.1
Adults
Atlanta
1.3
0.8
1.8
0.2
0.1
0.4
<0.1
0
0.1
Boston
0.9
0.5
1.3
0.1
0.1
0.2
<0.1
<0.1
<0.1
Dallas
1.4
0.7
2.1
0.2
0.1
0.3
<0.1
0
<0.1
Detroit
1.7
1.3
2.3
0.3
0.2
0.5
<0.1
0
<0.1
Philadelphia
1.2
1.1
1.4
0.2
0.1
0.2
<0.1
<0.1
<0.1
Phoenix
3.2
2.6
3.6
O
CO
0.2
0.4
<0.1
0
<0.1
Sacramento
1.1
0.9
1.3
0.1
0.1
0.2
<0.1
0
<0.1
St. Louis
1.7
1.2
2.1
0.2
0.1
0.4
<0.1
0
<0.1
Adults
with
Asthma
Atlanta
0.9
0.6
1.1
0.2
0.1
0.3
<0.1
0
<0.1
Boston
0.6
0.4
1.0
0.1
<0.1
0.1
<0.1
0
<0.1
Dallas
1.1
0.5
1.5
0.1
<0.1
0.1
<0.1
0
<0.1
Detroit
1.5
1.1
1.7
0.2
0.1
0.4
<0.1
0
<0.1
Philadelphia
1.0
0.7
1.2
0.1
0.1
0.1
0
0
0
Phoenix
2.7
2.3
3.0
0.2
0.1
0.4
0
0
0
Sacramento
0.9
0.7
1.2
0.1
0.1
0.1
<0.1
0
<0.1
St. Louis
1.3
1.0
1.8
0.2
<0.1
0.4
<0.1
0
<0.1
A Calculated percent is rounded to the nearest tenth decimal using conventional rounding. Values equal to zero are designated by "0" (there
are no individuals exposed at that level). Small, non-zero values that do not round upwards to 0.1 (i.e., <0.05) are given a value of "<0.1".
3D-107
-------
Table 3D-33. Percent of people estimated to experience at least two exposures at or above
benchmarks while at moderate or greater exertion, for the 75 ppb air quality
scenario.
Study
Group
Study Area
60 ppb Benchmark (7-hr)A
(% per Year)
70 ppb Benchmark (7-hr)A
(% per Year)
80 ppb Benchmark (7-hr)A
(% per Year)
Avg
Min
Max
Avg
Min
Max
Avg
Min
Max
Atlanta
2.5
1.1
4.0
0.2
<0.1
0.4
<0.1
0
<0.1
Boston
1.7
1.1
2.6
0.1
<0.1
0.2
<0.1
0
<0.1
Dallas
2.9
1.1
4.8
0.1
<0.1
0.3
<0.1
0
<0.1
Children
Detroit
4.0
2.5
5.8
0.2
<0.1
0.4
<0.1
0
<0.1
Philadelphia
2.8
2.5
3.0
0.1
0.1
0.2
<0.1
0
<0.1
Phoenix
8.0
6.0
9.9
0.3
0.1
r—
o
<0.1
0
<0.1
Sacramento
2.4
1.7
3.4
0.1
<0.1
0.2
0
0
0
St. Louis
3.9
2.7
5.4
0.2
<0.1
0.4
<0.1
0
<0.1
Atlanta
2.7
1.1
4.2
0.2
<0.1
0.4
<0.1
0
<0.1
Boston
2.0
1.3
3.0
0.1
<0.1
0.2
0
0
0
Children
with
Asthma
Dallas
2.9
1.0
4.8
0.1
0
0.3
0
0
0
Detroit
4.4
2.8
6.4
0.2
<0.1
0.4
0
0
0
Philadelphia
3.0
2.8
3.1
0.1
0.1
0.1
0
0
0
Phoenix
8.9
6.7
11.0
0.4
0.2
0.8
0
0
0
Sacramento
2.6
1.9
3.8
0.1
<0.1
0.2
0
0
0
St. Louis
3.6
2.5
4.9
0.1
0
0.3
0
0
0
Atlanta
0.2
0.1
0.4
<0.1
<0.1
<0.1
0
0
0
Boston
0.1
0.1
0.2
<0.1
<0.1
<0.1
0
0
0
Dallas
0.2
0.1
0.5
<0.1
<0.1
<0.1
0
0
0
Adults
Detroit
0.3
0.2
0.5
<0.1
<0.1
<0.1
0
0
0
Philadelphia
0.2
0.1
0.2
<0.1
<0.1
<0.1
0
0
0
Phoenix
1.0
0.7
1.2
<0.1
<0.1
<0.1
0
0
0
Sacramento
0.2
0.1
0.2
<0.1
<0.1
<0.1
0
0
0
St. Louis
0.3
0.2
0.5
<0.1
<0.1
<0.1
0
0
0
Atlanta
0.1
<0.1
0.2
<0.1
0
<0.1
0
0
0
Boston
0.1
<0.1
0.1
<0.1
0
<0.1
0
0
0
Adults
with
Asthma
Dallas
0.1
<0.1
0.2
0
0
0
0
0
0
Detroit
0.3
0.2
0.5
<0.1
0
<0.1
0
0
0
Philadelphia
0.1
<0.1
0.2
0
0
0
0
0
0
Phoenix
0.8
0.6
1.1
<0.1
0
<0.1
0
0
0
Sacramento
0.1
0.1
0.2
0
0
0
0
0
0
St. Louis
0.3
0.1
0.4
<0.1
0
<0.1
0
0
0
A Calculated percent is rounded to the nearest tenth decimal using conventional rounding. Values equal to zero are designated by "0" (there
are no individuals exposed at that level). Small, non-zero values that do not round upwards to 0.1 (i.e., <0.05) are given a value of "<0.1".
3D-108
-------
Table 3D-34. Percent of people estimated to experience at least four exposures at or above
benchmarks while at moderate or greater exertion, for the 75 ppb air quality
scenario.
Study
Group
Study Area
60 ppb Benchmark (7-hr)A
(% per Year)
70 ppb Benchmark (7-hr)A
(% per Year)
80 ppb Benchmark (7-hr)A
(% per Year)
Avg
Min
Max
Avg
Min
Max
Avg
Min
Max
Atlanta
0.4
0.1
0.9
<0.1
0
<0.1
0
0
0
Boston
0.1
0.1
0.2
<0.1
0
<0.1
0
0
0
Dallas
0.6
0.1
1.2
<0.1
0
<0.1
0
0
0
Children
Detroit
0.7
0.3
1.2
<0.1
0
<0.1
0
0
0
Philadelphia
0.4
0.3
0.6
<0.1
0
<0.1
0
0
0
Phoenix
3.0
2.0
4.1
<0.1
<0.1
<0.1
0
0
0
Sacramento
0.4
0.2
0.8
<0.1
0
<0.1
0
0
0
St. Louis
0.8
0.4
1.2
<0.1
0
<0.1
0
0
0
Atlanta
0.5
0.1
0.9
<0.1
0
<0.1
0
0
0
Boston
0.1
0.1
0.2
0
0
0
0
0
0
Children
with
Asthma
Dallas
0.5
0.1
1.1
0
0
0
0
0
0
Detroit
0.6
0.2
1.0
<0.1
0
<0.1
0
0
0
Philadelphia
0.4
0.3
0.6
0
0
0
0
0
0
Phoenix
3.3
2.2
4.4
<0.1
0
0.1
0
0
0
Sacramento
0.4
0.2
0.8
0
0
0
0
0
0
St. Louis
0.6
0.3
1.1
0
0
0
0
0
0
Atlanta
<0.1
0
<0.1
0
0
0
0
0
0
Boston
<0.1
<0.1
<0.1
0
0
0
0
0
0
Dallas
<0.1
0
<0.1
0
0
0
0
0
0
Adults
Detroit
<0.1
<0.1
<0.1
0
0
0
0
0
0
Philadelphia
<0.1
<0.1
<0.1
0
0
0
0
0
0
Phoenix
0.1
0.1
0.2
<0.1
0
<0.1
0
0
0
Sacramento
<0.1
<0.1
<0.1
0
0
0
0
0
0
St. Louis
<0.1
<0.1
<0.1
0
0
0
0
0
0
Atlanta
0
0
0
0
0
0
0
0
0
Boston
<0.1
0
<0.1
0
0
0
0
0
0
Adults
with
Asthma
Dallas
0
0
0
0
0
0
0
0
0
Detroit
<0.1
0
<0.1
0
0
0
0
0
0
Philadelphia
0
0
0
0
0
0
0
0
0
Phoenix
0.1
0.1
0.1
0
0
0
0
0
0
Sacramento
0
0
0
0
0
0
0
0
0
St. Louis
<0.1
0
0.1
0
0
0
0
0
0
A Calculated percent is rounded to the nearest tenth decimal using conventional rounding. Values equal to zero are designated by "0" (there
are no individuals exposed at that level). Small, non-zero values that do not round upwards to 0.1 (i.e., <0.05) are given a value of "<0.1".
3D-109
-------
3D.3.2.3 Additional Air Quality Scenario: 65 ppb
With increasing stringency (i.e., lowering) of the design value used to represent the air
quality scenario, there is a reduction in the percent and number of simulated individuals
experiencing 7-hr exposures at or above the benchmarks. Under the 65 ppb air quality scenario,
in 6 of the 8 study areas, there are no people estimated to experience at least one benchmark at or
above the 80-ppb benchmark (Table 3D-35). Exposures at or above the 70-ppb benchmark are
also limited, with at most 0.2% of children (and 0.3% of children with asthma) estimated
experience one such exposure during the worst air quality year. On average, between 0.4 to 2.3%
of children (and 0.5 to 2.5% of children with asthma) and between 0.1 to 0.4% of adults (and
<0.1 to 0.3% of adults with asthma) are estimated to experience at least one 7-hr O3 exposure at
or above the 60-ppb benchmark, while during the worst air quality year, upwards to 3.7% of
children (and 4.3% of children with asthma) would experience such an exposure.
Multiday exposures at or above the 70-ppb benchmark are nearly eliminated under the 65
ppb air quality scenario, with only three study areas having at most, <0.1% of children (and no
children with asthma) estimated to experience 7-hr exposures at or above that benchmark for at
least two days (Table 3D-36). When considering the worst air quality year, <0.5% of children
(and 0.6%) of children with asthma) and <0.1% of adults (and similarly for adults with asthma)
are estimated to experience at least two days with 7-hr O3 exposures at or above the 60-ppb
benchmark. There are no people in any of the study areas estimated to experience at least four
days with 7-hr O3 exposures at or above the 70-ppb benchmark (Table 3D-37), and there no
simulated individuals estimated to experience at least six days with 7-hr O3 exposures at or
above the 60-ppb benchmark in all but two study areas (Attachment 4).
3D-110
-------
Table 3D-35. Percent of people estimated to experience at least one exposure at or above
benchmarks while at moderate or greater exertion, for the 65 ppb air quality
scenario.
Study
Group
Study Area
60 ppb Benchmark (7-hr)A
(% per Year)
70 ppb Benchmark (7-hr)A
(% per Year)
80 ppb Benchmark (7-hr)A
(% per Year)
Avg
Min
Max
Avg
Min
Max
Avg
Min
Max
Atlanta
1.0
0.3
1.7
0.1
<0.1
0.1
<0.1
0
<0.1
Boston
1.8
1.1
2.5
0.2
0.1
0.2
<0.1
0
<0.1
Dallas
2.1
0.9
2.8
0.1
0.1
0.1
0
0
0
Children
Detroit
2.3
1.4
3.7
<0.1
<0.1
0.1
0
0
0
Philadelphia
1.5
1.4
1.6
<0.1
<0.1
<0.1
0
0
0
Phoenix
1.8
0.9
3.0
0
0
0
0
0
0
Sacramento
0.4
0.3
0.6
0
0
0
0
0
0
St. Louis
1.6
0.7
3.1
<0.1
0
<0.1
0
0
0
Atlanta
1.1
0.3
1.9
0.1
<0.1
0.2
<0.1
0
<0.1
Boston
2.1
1.3
3.1
0.2
0.1
0.3
<0.1
0
<0.1
Children
with
Asthma
Dallas
2.2
1.1
2.9
0.1
<0.1
0.1
0
0
0
Detroit
2.5
1.5
4.3
<0.1
0
<0.1
0
0
0
Philadelphia
1.6
1.3
1.9
<0.1
<0.1
<0.1
0
0
0
Phoenix
2.1
1.0
3.4
0
0
0
0
0
0
Sacramento
0.5
0.3
0.6
0
0
0
0
0
0
St. Louis
1.5
0.6
3.0
<0.1
0
<0.1
0
0
0
Atlanta
0.1
<0.1
0.3
<0.1
0
<0.1
0
0
0
Boston
0.2
0.1
0.3
<0.1
<0.1
<0.1
<0.1
0
<0.1
Dallas
0.3
0.1
0.4
<0.1
<0.1
<0.1
0
0
0
Adults
Detroit
0.4
0.2
0.6
<0.1
0
<0.1
0
0
0
Philadelphia
0.2
0.2
0.2
<0.1
<0.1
<0.1
0
0
0
Phoenix
0.3
0.2
0.4
0
0
0
0
0
0
Sacramento
0.1
<0.1
0.1
0
0
0
0
0
0
St. Louis
0.2
0.1
0.4
<0.1
0
<0.1
0
0
0
Atlanta
0.1
<0.1
0.1
0
0
0
0
0
0
Boston
0.1
<0.1
0.2
<0.1
0
<0.1
0
0
0
Adults
with
Asthma
Dallas
0.2
<0.1
0.3
<0.1
0
<0.1
0
0
0
Detroit
0.3
0.2
0.5
<0.1
0
<0.1
0
0
0
Philadelphia
0.1
0.1
0.2
0
0
0
0
0
0
Phoenix
0.2
0.1
0.4
0
0
0
0
0
0
Sacramento
<0.1
<0.1
0.1
0
0
0
0
0
0
St. Louis
0.2
<0.1
0.5
<0.1
0
<0.1
0
0
0
A Calculated percent is rounded to the nearest tenth decimal using conventional rounding. Values equal to zero are designated by "0" (there
are no individuals exposed at that level). Small, non-zero values that do not round upwards to 0.1 (i.e., <0.05) are given a value of "<0.1".
3D-111
-------
Table 3D-36. Percent of people estimated to experience at least two exposures at or above
benchmarks while at moderate or greater exertion, for the 65 ppb air quality
scenario.
Study
Group
Study Area
60 ppb Benchmark (7-hr)A
(% per Year)
70 ppb Benchmark (7-hr)A
(% per Year)
80 ppb Benchmark (7-hr)A
(% per Year)
Avg
Min
Max
Avg
Min
Max
Avg
Min
Max
Atlanta
0.1
<0.1
0.2
<0.1
0
<0.1
0
0
0
Boston
0.2
0.1
0.3
<0.1
0
<0.1
0
0
0
Dallas
0.3
<0.1
0.5
0
0
0
0
0
0
Children
Detroit
0.3
0.1
0.5
<0.1
0
<0.1
0
0
0
Philadelphia
0.1
0.1
0.2
0
0
0
0
0
0
Phoenix
0.3
0.1
0.5
0
0
0
0
0
0
Sacramento
<0.1
<0.1
<0.1
0
0
0
0
0
0
St. Louis
0.2
<0.1
0.4
0
0
0
0
0
0
Atlanta
0.1
0
0.2
0
0
0
0
0
0
Boston
0.2
0.1
0.3
0
0
0
0
0
0
Children
with
Asthma
Dallas
0.3
<0.1
0.5
0
0
0
0
0
0
Detroit
0.2
0.1
0.4
0
0
0
0
0
0
Philadelphia
0.2
0.1
0.2
0
0
0
0
0
0
Phoenix
0.3
0.1
0.6
0
0
0
0
0
0
Sacramento
<0.1
0
<0.1
0
0
0
0
0
0
St. Louis
0.1
<0.1
0.3
0
0
0
0
0
0
Atlanta
<0.1
0
<0.1
0
0
0
0
0
0
Boston
<0.1
<0.1
<0.1
0
0
0
0
0
0
Dallas
<0.1
<0.1
<0.1
0
0
0
0
0
0
Adults
Detroit
<0.1
<0.1
0.1
0
0
0
0
0
0
Philadelphia
<0.1
<0.1
<0.1
0
0
0
0
0
0
Phoenix
<0.1
<0.1
<0.1
0
0
0
0
0
0
Sacramento
<0.1
<0.1
<0.1
0
0
0
0
0
0
St. Louis
<0.1
0
<0.1
0
0
0
0
0
0
Atlanta
0
0
0
0
0
0
0
0
0
Boston
<0.1
0
<0.1
0
0
0
0
0
0
Adults
with
Asthma
Dallas
0
0
0
0
0
0
0
0
0
Detroit
<0.1
0
0.1
0
0
0
0
0
0
Philadelphia
0
0
0
0
0
0
0
0
0
Phoenix
<0.1
0
<0.1
0
0
0
0
0
0
Sacramento
0
0
0
0
0
0
0
0
0
St. Louis
<0.1
0
<0.1
0
0
0
0
0
0
A Calculated percent is rounded to the nearest tenth decimal using conventional rounding. Values equal to zero are designated by "0" (there
are no individuals exposed at that level). Small, non-zero values that do not round upwards to 0.1 (i.e., <0.05) are given a value of "<0.1".
3D-112
-------
Table 3D-37. Percent of people estimated to experience at least four exposures at or above
benchmarks while at moderate or greater exertion, for the 65 ppb air quality
scenario.
Study
Group
Study Area
60 ppb Benchmark (7-hr)A
(% per Year)
70 ppb Benchmark (7-hr)A
(% per Year)
80 ppb Benchmark (7-hr)A
(% per Year)
Avg
Min
Max
Avg
Min
Max
Avg
Min
Max
Atlanta
<0.1
0
<0.1
0
0
0
0
0
0
Boston
<0.1
0
<0.1
0
0
0
0
0
0
Dallas
<0.1
<0.1
<0.1
0
0
0
0
0
0
Children
Detroit
<0.1
<0.1
<0.1
0
0
0
0
0
0
Philadelphia
<0.1
0
<0.1
0
0
0
0
0
0
Phoenix
<0.1
<0.1
<0.1
0
0
0
0
0
0
Sacramento
0
0
0
0
0
0
0
0
0
St. Louis
<0.1
0
<0.1
0
0
0
0
0
0
Atlanta
<0.1
0
<0.1
0
0
0
0
0
0
Boston
0
0
0
0
0
0
0
0
0
Children
with
Asthma
Dallas
0
0
0
0
0
0
0
0
0
Detroit
<0.1
0
<0.1
0
0
0
0
0
0
Philadelphia
0
0
0
0
0
0
0
0
0
Phoenix
<0.1
0
0.1
0
0
0
0
0
0
Sacramento
0
0
0
0
0
0
0
0
0
St. Louis
0
0
0
0
0
0
0
0
0
Atlanta
0
0
0
0
0
0
0
0
0
Boston
0
0
0
0
0
0
0
0
0
Dallas
0
0
0
0
0
0
0
0
0
Adults
Detroit
0
0
0
0
0
0
0
0
0
Philadelphia
0
0
0
0
0
0
0
0
0
Phoenix
0
0
0
0
0
0
0
0
0
Sacramento
0
0
0
0
0
0
0
0
0
St. Louis
0
0
0
0
0
0
0
0
0
Atlanta
0
0
0
0
0
0
0
0
0
Boston
0
0
0
0
0
0
0
0
0
Adults
with
Asthma
Dallas
0
0
0
0
0
0
0
0
0
Detroit
0
0
0
0
0
0
0
0
0
Philadelphia
0
0
0
0
0
0
0
0
0
Phoenix
0
0
0
0
0
0
0
0
0
Sacramento
0
0
0
0
0
0
0
0
0
St. Louis
0
0
0
0
0
0
0
0
0
A Calculated percent is rounded to the nearest tenth decimal using conventional rounding. Values equal to zero are designated by "0" (there
are no individuals exposed at that level). Small, non-zero values that do not round upwards to 0.1 (i.e., <0.05) are given a value of "<0.1".
3D-113
-------
3D.3.2.4 Comparison with 2014 HREA Exposure Results
We compared the exposure results for the current exposure and risk analysis with those
generated for the 2014 HREA. Table 3D-38 presents the percent of children experiencing at least
one exposure at or above the three benchmarks for the two assessments and Table 3D-39
presents the similar comparison for two or more exposures. Results are presented for all study
areas, and for the seven study areas common to both assessments. In general, the comparison
indicates similarity between the two assessments, particularly for the highest benchmark and
when focusing on the summary for all areas in each assessment. Such a focus is appropriate
given the purpose of the assessments in providing estimates across a range of study areas to
inform decision making with regard to the exposures and risks that may occur across the U.S. in
areas that just meet the current standard. For the lower benchmarks and particularly in comparing
for the seven areas common to both assessments, the current assessment estimates are slightly
lower than the 2014 HREA results, most notably for the highest single year, likely reflecting the
greater variation in ambient air concentrations in some study areas in the 2014 HREA. This is
supported by recent analyses that show changes to the distribution of ambient air O3
concentrations over time occur primarily as reductions to the highest and lowest concentrations
(Downey et al., 2015; Simon et al., 2012).
In addition to generally lower baseline O3 concentrations and lower variability in the
concentrations in the three air quality scenarios for the current assessment compared to 2014
HREA, there were also two important differences in the exposure modeling approach. The first is
the use, in the current assessment, of an EVR distribution (17.32 ± 1.25 L/minute-m2) to indicate
when a simulated individual is at moderate or greater exertion (section 3D.2.2.3.3) rather than
using a lower value for all simulated individuals (13 L/minute-m2; 5th percentile). The current
approach would be expected to result in far fewer individuals reaching the exertion level
concomitant with the exposure level of interest, thus reducing the percent of the population at or
above benchmarks. The second difference is the focus on 7-hr average exposures (compared to
the benchmarks) in this assessment rather than 8-hr averages. With this change, it would be
expected that there would be more simulated individuals at or above a given benchmark
concentration. While these two changes to the exposure modeling approach compete in their
overall influence on the exposure results, it would be expected that the change to using the EVR
distribution would have a greater impact.
As suggested above, the difference between the two assessments in the highest year
estimates is likely a function of the baseline ambient air concentrations in the study areas. As a
reminder, the 2014 HREA used air quality scenarios developed from adjusting 2006-2010
ambient air concentrations, and some study areas had design values in that time period that were
well above the then-existing standard (and more so for the current standard). In the current
3D-114
-------
exposure analysis, we selected study areas that had 2015-2017 design values close to the current
standard, requiring less of an adjustment for the current standard (70 ppb) air quality scenario.
Table 3D-38. Comparison of current assessment to 2014 HREA for percent of children
estimated to experience at least one exposure at or above benchmarks while at
moderate or greater exertion.
Air Quality
Scenario
(DV, ppb)
Average Percent (%) of Simulated Children with at least One Day per Year
at or above Specified Benchmark Exposure Concentration
(highest in single season)
All areas A
7 common areas A
Current PAB
2014 H RE A c
Current PAB
2014 HREA c
Benchmark Exposure Concentration of 80 ppb
75
<0.1B -0.3 (0.6)
0 -0.3(1.1)
<0.1 -0.3(0.6)
0.1 -0.3 (1.1)
70
0 - <0.1 (0.1)
0 -0.1 (0.2)
0 - <0.1 (0.1)
0 B - 0.1 (0.2)
65
0 - <0.1 (<0.1)
0(0)
0 - <0.1 (<0.1)
0(0)
Benchmark Exposure Concentration of 70 ppb
75
1.1-2.0 (3.4)
0.6-3.3(8.1)
1.1-1.9(3.4)
1.6-3.3 (8.1)
70
0.2-0.6 (0.9)
0.1-1.2(3.2)
0.2-0.6(0.9)
0.4-1.2(3.2)
65
0 -0.2(0.2)
0 -0.2(0.5)
0B -0.2(0.2)
0.1-0.2 (0.5)
Benchmark Exposure Concentration of 60 ppb
75
6.6-15.7(17.9)
9.5-17.0(25.8)
6.6-11.0(13.9)
10.3-16.3 (25.8)
70
3.2-8.2 (10.6)
3.3-10.2(18.9)
3.2-6.7(9.2)
5.8-10.2(16.9)
65
0.4-2.3 (3.7)
0 -4.2(9.5)
0.4-2.3(3.7)
2.4-3.9 (7.6)
A Footnote 9 contains the names of the 15 study areas evaluated for the 2014 HREA. The seven study areas common
to both include the eight evaluated in this assessment with exception of Phoenix.
B For the current analysis, calculated percent is rounded to the nearest tenth decimal using conventional rounding.
Values equal to zero are designated by "0" (there are no individuals exposed at that level). Small, non-zero values that
do not round upwards to 0.1 (i.e., <0.05) are given a value of "<0.1"
c For the 2014 HREA, calculated percent was rounded to the nearest tenth decimal using conventional rounding.
Values that did not round upwards to 0.1 (i.e., <0.05) were given a value of "0".
3D-115
-------
Table 3D-39. Comparison of current assessment to 2014 HREA for percent of children
estimated to experience at least two exposure at or above benchmarks while
at moderate or greater exertion.
Air Quality
Scenario
(DV, ppb)
Average Percent (%) of Simulated Children with at least Two Days per
Year at or above Specified Benchmark Exposure Concentration
(highest in single season)
All areas A
7 common areas A
Current PAB
2014 H RE A c
Current PAB
2014 HREA c
Benchmark Exposure Concentration of 80 ppb
75
0 - <0.1 (<0.1)
0(0.1)
0 - <0.1 (<0.1)
0 (0.1)
70
0(0)
0(0)
0(0)
0(0)
65
0(0)
0(0)
0(0)
0(0)
Benchmark Exposure Concentration of 70 ppb
75
0.1-0.3 (0.7)
0.1-0.6(2.2)
0.1-0.2(0.4)
0.2-0.6 (2.2)
70
<0.1 (0.1)
0-0.1 (0.4)
<0.1 (0.1)
0-0.1 (0.4)
65
0 - <0.1 (<0.1)
0(0)
0 - <0.1 (<0.1)
0(0)
Benchmark Exposure Concentration of 60 ppb
75
1.7-8.0(9.9)
3.1-7.6(14.4)
1.7-4.0(5.8)
3.7-7.0 (13.8)
70
0.6-2.9 (4.3)
0.5-3.5(9.2)
0.6-1.7(2.8)
1.5-3.2(7.1)
65
<0.1-0.3 (0.5)
0-0.8 (2.8)
<0.1-0.3(0.5)
0.3-0.7(2.0)
A Footnote 9 contains the names of the 15 study areas evaluated for the 2014 HREA. The seven study areas common
to both include the eight evaluated in this assessment with exception of Phoenix.
B For the current analysis, calculated percent is rounded to the nearest tenth decimal using conventional rounding.
Values equal to zero are designated by "0" (there are no individuals exposed at that level). Small, non-zero values that
do not round upwards to 0.1 (i.e., <0.05) are given a value of "<0.1"
c For the 2014 HREA. calculated percent was rounded to the nearest tenth decimal using conventional rounding.
Values that did not round upwards to 0.1 (i.e., <0.05) were given a value of "0".
3D.3.3 Lung Function Risk
As described above, lung function risk was estimated using two approaches. The first, a
population-based risk approach (i.e., using E-R functions, section 3D.2.8.2.1), combined the
population distribution of daily maximum 7-hr exposures occurring while at moderate or greater
exertion with continuous E-R functions derived from the controlled human exposure study data
(Table 3D-20 and Figure 3D-12). Note that the E-R function risk approach uses the full
distribution of daily maximum 7-hr exposures, from the minimum to the maximum exposures
(i.e., not simply including the upper level exposures or benchmarks). It is, however, necessary
that the daily maximum exposure did occur at a 7-hr EVR >17.32 ± 1.25 L/min-m2. The results
for the population-based (E-R function) risk approach, represented as percent (or counts) of the
population estimated to experience lung function decrements (i.e., >10%, >15%, and >20%
3D-116
-------
reduction in FEVi) is provided in section 3D.3.3.1. A similar format to that provided for the
benchmark results above is followed, focusing largely on the air quality scenario for just meeting
the current standard and presenting the percent (and counts) of the population estimated to
experience lung function decrements while at elevated exertion.
The second risk approach, an individual-based risk approach (i.e., the MSS model,
section 3D.2.8.2.2), calculates the decrements in lung function continuously for each simulated
person using their unique time-series of O3 exposures, simultaneously occurring breathing rates,
and personal attributes (e.g., age, body mass). Note that when using the MSS model risk
approach, the estimated reduction in FEVi considers the prior and current exposures/breathing
rates and has no hard restriction on either the exposure or exertion level. As such, lung function
decrements could also occur at exposures and/or breathing rates below that observed in the
controlled human exposure studies. The results for the individual-based (MSS model) risk
approach are found in section 3D.3.3.2. The complete results for both of the risk approaches can
be found in Attachment 4.
3D.3.3.1 Population-based (E-R Function) Risk Approach
As was observed with the exposure benchmarks and considering any of the air quality
scenarios, a smaller percent (and number) of adults are estimated to experience lung function
decrements when compared to children (Table 3D-40 to Table 3D-51). Again, this is driven
largely by the difference in time spent outdoors at elevated exertion. Even though there is limited
variability across the eight study areas, Detroit, Phoenix, and St. Louis generally exhibited higher
risk estimates relative to the other study areas for instances where risk estimates were above 1%
(e.g., where FEVi reductions >10%). This is expected given the observation made above
regarding the results for the exposure to benchmark comparison and its relationship with the
overall distribution of O3 concentrations in ambient air (Figure 3D-7).
In general, when comparing E-R function risk estimates to the benchmark results, the
attenuation of the percent estimated to experience lung function decrements is at a lesser rate
than that observed for the percent of the population at or above the benchmark levels, with
increasing stringency of the design values, and when considering the number of times per year
either might occur. For example, while as much as 0.9% of children (and 1.0% of children with
asthma) are estimated to experience at least one FEVi reduction >15% while at elevated exertion
with air quality just meeting the current standard (Table 3D-40), on average between 0.2 to 0.4%
of children (and similarly for children with asthma) in all 8 study areas are estimated to
experience at least four such decrements (Table 3D-44) when considering the same air quality
scenario. For comparison, while as much as 0.9% of children (and 1.0% of children with asthma)
are estimated to experience at least one exposure at or above the 70 ppb benchmark while at
3D-117
-------
elevated exertion for air quality just meeting the current standard (Table 3D-26), there are no
children (and similarly for children with asthma) estimated to experience at least four such
exposures in all but one study area (Table 3D-30) when considering the same air quality
scenario. This relative decreased rate of change observed for the E-R function risk results is
likely a function of the broader range (and low level) of exposures used in the calculation
compared to that represented by the exposure benchmarks.
The risks of lung function decrements in the 75 ppb air quality scenario, which allows
higher O3 concentrations, are of course greater (Table 3D-46 through Table 3D-48) than those
for air quality adjusted to just meet the current standard (Table 3D-40 through Table 3D-45),
differing by at most a few tenths of a percentage point for both the 15% and 20% reduction in
FEVi. A similar pattern is exhibited when comparing the lung function results for the current
standard to those for the 65 ppb air quality scenario (Table 3D-49 through Table 3D-51). A few
tenths of a percentage point lower risks are estimated for the lower design value scenario
compared to those estimated for the current standard.
3D-118
-------
Table 3D-40. Percent of people estimated to experience at least one lung function
decrement at or above the indicated level, for air quality adjusted to just meet
the current standard, using the population-based (E-R function) risk
approach.
Study
Group
Study Area
>10% reduction in FEViA
(% per Year)
>15% reduction in FEViA
(% per Year)
>20% reduction in FEViA
(% per Year)
Avg
Min
Max
Avg
Min
Max
Avg
Min
Max
Children
Atlanta
2.2
1.9
2.5
0.5
0.4
0.6
0.2
0.1
0.2
Boston
2.2
2.0
2.3
0.5
0.5
0.6
0.2
0.2
0.2
Dallas
2.4
2.1
2.6
0.6
0.5
r—
o
0.2
0.2
0.3
Detroit
2.5
2.3
2.8
0.7
0.6
0.8
0.3
0.2
0.3
Philadelphia
2.3
2.2
2.4
0.6
0.5
0.6
0.2
0.2
0.2
Phoenix
3.1
2.9
3.3
0.8
r—
o
0.9
0.3
0.3
0.4
Sacramento
2.2
2.2
2.3
0.5
0.5
0.6
0.2
0.2
0.2
St. Louis
2.5
2.3
2.8
0.7
0.6
0.8
0.2
0.2
0.3
Children
with
Asthma
Atlanta
2.3
2.0
2.6
0.6
0.5
0.7
0.2
0.1
0.3
Boston
2.4
2.2
2.6
0.6
0.6
f—
O
0.2
0.2
0.3
Dallas
2.6
2.3
2.8
0.7
0.5
0.8
0.2
0.2
0.3
Detroit
2.7
2.6
3.0
0.7
0.6
0.8
0.3
0.2
0.3
Philadelphia
2.4
2.4
2.5
0.6
0.6
0.6
0.2
0.2
0.2
Phoenix
3.3
3.1
3.6
0.9
0.8
1.0
0.3
0.3
0.4
Sacramento
2.3
2.3
2.4
0.5
0.5
0.6
0.2
0.2
0.2
St. Louis
2.6
2.3
2.8
r—
o
0.6
0.8
0.2
0.2
0.3
Adults
Atlanta
0.6
0.6
0.7
0.1
0.1
0.2
<0.1
<0.1
0.1
Boston
0.6
0.5
0.6
0.1
0.1
0.1
<0.1
<0.1
<0.1
Dallas
0.7
0.6
0.7
0.2
0.1
0.2
0.1
<0.1
0.1
Detroit
0.6
0.6
0.7
0.2
0.1
0.2
0.1
<0.1
0.1
Philadelphia
0.6
0.6
0.6
0.1
0.1
0.1
<0.1
<0.1
<0.1
Phoenix
0.9
0.9
1.0
0.2
0.2
0.2
0.1
0.1
0.1
Sacramento
0.6
0.6
0.6
0.1
0.1
0.1
<0.1
<0.1
<0.1
St. Louis
0.7
0.6
0.7
0.1
0.1
0.2
0.1
<0.1
0.1
Adults
with
Asthma
Atlanta
0.5
0.4
0.5
0.1
0.1
0.1
<0.1
<0.1
<0.1
Boston
0.4
0.4
0.5
0.1
0.1
0.1
<0.1
<0.1
<0.1
Dallas
0.6
0.5
0.6
0.1
0.1
0.1
<0.1
<0.1
<0.1
Detroit
0.5
0.5
0.6
0.1
0.1
0.1
<0.1
<0.1
<0.1
Philadelphia
0.5
0.5
0.5
0.1
0.1
0.1
<0.1
<0.1
<0.1
Phoenix
0.7
0.7
0.8
0.2
0.2
0.2
0.1
0.1
0.1
Sacramento
0.5
0.4
0.5
0.1
0.1
0.1
<0.1
<0.1
<0.1
St. Louis
0.6
0.5
0.6
0.1
0.1
0.1
<0.1
<0.1
0.1
A Calculated percent is rounded to the nearest tenth decimal using conventional rounding. Values equal to zero are designated by "0" (there
are no individuals experiencing decrements at that level). Small, non-zero values that do not round upwards to 0.1 (i.e., <0.05) are given a
value of "<0.1".
3D-119
-------
Table 3D-41. Number of people estimated to experience at least one lung function
decrement at or above the indicated level, for air quality adjusted to just meet
the current standard, using the population-based (E-R function) risk
approach.
Study
Group
Study Area
>10% reduction in FEViA
(# per Year)
>15% reduction in FEViA
(# per Year)
>20% reduction in FEViA
(# per Year)
Avg
Min
Max
Avg
Min
Max
Avg
Min
Max
Children
Atlanta
26149
22779
29781
6369
5064
7768
2273
1634
2966
Boston
29437
27715
31856
7433
6804
8442
2746
2457
3254
Dallas
34128
30101
37100
8615
7070
9837
3153
2412
3760
Detroit
26489
24402
28928
6978
6122
8030
2642
2220
3174
Philadelphia
30134
28919
31014
7406
7050
7661
2655
2510
2750
Phoenix
26169
24400
28193
6930
6199
7770
2614
2250
3029
Sacramento
10458
10047
10800
2484
2321
2632
859
784
932
St. Louis
13912
12540
15144
3594
3069
4143
1345
1093
1630
Children
with
Asthma
Atlanta
3322
2885
3793
814
646
989
289
202
383
Boston
4027
3686
4323
1024
910
1160
387
341
455
Dallas
3389
2956
3712
859
686
993
315
236
378
Detroit
3208
2931
3503
844
CO
r—
971
318
260
382
Philadelphia
3594
3448
3732
880
829
917
320
306
327
Phoenix
2684
2463
2901
713
623
807
269
226
311
Sacramento
1043
1009
1095
246
233
264
85
78
93
St. Louis
1439
1302
1530
370
319
419
137
109
164
Adults
Atlanta
26671
24018
29934
5658
4789
6691
1808
1409
2254
Boston
33036
30818
35514
7011
6261
7925
2218
1859
2642
Dallas
32817
29848
35083
7215
6095
8126
2370
1875
2813
Detroit
25452
23857
27527
5921
5309
6816
2054
1770
2491
Philadelphia
32243
30936
33288
6826
6449
7146
2150
2004
2266
Phoenix
28046
26622
29304
6639
6109
7102
2284
2036
2483
Sacramento
10719
10490
10891
2239
2144
2315
677
629
715
St. Louis
14271
12662
15165
3207
2683
3577
1073
858
1252
Adults
with
Asthma
Atlanta
1714
1550
1902
352
282
423
117
70
141
Boston
2870
2544
3131
587
489
685
196
196
196
Dallas
1953
1797
2110
443
391
469
130
78
156
Detroit
2338
2163
2491
524
459
590
175
131
197
Philadelphia
2585
2527
2701
552
523
610
174
174
174
Phoenix
2020
1937
2086
480
447
497
166
149
199
Sacramento
629
600
657
133
114
143
29
29
29
St. Louis
1132
1037
1180
250
215
286
84
72
107
A These values represent the population of individuals exposed in each study area. Values equal to zero are indicated by "0" (there are no
individuals experiencing decrements at the level).
3D-120
-------
Table 3D-42. Percent of people estimated to experience at least two lung function
decrements at or above the indicated level, for air quality adjusted to just
meet the current standard, using the population-based (E-R function) risk
approach.
Study
Group
Study Area
>10% reduction in FEViA
(% per Year)
>15% reduction in FEViA
(% per Year)
>20% reduction in FEViA
(% per Year)
Avg
Min
Max
Avg
Min
Max
Avg
Min
Max
Children
Atlanta
1.4
1.3
1.6
0.3
0.3
0.4
0.1
0.1
0.1
Boston
1.3
1.3
1.4
0.3
0.3
0.3
0.1
0.1
0.1
Dallas
1.6
1.5
1.7
0.4
0.3
0.4
0.1
0.1
0.1
Detroit
1.6
1.5
1.8
0.4
O
CO
0.4
0.1
0.1
0.1
Philadelphia
1.5
1.4
1.6
O
CO
0.3
0.3
0.1
0.1
0.1
Phoenix
2.2
2.1
2.4
0.5
0.5
0.6
0.2
0.2
0.2
Sacramento
1.5
1.5
1.6
0.3
0.3
0.3
0.1
0.1
0.1
St. Louis
1.7
1.5
1.8
0.4
0.3
0.4
0.1
0.1
0.1
Children
with
Asthma
Atlanta
1.6
1.4
1.7
0.3
0.3
0.4
0.1
0.1
0.1
Boston
1.5
1.4
1.6
0.3
0.3
0.4
0.1
0.1
0.1
Dallas
1.7
1.6
1.9
0.4
O
CO
0.4
0.1
0.1
0.1
Detroit
1.8
1.6
1.9
0.4
0.4
0.4
0.1
0.1
0.1
Philadelphia
1.6
1.6
1.7
0.4
o
CO
0.4
0.1
0.1
0.1
Phoenix
2.4
2.2
2.6
0.6
0.5
0.6
0.2
0.2
0.2
Sacramento
1.6
1.6
1.6
0.3
o
CO
0.4
0.1
0.1
0.1
St. Louis
1.7
1.5
1.8
0.4
0.3
0.4
0.1
0.1
0.1
Adults
Atlanta
0.4
0.3
0.4
0.1
0.1
0.1
<0.1
<0.1
<0.1
Boston
0.3
0.3
0.3
0.1
0.1
0.1
<0.1
<0.1
<0.1
Dallas
0.4
0.4
0.4
0.1
0.1
0.1
<0.1
<0.1
<0.1
Detroit
0.4
0.3
0.4
0.1
0.1
0.1
<0.1
<0.1
<0.1
Philadelphia
0.4
0.3
0.4
0.1
0.1
0.1
<0.1
<0.1
<0.1
Phoenix
0.6
0.6
0.6
0.1
0.1
0.1
<0.1
<0.1
<0.1
Sacramento
0.4
0.4
0.4
0.1
0.1
0.1
<0.1
<0.1
<0.1
St. Louis
0.4
0.3
0.4
0.1
0.1
0.1
<0.1
<0.1
<0.1
Adults
with
Asthma
Atlanta
0.3
0.3
0.3
0.1
<0.1
0.1
<0.1
<0.1
<0.1
Boston
0.3
0.2
0.3
<0.1
<0.1
<0.1
<0.1
<0.1
<0.1
Dallas
0.3
0.3
0.4
0.1
<0.1
0.1
<0.1
<0.1
<0.1
Detroit
0.3
0.3
0.3
0.1
0.1
0.1
<0.1
<0.1
<0.1
Philadelphia
0.3
0.3
0.3
0.1
0.1
0.1
<0.1
<0.1
<0.1
Phoenix
0.5
0.4
0.5
0.1
0.1
0.1
<0.1
<0.1
<0.1
Sacramento
0.3
0.3
0.3
<0.1
<0.1
0.1
<0.1
<0.1
<0.1
St. Louis
0.3
0.3
0.3
0.1
0.1
0.1
<0.1
<0.1
<0.1
A Calculated percent is rounded to the nearest tenth decimal using conventional rounding. Values equal to zero are designated by "0" (there
are no individuals experiencing decrements at that level). Small, non-zero values that do not round upwards to 0.1 (i.e., <0.05) are given a
value of "<0.1".
3D-121
-------
Table 3D-43. Number of people estimated to experience at least two lung function
decrements at or above the indicated level, for air quality adjusted to just
meet the current standard, using the population-based (E-R function) risk
approach.
Study
Group
Study Area
>10% reduction in FEViA
(# per Year)
>15% reduction in FEViA
(# per Year)
>20% reduction in FEViA
(# per Year)
Avg
Min
Max
Avg
Min
Max
Avg
Min
Max
Children
Atlanta
17291
15395
19450
3632
3047
4277
1110
868
1372
Boston
18378
17430
19432
3891
3572
4278
1206
1069
1388
Dallas
22897
20572
24757
5036
4256
5722
1624
1277
1939
Detroit
17100
15973
18384
3896
3503
4370
1295
1127
1509
Philadelphia
19992
18901
20909
4263
3972
4518
1324
1222
1419
Phoenix
18937
17748
20310
4529
4076
5039
1566
1359
1797
Sacramento
7200
6972
7415
1511
1436
1592
461
427
497
St. Louis
9222
8351
9790
2076
1803
2295
680
565
783
Children
with
Asthma
Atlanta
2219
1977
2462
464
383
545
148
121
182
Boston
2526
2321
2662
539
478
592
159
137
182
Dallas
2278
2034
2483
496
402
567
158
118
189
Detroit
2064
1890
2220
468
416
520
156
139
173
Philadelphia
2416
2292
2532
517
480
546
160
153
175
Phoenix
1948
1797
2109
467
410
524
161
142
184
Sacramento
717
699
753
153
148
163
49
47
54
St. Louis
941
856
1002
210
182
228
67
55
73
Adults
Atlanta
15542
14157
17468
2841
2465
3310
751
634
916
Boston
18654
17904
19274
3326
3131
3522
848
783
881
Dallas
19091
17893
19925
3542
3204
3829
990
859
1094
Detroit
14135
13567
14747
2731
2556
2949
765
721
852
Philadelphia
18939
18126
19607
3457
3224
3660
930
871
959
Phoenix
17781
16986
18625
3708
3427
3973
1142
1043
1242
Sacramento
6536
6489
6574
1182
1172
1201
305
286
314
St. Louis
8203
7261
8870
1586
1323
1753
453
358
501
Adults
with
Asthma
Atlanta
1010
916
1127
188
141
211
70
70
70
Boston
1631
1468
1761
261
196
294
98
98
98
Dallas
1120
1094
1172
208
156
234
78
78
78
Detroit
1311
1245
1376
262
262
262
66
66
66
Philadelphia
1452
1394
1569
261
261
261
87
87
87
Phoenix
1275
1192
1341
265
248
298
83
50
99
Sacramento
391
372
400
67
57
86
29
29
29
St. Louis
656
608
715
131
107
143
36
36
36
A These values represent the population of individuals exposed in each study area. Values equal to zero are indicated by "0" (there are no
individuals experiencing decrements at the level).
3D-122
-------
Table 3D-44. Percent of people estimated to experience at least four lung function
decrements at or above the indicated level, for air quality adjusted to just
meet the current standard, using the population-based (E-R function) risk
approach.
Study
Group
Study Area
>10% reduction in FEViA
(% per Year)
>15% reduction in FEViA
(% per Year)
>20% reduction in FEViA
(% per Year)
Avg
Min
Max
Avg
Min
Max
Avg
Min
Max
Children
Atlanta
1.0
0.9
1.1
0.2
0.2
0.2
<0.1
<0.1
0.1
Boston
0.8
0.8
0.9
0.2
0.1
0.2
<0.1
<0.1
<0.1
Dallas
1.1
1.0
1.1
0.2
0.2
0.2
0.1
0.1
0.1
Detroit
1.0
1.0
1.1
0.2
0.2
0.2
0.1
0.1
0.1
Philadelphia
1.0
0.9
1.1
0.2
0.2
0.2
0.1
<0.1
0.1
Phoenix
1.6
1.5
1.7
0.4
0.3
0.4
0.1
0.1
0.1
Sacramento
1.1
1.0
1.1
0.2
0.2
0.2
0.1
0.1
0.1
St. Louis
1.1
1.0
1.2
0.2
0.2
0.2
0.1
0.1
0.1
Children
with
Asthma
Atlanta
1.0
1.0
1.2
0.2
0.2
0.2
0.1
<0.1
0.1
Boston
0.9
0.9
1.0
0.2
0.2
0.2
<0.1
<0.1
0.1
Dallas
1.2
1.1
1.2
0.2
0.2
0.3
0.1
0.1
0.1
Detroit
1.1
1.0
1.2
0.2
0.2
0.2
0.1
0.1
0.1
Philadelphia
1.1
1.0
1.1
0.2
0.2
0.2
0.1
<0.1
0.1
Phoenix
1.7
1.6
1.8
0.4
0.4
0.4
0.1
0.1
0.1
Sacramento
1.1
1.1
1.1
0.2
0.2
0.2
0.1
0.1
0.1
St. Louis
1.1
1.0
1.2
0.2
0.2
0.2
0.1
<0.1
0.1
Adults
Atlanta
0.2
0.2
0.2
<0.1
<0.1
<0.1
<0.1
<0.1
<0.1
Boston
0.2
0.2
0.2
<0.1
<0.1
<0.1
<0.1
<0.1
<0.1
Dallas
0.2
0.2
0.2
<0.1
<0.1
<0.1
<0.1
<0.1
<0.1
Detroit
0.2
0.2
0.2
<0.1
<0.1
<0.1
<0.1
<0.1
<0.1
Philadelphia
0.2
0.2
0.2
<0.1
<0.1
<0.1
<0.1
<0.1
<0.1
Phoenix
0.4
0.4
0.4
0.1
0.1
0.1
<0.1
<0.1
<0.1
Sacramento
0.2
0.2
0.2
<0.1
<0.1
<0.1
<0.1
<0.1
<0.1
St. Louis
0.2
0.2
0.2
<0.1
<0.1
<0.1
<0.1
<0.1
<0.1
Adults
with
Asthma
Atlanta
0.2
0.2
0.2
<0.1
<0.1
<0.1
0
0
0
Boston
0.1
0.1
0.2
<0.1
<0.1
<0.1
0
0
0
Dallas
0.2
0.2
0.2
<0.1
<0.1
<0.1
0
0
0
Detroit
0.2
0.2
0.2
<0.1
<0.1
<0.1
0
0
0
Philadelphia
0.2
0.2
0.2
<0.1
<0.1
<0.1
0
0
0
Phoenix
0.3
0.3
0.3
0.1
0.1
0.1
<0.1
<0.1
<0.1
Sacramento
0.2
0.2
0.2
<0.1
<0.1
<0.1
0
0
0
St. Louis
0.2
0.2
0.2
<0.1
<0.1
<0.1
0
0
0
A Calculated percent is rounded to the nearest tenth decimal using conventional rounding. Values equal to zero are designated by "0" (there
are no individuals experiencing decrements at that level). Small, non-zero values that do not round upwards to 0.1 (i.e., <0.05) are given a
value of "<0.1".
3D-123
-------
Table 3D-45. Number of people estimated to experience at least four lung function
decrements at or above the indicated level, for air quality adjusted to just
meet the current standard, using the population-based (E-R function) risk
approach.
Study
Group
Study Area
>10% reduction in FEViA
(# per Year)
>15% reduction in FEViA
(# per Year)
>20% reduction in FEViA
(# per Year)
Avg
Min
Max
Avg
Min
Max
Avg
Min
Max
Children
Atlanta
11501
10371
12953
2179
1876
2542
592
484
726
Boston
11476
11127
11673
2131
2025
2207
561
523
592
Dallas
15259
14022
16197
3011
2648
3334
851
709
993
Detroit
10816
10180
11429
2174
1994
2359
636
572
711
Philadelphia
13227
12310
13969
2539
2314
2728
698
611
764
Phoenix
13597
12823
14564
2972
2703
3284
948
835
1076
Sacramento
4935
4814
5023
944
908
978
256
241
272
St. Louis
6041
5446
6411
1214
1056
1302
352
291
382
Children
with
Asthma
Atlanta
1486
1352
1654
282
242
323
81
61
101
Boston
1585
1502
1661
296
273
319
83
68
91
Dallas
1529
1395
1632
299
260
331
87
71
95
Detroit
1306
1197
1387
260
243
277
75
69
87
Philadelphia
1593
1484
1702
306
284
327
80
65
87
Phoenix
1406
1316
1500
311
283
340
99
85
113
Sacramento
489
474
512
96
93
101
26
23
31
St. Louis
619
565
674
124
109
137
33
27
36
Adults
Atlanta
8969
8382
10072
1455
1338
1690
329
282
423
Boston
10534
10175
10762
1630
1565
1663
359
294
391
Dallas
10965
10548
11252
1823
1719
1875
417
391
469
Detroit
7865
7537
8127
1311
1245
1376
328
328
328
Philadelphia
10922
10457
11241
1772
1656
1830
407
349
436
Phoenix
11093
10629
11672
2069
1937
2235
563
497
646
Sacramento
3992
3916
4059
648
629
657
143
143
143
St. Louis
4626
4113
5043
775
680
858
179
143
215
Adults
with
Asthma
Atlanta
587
563
634
70
70
70
0
0
0
Boston
913
783
978
131
98
196
0
0
0
Dallas
651
625
703
78
78
78
0
0
0
Detroit
721
655
786
131
131
131
0
0
0
Philadelphia
813
784
871
116
87
174
0
0
0
Phoenix
778
745
844
149
149
149
50
50
50
Sacramento
229
229
229
29
29
29
0
0
0
St. Louis
358
322
393
60
36
72
0
0
0
A These values represent the population of individuals exposed in each study area. Values equal to zero are indicated by "0" (there are no
individuals experiencing decrements at the level).
3D-124
-------
Table 3D-46. Percent of people estimated to experience at least one lung function
decrement at or above the indicated level, for the 75 ppb air quality scenario,
using the population-based (E-R function) risk approach.
Study
Group
Study Area
>10% reduction in FEViA
(% per Year)
>15% reduction in FEViA
(% per Year)
>20% reduction in FEViA
(% per Year)
Avg
Min
Max
Avg
Min
Max
Avg
Min
Max
Children
Atlanta
2.8
2.4
3.2
0.8
0.6
1.0
0.3
0.2
0.4
Boston
2.4
2.3
2.7
0.7
0.6
0.8
0.3
0.2
0.3
Dallas
2.8
2.4
3.1
0.8
0.6
0.9
0.3
0.2
0.4
Detroit
3.0
2.7
3.4
0.9
0.8
1.0
0.4
0.3
0.5
Philadelphia
2.9
2.7
3.0
0.8
0.8
0.8
0.3
0.3
0.3
Phoenix
3.8
3.5
4.1
1.1
1.0
1.3
0.5
0.4
0.6
Sacramento
2.8
2.7
2.9
0.8
0.7
0.8
0.3
0.3
0.3
St. Louis
3.1
2.7
3.3
0.9
0.8
1.0
0.4
0.3
0.4
Children
with
Asthma
Atlanta
3.0
2.6
3.5
0.9
0.7
1.1
0.4
0.2
0.5
Boston
2.7
2.5
3.0
0.8
0.6
0.9
0.3
0.2
0.4
Dallas
3.0
2.6
3.3
0.8
0.7
1.0
0.3
0.3
0.4
Detroit
3.3
3.0
3.6
1.0
0.8
1.1
0.4
0.3
0.5
Philadelphia
3.1
2.9
3.1
0.9
0.8
0.9
0.3
0.3
0.4
Phoenix
4.1
3.7
4.4
1.2
1.1
1.4
0.5
0.4
0.6
Sacramento
2.9
2.8
2.9
0.8
0.7
0.8
0.3
0.3
0.3
St. Louis
3.1
2.7
3.4
0.9
0.7
1.0
0.4
0.3
0.4
Adults
Atlanta
0.8
0.7
0.9
0.2
0.2
0.2
0.1
0.1
0.1
Boston
0.6
0.6
0.7
0.1
0.1
0.2
<0.1
<0.1
0.1
Dallas
0.8
0.7
0.9
0.2
0.2
0.2
0.1
0.1
0.1
Detroit
0.7
0.7
0.8
0.2
0.2
0.2
0.1
0.1
0.1
Philadelphia
0.7
0.7
0.8
0.2
0.2
0.2
0.1
0.1
0.1
Phoenix
1.1
1.0
1.2
0.3
0.3
0.3
0.1
0.1
0.1
Sacramento
0.7
0.7
0.8
0.2
0.2
0.2
0.1
0.1
0.1
St. Louis
0.8
0.7
0.8
0.2
0.2
0.2
0.1
0.1
0.1
Adults
with
Asthma
Atlanta
0.6
0.5
0.7
0.1
0.1
0.2
<0.1
<0.1
0.1
Boston
0.5
0.4
0.5
0.1
0.1
0.1
<0.1
<0.1
<0.1
Dallas
0.6
0.6
0.7
0.2
0.1
0.2
0.1
<0.1
0.1
Detroit
0.6
0.6
0.7
0.2
0.1
0.2
0.1
<0.1
0.1
Philadelphia
0.6
0.6
0.6
0.1
0.1
0.2
0.1
0.1
0.1
Phoenix
0.9
0.8
0.9
0.2
0.2
0.2
0.1
0.1
0.1
Sacramento
0.6
0.5
0.6
0.1
0.1
0.1
<0.1
<0.1
<0.1
St. Louis
0.6
0.6
0.7
0.2
0.1
0.2
0.1
0.1
0.1
A Calculated percent is rounded to the nearest tenth decimal using conventional rounding. Values equal to zero are designated by "0" (there
are no individuals experiencing decrements at that level). Small, non-zero values that do not round upwards to 0.1 (i.e., <0.05) are given a
value of "<0.1".
3D-125
-------
Table 3D-47. Percent of people estimated to experience at least two lung function
decrements at or above the indicated level, for the 75 ppb air quality scenario,
using the population-based (E-R function) risk approach.
Study
Group
Study Area
>10% reduction in FEViA
(% per Year)
>15% reduction in FEViA
(% per Year)
>20% reduction in FEViA
(% per Year)
Avg
Min
Max
Avg
Min
Max
Avg
Min
Max
Children
Atlanta
1.8
1.6
2.0
0.4
0.3
0.5
0.1
0.1
0.2
Boston
1.5
1.4
1.6
0.3
0.3
0.4
0.1
0.1
0.1
Dallas
1.9
1.6
2.1
0.4
0.4
0.5
0.2
0.1
0.2
Detroit
1.9
1.7
2.1
0.5
0.4
0.5
0.2
0.1
0.2
Philadelphia
1.9
1.8
1.9
0.4
0.4
0.5
0.2
0.1
0.2
Phoenix
2.7
2.5
2.9
0.7
0.6
0.8
0.3
0.2
0.3
Sacramento
1.9
1.8
1.9
0.4
0.4
0.5
0.1
0.1
0.2
St. Louis
2.0
1.8
2.1
0.5
0.4
0.5
0.2
0.1
0.2
Children
with
Asthma
Atlanta
1.9
1.7
2.2
0.5
0.4
0.5
0.2
0.1
0.2
Boston
1.7
1.5
1.8
0.4
0.3
0.4
0.1
0.1
0.2
Dallas
2.0
1.8
2.2
0.5
0.4
0.6
0.2
0.1
0.2
Detroit
2.0
1.9
2.2
0.5
0.5
0.6
0.2
0.2
0.2
Philadelphia
2.0
1.9
2.1
0.5
0.4
0.5
0.2
0.1
0.2
Phoenix
2.9
2.7
3.1
0.8
0.7
0.9
0.3
0.3
0.3
Sacramento
1.9
1.9
2.0
0.5
0.4
0.5
0.1
0.1
0.2
St. Louis
2.0
1.8
2.1
0.5
0.4
0.5
0.2
0.1
0.2
Adults
Atlanta
0.4
0.4
0.5
0.1
0.1
0.1
<0.1
<0.1
<0.1
Boston
0.3
0.3
0.3
0.1
0.1
0.1
<0.1
<0.1
<0.1
Dallas
0.4
0.4
0.5
0.1
0.1
0.1
<0.1
<0.1
<0.1
Detroit
0.4
0.4
0.4
0.1
0.1
0.1
<0.1
<0.1
<0.1
Philadelphia
0.4
0.4
0.4
0.1
0.1
0.1
<0.1
<0.1
<0.1
Phoenix
0.7
0.6
0.7
0.2
0.1
0.2
0.1
<0.1
0.1
Sacramento
0.4
0.4
0.4
0.1
0.1
0.1
<0.1
<0.1
<0.1
St. Louis
0.4
0.4
0.5
0.1
0.1
0.1
<0.1
<0.1
<0.1
Adults
with
Asthma
Atlanta
0.3
0.3
0.3
0.1
0.1
0.1
<0.1
<0.1
<0.1
Boston
0.3
0.2
0.3
0.1
<0.1
0.1
<0.1
<0.1
<0.1
Dallas
0.4
0.4
0.4
0.1
0.1
0.1
<0.1
<0.1
<0.1
Detroit
0.3
0.3
0.3
0.1
0.1
0.1
<0.1
<0.1
<0.1
Philadelphia
0.3
0.3
0.4
0.1
0.1
0.1
<0.1
<0.1
<0.1
Phoenix
0.5
0.5
0.6
0.1
0.1
0.1
<0.1
<0.1
0.1
Sacramento
0.3
0.3
0.3
0.1
0.1
0.1
<0.1
<0.1
<0.1
St. Louis
0.4
0.3
0.4
0.1
0.1
0.1
<0.1
<0.1
<0.1
A Calculated percent is rounded to the nearest tenth decimal using conventional rounding. Values equal to zero are designated by "0" (there
are no individuals experiencing decrements at that level). Small, non-zero values that do not round upwards to 0.1 (i.e., <0.05) are given a
value of "<0.1".
3D-126
-------
Table 3D-48. Percent of people estimated to experience at least four lung function
decrements at or above the indicated level, for the 75 ppb air quality scenario,
using the population-based (E-R function) risk approach.
Study
Group
Study Area
>10% reduction in FEViA
(% per Year)
>15% reduction in FEViA
(% per Year)
>20% reduction in FEViA
(% per Year)
Avg
Min
Max
Avg
Min
Max
Avg
Min
Max
Atlanta
1.2
1.0
1.3
0.2
0.2
0.3
0.1
0.1
0.1
Boston
0.9
0.9
0.9
0.2
0.2
0.2
<0.1
<0.1
0.1
Dallas
1.2
1.1
1.3
0.3
0.2
0.3
0.1
0.1
0.1
Children
Detroit
1.2
1.1
1.3
0.3
0.2
0.3
0.1
0.1
0.1
Philadelphia
1.2
1.1
1.3
0.3
0.2
O
CO
0.1
0.1
0.1
Phoenix
1.9
1.8
2.0
0.5
0.4
0.5
0.2
0.1
0.2
Sacramento
1.3
1.2
1.3
0.3
0.2
0.3
0.1
0.1
0.1
St. Louis
1.3
1.1
1.4
0.3
0.2
0.3
0.1
0.1
0.1
Atlanta
1.3
1.2
1.4
0.3
0.2
0.3
0.1
0.1
0.1
Boston
1.0
1.0
1.1
0.2
0.2
0.2
0.1
0.1
0.1
Children
with
Asthma
Dallas
1.3
1.2
1.4
O
CO
0.2
0.3
0.1
0.1
0.1
Detroit
1.2
1.2
1.3
0.3
0.2
0.3
0.1
0.1
0.1
Philadelphia
1.3
1.2
1.4
0.3
0.3
0.3
0.1
0.1
0.1
Phoenix
2.0
1.9
2.2
0.5
0.4
0.6
0.2
0.2
0.2
Sacramento
1.3
1.3
1.3
0.3
0.3
0.3
0.1
0.1
0.1
St. Louis
1.3
1.1
1.4
0.3
0.2
0.3
0.1
0.1
0.1
Atlanta
0.2
0.2
0.3
<0.1
<0.1
<0.1
<0.1
<0.1
<0.1
Boston
0.2
0.2
0.2
<0.1
<0.1
<0.1
<0.1
<0.1
<0.1
Dallas
0.3
0.2
0.3
<0.1
<0.1
<0.1
<0.1
<0.1
<0.1
Adults
Detroit
0.2
0.2
0.2
<0.1
<0.1
<0.1
<0.1
<0.1
<0.1
Philadelphia
0.2
0.2
0.2
<0.1
<0.1
<0.1
<0.1
<0.1
<0.1
Phoenix
0.4
0.4
0.4
0.1
0.1
0.1
<0.1
<0.1
<0.1
Sacramento
0.3
0.3
0.3
<0.1
<0.1
<0.1
<0.1
<0.1
<0.1
St. Louis
0.2
0.2
0.3
<0.1
<0.1
<0.1
<0.1
<0.1
<0.1
Atlanta
0.2
0.2
0.2
<0.1
<0.1
<0.1
0
0
0
Boston
0.1
0.1
0.2
<0.1
<0.1
<0.1
0
0
0
Adults
with
Asthma
Dallas
0.2
0.2
0.2
<0.1
<0.1
<0.1
0
0
0
Detroit
0.2
0.2
0.2
<0.1
<0.1
<0.1
<0.1
0
<0.1
Philadelphia
0.2
0.2
0.2
<0.1
<0.1
<0.1
0
0
0
Phoenix
0.3
0.3
0.3
0.1
0.1
0.1
<0.1
<0.1
<0.1
Sacramento
0.2
0.2
0.2
<0.1
<0.1
<0.1
0
0
0
St. Louis
0.2
0.2
0.2
<0.1
<0.1
<0.1
<0.1
0
<0.1
A Calculated percent is rounded to the nearest tenth decimal using conventional rounding. Values equal to zero are designated by "0" (there
are no individuals experiencing decrements at that level). Small, non-zero values that do not round upwards to 0.1 (i.e., <0.05) are given a
value of "<0.1".
3D-127
-------
Table 3D-49. Percent of people estimated to experience at least one lung function
decrement at or above the indicated level, for the 65 ppb air quality scenario,
using the population-based (E-R function) risk approach.
Study
Group
Study Area
>10% reduction in FEViA
(% per Year)
>15% reduction in FEViA
(% per Year)
>20% reduction in FEViA
(% per Year)
Avg
Min
Max
Avg
Min
Max
Avg
Min
Max
Children
Atlanta
1.7
1.5
1.9
0.4
0.3
0.4
0.1
0.1
0.1
Boston
1.8
1.7
1.9
0.4
0.4
0.4
0.1
0.1
0.2
Dallas
2.0
1.8
2.1
0.5
0.4
0.5
0.2
0.1
0.2
Detroit
2.0
1.9
2.2
0.5
0.4
0.5
0.2
0.1
0.2
Philadelphia
1.9
1.8
2.0
0.4
0.4
0.4
0.1
0.1
0.1
Phoenix
2.4
2.3
2.6
0.6
0.5
0.6
0.2
0.2
0.2
Sacramento
1.7
1.7
1.7
0.3
0.3
0.4
0.1
0.1
0.1
St. Louis
2.0
1.8
2.1
0.4
0.4
0.5
0.1
0.1
0.2
Children
with
Asthma
Atlanta
1.8
1.6
2.0
0.4
0.3
0.5
0.1
0.1
0.2
Boston
2.0
1.9
2.1
0.4
0.4
0.5
0.1
0.1
0.2
Dallas
2.2
2.0
2.3
0.5
0.4
0.5
0.2
0.1
0.2
Detroit
2.2
2.1
2.4
0.5
0.5
0.6
0.2
0.2
0.2
Philadelphia
2.0
2.0
2.1
0.5
0.4
0.5
0.1
0.1
0.1
Phoenix
2.6
2.4
2.8
0.6
0.5
0.7
0.2
0.2
0.2
Sacramento
1.7
1.7
1.8
0.4
0.3
0.4
0.1
0.1
0.1
St. Louis
2.0
1.8
2.2
0.5
0.4
0.5
0.1
0.1
0.2
Adults
Atlanta
0.5
0.5
0.6
0.1
0.1
0.1
<0.1
<0.1
<0.1
Boston
0.5
0.5
0.5
0.1
0.1
0.1
<0.1
<0.1
<0.1
Dallas
0.6
0.6
0.6
0.1
0.1
0.1
<0.1
<0.1
<0.1
Detroit
0.5
0.5
0.6
0.1
0.1
0.1
<0.1
<0.1
<0.1
Philadelphia
0.5
0.5
0.5
0.1
0.1
0.1
<0.1
<0.1
<0.1
Phoenix
0.8
0.7
0.8
0.2
0.2
0.2
<0.1
<0.1
0.1
Sacramento
0.5
0.5
0.5
0.1
0.1
0.1
<0.1
<0.1
<0.1
St. Louis
0.6
0.5
0.6
0.1
0.1
0.1
<0.1
<0.1
<0.1
Adults
with
Asthma
Atlanta
0.4
0.4
0.4
0.1
0.1
0.1
<0.1
<0.1
<0.1
Boston
0.4
0.4
0.4
0.1
0.1
0.1
<0.1
<0.1
<0.1
Dallas
0.5
0.5
0.5
0.1
0.1
0.1
<0.1
<0.1
<0.1
Detroit
0.5
0.5
0.5
0.1
0.1
0.1
<0.1
<0.1
<0.1
Philadelphia
0.4
0.4
0.5
0.1
0.1
0.1
<0.1
<0.1
<0.1
Phoenix
0.6
0.6
0.6
0.1
0.1
0.1
<0.1
<0.1
<0.1
Sacramento
0.4
0.4
0.4
0.1
0.1
0.1
<0.1
<0.1
<0.1
St. Louis
0.5
0.4
0.5
0.1
0.1
0.1
<0.1
<0.1
<0.1
A Calculated percent is rounded to the nearest tenth decimal using conventional rounding. Values equal to zero are designated by "0" (there
are no individuals experiencing decrements at that level). Small, non-zero values that do not round upwards to 0.1 (i.e., <0.05) are given a
value of "<0.1".
3D-128
-------
Table 3D-50. Percent of people estimated to experience at least two lung function
decrements at or above the indicated level, for the 65 ppb air quality scenario,
using the population-based (E-R function) risk approach.
Study
Group
Study Area
>10% reduction in FEViA
(% per Year)
>15% reduction in FEViA
(% per Year)
>20% reduction in FEViA
(% per Year)
Avg
Min
Max
Avg
Min
Max
Avg
Min
Max
Children
Atlanta
1.1
1.0
1.3
0.2
0.2
0.2
0.1
<0.1
0.1
Boston
1.1
1.1
1.2
0.2
0.2
0.2
0.1
0.1
0.1
Dallas
1.4
1.3
1.5
0.3
0.2
0.3
0.1
0.1
0.1
Detroit
1.4
1.3
1.4
0.3
0.3
0.3
0.1
0.1
0.1
Philadelphia
1.3
1.2
1.4
0.3
0.2
O
CO
0.1
0.1
0.1
Phoenix
1.8
1.7
1.9
0.4
0.4
0.4
0.1
0.1
0.1
Sacramento
1.2
1.2
1.2
0.2
0.2
0.2
0.1
0.1
0.1
St. Louis
1.4
1.2
1.4
0.3
0.2
0.3
0.1
0.1
0.1
Children
with
Asthma
Atlanta
1.2
1.1
1.4
0.2
0.2
0.3
0.1
0.1
0.1
Boston
1.3
1.2
1.3
0.2
0.2
0.3
0.1
0.1
0.1
Dallas
1.5
1.4
1.6
0.3
0.3
0.3
0.1
0.1
0.1
Detroit
1.5
1.4
1.5
0.3
0.3
0.3
0.1
0.1
0.1
Philadelphia
1.4
1.3
1.5
0.3
0.3
0.3
0.1
0.1
0.1
Phoenix
1.9
1.8
2.1
0.4
0.4
0.5
0.1
0.1
0.1
Sacramento
1.2
1.2
1.3
0.2
0.2
0.2
0.1
0.1
0.1
St. Louis
1.4
1.2
1.4
0.3
0.2
0.3
0.1
0.1
0.1
Adults
Atlanta
0.3
0.3
0.3
0.1
<0.1
0.1
<0.1
<0.1
<0.1
Boston
0.3
0.3
0.3
<0.1
<0.1
0.1
<0.1
<0.1
<0.1
Dallas
0.4
0.3
0.4
0.1
0.1
0.1
<0.1
<0.1
<0.1
Detroit
0.3
0.3
0.3
0.1
0.1
0.1
<0.1
<0.1
<0.1
Philadelphia
0.3
0.3
0.3
0.1
0.1
0.1
<0.1
<0.1
<0.1
Phoenix
0.5
0.5
0.5
0.1
0.1
0.1
<0.1
<0.1
<0.1
Sacramento
0.3
0.3
0.3
0.1
0.1
0.1
<0.1
<0.1
<0.1
St. Louis
0.3
0.3
0.4
0.1
0.1
0.1
<0.1
<0.1
<0.1
Adults
with
Asthma
Atlanta
0.2
0.2
0.3
<0.1
<0.1
<0.1
<0.1
0
<0.1
Boston
0.2
0.2
0.2
<0.1
<0.1
<0.1
<0.1
<0.1
<0.1
Dallas
0.3
0.3
0.3
<0.1
<0.1
<0.1
<0.1
<0.1
<0.1
Detroit
0.3
0.3
0.3
<0.1
<0.1
<0.1
<0.1
<0.1
<0.1
Philadelphia
0.3
0.2
0.3
<0.1
<0.1
0.1
<0.1
<0.1
<0.1
Phoenix
0.4
0.4
0.4
0.1
0.1
0.1
<0.1
<0.1
<0.1
Sacramento
0.2
0.2
0.2
<0.1
<0.1
<0.1
0
0
0
St. Louis
0.3
0.2
0.3
<0.1
<0.1
0.1
<0.1
<0.1
<0.1
A Calculated percent is rounded to the nearest tenth decimal using conventional rounding. Values equal to zero are designated by "0" (there
are no individuals experiencing decrements at that level). Small, non-zero values that do not round upwards to 0.1 (i.e., <0.05) are given a
value of "<0.1".
3D-129
-------
Table 3D-51. Percent of people estimated to experience at least four lung function
decrements at or above the indicated level, for the 65 ppb air quality scenario,
using the population-based (E-R function) risk approach.
Study
Group
Study Area
>10% reduction in FEViA
(% per Year)
>15% reduction in FEViA
(% per Year)
>20% reduction in FEViA
(% per Year)
Avg
Min
Max
Avg
Min
Max
Avg
Min
Max
Children
Atlanta
0.8
0.7
0.9
0.1
0.1
0.2
<0.1
<0.1
<0.1
Boston
0.7
0.7
0.7
0.1
0.1
0.1
<0.1
<0.1
<0.1
Dallas
0.9
0.9
1.0
0.2
0.2
0.2
<0.1
<0.1
0.1
Detroit
0.9
0.8
0.9
0.2
0.2
0.2
<0.1
<0.1
<0.1
Philadelphia
0.9
0.8
0.9
0.2
0.1
0.2
<0.1
<0.1
<0.1
Phoenix
1.3
1.2
1.4
0.3
0.2
0.3
0.1
0.1
0.1
Sacramento
0.8
0.8
0.9
0.1
0.1
0.2
<0.1
<0.1
<0.1
St. Louis
0.9
0.8
1.0
0.2
0.1
0.2
<0.1
<0.1
<0.1
Children
with
Asthma
Atlanta
0.8
0.8
0.9
0.1
0.1
0.2
<0.1
<0.1
<0.1
Boston
0.8
0.8
0.8
0.1
0.1
0.2
<0.1
<0.1
<0.1
Dallas
1.0
0.9
1.1
0.2
0.2
0.2
<0.1
<0.1
0.1
Detroit
0.9
0.9
1.0
0.2
0.2
0.2
<0.1
<0.1
<0.1
Philadelphia
0.9
0.9
1.0
0.2
0.1
0.2
<0.1
<0.1
<0.1
Phoenix
1.4
1.3
1.5
0.3
0.3
0.3
0.1
0.1
0.1
Sacramento
0.9
0.9
0.9
0.2
0.2
0.2
<0.1
<0.1
<0.1
St. Louis
0.9
0.8
1.0
0.2
0.1
0.2
<0.1
<0.1
<0.1
Adults
Atlanta
0.2
0.2
0.2
<0.1
<0.1
<0.1
<0.1
<0.1
<0.1
Boston
0.2
0.2
0.2
<0.1
<0.1
<0.1
<0.1
<0.1
<0.1
Dallas
0.2
0.2
0.2
<0.1
<0.1
<0.1
<0.1
<0.1
<0.1
Detroit
0.2
0.2
0.2
<0.1
<0.1
<0.1
<0.1
<0.1
<0.1
Philadelphia
0.2
0.2
0.2
<0.1
<0.1
<0.1
<0.1
<0.1
<0.1
Phoenix
0.3
0.3
0.3
0.1
0.1
0.1
<0.1
<0.1
<0.1
Sacramento
0.2
0.2
0.2
<0.1
<0.1
<0.1
<0.1
<0.1
<0.1
St. Louis
0.2
0.2
0.2
<0.1
<0.1
<0.1
<0.1
<0.1
<0.1
Adults
with
Asthma
Atlanta
0.1
0.1
0.1
<0.1
<0.1
<0.1
0
0
0
Boston
0.1
0.1
0.1
<0.1
<0.1
<0.1
0
0
0
Dallas
0.2
0.2
0.2
<0.1
<0.1
<0.1
0
0
0
Detroit
0.2
0.1
0.2
<0.1
<0.1
<0.1
0
0
0
Philadelphia
0.1
0.1
0.2
<0.1
<0.1
<0.1
0
0
0
Phoenix
0.2
0.2
0.3
<0.1
<0.1
<0.1
<0.1
<0.1
<0.1
Sacramento
0.1
0.1
0.1
<0.1
<0.1
<0.1
0
0
0
St. Louis
0.2
0.1
0.2
<0.1
<0.1
<0.1
0
0
0
A Calculated percent is rounded to the nearest tenth decimal using conventional rounding. Values equal to zero are designated by "0" (there
are no individuals experiencing decrements at that level). Small, non-zero values that do not round upwards to 0.1 (i.e., <0.05) are given a
value of "<0.1".
3D-130
-------
3D.3.3.2 Individual-based (MSS Model) Risk Approach
Lung function decrements estimated using the individual-based (MSS model) risk
approach are about a factor of four or greater than those estimated using the population-based (E-
R function) risk approach (Table 3D-52 through Table 3D-63). The estimated risk of at least one
lung function decrement at or above 15% could be as high as 7.8% of children (and 8.7% of
children with asthma) considering the worst year air quality and air quality just meeting the
current standard, with the average across the 3-year period ranging from about 4.1% to 7.1% of
children (and 4.5 to 8.2% of children with asthma) across the eight study areas (Table 3D-52).
Recall that when using the E-R approach for the same air quality scenario, only about 1% of
children were estimated to experience a decrement at or above 15% in the worst single year,
worst area, and between 0.5 to 0.9% on average across the three years. This difference in
estimated risks is generally similar to the comparison of the two approaches provided in the 2014
HREA (2014 HREA, Table 6-8) and is directly a result of the differences that exist between the
approaches. While both of these risk approaches allow for exposures at and below that observed
in the controlled human exposure studies, the MSS model does not have a strict restriction
regarding the magnitude of the ventilation rate or its duration. The impact of these important
model inputs (i.e., exposure, ventilation rate, and their duration) on the E-R and MSS risk results
is discussed further in section 3D.3.4.
3D-131
-------
Table 3D-52. Percent of people estimated to experience at least one lung function
decrement at or above the indicated level, for air quality adjusted to just meet
the current standard, using the individual-based (MSS model) risk approach.
Study
Group
Study Area
>10% reduction in FEViA
(% per Year)
>15% reduction in FEViA
(% per Year)
>20% reduction in FEViA
(% per Year)
Avg
Min
Max
Avg
Min
Max
Avg
Min
Max
Children
Atlanta
13.2
11.7
15.1
4.1
3.4
5.0
1.7
1.3
2.1
Boston
13.2
12.4
14.1
4.4
4.0
5.0
1.9
1.6
2.3
Dallas
14.6
13.1
15.7
4.9
4.0
5.4
2.1
1.6
2.5
Detroit
15.6
14.4
16.9
5.4
4.8
6.1
2.4
2
2.7
Philadelphia
14.5
13.6
15.0
4.6
4.3
4.8
1.9
1.8
1.9
Phoenix
20.4
19.4
21.8
7.1
6.4
7.8
3.1
2.7
3.6
Sacramento
14.3
13.8
14.7
4.4
4.3
4.7
1.8
1.7
2
St. Louis
15.4
14.0
16.3
5.2
4.5
5.9
2.2
1.9
2.7
Children
with
Asthma
Atlanta
14.4
12.5
16.6
4.5
3.4
5.9
1.9
1.5
2.6
Boston
13.9
12.9
14.7
4.8
4.4
5.4
2
1.7
2.4
Dallas
15.7
13.6
16.9
5.4
4.5
5.9
2.5
1.8
2.8
Detroit
16.8
15.3
18.4
6.2
5.7
6.9
2.7
2.3
3.3
Philadelphia
15.2
15.0
15.5
4.8
4.6
5.3
1.9
1.8
2.1
Phoenix
22.0
20.4
23.3
8.2
7.6
8.7
3.5
3
3.9
Sacramento
14.7
14.2
15.0
4.5
4.3
4.8
1.8
1.6
2.1
St. Louis
15.8
14.5
16.5
5.4
4.7
5.8
2.4
2
2.8
Adults
Atlanta
2.5
2.3
2.8
0.7
0.6
0.8
0.3
0.2
0.4
Boston
2.3
2.1
2.5
0.6
0.6
0.7
0.3
0.2
0.3
Dallas
2.9
2.6
3.1
0.8
0.7
1.0
0.3
0.3
0.4
Detroit
2.6
2.5
2.8
0.8
0.7
0.9
0.3
0.3
0.4
Philadelphia
2.5
2.4
2.6
0.7
0.7
0.7
0.3
0.3
0.3
Phoenix
4.4
4.1
4.8
1.4
1.3
1.5
0.6
0.6
0.6
Sacramento
2.6
2.6
2.6
0.7
0.6
0.7
0.3
0.3
0.3
St. Louis
2.7
2.3
2.9
0.8
0.7
0.9
0.3
0.3
0.4
Adults
with
Asthma
Atlanta
2.3
2.2
2.4
0.6
0.6
0.7
0.2
0.1
0.3
Boston
2.0
1.8
2.4
0.5
0.4
0.6
0.2
0.1
0.3
Dallas
2.5
2.1
2.9
0.7
0.5
1.0
0.3
0.1
0.5
Detroit
2.5
2.2
2.6
0.7
0.6
0.8
0.4
0.2
0.5
Philadelphia
2.2
2.1
2.4
0.6
0.4
0.7
0.2
0.1
0.3
Phoenix
3.5
3.1
3.8
1.1
1.0
1.2
0.5
0.4
0.6
Sacramento
2.0
2.0
2.1
0.6
0.5
0.6
0.2
0.2
0.3
St. Louis
2.6
2.2
2.9
0.7
0.6
0.8
0.3
0.2
0.4
A Calculated percent is rounded to the nearest tenth decimal using conventional rounding. Values equal to zero are designated by "0" (there
are no individuals experiencing decrements at that level). Small, non-zero values that do not round upwards to 0.1 (i.e., <0.05) are given a
value of "<0.1".
3D-132
-------
Table 3D-53. Number of people estimated to experience at least one lung function
decrement at or above the indicated level, for air quality adjusted to just meet
the current standard, using the individual-based (MSS model) risk approach.
Study
Group
Study Area
>10% reduction in FEViA
(# per Year)
>15% reduction in FEViA
(# per Year)
>20% reduction in FEViA
(# per Year)
Avg
Min
Max
Avg
Min
Max
Avg
Min
Max
Children
Atlanta
159429
141680
182558
49769
40676
60328
20378
15233
25685
Boston
179806
168747
192821
60125
55225
68218
26251
22368
31924
Dallas
207221
185830
222622
68911
57317
76942
29588
22747
34853
Detroit
162690
149480
176362
56695
50434
63632
24708
21037
28547
Philadelphia
189412
178688
196192
60159
56856
62400
24692
23615
25449
Phoenix
173383
164589
185182
60104
54688
65827
26306
22773
30302
Sacramento
66574
64364
68293
20665
20024
21871
8473
7904
9286
St. Louis
84339
76632
89017
28337
24351
32356
12185
10309
14507
Children
with
Asthma
Atlanta
20513
17776
23929
6356
4802
8434
2670
2078
3733
Boston
23353
21366
24529
8002
7259
9056
3390
2844
3959
Dallas
20485
17781
22298
7093
5911
7779
3208
2388
3689
Detroit
19702
17603
21662
7226
6521
8151
3197
2653
3902
Philadelphia
22393
21869
23135
7115
6701
7923
2859
2575
3099
Phoenix
17885
16460
18994
6690
6114
7119
2854
2434
3199
Sacramento
6625
6328
6972
2058
1941
2244
807
699
978
St. Louis
8852
8278
9234
3008
2650
3206
1327
1157
1512
Adults
Atlanta
105509
97903
117483
29535
25779
35639
11692
8804
15284
Boston
135567
121022
149395
37732
32286
41874
16110
12229
18295
Dallas
133978
119705
144083
39563
33442
44694
15680
13127
19222
Detroit
104123
97067
110175
30411
26872
33426
11994
10159
13764
Philadelphia
131672
124004
135506
36048
35118
37123
14175
14030
14466
Phoenix
131520
121636
143440
41440
40132
43658
17367
16887
18327
Sacramento
43953
43734
44306
11643
10948
12291
4840
4802
4888
St. Louis
57287
48965
62593
17168
13985
19207
6974
5508
8226
Adults
with
Asthma
Atlanta
8123
7677
8875
2278
2043
2395
751
493
1127
Boston
12980
11447
15067
3229
2348
3816
1337
881
1663
Dallas
8413
6876
9611
2344
1563
3360
912
391
1641
Detroit
10508
9241
11273
3059
2359
3474
1529
852
1966
Philadelphia
11241
10631
12026
2905
1917
3399
1220
436
1656
Phoenix
9520
8543
10232
2980
2632
3328
1358
1093
1738
Sacramento
2811
2716
2887
762
715
800
305
229
343
St. Louis
5258
4435
5687
1503
1288
1610
548
322
715
A These values represent the population of individuals exposed in each study area. Values equal to zero are indicated by "0" (there are no
individuals experiencing decrements at the level).
3D-133
-------
Table 3D-54. Percent of people estimated to experience at least two lung function
decrements at or above the indicated level, for air quality adjusted to just
meet the current standard, using the individual-based (MSS model) risk
approach.
Study
Group
Study Area
>10% reduction in FEViA
(% per Year)
>15% reduction in FEViA
(% per Year)
>20% reduction in FEViA
(% per Year)
Avg
Min
Max
Avg
Min
Max
Avg
Min
Max
Children
Atlanta
7.7
6.7
9.1
2.1
1.7
2,6
0.8
0.6
1.0
Boston
7.4
6.9
7.9
2.1
1.9
2.4
0.8
0.7
0.9
Dallas
8.8
7.8
9.5
2.6
2.0
3.0
1.0
0.7
1.2
Detroit
9.4
8.5
10.3
2.9
2.5
3.3
1.1
0.9
1.3
Philadelphia
8.7
8.0
9.1
2.4
2.3
2.5
0.9
0.8
0.9
Phoenix
13.6
12.8
14.8
4.3
3.8
4.9
1.7
1.5
2
Sacramento
8.7
8.3
8.9
2.4
2.3
2.5
0.8
0.8
0.9
St. Louis
9.3
8.2
10.0
2.8
2.3
3.1
1.1
0.9
1.2
Children
with
Asthma
Atlanta
8.3
6.9
10.2
2.2
1.7
3.3
0.8
0.6
1.2
Boston
8.0
7.7
8.6
2.3
2.1
2.5
0.9
0.9
0.9
Dallas
9.6
8.1
10.5
3.1
2.4
3.5
1.1
0.8
1.4
Detroit
10.3
9.3
11.5
CO
CO
2.9
3.9
1.3
1.1
1.5
Philadelphia
9.3
8.6
9.7
2.5
2.3
2.6
0.9
0.7
1
Phoenix
14.9
13.7
16.0
4.9
4.4
5.3
2.1
1.8
2.5
Sacramento
8.9
8.4
9.3
2.5
2.2
2.8
0.8
0.5
1.2
St. Louis
9.4
8.5
9.9
2.9
2.5
3.1
1.1
0.9
1.4
Adults
Atlanta
1.2
1.0
1.4
0.3
0.2
0.4
0.1
0.1
0.1
Boston
1.0
0.9
1.1
0.3
0.2
0.3
0.1
0.1
0.1
Dallas
1.4
1.2
1.5
0.3
0.3
0.4
0.1
0.1
0.1
Detroit
1.2
1.1
1.3
0.3
0.3
0.4
0.1
0.1
0.1
Philadelphia
1.2
1.2
1.2
0.3
0.3
0.3
0.1
0.1
0.1
Phoenix
2.4
2.3
2.6
0.7
0.7
r—
o
0.3
0.2
0.3
Sacramento
1.2
1.2
1.2
O
CO
0.3
0.3
0.1
0.1
0.1
St. Louis
1.3
1.0
1.4
0.3
0.3
0.4
0.1
0.1
0.1
Adults
with
Asthma
Atlanta
1.1
0.9
1.2
0.2
0.2
0.3
0.1
<0.1
0.2
Boston
0.8
0.7
0.9
0.2
0.2
0.3
0.1
<0.1
0.1
Dallas
1.1
0.9
1.4
0.3
0.1
0.4
0.1
<0.1
0.3
Detroit
1.1
1.0
1.2
0.3
0.2
0.4
0.1
0.1
0.2
Philadelphia
1.0
0.9
1.2
0.2
0.1
0.3
0.1
0.1
0.1
Phoenix
2.0
1.8
2.1
0.6
0.5
0.7
0.2
0.2
0.2
Sacramento
1.0
0.9
1.0
0.2
0.2
0.3
0.1
0.1
0.1
St. Louis
1.1
0.9
1.3
0.3
0.2
0.4
0.1
<0.1
0.1
A Calculated percent is rounded to the nearest tenth decimal using conventional rounding. Values equal to zero are designated by "0" (there
are no individuals experiencing decrements at that level). Small, non-zero values that do not round upwards to 0.1 (i.e., <0.05) are given a
value of "<0.1".
3D-134
-------
Table 3D-55. Number of people estimated to experience at least two lung function
decrements at or above the indicated level, for air quality adjusted to just
meet the current standard, using the individual-based (MSS model) risk
approach.
Study
Group
Study Area
>10% reduction in FEViA
(# per Year)
>15% reduction in FEViA
(# per Year)
>20% reduction in FEViA
(# per Year)
Avg
Min
Max
Avg
Min
Max
Avg
Min
Max
Children
Atlanta
92853
80585
110689
25537
20338
32040
9308
7042
11904
Boston
100764
94590
107742
29058
25849
32471
10816
9079
12697
Dallas
124935
109975
134637
36832
28658
42869
13738
10475
16410
Detroit
97682
87982
107267
29946
26015
33819
11643
9729
13302
Philadelphia
113291
104546
118929
31574
29508
32913
11226
10324
11677
Phoenix
115472
108372
125399
36615
32468
41342
14766
12596
17210
Sacramento
40342
38712
41406
10991
10559
11848
3822
3540
4371
St. Louis
50799
44731
54612
15129
12704
17175
5795
4826
6775
Children
with
Asthma
Atlanta
11850
9725
14608
3208
2361
4681
1123
807
1715
Boston
13433
12788
14267
3846
3390
4210
1517
1434
1570
Dallas
12532
10570
13785
4051
3121
4587
1490
1017
1773
Detroit
12036
10649
13510
3850
3295
4544
1515
1214
1821
Philadelphia
13612
12572
14449
3660
3318
3841
1317
1091
1484
Phoenix
12091
11025
13049
4015
3552
4331
1703
1429
1996
Sacramento
4040
3773
4294
1121
994
1320
368
233
536
St. Louis
5270
4863
5509
1627
1421
1739
619
501
747
Adults
Atlanta
48670
43528
58037
12091
10354
15495
4226
3029
6198
Boston
59908
55473
63495
15426
12621
17513
5479
4305
6164
Dallas
63629
57118
68916
16096
13908
18675
5912
5391
6720
Detroit
48523
44634
52368
13130
11142
14681
4566
3867
5571
Philadelphia
61406
60128
62830
14146
13856
14640
5025
4706
5316
Phoenix
72697
68690
78624
21043
20513
21655
7665
7152
8642
Sacramento
21086
21038
21152
5203
5088
5260
1906
1887
1944
St. Louis
27183
21174
30545
6927
5437
7762
2528
1896
3040
Adults
with
Asthma
Atlanta
3804
3381
4367
845
634
986
305
141
563
Boston
5316
4598
5870
1370
978
1859
424
196
587
Dallas
3620
2891
4532
938
469
1485
416
156
859
Detroit
4719
4064
5047
1442
983
1704
546
328
852
Philadelphia
5170
4357
6013
1220
610
1569
378
261
436
Phoenix
5281
4818
5612
1639
1341
1788
596
497
646
Sacramento
1324
1258
1429
324
257
400
153
143
172
St. Louis
2313
1860
2611
632
501
787
179
72
286
A These values represent the population of individuals exposed in each study area. Values equal to zero are indicated by "0" (there are no
individuals experiencing decrements at the level).
3D-135
-------
Table 3D-56. Percent of people estimated to experience at least four lung function
decrements at or above the indicated level, for air quality adjusted to just
meet the current standard, using the individual-based (MSS model) risk
approach.
Study
Group
Study Area
>10% reduction in FEViA
(% per Year)
>15% reduction in FEViA
(% per Year)
>20% reduction in FEViA
(% per Year)
Avg
Min
Max
Avg
Min
Max
Avg
Min
Max
Children
Atlanta
4.3
3.6
5.2
1.0
0.8
1.3
0.3
0.2
0.4
Boston
3.9
3.7
4.0
1.0
0.9
1.1
0.3
0.2
0.3
Dallas
5.1
4.4
5.4
1.3
1.0
1.5
0.4
0.3
0.5
Detroit
5.2
4.7
5.7
1.4
1.2
1.5
0.5
0.4
0.5
Philadelphia
4.8
4.3
5.1
1.2
1.0
1.3
0.4
0.3
0.4
Phoenix
8.8
8.2
9.7
2.5
2.2
2.9
0.9
0.8
1.1
Sacramento
5.0
4.8
5.2
1.2
1.1
1.3
0.4
0.3
0.4
St. Louis
5.3
4.5
5.8
1.4
1.2
1.6
0.5
0.4
0.5
Children
with
Asthma
Atlanta
4.7
4.0
6.0
1.2
0.9
1.7
0.3
0.2
0.6
Boston
4.3
4.1
4.4
1.1
0.9
1.3
0.3
0.3
0.4
Dallas
5.7
4.9
6.2
1.4
1.1
1.7
0.5
0.3
0.6
Detroit
5.8
5.4
6.3
1.6
1.4
1.8
0.5
0.4
0.7
Philadelphia
5.1
4.9
5.6
1.3
1.2
1.5
0.4
0.3
0.5
Phoenix
9.8
9.2
10.5
2.9
2.6
3.3
1.1
0.9
1.3
Sacramento
5.1
4.9
5.3
1.2
1.1
1.5
0.5
0.3
0.7
St. Louis
5.4
4.8
5.7
1.5
1.3
1.8
0.5
0.4
0.6
Adults
Atlanta
0.5
0.4
0.6
0.1
0.1
0.2
<0.1
<0.1
0.1
Boston
0.4
0.4
0.5
0.1
0.1
0.1
<0.1
<0.1
<0.1
Dallas
0.6
0.5
0.7
0.1
0.1
0.1
<0.1
<0.1
0.1
Detroit
0.5
0.5
0.6
0.1
0.1
0.1
<0.1
<0.1
<0.1
Philadelphia
0.5
0.5
0.5
0.1
0.1
0.1
<0.1
<0.1
<0.1
Phoenix
1.3
1.2
1.4
O
CO
O
CO
0.4
0.1
0.1
0.1
Sacramento
0.6
0.6
0.6
0.1
0.1
0.1
<0.1
<0.1
<0.1
St. Louis
0.5
0.4
0.6
0.1
0.1
0.1
<0.1
<0.1
0.1
Adults
with
Asthma
Atlanta
0.5
0.4
0.6
0.1
<0.1
0.2
<0.1
0
0.1
Boston
0.3
0.2
0.4
0.1
<0.1
0.1
<0.1
0
<0.1
Dallas
0.5
0.3
0.8
0.2
0.1
0.3
<0.1
<0.1
0.1
Detroit
0.5
0.3
0.6
0.1
<0.1
0.2
0.1
<0.1
0.1
Philadelphia
0.5
0.3
0.6
0.1
<0.1
0.1
<0.1
0
0.1
Phoenix
1.0
0.8
1.0
0.3
0.2
O
CO
0.1
<0.1
0.1
Sacramento
0.5
0.4
0.5
0.1
0.1
0.1
<0.1
<0.1
0.1
St. Louis
0.5
0.4
0.5
0.1
0.1
0.1
<0.1
<0.1
0.1
A Calculated percent is rounded to the nearest tenth decimal using conventional rounding. Values equal to zero are designated by "0" (there
are no individuals experiencing decrements at that level). Small, non-zero values that do not round upwards to 0.1 (i.e., <0.05) are given a
value of "<0.1".
3D-136
-------
Table 3D-57. Number of people estimated to experience at least four lung function
decrements at or above the indicated level, for air quality adjusted to just
meet the current standard, using the individual-based (MSS model) risk
approach.
Study
Group
Study Area
>10% reduction in FEViA
(# per Year)
>15% reduction in FEViA
(# per Year)
>20% reduction in FEViA
(# per Year)
Avg
Min
Max
Avg
Min
Max
Avg
Min
Max
Children
Atlanta
51699
44106
63455
12355
9967
15536
3780
2764
5206
Boston
53018
51152
54929
13205
11832
14358
3921
3390
4323
Dallas
71709
62140
77108
18302
14282
21352
6101
4540
7094
Detroit
54058
48613
59001
14453
12574
16025
4839
3885
5584
Philadelphia
62298
56943
66547
15358
13576
16544
4780
4038
5282
Phoenix
74522
69479
81962
21334
18895
24698
7954
6893
9469
Sacramento
23088
22182
24045
5531
5311
5955
1703
1522
1988
St. Louis
28838
24579
31582
7625
6356
8587
2504
2158
2732
Children
with
Asthma
Atlanta
6699
5710
8636
1688
1291
2421
450
242
847
Boston
7190
6849
7372
1821
1525
2230
508
432
592
Dallas
7456
6432
8158
1852
1442
2128
702
426
851
Detroit
6810
6226
7423
1902
1596
2151
613
468
780
Philadelphia
7515
7093
8272
1906
1746
2183
597
458
786
Phoenix
7973
7445
8534
2359
2081
2703
887
750
1033
Sacramento
2290
2174
2453
561
481
683
207
140
303
St. Louis
2996
2741
3151
859
747
965
289
228
319
Adults
Atlanta
21999
18947
26553
4625
3029
7325
1432
634
2395
Boston
25307
22111
28079
5609
4403
6457
1859
1174
2544
Dallas
28181
23519
31801
5730
4923
6563
1823
1485
2422
Detroit
20689
17893
22743
4828
3801
5571
1486
1114
1704
Philadelphia
27218
26840
27450
5810
5403
6100
1743
1394
1917
Phoenix
38294
36555
41423
9553
8791
10828
3311
2732
4172
Sacramento
9862
9547
10119
2201
1972
2344
610
486
715
St. Louis
11708
9156
13520
2647
2182
3076
954
680
1109
Adults
with
Asthma
Atlanta
1714
1409
2113
376
141
704
164
0
423
Boston
2218
1468
2739
391
294
489
131
0
196
Dallas
1693
938
2657
521
313
938
156
78
234
Detroit
2141
1376
2687
546
197
852
240
131
328
Philadelphia
2382
1656
2789
436
174
610
116
0
349
Phoenix
2599
2285
2781
679
497
844
232
50
397
Sacramento
648
515
772
152
114
200
48
29
86
St. Louis
978
751
1109
191
107
250
72
36
143
A These values represent the population of individuals exposed in each study area. Values equal to zero are indicated by "0" (there are no
individuals experiencing decrements at the level).
3D-137
-------
Table 3D-58. Percent of people estimated to experience at least one lung function
decrement at or above the indicated level, for the 75 ppb air quality scenario,
using the individual-based (MSS model) risk approach.
Study
Group
Study Area
>10% reduction in FEViA
(% per Year)
>15% reduction in FEViA
(% per Year)
>20% reduction in FEViA
(% per Year)
Avg
Min
Max
Avg
Min
Max
Avg
Min
Max
Children
Atlanta
16.4
14.5
18.7
5.8
4.7
7.1
2.6
2.0
3.3
Boston
14.7
13.6
15.9
5.2
4.7
5.9
2.4
2.0
2.9
Dallas
16.7
14.9
18.2
6.0
5.0
6.8
2.7
2.1
3.2
Detroit
17.8
16.2
19.5
6.7
5.9
7.7
3.1
2.6
3.7
Philadelphia
17.5
16.6
18.1
6.2
5.9
6.4
2.8
2.7
2.9
Phoenix
23.6
22.4
25.1
9.0
8.1
9.9
4.2
3.7
4.8
Sacramento
17.2
16.6
17.7
6.0
5.7
6.3
2.7
2.5
2.9
St. Louis
17.8
16.2
18.8
6.6
5.7
7.5
3.0
2.5
3.6
Children
with
Asthma
Atlanta
17.6
15.3
20.2
6.3
4.8
8.2
2.8
2.2
3.8
Boston
15.6
14.2
16.8
5.6
4.9
6.4
2.6
2.1
3.2
Dallas
17.9
15.6
19.7
6.7
5.6
7.5
3.1
2.4
3.6
Detroit
19.1
17.5
20.9
7.6
6.7
8.8
3.5
2.9
4.3
Philadelphia
18.4
18.0
18.7
6.7
6.5
6.9
2.9
2.8
3.1
Phoenix
25.1
23.5
26.4
10.2
9.3
11.0
4.9
4.2
5.3
Sacramento
17.5
16.7
18.2
6.2
6.0
6.4
2.6
2.4
2.8
St. Louis
18.1
16.5
19.0
6.8
6.0
7.3
3.1
2.6
3.6
Adults
Atlanta
3.2
2.9
3.6
1.0
0.8
1.2
0.4
0.3
0.5
Boston
2.6
2.3
2.9
r—
o
0.6
0.8
0.3
0.3
0.4
Dallas
3.3
2.9
3.6
1.0
0.9
1.2
0.4
0.4
0.5
Detroit
3.1
2.8
3.3
1.0
0.8
1.1
0.4
0.3
0.5
Philadelphia
3.1
2.9
3.3
0.9
0.9
1.0
0.4
0.4
0.4
Phoenix
5.2
4.8
5.7
1.7
1.7
1.9
0.8
0.7
0.8
Sacramento
3.2
3.1
3.2
0.9
0.9
1.0
0.4
0.4
0.4
St. Louis
3.1
2.7
3.4
1.0
0.8
1.1
0.4
0.3
0.5
Adults
with
Asthma
Atlanta
2.9
2.7
3.3
0.9
0.8
1.0
0.4
0.3
0.4
Boston
2.2
1.9
2.5
0.6
0.4
f—
O
0.2
0.1
0.3
Dallas
3.0
2.5
3.2
0.9
r—
o
1.2
0.4
0.2
0.6
Detroit
2.9
2.5
3.1
0.9
r—
o
1.0
0.4
0.3
0.5
Philadelphia
2.8
2.7
3.1
0.9
0.8
0.9
0.3
0.1
0.4
Phoenix
4.1
3.7
4.4
1.4
1.3
1.5
0.7
0.6
0.7
Sacramento
2.5
2.5
2.6
0.8
r—
o
0.9
0.3
0.2
0.3
St. Louis
3.0
2.5
3.3
1.0
0.8
1.0
0.4
0.2
0.5
A Calculated percent is rounded to the nearest tenth decimal using conventional rounding. Values equal to zero are designated by "0" (there
are no individuals experiencing decrements at that level). Small, non-zero values that do not round upwards to 0.1 (i.e., <0.05) are given a
value of "<0.1".
3D-138
-------
Table 3D-59. Percent of people estimated to experience at least two lung function
decrements at or above the indicated level, for the 75 ppb air quality scenario,
using the individual-based (MSS model) risk approach.
Study
Group
Study Area
>10% reduction in FEViA
(% per Year)
>15% reduction in FEViA
(% per Year)
>20% reduction in FEViA
(% per Year)
Avg
Min
Max
Avg
Min
Max
Avg
Min
Max
Children
Atlanta
10.0
8.7
11.8
3.1
2.5
3.9
1.3
1.0
1.6
Boston
8.4
7.8
9.0
2.6
2.3
2.9
1.0
0.8
1.2
Dallas
10.3
9.0
11.4
3.3
2.6
4.0
1.3
1.0
1.7
Detroit
10.9
9.6
12.2
3.7
3.1
4.2
1.5
1.2
1.8
Philadelphia
10.8
9.9
11.3
3.4
3.2
3.5
1.4
1.2
1.4
Phoenix
16.1
15.2
17.5
5.6
5.0
6.3
2.5
2.1
2.9
Sacramento
10.8
10.4
11.2
3.4
3.2
3.6
1.3
1.2
1.5
St. Louis
11.1
9.7
11.9
3.6
3.0
4.1
1.5
1.3
1.8
Children
with
Asthma
Atlanta
10.8
9.2
13.1
3.4
2.6
4.6
1.3
1.0
1.9
Boston
9.0
8.4
9.9
2.8
2.5
3.1
1.1
1.0
1.2
Dallas
11.1
9.4
12.3
3.7
3.0
4.3
1.5
1.1
1.8
Detroit
11.8
10.5
13.2
4.1
3.5
4.7
1.8
1.3
2.3
Philadelphia
11.5
10.9
11.9
3.5
3.3
3.6
1.4
1.3
1.5
Phoenix
17.4
16.1
18.8
6.4
5.9
6.8
2.9
2.4
3.4
Sacramento
11.1
10.4
11.7
3.5
3.4
3.8
1.3
1.1
1.6
St. Louis
11.2
9.9
11.9
3.8
3.4
4.2
1.7
1.4
2.0
Adults
Atlanta
1.5
1.4
1.8
0.4
0.3
0.5
0.2
0.1
0.2
Boston
1.1
1.0
1.2
0.3
0.3
0.3
0.1
0.1
0.1
Dallas
1.6
1.4
1.7
0.4
0.4
0.5
0.2
0.1
0.2
Detroit
1.4
1.3
1.6
0.4
O
CO
0.5
0.2
0.1
0.2
Philadelphia
1.5
1.4
1.5
0.4
0.4
0.4
0.1
0.1
0.2
Phoenix
2.9
2.7
3.1
0.9
0.9
1.0
0.4
0.3
0.4
Sacramento
1.6
1.6
1.6
0.4
0.4
0.4
0.2
0.2
0.2
St. Louis
1.5
1.2
1.7
0.4
0.3
0.5
0.2
0.1
0.2
Adults
with
Asthma
Atlanta
1.3
1.2
1.5
0.4
0.3
0.4
0.1
0.1
0.2
Boston
0.9
0.8
1.1
0.2
0.2
0.3
0.1
<0.1
0.1
Dallas
1.3
1.1
1.6
0.3
0.2
0.5
0.2
0.1
0.3
Detroit
1.3
1.1
1.4
0.4
0.3
0.5
0.2
0.1
0.3
Philadelphia
1.3
1.1
1.4
0.3
0.3
0.4
0.1
0.1
0.2
Phoenix
2.3
2.1
2.5
0.7
0.6
0.8
0.3
0.2
0.4
Sacramento
1.2
1.1
1.3
0.3
0.3
0.4
0.1
0.1
0.2
St. Louis
1.4
1.1
1.6
0.4
0.3
0.5
0.1
0.1
0.2
A Calculated percent is rounded to the nearest tenth decimal using conventional rounding. Values equal to zero are designated by "0" (there
are no individuals experiencing decrements at that level). Small, non-zero values that do not round upwards to 0.1 (i.e., <0.05) are given a
value of "<0.1".
3D-139
-------
Table 3D-60. Percent of people estimated to experience at least four lung function
decrements at or above the indicated level, for the 75 ppb air quality scenario,
using the individual-based (MSS model) risk approach.
Study
Group
Study Area
>10% reduction in FEViA
(% per Year)
>15% reduction in FEViA
(% per Year)
>20% reduction in FEViA
(% per Year)
Avg
Min
Max
Avg
Min
Max
Avg
Min
Max
Children
Atlanta
5.8
4.9
7.1
1.6
1.3
2.1
0.6
0.4
0.7
Boston
4.5
4.2
4.7
1.2
1.1
1.3
0.4
0.3
0.4
Dallas
6.0
5.1
6.6
1.7
1.3
2
0.6
0.4
0.7
Detroit
6.2
5.4
6.9
1.8
1.6
2.1
0.7
0.5
0.8
Philadelphia
6.2
5.6
6.7
1.7
1.5
1.9
0.6
0.5
0.7
Phoenix
10.6
9.8
11.7
3.4
3.1
3.9
1.4
1.2
1.7
Sacramento
6.4
6.1
6.7
1.8
1.7
1.9
0.6
0.6
0.7
St. Louis
6.5
5.6
7.0
1.9
1.6
2.1
0.7
0.6
0.8
Children
with
Asthma
Atlanta
6.4
5.3
8.0
1.8
1.4
2.6
0.6
0.4
1.0
Boston
5.0
4.7
5.2
1.3
1.2
1.6
0.4
0.3
0.5
Dallas
6.6
5.5
7.3
2.1
1.6
2.4
0.7
0.5
0.9
Detroit
6.9
6.2
7.7
2.1
1.7
2.5
0.8
0.5
0.9
Philadelphia
6.7
6.4
7.3
1.8
1.7
2.0
0.7
0.6
0.7
Phoenix
11.6
10.7
12.6
4.0
3.6
4.5
1.6
1.4
2.0
Sacramento
6.6
6.3
7.0
1.8
1.5
2.1
0.6
0.5
0.9
St. Louis
6.4
5.8
6.8
2.1
1.8
2.4
0.7
0.6
0.9
Adults
Atlanta
0.7
0.6
0.9
0.2
0.1
0.2
0.1
<0.1
0.1
Boston
0.5
0.4
0.5
0.1
0.1
0.1
<0.1
<0.1
0.1
Dallas
0.7
0.6
0.8
0.2
0.1
0.2
0.1
<0.1
0.1
Detroit
0.6
0.5
0.7
0.1
0.1
0.2
0.1
<0.1
0.1
Philadelphia
0.7
0.7
0.7
0.2
0.1
0.2
<0.1
<0.1
0.1
Phoenix
1.6
1.5
1.7
0.4
0.4
0.4
0.2
0.1
0.2
Sacramento
0.7
0.7
0.7
0.2
0.2
0.2
0.1
0.1
0.1
St. Louis
0.7
0.5
0.8
0.2
0.1
0.2
0.1
<0.1
0.1
Adults
with
Asthma
Atlanta
0.7
0.6
0.8
0.1
0.1
0.3
0.1
<0.1
0.1
Boston
0.4
0.2
0.4
0.1
0.1
0.1
<0.1
0
<0.1
Dallas
0.6
0.3
0.9
0.2
0.1
0.4
0.1
<0.1
0.1
Detroit
0.5
0.4
0.7
0.2
0.1
0.2
0.1
<0.1
0.1
Philadelphia
0.6
0.5
0.7
0.1
0.1
0.2
<0.1
<0.1
0.1
Phoenix
1.2
1.1
1.2
0.4
0.3
0.4
0.1
<0.1
0.2
Sacramento
0.6
0.5
0.7
0.2
0.1
0.3
0.1
0.1
0.1
St. Louis
0.6
0.4
0.7
0.1
0.1
0.1
0.1
<0.1
0.1
A Calculated percent is rounded to the nearest tenth decimal using conventional rounding. Values equal to zero are designated by "0" (there
are no individuals experiencing decrements at that level). Small, non-zero values that do not round upwards to 0.1 (i.e., <0.05) are given a
value of "<0.1".
3D-140
-------
Table 3D-61. Percent of people estimated to experience at least one lung function
decrement at or above the indicated level, for the 65 ppb air quality scenario,
using the individual-based (MSS model) risk approach.
Study
Group
Study Area
>10% reduction in FEViA
(% per Year)
>15% reduction in FEViA
(% per Year)
>20% reduction in FEViA
(% per Year)
Avg
Min
Max
Avg
Min
Max
Avg
Min
Max
Children
Atlanta
10.3
9.2
11.7
2.8
2.3
3.4
1.0
0.8
1.3
Boston
10.8
10.4
11.4
3.3
3.1
3.8
1.3
1.2
1.6
Dallas
12.4
11.2
13.0
3.8
3.1
4.1
1.5
1.2
1.6
Detroit
12.9
12.0
13.9
4.0
3.6
4.5
1.6
1.4
1.8
Philadelphia
12.1
11.5
12.6
3.4
3.3
3.6
1.3
1.2
1.3
Phoenix
16.9
16.0
18.1
5.2
4.7
5.7
2.1
1.8
2.4
Sacramento
10.8
10.5
11.1
2.8
2.8
3.0
1.0
0.9
1.1
St. Louis
12.3
11.2
13.2
3.6
3.1
4.2
1.4
1.2
1.7
Children
with
Asthma
Atlanta
11.2
9.6
13.2
3.1
2.5
4.0
1.1
0.8
1.6
Boston
11.6
11.2
12.0
3.6
3.3
4.2
1.4
1.3
1.5
Dallas
13.5
11.8
14.3
4.3
3.6
4.6
1.8
1.3
2.1
Detroit
14.0
12.9
15.3
4.6
4.3
5.2
1.8
1.5
2.2
Philadelphia
12.8
12.6
13.0
3.5
3.3
3.7
1.4
1.3
1.6
Phoenix
18.5
17.1
19.7
6.0
5.4
6.4
2.4
2.1
2.9
Sacramento
11.2
11.0
11.3
2.9
2.8
3.0
1.0
0.7
1.3
St. Louis
12.7
11.8
13.2
3.8
3.2
4^
CO
1.5
1.3
1.8
Adults
Atlanta
1.9
1.8
2.1
0.5
0.4
0.6
0.2
0.1
0.2
Boston
1.9
1.8
2.1
0.5
0.4
0.5
0.2
0.2
0.2
Dallas
2.4
2.2
2.5
0.6
0.6
0.7
0.2
0.2
0.3
Detroit
2.1
2.0
2.3
0.6
0.5
0.6
0.2
0.2
0.2
Philadelphia
2.1
2.0
2.1
0.5
0.5
0.5
0.2
0.2
0.2
Phoenix
3.6
3.3
3.9
1.0
1.0
1.0
0.4
0.4
0.4
Sacramento
1.9
1.9
1.9
0.4
0.4
0.5
0.2
0.2
0.2
St. Louis
2.1
1.8
2.3
0.5
0.4
0.6
0.2
0.2
0.2
Adults
with
Asthma
Atlanta
1.7
1.5
1.8
0.5
0.4
0.5
0.1
0.1
0.2
Boston
1.6
1.5
1.9
0.4
0.3
0.5
0.2
0.1
0.2
Dallas
2.0
1.7
2.5
0.6
0.4
0.8
0.3
0.1
0.5
Detroit
2.0
1.8
2.1
0.5
0.4
0.6
0.3
0.2
0.3
Philadelphia
1.9
1.7
2.0
0.4
0.3
0.5
0.2
0.1
0.2
Phoenix
2.9
2.5
3.1
0.8
0.8
0.9
0.3
0.2
0.3
Sacramento
1.6
1.5
1.7
0.4
0.3
0.4
0.1
0.1
0.2
St. Louis
2.0
1.8
2.2
0.5
0.4
0.6
0.2
0.1
0.3
A Calculated percent is rounded to the nearest tenth decimal using conventional rounding. Values equal to zero are designated by "0" (there
are no individuals experiencing decrements at that level). Small, non-zero values that do not round upwards to 0.1 (i.e., <0.05) are given a
value of "<0.1".
3D-141
-------
Table 3D-62. Percent of people estimated to experience at least two lung function
decrements at or above the indicated level, for the 65 ppb air quality scenario,
using the individual-based (MSS model) risk approach.
Study
Group
Study Area
>10% reduction in FEViA
(% per Year)
>15% reduction in FEViA
(% per Year)
>20% reduction in FEViA
(% per Year)
Avg
Min
Max
Avg
Min
Max
Avg
Min
Max
Children
Atlanta
5.7
5.0
6.8
1.4
1.1
1.6
0.4
0.3
0.5
Boston
5.9
5.7
6.1
1.5
1.4
1.7
0.5
0.4
0.6
Dallas
7.3
6.5
7.7
1.9
1.5
2.2
0.7
0.5
0.7
Detroit
7.4
6.8
8.0
2.0
1.8
2.2
0.7
0.6
0.8
Philadelphia
7.0
6.5
7.3
1.7
1.6
1.8
0.6
0.5
0.6
Phoenix
11
10.2
12.0
3.1
2.8
3.5
1.1
1.0
1.3
Sacramento
6.2
6.0
6.5
1.4
1.3
1.5
0.4
0.4
0.5
St. Louis
7.1
6.2
7.7
1.8
1.6
2.1
0.6
0.5
0.7
Children
with
Asthma
Atlanta
6.2
5.3
7.5
1.4
1.1
2.0
0.4
0.3
0.7
Boston
6.3
6.0
6.8
1.6
1.4
1.8
0.6
0.6
0.6
Dallas
8.1
6.9
9.0
2.3
1.9
2.6
0.8
0.5
1.0
Detroit
8.3
7.6
9.0
2.3
2.0
2.8
0.9
0.7
1.0
Philadelphia
7.6
7.1
8.1
1.8
1.7
1.9
0.6
0.6
0.7
Phoenix
12.1
11.2
12.9
3.7
3.4
4.0
1.3
1.1
1.5
Sacramento
6.4
6.1
6.6
1.4
1.2
1.6
0.4
0.3
0.6
St. Louis
7.3
6.7
7.6
2.0
1.7
2.2
0.7
0.6
0.7
Adults
Atlanta
0.8
0.8
1.0
0.2
0.1
0.3
0.1
<0.1
0.1
Boston
0.8
0.8
0.9
0.2
0.2
0.2
0.1
0.1
0.1
Dallas
1.1
1.0
1.2
O
CO
0.2
0.3
0.1
0.1
0.1
Detroit
1.0
0.9
1.0
0.2
0.2
O
CO
0.1
0.1
0.1
Philadelphia
0.9
0.9
1.0
0.2
0.2
0.2
0.1
0.1
0.1
Phoenix
1.9
1.9
2.1
0.5
0.4
0.5
0.2
0.1
0.2
Sacramento
0.9
0.9
0.9
0.2
0.2
0.2
0.1
0.1
0.1
St. Louis
1.0
0.8
1.1
0.2
0.2
0.2
0.1
0.1
0.1
Adults
with
Asthma
Atlanta
0.7
0.6
0.9
0.2
0.1
0.3
0.1
<0.1
0.1
Boston
0.7
0.6
0.8
0.2
0.1
0.2
0.1
<0.1
0.1
Dallas
0.8
0.7
1.1
0.2
0.1
0.4
0.1
0
0.1
Detroit
0.9
0.7
1.0
0.3
0.2
0.3
0.1
0.1
0.2
Philadelphia
0.8
0.6
0.9
0.2
0.1
0.3
<0.1
<0.1
0.1
Phoenix
1.5
1.5
1.6
0.4
0.3
0.5
0.1
0.1
0.2
Sacramento
0.7
0.6
0.8
0.2
0.1
0.2
<0.1
<0.1
0.1
St. Louis
0.9
0.7
1.0
0.2
0.1
0.3
<0.1
<0.1
0.1
A Calculated percent is rounded to the nearest tenth decimal using conventional rounding. Values equal to zero are designated by "0" (there
are no individuals experiencing decrements at that level). Small, non-zero values that do not round upwards to 0.1 (i.e., <0.05) are given a
value of "<0.1".
3D-142
-------
Table 3D-63. Percent of people estimated to experience at least four lung function
decrements at or above the indicated level, for the 65 ppb air quality scenario,
using the individual-based (MSS model) risk approach.
Study
Group
Study Area
>10% reduction in FEViA
(% per Year)
>15% reduction in FEViA
(% per Year)
>20% reduction in FEViA
(% per Year)
Avg
Min
Max
Avg
Min
Max
Avg
Min
Max
Children
Atlanta
3.0
2.6
3.7
0.6
0.5
0.8
0.2
0.1
0.2
Boston
3.0
2.9
3.0
0.6
0.6
r—
o
0.2
0.2
0.2
Dallas
4.0
3.6
4.3
0.9
0.7
1.0
0.3
0.2
0.3
Detroit
4.0
3.6
4.3
0.9
0.8
1.0
0.3
0.2
0.3
Philadelphia
3.7
3.5
4.0
0.8
r—
o
0.9
0.2
0.2
0.2
Phoenix
6.8
6.4
7.5
1.7
1.5
2.0
0.5
0.5
0.6
Sacramento
3.4
3.3
3.5
0.6
0.6
0.7
0.2
0.1
0.2
St. Louis
3.9
3.3
4.2
0.9
0.7
1.0
0.2
0.2
0.3
Children
with
Asthma
Atlanta
3.2
2.7
4.3
0.7
0.4
1.1
0.2
0.1
0.2
Boston
3.4
3.3
3.5
0.8
0.6
0.9
0.2
0.2
0.2
Dallas
4.6
4.0
4.9
1.1
0.8
1.3
0.3
0.2
0.4
Detroit
4.5
4.2
4.7
1.1
0.9
1.1
0.3
0.2
0.3
Philadelphia
3.9
3.6
4.1
0.9
0.8
1.0
0.2
0.2
0.3
Phoenix
7.6
7.2
8.2
2.1
1.9
2.3
0.7
0.5
0.9
Sacramento
3.6
3.3
3.7
0.7
0.6
0.9
0.2
0.1
0.3
St. Louis
4.2
3.8
4.4
1.0
0.8
1.1
0.3
0.2
0.3
Adults
Atlanta
0.4
0.3
0.4
0.1
<0.1
0.1
<0.1
<0.1
<0.1
Boston
0.3
0.3
0.4
0.1
<0.1
0.1
<0.1
<0.1
<0.1
Dallas
0.5
0.4
0.5
0.1
0.1
0.1
<0.1
<0.1
<0.1
Detroit
0.4
0.4
0.4
0.1
0.1
0.1
<0.1
<0.1
<0.1
Philadelphia
0.4
0.4
0.4
0.1
0.1
0.1
<0.1
<0.1
<0.1
Phoenix
1.0
1.0
1.0
0.2
0.2
0.3
0.1
0.1
0.1
Sacramento
0.4
0.4
0.4
0.1
0.1
0.1
<0.1
<0.1
<0.1
St. Louis
0.4
0.3
0.4
0.1
0.1
0.1
<0.1
<0.1
<0.1
Adults
with
Asthma
Atlanta
0.3
0.2
0.4
0.1
<0.1
0.2
<0.1
0
0.1
Boston
0.3
0.2
0.3
<0.1
<0.1
<0.1
<0.1
0
<0.1
Dallas
0.4
0.2
0.6
0.1
<0.1
0.3
<0.1
0
<0.1
Detroit
0.4
0.3
0.5
0.1
<0.1
0.2
<0.1
<0.1
0.1
Philadelphia
0.4
0.3
0.4
<0.1
0
0.1
<0.1
0
<0.1
Phoenix
0.7
0.6
0.8
0.2
0.1
0.3
0.1
0
0.1
Sacramento
0.3
0.3
0.4
0.1
<0.1
0.1
<0.1
0
<0.1
St. Louis
0.3
0.3
0.4
0.1
<0.1
0.1
<0.1
0
<0.1
A Calculated percent is rounded to the nearest tenth decimal using conventional rounding. Values equal to zero are designated by "0" (there
are no individuals experiencing decrements at that level). Small, non-zero values that do not round upwards to 0.1 (i.e., <0.05) are given a
value of "<0.1".
3D-143
-------
3D.3.4 Uncertainty Characterization
While it may be possible to estimate a range of O3 exposures or risks by accounting for
variability inherent to influential factors, the true exposure or risk for any given individual
residing within a study area is unknown. To characterize health risks, risk assessors commonly
use an iterative process of gathering data, developing models, and estimating exposures and
risks, which is based upon 1) the goals of the assessment, 2) evaluating results for correctness
and identifying areas for potential improvement, 3) scale and complexity of the assessment
performed, and 4) availability and limitations of the input data and information. Uncertainty can
still remain following each iteration and emphasis is then placed on characterizing the nature and
magnitude of that uncertainty and its impact on exposure and risk estimates. A summary of the
overall characterization of uncertainty for the current O3 exposure and risk analysis is provided
in section 3D.3.4.1. The summary is followed by APEX sensitivity analyses in section 3D.3.4.2
that provide additional support to the uncertainty characterization regarding the influence a
number of factors (e.g., contribution of low exposures) have on estimating lung function risk
resulting from O3 exposure.
3D.3.4.1 Summary of the Uncertainty Characterization
The REAs for the previous reviews of the O3, NO2, SO2, and CO NAAQS characterized
uncertainty in exposure and risk modeling (Langstaff, 2007; U.S. EPA, 2008, 2009, 2010, 2014,
2018). The mainly qualitative approach used in this and other REAs, also informed by
quantitative sensitivity analyses, is described by WHO (2008). Briefly, we identified key aspects
of the assessment approach that may contribute to uncertainty in the exposure and risk estimates
and provided the rationale for their inclusion. Then, we characterized the magnitude and
direction of the influence on the assessment for each of these identified sources of uncertainty.
Consistent with the WHO (2008) guidance, we scaled the overall impact of the
uncertainty by considering the degree of uncertainty as implied by the relationship between the
source of uncertainty and the exposure and risk estimates. A qualitative characterization of low,
moderate, and high was assigned to the magnitude of influence and knowledge base uncertainty
descriptors, using quantitative observations relating to understanding the uncertainty, where
possible. Where the magnitude of uncertainty was rated low, it was judged that large changes
within the source of uncertainty would have only a small effect on the assessment results (e.g.,
an impact of few percentage points upwards to a factor of two). A designation of medium implies
that a change within the source of uncertainty would likely have a moderate (or proportional)
effect on the results (e.g., a factor of two or more). A characterization of high implies that a
change in the source would have a large effect on results (e.g., an order of magnitude). We also
included the direction of influence, whether the source of uncertainty was judged to potentially
3D-144
-------
over-estimate ("over"), under-estimate ("wwder"), or have an unknown impact to exposure/risk
estimates. A summary of the key findings of the prior uncertainty characterizations that are most
relevant to the current O3 exposure and risk analysis are also provided in Table 3D-64.
3D-145
-------
Table 3D-64. Characterization of key uncertainties in exposure and risk analyses using APEX.
Sources of Uncertainty
Uncertainty Characterization
Newly
Characterized/
New
Information?
Influence of Uncertainty on
Exposure | Risk Estimates
Knowledge-
base
Uncertainty
Comments
Category
Element
Direction
Magnitude
General Aspects of
Assessment Design
Representation of
population groups
with asthma
Unknown
Low - Medium
Medium
Consistent with the ISA identification of people with asthma (and children with asthma in
particular) as an important at-risk population for O3 in ambient air, risk estimates are
developed for people with asthma and are reported separately for children and adults.
Exposure and risk were not estimated for more targeted population groups with asthma
based on additional personal attributes associated with increased asthma prevalence (e.g.,
obesity or African American or Hispanic ethnicity) generally due to limitations in the data
needed to simulate these population groups. Such data limitations affect our ability to
characterize O3 exposure and associated health risks for different population subgroups of
children and adults with asthma, some of which may have higher exposure/risk and others
lower.
Yes. Newly
identified element
of uncertainty
Representation of
population groups
having health
conditions other than
asthma
Unknown
Unknown
Medium
Individuals having health conditions other than asthma have not been explicitly represented
in this exposure and risk assessment as the evidence has not indicated any other
population groups with a health condition that places them at increased risk (ISA, Table IS-
11). Additionally, exposure/risk modeling for other such groups is hampered by data
limitations in accurately defining the size of a particular population group, assigning
appropriate activity pattern data, and estimating how responses observed in the controlled
human exposure study data would quantitatively relate to simulated individuals. For
example, the likelihood of individuals having a health condition such as chronic obstructive
pulmonary disorder exercising for sufficient duration at a ventilation rate needed (e.g., EVR
>17.32 ± 1.25 L/min-m2) to receive a dose that would elicit a response is unknown.
Yes. Newly
identified element
of uncertainty
Representation of
older adults
Neither
Low
Low
In the current exposure and risk analysis, older adults (ages 65-95) were simulated as part
of the all adult groups (ages 18-90) and not as a separate population subgroup. In the 2014
HREA, exposures and risks were estimated separately for older adults (2014 HREA,
section 5.6). In those 2014 HREA results, the percent of older adults experiencing
exposures at or above any benchmark tended to be lower than the comparable percentage
for all adults or all adults with asthma, by a few percentage points or less. A similar pattern
would be expected if this group were to have been included in the current analysis.
Yes. Newly
identified element
of uncertainty
3D-146
-------
Sources of Uncertainty
Uncertainty Characterization
Newly
Characterized/
New
Information?
Influence of Uncertainty on
Exposure | Risk Estimates
Knowledge-
base
Uncertainty
Comments
Category
Element
Direction
Magnitude
Representation of
outdoor workers
Under
Low - Medium
Medium
In the current exposure and risk analysis, outdoor workers were not evaluated as a
separate population subgroup. In the 2014 HREA, limited analyses were conducted for this
subgroup because of appreciable data limitations and associated uncertainty. The
exposures and risk estimates for this subgroup for the single study area and air quality
scenario assessed indicated a greater percentage of outdoor workers experience single
and multi-day exposures than that estimated for the full adult population, differing by about
a factor of 5 or more depending on the benchmark level and number of days per year
(2014 HREA, section 5.4.3.2). These limited results suggest that results for the full adult
population would likely underestimate exposures and risks for outdoor workers. Important
uncertainties exist in generating the simulated activity patterns for this group, including the
limited number of CHAD diary days available for outdoor workers, assignment of diaries to
proper occupation categories, approximating number of days/week and hours/day
outdoors, etc.
Yes. Newly
identified element
of uncertainty
Ambient Air Monitor
Concentrations
Ambient Air O3
Measurements
Both
Low
Low
Ozone measurements are assumed to be accurate to within % of the instrument's Method
Detection Limit (MDL), which is 2.5 ppb for most instruments. The EPA requires that
routine quality assurance checks are performed on all instruments. There is a known
tendency for smoke produced from wildfires to cause interference in O3 instruments.
Measurements collected by O3 analyzers were reported to be biased high by 5.1-6.6 ppb
per 100 |jg/m3 of PM2.5from wildfire smoke (U.S. EPA, 2007b). However, smoke
concentrations high enough to cause significant interferences are infrequent and the overall
impact is expected to be minimal.
No
Air Quality System
(AQS) Database
Quality
Both
Low
Low
All ambient air pollutant measurements available from AQS are certified by both the
monitoring agency and the corresponding EPA regional office. Monitor malfunctions
sometimes occur causing periods of missing data or poor data quality. Monitoring data
affected by malfunctions are usually flagged by the monitoring agency and removed from
AQS. In addition, the AQS database managers run several routines to identify suspicious
data for potential removal.
Yes. Recent year
data used (2015 -
2017)
Temporal
Representation
Both
Low
Low
The temporal scale (hourly) is appropriate for analysis performed. Required O3 monitoring
seasons are used to define the duration of the exposure and risk analyses in each study
area. Monitor data are screened for temporal completeness and considered appropriate
when calculating design values (and used for adjustments needed to meet air quality
scenarios). While some monitoring data used in developing the air quality surface were not
screened for temporal completeness, the inclusion of monitor data somewhat less than
complete is considered a holistic approach that improves the filling of both temporal and
spatial gaps that exist, where present.
No
3D-147
-------
Sources of Uncertainty
Uncertainty Characterization
Newly
Characterized/
New
Information?
Influence of Uncertainty on
Exposure | Risk Estimates
Knowledge-
base
Uncertainty
Comments
Category
Element
Direction
Magnitude
Spatial
Representation
Both
Low - Medium
Low - Medium
Overall, the eight study areas have reasonably dense ambient monitoring networks but
vary in size and geographic location. They are however considered adequate to capture
spatial gradients in O3 concentrations that occur in urban areas.
No
Adjusted O3
Concentrations for Air
Quality Scenarios
Modeled
atmospheric state
(CAMx)
Both
Medium
Medium
In the rollback adjustment framework applied in this assessment, the CAMx air quality
model is used to calculate the chemical state of the atmosphere so that the Higher
Decoupled Direct Method (HDDM) tool can archive O3 responsiveness to precursors at all
times and locations within the model domain. Model predictions from CAMx, like all
deterministic photochemical models, have both parametric and structural uncertainty
associated with them. CAMx is regularly updated to include state-of-the-science
parameterizations and processes relevant for atmospheric chemistry and physics. CAMx
model performance is also routinely evaluated against available observational datasets
(PA, Appendix 3C).
Yes. Recent year
meteorology and
emissions inputs
used (2016)
Ozone response
sensitivities (HDDM)
Both
Medium
Medium
The HDDM approach allows for the approximation of O3 response to alternate emissions
scenarios without re-running the model simulation multiple times using different emissions
inputs. This approximation becomes less accurate for larger emissions perturbations
especially under nonlinear chemistry conditions. However, even at 90% NOx cut
conditions, mean error in predicted O3 using HDDM sensitivities was within 2 ppb across all
urban study areas compared to the brute force simulation (PA, Appendix 3C).
Yes. Recent year
sensitivities used
(from 2016
simulation)
Voronoi Neighbor
Averaging (VNA)
spatial interpolation
Both
Low - Medium
Low - Medium
The VNA estimates are weighted based on distance from neighboring monitoring sites,
thus the amount of uncertainty tends to increase with distance from the monitoring sites.
Areas having a relatively less dense monitoring network (e.g., Atlanta, St. Louis) may have
greater uncertainty in the air quality surface than areas with a denser network (e.g., Boston,
Philadelphia).
No
APEX: General Input
Databases
Population
Demographics and
Commuting
Both
Low
Low
The U.S. Census data are comprehensive and subject to quality control. Differences in
2010 population data versus modeled years (2015-2017) are likely small when estimating
percent of population exposed. While population counts in most areas have likely increased
(and thus total number exposed and at risk is likely underestimated), it is likely that there
have not been substantive changes to the demographic distributions and commuting
patterns in each study area, thus having minimal impact to the percent of the population
exposed or experiencing lung function decrements.
Yes. Most recent
year data used
(2010)
3D-148
-------
Sources of Uncertainty
Uncertainty Characterization
Newly
Characterized/
New
Information?
Influence of Uncertainty on
Exposure | Risk Estimates
Knowledge-
base
Uncertainty
Comments
Category
Element
Direction
Magnitude
Activity Patterns
(CHAD)
Both
Low - Medium
Low - Medium
The CHAD data are comprehensive and subject to quality control. The current version of
CHAD contains an increased number of diaries used to estimate exposure from 2014
HREA. Previously, we evaluated trends and patterns in historical activity pattern data - no
major issues noted with use of historical data to represent current patterns (2014 HREA,
Appendix G, Figures 5G-1 and 5G-2). Compared outdoor event participation and outdoor
time of the larger American Time Use Survey (ATUS) data with all other survey data.
Participation rate in outdoor events by ATUS is lower, likely due to ATUS survey methods
(i.e., a lack of distinction of time spent inside or outside of residence). This finding would
primarily apply to adults (ATUS subjects are ages 16 and older). Comparison of activity
data (outdoor events and exertion level) for people with asthma generally similar to
individuals without asthma (section 3D.2.5.3, Table 3D-10) (see also 2014 HREA,
Appendix G, Tables 5G2 to 5G-5). There is little indication of differences in time spent
outdoors comparing activity patterns across U.S. regions, though sample size may be a
limiting factor in drawing significant conclusions (2014 HREA, section 5.4.1.6). Remaining
uncertainty exists for other influential factors that cannot be accounted for (e.g., SES,
region/local participation in outdoor events and associated amount of time).
Yes. New data
added to CHAD
(ATUS 2003-
2013) (U.S. EPA,
2019c,
Attachment 3)
Meteorological Data
Both
Low
Low
The NOAAISH data are comprehensive and subject to quality control, having very few
missing values. Limited use in selecting CHAD diaries for simulated individuals and AERs
that may vary with temperature. However, while using three years of varying meteorological
conditions, the 2015-2017 MET data set may not reflect the full suite of conditions that
could exist in future hypothetical air quality scenarios or across periods greater than 3-
years.
Yes. Recent year
data used (2015-
2017) (section
3D.2.4)
3D-149
-------
Sources of Uncertainty
Uncertainty Characterization
Newly
Characterized/
New
Information?
Influence of Uncertainty on
Exposure | Risk Estimates
Knowledge-
base
Uncertainty
Comments
Category
Element
Direction
Magnitude
Asthma Prevalence:
Selection of "Still"
Rather than "Ever"
Questionnaire
Response
Under
Low
Low
One of the two datasets used to estimate asthma prevalence is 2013-2017 NHIS data. The
NHIS dataset includes several categories describing whether a surveyed individual has
asthma based on a series of questions (Attachment 1). The first question inquires whether
a doctor has "Ever" told the individual they have asthma. This is followed by a question as
to whether they "Still" have asthma. In all instances, those responding "Yes" to the "Still"
question is a subset of those responding "Yes" to the "Ever" question. For estimating
asthma prevalence for the simulated populations, we focused on the dataset for those
answering they "Still" have asthma, consistent with the characterization of asthma
prevalence in the ISA (ISA, Table IS-11), and concluding that this approach would provide
us with the most appropriate estimate of the population of individuals that have asthma and
accordingly (based on the at-risk status of this population group) would likely be at
increased risk of response to O3 exposures. To the extent that some individuals answering
"No" to the "Still" question are at increased risk, our approach would underestimate the at-
risk group. The answers to subsequent questions in the NHIS dataset (regarding whether
the respondent had an asthma attack or asthma-related ER visit during past year) indicate
that the extent to which focusing on "Still" have asthma may underestimate this population
group is likely small. This conclusion is based on the findings that nearly 95% of those
answering "No" to the "Still" have question also did not have an asthma attack or asthma-
related ER visit in the past year, while nearly half of those answering "Yes" did have such
an experience, as well as the fact that nearly 95% of the survey respondents that indicated
they had had an asthma attack or asthma-related ER visit over the past year are captured
by the "Still" have asthma category (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.
Yes. Newly
identified element
of uncertainty
3D-150
-------
Sources of Uncertainty
Uncertainty Characterization
Newly
Characterized/
New
Information?
Influence of Uncertainty on
Exposure | Risk Estimates
Knowledge-
base
Uncertainty
Comments
Category
Element
Direction
Magnitude
Asthma Prevalence:
Weighted by Family
Income
Both
Low
Low - Medium
Data used are from peer-reviewed quality-controlled sources. Use of these data accounts
for variability in important influential variables (poverty status, as well as age, sex, and
region). Regional prevalence from NHIS were adjusted to reflect state-level prevalence
from BRFSS, improving local representation. It is possible however that variability in
microscale prevalence is not entirely represented when considering other potentially
influential variables such as race and obesity, two attributes that can influence asthma
prevalence and can vary spatially (U.S. EPA, 2018, section 4.1.2). Family income level was
used to represent spatial variability in asthma prevalence and may, in some instances,
capture spatial variability in race and obesity (Ogden et al„ 2010), and thus to some extent,
reasonably represent the potential influence race and obesity have on asthma prevalence.
Flowever, instances where these influential variables are not fully represented in simulating
the at-risk population, and where populations identified by such variables are associated
with increased asthma prevalence that may spatially intersect with the highest ambient air
concentrations, could lead to uncertainty in estimated exposures and health risk. Further
characterization could be appropriate by comparing with local prevalence rates stratified by
a similar collection of influential variables, where such data exist.
Yes. Recent year
data used (2013-
2017)
(Attachment 1)
APEX:
Microenvironmental
Concentrations
Outdoor Near-road
and Vehicle PE and
PROX Factors
Both
Low
Low - Medium
Uncertainty in mean PROX value used is approximately 15 percentage points (Figure 10
and Table 7 of Langstaff (2007)). Factor may be of greater importance in certain study
areas or under varying conditions, though even with this mean difference, in-vehicle
penetration/decay decreases exposures and hence the importance of in-vehicle
microenvironments. Further, considering that the exposures of interest need to be
concomitant with elevated exertion, the accurate estimation of exposures occurring inside
vehicles is considered relatively unimportant. This uncertainty could be important for
exposure events that occur outdoors near roads (i.e., PE factor = 1) and when simulated
individuals might be at elevated exertion for long durations. That said, the frequency of
these specific events is likely low, but nevertheless unquantified at this time.
No
Indoor: Air Exchange
Rates
Both
Low
Medium
Uncertainty due to random sampling variation via bootstrap distribution analysis indicated
the AER geometric mean (GM) and standard deviation (GSD) uncertainty for a given study
area tends to range from ±1.0 GM and ± 0.5 GSD hr1 (Langstaff, 2007). Some of the eight
study areas used AER from a geographically similar city. Non-representativeness remains
an important issue as city-to-city variability can be wide ranging (GM/GSD pairs can vary
by factors of 2-3) and data available for city-specific evaluation are limited (Langstaff,
2007). The restaurant and school AER distributions are derived from small samples and
may not be representative of all possible types of restaurants and schools, in general. That
said, indoor microenvironments are considered less likely to contribute to an individual's
daily maximum 7-hr average O3 exposure while at elevated exertion levels and likely does
not contribute substantially to uncertainty in the exposure and risk estimates.
Yes. New
distribution used
for restaurant and
school ME
(section
3D.2.6.1.1).
3D-151
-------
Sources of Uncertainty
Uncertainty Characterization
Newly
Characterized/
New
Information?
Influence of Uncertainty on
Exposure | Risk Estimates
Knowledge-
base
Uncertainty
Comments
Category
Element
Direction
Magnitude
Indoor-Residence:
A/C Prevalence
Both
Low
Low
Data were obtained from a reliable source, are comprehensive, and subject to quality
control (U.S. Census Bureau, 2019). For six of the of the eight study areas, A/C prevalence
was available for 2015 and 2017, while for the remaining two study areas (Sacramento and
St. Louis), the most recent year data available was 2011. There is uncertainty associated
with the use of an A/C prevalence derived from a different year than the years simulated in
the two study areas due to changes in housing stock that may occur over time. That said,
indoor microenvironments are considered less likely to contribute to an individual's daily
maximum 7-hr average O3 exposure while at elevated exertion levels and likely do not
contribute substantially to uncertainty in the exposure and risk estimates.
No
Indoor: Removal
Rate
Both
Low
Medium
Greatest uncertainty in the input distribution regarded representativeness, though
estimated as unbiased but correct to within 10% (Langstaff, 2007).
No
APEX: Simulated
Activity Profiles
Longitudinal Profiles
Under
Low - Medium
Medium
The magnitude of potential influence for this uncertainty would be mostly directed toward
estimates of multiday exposures. Simulations indicate the number of single day and
multiday exposures of interest can vary based on the longitudinal approach selected (Che
et al„ 2014). As discussed in section 3D.2.5.4, the D&A method provides a reasonable
balance of this exposure feature. Note however, long-term diary profiles (i.e., monthly,
annual) do not exist for a population, thus limiting the evaluation. Further, the general
population-based modeling approach used for main body results does not assign rigid
schedules, for example explicitly representing a 5-day work week for employed people.
No
Commuting
Both
Low
Medium
Method used in this assessment (and used previously in the 2014 HREA) is designed to
link Census commute distances with CFIAD vehicle drive times. This is considered an
improvement over the former approach that did not match commute distance and activity
time. While vehicle time is accounted for through diary selection, it is not rigidly scheduled.
Flowever, accurate estimation of exposures occurring while inside vehicles is considered
relatively unimportant because it is unlikely to occur while at elevated exertion.
No
Activity Patterns for
Simulated At-Risk
Population
Both
Low
Low - Medium
Analyses of activity patterns of people with asthma are similar to that of individuals not
having asthma regarding participation rate in outdoor activities and exertion level (section
3D.2.5.3; see also 2014 HREA, Appendix G, Tables 5G-2 to 5G-5 ).
No
APEX: Physiological
Processes
Body Weight
(NHANES)
Unknown
Low
Low
Comprehensive and subject to quality control, appropriate years selected for simulated
population, though possible small regional variation is possibly not well-represented by
national data (U.S. EPA, 2018, Appendix G.)
Yes. Recent year
data used (2009-
2014) (U.S. EPA
(2018), Appendix
G)
NVCfemax
Unknown
Low
Low
Upper bound control for unrealistic activity levels rarely used by model, thus likely not very
influential.
No
3D-152
-------
Sources of Uncertainty
Uncertainty Characterization
Newly
Characterized/
New
Information?
Influence of Uncertainty on
Exposure | Risk Estimates
Knowledge-
base
Uncertainty
Comments
Category
Element
Direction
Magnitude
Resting Metabolic
Rate (RMR)
Unknown
Low
Low
New, improved algorithm used for the current O3 exposure and risk analysis (U.S. EPA
2018, Appendix H). Comprehensive literature review resulted in construction of large data
base used to derive new RMR equations. Equations consider variables most influential to
RMR (i.e., age, body weight, and sex). There are other factors that could affect intra-
personal variability in RMR such as time-of-day (Haugen et al„ 2003) or
seasonal/temperature influences (van Ooijen etal., 2004;Leonard etal., 2014). Variability
from these and other potentially influential factors may be indirectly accounted for by the
residual error term used in the RMR Equation 3D-2 depending on the extent to which these
influential factors varied across the clinical study data that were used to create the RMR
analytical data set. However, because there is inadequate information regarding the
presence of multiple RMR measurements for individual study subjects, we could not
estimate intra-individual variability nor could we use these influential factors, other than age
and sex, as explanatory variables in the RMR equation. Therefore, any influences on
spatial variability in RMR, both within and among the eight study areas, would largely be
driven by the spatial distribution of age and sex.
Yes. Recent data
and new
equations (U.S.
EPA (2018),
Appendix H).
METS Distributions
Over
Low - Medium
Medium
In a prior characterization of uncertainty in METs, APEX estimated daily mean METs range
from about 0.1 to 0.2 units (between about 5-10%) higher than independent literature
reported values (Table 15 of Langstaff, 2007). Some of the diary activities in CHAD
encompassed broad categories (e.g., 'play sports', 'travel, general') and as such METs
distributions were developed using multiple activities, some of which that could vary greatly
in magnitude. Since the 2014 HREA, the list of CHAD activities (and corresponding METs
values) were expanded from 142 to 320, by evaluating available diary comment details and
disaggregating the originally assigned broad activities to more specific activities (see
Attachment 3 and U.S. EPA, 2019c. New distributions were developed using METs
estimates provided by Ainsworth et al. (2011). It is expected that the added specificity and
redevelopment of METs distributions would more realistically estimate activity-specific
energy expenditure. Two important uncertainties remain: the application of literature
provided longer-term average METs values to short-term events and the extrapolation of
METs data provided for adults to children. However, shorter-term values are of greater
importance in this assessment, thus METs could be better characterized where short-term
METS data are available.
Yes. New activity
codes and MET
distributions(U.S.
EPA, 2019a, U.S.
EPA, 2019b)
3D-153
-------
Sources of Uncertainty
Uncertainty Characterization
Newly
Characterized/
New
Information?
Influence of Uncertainty on
Exposure | Risk Estimates
Knowledge-
base
Uncertainty
Comments
Category
Element
Direction
Magnitude
Ventilation Rates
Unknown
Low
Low - Medium
Predictions made using the prior algorithm showed excellent agreement with independent
measurement data, particularly when considering simulated study group (Graham and
McCurdy, 2009; 2014 HREA Figure 5-23 and Figure 5-24). New algorithm derived using
the same data observed to have improved predictability (U.S. EPA, 2018, Appendix H).
Flowever, a shorter-term comparison (a single hour rather than daily) of predicted versus
measured ventilation rates, while more informative, cannot be performed due to lack of
ventilation rate data at this duration and considering influential factors (e.g., age, particular
activity performed).
Yes. New
equation (U.S.
EPA, 2018,
Appendix H).
Exposure-based risk
EVR
Characterization of
Moderate or Greater
Exertion
Both
Low
Low - Medium
The 2014 HREA recognized that the simulated number of people achieving this level of
exertion could be moderately overestimated, affecting the results for comparison to
benchmarks and the population-based E-R approach used to estimate lung function risk. A
new approach to identifying when individuals may be at moderate or greater exertion was
developed to better address inter-personal variability observed in the controlled human
exposure study subjects (Attachment 2). Uncertainty remains in the extrapolation of the
observations made from adults and proportionally applied to children.
Yes. New
distribution-based
approach
(Attachment 2).
Benchmark
Concentration
Levels for Population
Study Groups
Under
Low
Medium
There is only very limited evidence from controlled human exposure studies of population
groups potentially at greater risk. Compared to the healthy young adults included in the
controlled human exposure studies, members of some populations (e.g., children with
asthma) are considered more likely to experience adverse effects following exposures to
lower O3 concentrations (80 FR 65322, 65346, October 26, 2015; Frey, 2014, p. 7).
Although not directly characterized in the 2014 HREA, the benchmark levels derived from
the controlled human exposure studies may not be entirely representative of effects likely
to be exhibited by the simulated population and could underestimate the size of the
population at risk and/or the magnitude of adverse effects.
No
Exposure Duration
Under
Low
Low
The exposure duration for the studies from which the benchmark concentrations are drawn
is 6.6-hr (six 50-min exercise periods separated by 10-min rest periods, and with a 35-min
lunch after 3rd hour). For practical reasons, daily maximum exposures were time averaged
over 7 hours (rather than 8 hours previously used) to better relate to the concentrations
used for the controlled human exposure study subjects. The whole number 7, was used
(rather than 6.6) due to logistical and timeline constraints on implementation of a 6.6-hr
duration in the exposure model. Use of 7 hours, while more accurately reflecting the
exposure duration in the controlled human exposure studies, would likely underestimate
risk relative to directly using 6.6 hours, albeit to a limited extent. Use of 7 hours reduces the
magnitude of risk underestimation compared to use of 8 hours (as was done in the prior
REAs).
Yes. A 7-hr
duration for
averaging
exposure
concentrations
was used to
better represent
6.6-hr exposures
(section 3D.2.8.1)
3D-154
-------
Sources of Uncertainty
Uncertainty Characterization
Newly
Characterized/
New
Information?
Influence of Uncertainty on
Exposure | Risk Estimates
Knowledge-
base
Uncertainty
Comments
Category
Element
Direction
Magnitude
Lung Function Risk
Estimation
Contribution to Risk
of Exposures at or
Below 40 ppb
Over
Low - Medium
Low - Medium
While there is limited support for O3 being causally linked to lung function responses at the
lowest tested exposure level (i.e.,40 ppb exposures), there are no observations at lower
exposures. Data available at 40 ppb are limited to two studies, in one of which O3 was
administered by facemask and had the only positive response. Because the lung function
risk analysis assumes there is an exposure response relationship at exposures below 40
ppb, the influence of this source of uncertainty could possibly contribute to the
overestimation of risk when including risk resulting from low exposures. The magnitude of
influence appears to be greater for the MSS model estimates when compared to the E-R
function estimates.
Yes. New
evaluation of the
contribution of risk
from low
exposures,
(section 3D.3.4.3)
Extrapolation of E-R
Data from Healthy
Subjects to
Simulated People
with Asthma
Under
Low
Low - Medium
Subjects with asthma in controlled human exposure studies appear to be at least as
sensitive to acute effects of O3 in terms of FEV1 and inflammatory responses as healthy
non-asthmatic subjects (2013 ISA, section 8.2.2). Note however, study subjects with
asthma are typically characterized as having a "mild" condition, thus, there is uncertainty in
how others expressing a more severe condition would respond to similar O3 exposures. In
addition, many epidemiologic studies report greater risk of health effects among individuals
with asthma. Considering each of these elements, a direct extrapolation could understate
the at-risk population.
No
Extrapolation of E-R
Data from Adults 18-
35) to Children and
to Older Adults
Both
Low - Medium
Low
Because the vast majority of controlled human exposure studies investigating lung function
responses were conducted with adult subjects, the lung function risk estimates for children,
ages 5-18, is based on E-R data from adult subjects to estimate responses in children aged
5-18. However, the few available studies of 03-related lung function decrement in children
indicate that children's FEV1 responses are similar to those observed in adults 18-35 years
old (e.g. McDonnell et al„ 1985). Regarding older adults, the evidence indicates a decline
in responsiveness with increasing age 18 to 35, followed by a rate of decrease that
dampens for ages 36 to 55, and ultimately leads to limited responsiveness in adults >55
years old (2013 ISA, p. 6-22). A newly available study, Arjomandi etal., 2018, found a
statistically significant reduction in FEV1 (group mean of 1.7%) for older adults (mean age
59.9) following 3-hr exposures of 120 ppb O3 with exercise (EVR of 15-17 L/min-m2),
although statistical significance was not found for 3-hr exposures of 70 ppb. Given the 7-hr
focus of the current assessment as well as the fact that the air quality scenario for the
current standard includes no hours with an ambient air concentration at or above 120 ppb
(Appendix 3C, Figures 3C-67 to 3C-74), the setting of the age term at zero for older adults
appears to remain appropriate for the simulated exposure conditions.
No
3D-155
-------
Sources of Uncertainty
Uncertainty Characterization
Newly
Characterized/
New
Information?
Influence of Uncertainty on
Exposure | Risk Estimates
Knowledge-
base
Uncertainty
Comments
Category
Element
Direction
Magnitude
Assumed No
Interaction of other
Co-pollutants on O3-
related Lung
Function Responses
Under
Low
Medium
There are a few studies regarding the potential for an increased response to O3 when
exposure is in the presence of other common pollutants such as particulate matter
(potentially including particulate sulfur compounds), nitrogen dioxide, and sulfur dioxide,
although the studies are limited (e.g., with regard to relevance to ambient air exposure
concentrations) and/or provide inconsistent results.
No
Statistical model for
E-R Function
Both
Low
Low
The selection of statistical model to best reflect the E-R relationship can influence risk
estimates, particularly for instances when large proportions of the simulated population are
exposed to low-level concentrations. The 90/10 logit/linear model (section 3D-2.8.2.1)
yielding an E-R function similar in shape to an E-R function developed using a probit link (a
commonly used fitting method), would tend to estimate lower risk than a function based on
a logistic fit (which would have a relatively higher response at low level exposures). Overall,
the relatively low contribution of low-level exposures to risk when using the E-R function
approach indicates the selection of the 90/10 logit/linear fit to have a limited impact on
uncertainty in risk estimates.
Yes. Section
3D.3.4.2.1
Statistical
Uncertainty in E-R
Function
Both
Low - Medium
Low
A BMCMC approach was used to iteratively generate 9,000 unique E-R functions section
3D.2.8.2.1). We used the median (50th percentile) function for generating population-based
(E-R function) lung function risk in the main body results. A 95% confidence interval for risk
estimates was generated using the 2.5th and 97.5th percentile E-R functions. Overall, the
range of risk estimates using the confidence intervals was small, on the order of a few
percentage points, but increased in relative magnitude when considering the larger lung
function decrements.
Yes. section
3D.3.4.2.2
Contribution of Low-
level Ventilation
Rate in MSS model
Estimated Risk
Over
Low - Medium
Low - Medium
We evaluated the role of ventilation rate in estimating risk with the MSS model approach
(section 3D.2.8.2.2) by comparing risk generated using either of two model conditions: risk
for when simulated individuals experienced decrements at any ventilation rate, or risk for
when ventilation rate was at moderate or greater exertion (the latter reflects the E-R
function risk approach). The MSS model risk estimates were about 20-40% lower when
selecting for simulated individuals at moderate or greater exertion compared with MSS
model risk estimates using individuals at any ventilation rate (Table 3D-69). Even when
including only individuals at higher exertion rates, the MSS model risk estimates are still a
factor of three or more higher than risks estimated using the E-R function risk approach.
Given that the controlled human exposure studies indicate an importance of elevated
ventilation in combination with the studied exposure concentrations, the MSS model likely
overestimates risk.
Yes. section
3D.3.4.2.4
3D-156
-------
Sources of Uncertainty
Uncertainty Characterization
Newly
Characterized/
New
Information?
Influence of Uncertainty on
Exposure | Risk Estimates
Knowledge-
base
Uncertainty
Comments
Category
Element
Direction
Magnitude
Variability Parameter
Setting in MSS
Model
Over
Medium
Low - Medium
The value of the MSS model variable U (Equations 3D-15 and 3D-16) is randomly assigned
from a distribution to simulated individuals and is meant to address inter-individual
variability not accounted for by the other MSS model variables. The influence of Uwas
qualitatively evaluated by examining example time series for two children simulated with
different values for UanA for which similar-sized lung function decrements are predicted.
While both children had similar exposure profiles in terms of duration exposed to elevated
concentrations, the ventilation duration at peak concentrations differed. The difference
observed (Figure 3D-18) suggests that random assignment of high lvalues leads to
simulated individuals being predicted to experience lung function decrements at relatively
lower time-averaged breathing rates as those with a lower lvalue. Given the difference of
these exposure conditions from those in which such decrements are observed in controlled
human exposure studies, it is likely that the risk is overestimated, and the amount of
overestimation may be similar to that described for ventilation rate in the preceding entry. A
second variable vi, a constant, is used by the MSS model to address intra-individual
variability (Equation 3D-16). Because the vi is described as representing the non-ozone
related contribution to response variability in the study observations (McDonnell et al„
2013), a non-zero setting may contribute to over estimates in risk. We found estimated
risks using vi set to zero to be about 20-35% lower than when using the default parameter
setting (the setting used for the main results in section 3D.3.3.2).
Yes. section
3D.3.4.2.5
3D-157
-------
Sources of Uncertainty
Uncertainty Characterization
Newly
Characterized/
New
Information?
Influence of Uncertainty on
Exposure | Risk Estimates
Knowledge-
base
Uncertainty
Comments
Category
Element
Direction
Magnitude
Statistical and
Model Uncertainty in
MSS model
Both
Low
Low
Glasgow and Smith (2017) evaluated statistical uncertainty in the MSS model employed by
APEX. Multiple sets of lung function risk results were generated using random draws of the
MSS model coefficients (considering their standard errors) and performing APEX
simulations for children ages 5-17 and for 2010 air quality just meeting a design value of 75
ppb in Atlanta. Calculated bounds on the risk estimates could extend to as low as 0% and
>35% of children experiencing at least one decrement >10% (Glasgow and Smith (2017),
Figure 1). While the bounds were wide ranging (and affecting mostly the lowest decrement
size), the reported median risk estimate (18.1%) is similar to that estimated in the 2014
HREA. Note, these central tendency risk values are based on using the best estimates of
the MSS model coefficients and are derived from the existing controlled human exposure
study data. It is possible that with new controlled human study observations, these model
coefficients (and associated standard errors) could possibly change, resulting in a shift of
central tendency risk estimates in either direction (greater or lower frequency of lung
function decrements) while also changing the outer bounds (increasing or decreasing the
confidence intervals). Even so, the outer bounds of any risk estimates, based on the
current MSS model or a newly derived MSS model, are generated by using a distribution of
coefficient values, the bounds of which have a lesser probability of occurrence compared to
those generated using the central tendency values. Further, Glasgow and Smith (2017)
also evaluated MSS model uncertainty using two different parameterizations (one including
BMI as an explanatory variable, the other not). Comparison of median risk values for the
two parameterizations ranged from fractions to a few percentage points, with the largest
difference reported for the lowest decrement and overall, lower values were reported for
the MSS model that includes a BMI variable. We note the risk results generated in our
assessment are based on the MSS model that includes a BMI variable. While uncertainty in
the MSS model risk estimates could be further characterized (e.g., including the type of
analyses reported by Glasgow and Smith, 2017), that would not be expected to change the
overall conclusion that there is relatively greater uncertainty associated with the MSS
model estimates than with the E-R model estimates.
Yes (Glasgow
and Smith (2017))
3D-158
-------
3D.3.4.2 Targeted Evaluations of Lung Function Risk Models
The intent of the following targeted evaluations is to provide insight into a few of the
important uncertainties identified in section 3D.3.4.1 concerning the lung function risk estimates.
Analyses were designed to inform how the uncertainties may influence the exposure and risks
reported in section 3D.3. Results or estimates generated in these targeted evaluations do not
replace (nor supplement) the results in section 3D.3, nor do they address all aspects of the
exposure and risk assessment. Further, because the main body results indicated children were
estimated to experience lung function decrements more frequently than adults, we focused these
targeted evaluations on simulations with children.
Briefly, we performed five targeted evaluations with each discussed in the following
sections. The first section discusses the statistical model used to represent the E-R function
(section 3D.3.4.2.1). The next section discusses the development and interpretation of confidence
intervals for the lung function risk estimates generated using the population-based E-R function
(section 3D.3.4.2.2). This is followed by a section describing an evaluation of the contribution of
low-level exposures to risk estimated using the population-based E-R function and the individual
MSS model lung function risk approaches (section 3D.3.4.2.3). Section 3D.3.4.2.4 evaluates the
role moderate or greater ventilation has in estimating risk using the MSS model. And finally, a
discussion and evaluation of variability parameters used in the MSS model is presented in
section 3D.3.4.2.5.
3D.3.4.2.1 Statistical Model Used for the E-R Function
There are several approaches available to fit data to a continuous E-R function, for
example, regression models (linear, logistic) and use of curve smoothing techniques (moving
averages, polynomial splines). Logistic regression is commonly used for concentration-,
exposure-, and dose-response relationships when study observations contain a binary dependent
variable (e.g., either yes/no response). A logistic regression can be fit using a varied linking
approach, such as logit or probit, the selection of which can depend on assumptions made
regarding the distribution of responses (logistic or inverse normal, respectively)82 among other
factors (e.g., model fit statistics).
The statistical model selected for the E-R function and used to estimate the frequency of
lung function decrements in this exposure and risk analysis is the same as that used in the 2014
and 2007 REAs and was based on combining logistic (with a logit link) and two-piece linear
82 For example, regarding the development of an E-R function describing lung function decrements associated with
exposure to SO2 for the risk assessment performed in the 2012 review of the SO2 NAAQS (U.S. EPA, 2009), the
CASAC for that review (Samet, 2009) suggested that the distribution of individual response thresholds supported
use of a probit function rather than a logistic function (pp. 14 and 60-63).
3D-159
-------
forms in a 90/10 percent proportion, respectively, using a Bayesian Markov Chain Monte Carlo
(BMCMC) modeling approach (section 3D.2.8.2.1). The selection of this model was based on
advice received from the CASAC review in the 2008 O3 NAAQS review (Henderson, 2006) and
evaluation of the curve fit statistics for this function (U.S. EPA, 2007a; 2014 HREA).
Of practical importance for this assessment is how the response curve is extrapolated
from the lowest observed exposure to zero exposure/response. For general context, in comparing
a probit to a logit link in a logistic regression, the probit link would yield a relatively lower
response at the lowest level exposures. The two-piece linear model used in part for developing
the current E-R function resembles a hockey stick, with the paddle representing a zero response
for the lowest exposures, and the handle representing the increased response that coincides with
increasing exposures, beginning at the junction between the paddle and the stick.83 Based on this
statistical form, when combining the logit linked logistic model with a hockey stick type model
(as done for this assessment), it was assumed the 90/10 percent proportion logit/linear E-R form
would have a response curve shape for low-level exposures similar to that using a probit link. To
better evaluate these E-R functions, we fit the controlled human exposure study data (Figure 3D-
12) using a probit link and compared that with the 90/10 logit/linear curve.
As an example, Figure 3D-14 illustrates the E-R functions fit from these two approaches,
using the >15% lung function decrement observations. Plotted for the probit approach is the
curve derived from the best estimate of the model coefficients, along with 95% confidence
intervals derived from the model coefficient variability. For the 90/10 logit/linear approach, the
median (50th percentile) function is plotted, along with 95% confidence intervals derived from
the 2.5th and 97.5th percentile E-R functions obtained from the 9,000 BMCMC model iterations.
As expected, the probit curve is very similar to the 90/10 logit/linear curve, albeit with the
former having a response just below the latter for the lower exposures. The opposite occurs for
exposures above 55 ppb; for those higher exposures, a relatively greater response is indicated
using the 90/10 logit/linear E-R function. Based on there being little difference between the two
curves and only slight off-setting of the response at different exposure levels, it is likely that
applying a probit fit for the E-R function to the population distribution of daily maximum 7-hr
exposures would result in little to no difference from the risk estimates derived with the 90/10
logit/linear E-R function.84
83 The combined two-piece linear/logistic E-R function is used, as described in section 3D.2.8.2.1 above, because of
the limited controlled human exposure study data, and associated uncertainty regarding the response, at low level
exposures (i.e., <60 ppb). Note, the two-piece linear model has a lower percent contribution (10%) compared to
that of the non-threshold logistic model (90%) in deriving the combined E-R function.
84 Evaluation of the E-R functions fit for the 10% decrement indicated that the 90/10 logit/linear curve had a
somewhat higher response than the probit curve at the low-level exposures (and lower response at exposures >55
3D-160
-------
30%
i 20%
-Q
X!
O
Q.
0)
(/)
C
o
CL
W
15% (based on data in Table 3D-20). Confidence
intervals for the probit model reflect variability in the regression model
coefficients.
3D.3.4.2.2 Confidence Intervals for the Population-based E-R Function
and Effect on Lung Function Risk Estimates
To estimate lung function risk using the population-based E-R function approach, the
results of which are presented in the main body of this document (section 3D.3.3.1), we selected
the median (50th percentile) E-R function originally developed as part of the 2014 HREA. This
E-R function was derived from Bayesian Markov Chain Monte Carlo (BMCMC) approach that
iteratively combined logistic and linear E-R functions fit to the controlled human exposure study
data in Table 3D-20 (section 3D.2.8.2.1). The selection of the median E-R function to estimate
risk in the current assessment generally assumes the simulated at-risk population is comprised of
ppb). For the 20% decrement the probit curve was similar to the 90/10 logit/linear curve at low-level exposures,
but slightly higher for exposures between 50 and 70 ppb. Given the smallness of the difference and limited
contribution of the lower exposures to the risk estimates (Table 3D-66), these finding does not imply significant
uncertainty or support generation of new simulations and risk estimates using the probit E-R function.
3D-161
-------
individuals that have a similar response frequency as that of the general collection of controlled
human exposure study test subjects.
Because there were two or more studies reporting observed responses at most of the
exposure levels and the BMCMC approach generates numerous E-R functions, statistical
uncertainty in the E-R function can be used to approximate lower and upper bounds to the lung
function risk estimates. To evaluate such bounds here, a 95% confidence interval for lung
function risk was estimated by combining the population distribution of daily maximum 7-hr
exposures (occurring while at moderate or greater exertion) for simulated children in each study
area with the 2.5th and 97.5th percentile population-based E-R functions (2014 HREA, Appendix
6A, Table 6A-1). Lung function risk estimates based on these lower and upper percentile
functions and those based on the median function (for air quality just meeting the existing
standard) are presented, in terms of the minimum and maximum year results, in Table 3D-65 for
each of the three lung function decrements (i.e., FEVi decrements >10, 15, and 20%). The
estimates for the median E-R function are drawn from Table 3D-40. The estimates for the best
and worst air quality years yield the minimum and maximum estimates for each of the three
functions providing a range for estimates based on each of the particular E-R functions.
The range of values for the estimated risk generated by each of the E-R functions as a
result of using different air quality years (i.e., the distance in estimated risk between the
minimum and maximum values) is small, on the order of a few tenths of a percentage point, with
the smallest range of values associated with the largest lung function decrement (Table 3D-65).
In general, the range of the overall 95% confidence interval (i.e., the distance in estimated risk
between the 2.5th and 97.5th values) is also small when considering percentage points. For
example, the lower bound percent of children estimated to experience at least one lung function
decrement >10% for the Atlanta study area is about 1.5% and the upper bound value is about
4.0% (median E-R function value ~ 2.4%). With increasing magnitude of the decrement, the
range of percentage points becomes smaller (e.g., a >20% decrement has a range of about 0.6
percentage points for Atlanta). In terms of relative magnitude, one might consider this range
large (i.e., a factor of 6 or more), but because there are so few children estimated to experience
these large lung function decrements, this interpretation of the confidence interval would be
inappropriate. Further, it would be unreasonable to simply assume that use of the lower and
upper bounds of the E-R functions would appropriately estimate lower and upper bounds of risk
without additional context regarding the controlled human exposure study data, the interpretation
of the bounds on the E-R function, and how these might relate to statistical uncertainty in the risk
estimates.
3D-162
-------
Table 3D-65. Percent of children estimated to experience at least one lung function
decrement at or above the indicated level, for air quality adjusted to just meet
the current standard, using the population-based (E-R function) risk
approach.
Percent of Children Estimated to Experience at Least One
FEVi
Decrement
Decrement per Year usinq Specified E
-R Functions A
Study Area
Lower Bound (2.5%)
E-R Function
Median (50%)
E-R FunctionB
Upper Bound (97.5%)
E-R Function
minc
maxc
min
max
min
max
Atlanta
1.0
1.5
1.9
2.5
3.1
4.0
Boston
1.2
1.4
2.0
2.3
3.3
3.8
Dallas
1.2
1.6
2.1
2.6
3.5
4.3
>10%
Detroit
1.4
1.8
2.3
2.8
3.9
4.5
Philadelphia
1.3
1.4
2.2
2.4
3.6
3.9
Phoenix
1.8
2.2
2.9
3.3
4.8
5.4
Sacramento
1.2
1.4
2.2
2.3
3.6
3.8
St. Louis
1.4
1.8
2.3
2.8
3.8
4.5
Atlanta
>0.1
0.2
0.4
0.6
0.7
1.1
Boston
0.1
0.2
0.5
0.6
0.8
1.0
Dallas
0.1
0.2
0.5
0.7
0.8
1.1
>15%
Detroit
0.1
0.3
0.6
0.8
1.0
1.2
Philadelphia
0.1
0.1
0.5
0.6
0.9
1.0
Phoenix
0.2
0.3
0.7
0.9
1.2
1.5
Sacramento
0.1
0.1
0.5
0.6
0.9
1.0
St. Louis
0.1
0.3
0.6
0.8
0.9
1.2
Atlanta
>0.1
0.1
0.1
0.2
0.4
0.7
Boston
>0.1
0.1
0.2
0.2
0.5
0.7
Dallas
>0.1
0.1
0.2
0.3
0.5
0.8
>20%
Detroit
0.1
0.1
0.2
0.3
0.7
0.9
Philadelphia
>0.1
>0.1
0.2
0.2
0.6
0.6
Phoenix
0.1
0.1
0.3
0.4
0.8
1.1
Sacramento
>0.1
>0.1
0.2
0.2
0.5
0.6
St. Louis
>0.1
0.1
0.2
0.3
0.6
0.9
A Calculated percent is rounded to the nearest tenth decimal using conventional rounding. Values equal to zero are designated
by "0" (there are no individuals experiencing decrements at that level). Small, non-zero values that do not round upwards to 0.1
(i.e., <0.05) are given a value of "<0.1".
B The median function is used to generate E-R function risk estimates reported in the main body results. Note, these are
identical to the results reported in Table 3D-40.
c The minimum (min) are results for the best air quality year and the maximum (max) are results for the worst air quality year of
the three years simulated.
As a reminder, while the controlled human exposure study subjects are volunteers (and
assumed to be selected at random), it is important to note there are important fundamental biases
in their collective composition: none of the individuals have known preexisting health conditions
3D-163
-------
(e.g., cardiovascular disease, asthma) and all of the subjects are required to be physically fit
enough to meet a study's exercise target levels. Clearly, not every member of the simulated
population has these attributes, but the risk approach does select for when simulated individuals
are at moderate or greater exertion while exposed, as was done for the controlled human
exposure study subjects. Therefore, representation of potentially sensitive individuals (i.e., those
with pre-existing health conditions) in the study data and thus, in the derived E-R functions, is
absent.
In addition, use of this type of statistical approach to estimate lower and upper bounds of
lung function risk does not suggest that the range of functions could be equally applied to the
simulated population as a whole (e.g., that the entire population could have a risk as low as Xor
as high as 7, based on the 2.5th and 97.5th percentile functions selected, respectively) nor does the
range of E-R functions likely represent individuals that are least sensitive, or more importantly
(given the NAAQS review context for these analyses), those most sensitive to O3 exposure. The
variability in the observed response in study subjects at given O3 exposures can be due to many
factors (e.g., uncertainties in exposure conditions, response/concentration measurements, study
subject sensitivity for healthy individuals, number of subjects per study, etc.). When used in such
an analysis here, one might suggest the lower and upper bounds account for some of these
uncertainties, however, they would be bounded by their collective representativeness and actual
weighting of these uncertainties present in the study observations. In its application, it would be
assumed that the distribution of the features of the study subjects are similarly reflected in the
simulated population, which as described above, is not entirely the case.
Further, the range of functions used to represent lower and upper bounds is derived from
a distribution of functions. If there were a perfect matching of the study subject attributes with
those of the simulated population, the risks estimated using either end of the 95% interval for the
E-R function would certainly have a much smaller likelihood and better apply to a smaller
proportion of the population, than those estimated from using the median E-R function. That
said, even in the absence of perfectly matching the attributes of the controlled human exposure
study subjects with those of the simulated population, the median E-R function may be most
appropriate to estimate lung function risk for the simulated population as a whole. Still, the
median E-R function may be underestimating the number/percent of individuals experiencing
decrements to the extent that the general population includes individuals that would experience
greater decrements than experienced by individuals represented in the controlled human
exposure studies. Further, as recognized in Chapter 3 of the PA, similarly sized decrements in
individuals with compromised respiratory function or in individuals with asthma may be more
likely to elicit other, perhaps more significant, health outcomes.
3D-164
-------
3D.3.4.2.3 Contribution of Low-Level Exposures to Lung Function Risk
Estimates
The two approaches used to estimate lung function risk were evaluated to better
understand how the distribution of exposures influences the estimated risk. For the first approach
that used the population-based E-R function to estimate risk, we evaluated the risk contribution
resulting from each of the daily maximum 7-hr exposure levels that occur while at elevated risk.
Because the continuous function used is extrapolated from the lowest observed exposure (40
ppb) in the controlled human exposure studies to zero, of particular interest here were the
contributions from low exposures to the estimated risk where there are no controlled human
study data available (i.e., O3 exposures <40 ppb, 6.6-hrs). Further, because there were only two
studies that included exposures of 40 ppb (one that elicited decrements between 10 and 15% in
two study subjects, with no statistical significance at the group mean level, the other eliciting no
decrement of at least 10% in any subjects), we also evaluated the contribution to estimated risk
resulting from exposures >50 ppb and >60 ppb.
The APEX exposure output for the E-R function approach that were the basis for the
main results reported in section 3D.3.3 are 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, given the objectives for this evaluation,
time limitations on it, and that new simulations were required to evaluate the MSS model
approach (see below), 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 a single year (2016). The results for this evaluation are provided in Table 3D-66
for the three study areas, three air quality scenarios using 2016 data, and focusing on the risk
contribution to lung function decrements occurring at least one and two days per year. Figure
3D-15 illustrates the same results, but for air quality just meeting the current standard.
There is variability in the risk contribution across the three study areas, variability which
increases with increasing magnitude of the lung function decrement and increasing O3 exposures
across the three air quality scenarios. The risk estimated from 7-hr average exposures below 40
ppb is generally low and is lower for higher magnitudes of the lung function decrement and
higher air quality scenario design value. That said, the majority of risk (84 to 98%) is attributed
to 7-hr average exposures >40 ppb for any of the air quality scenarios. The risk contribution
attributed to 7-hr average exposures >60 ppb varies greatly across the study areas, the magnitude
of the decrements, and the air quality scenarios. For example, on average about 37% of the risk is
contributed by 7-hr average exposures >60 ppb. But in Dallas, the contribution from these
exposures is much less (on average about 22%), while in St. Louis, the contribution is much
more (on average about 50%).
3D-165
-------
Table 3D-66. Estimated lung function risk contribution resulting from selected 7-hr
average O3 exposures in children, using the E-R function risk approach, 2016.
Air Quality
Scenario
7-hr
Exposure
Study
Area
Risk Contribution from Indicated 7-hr Exposure, E-R Function Approach
One Decrement/FEVi
Reduction
Two Decrements/FEVi Reduction
>10%
>15%
>20%
>10%
>15%
>20%
Atlanta
6.2%
3.5%
1.8%
10.3%
6.7%
3.8%
<30 ppb
Dallas
6.7%
3.9%
2.1%
10.8%
7.1%
4.2%
St. Louis
5.4%
2.9%
1.4%
9.1%
5.6%
3.0%
Atlanta
19.8%
13.4%
8.1%
30.4%
23.3%
16.2%
<40 ppb
Dallas
20.9%
14.7%
9.2%
31.3%
24.3%
17.2%
65 ppb
St. Louis
16.5%
10.6%
6.1%
25.6%
18.5%
12.0%
Atlanta
54.3%
43.1%
32.2%
75.4%
67.7%
58.2%
<50 ppb
Dallas
58.1%
48.1%
37.6%
76.6%
69.5%
60.6%
St. Louis
43.4%
32.7%
23.3%
62.6%
53.1%
42.7%
Atlanta
88.7%
81.9%
74.4%
98.5%
97.4%
95.9%
<60 ppb
Dallas
94.2%
90.2%
85.6%
99.8%
99.6%
99.4%
St. Louis
83.3%
75.2%
67.7%
96.7%
94.5%
91.8%
Atlanta
4.2%
2.0%
0.9%
7.3%
4.2%
2.1%
<30 ppb
Dallas
5.3%
2.8%
1.4%
8.8%
5.4%
2.9%
St. Louis
3.7%
1.7%
0.7%
6.6%
3.6%
1.7%
Atlanta
12.9%
7.5%
3.9%
21.0%
14.4%
8.8%
Current
Standard
(70 ppb)
<40 ppb
Dallas
16.3%
10.5%
6.0%
25.1%
18.2%
11.9%
St. Louis
11.1%
6.1%
3.1%
18.0%
11.5%
6.6%
Atlanta
35.9%
24.5%
15.7%
53.9%
43.3%
32.9%
<50 ppb
Dallas
45.3%
34.4%
24.5%
63.4%
54.0%
43.7%
St. Louis
29.0%
18.7%
11.6%
44.9%
33.7%
23.8%
Atlanta
72.3%
59.7%
48.7%
91.7%
86.6%
81.2%
<60 ppb
Dallas
85.8%
77.9%
69.8%
97.0%
95.0%
92.6%
St. Louis
61.1%
48.3%
38.5%
82.9%
74.5%
66.5%
Atlanta
2.9%
1.2%
0.4%
5.2%
2.6%
1.2%
<30 ppb
Dallas
4.3%
2.1%
0.9%
7.4%
4.3%
2.2%
St. Louis
2.8%
1.2%
0.4%
5.2%
2.6%
1.1%
Atlanta
8.6%
4.3%
1.9%
14.7%
8.8%
4.7%
<40 ppb
Dallas
13.1%
7.7%
4.0%
21.0%
14.4%
8.8%
75 ppb
St. Louis
8.4%
4.1%
1.9%
14.1%
8.1%
4.2%
Atlanta
23.6%
13.7%
7.6%
37.6%
26.5%
17.6%
<50 ppb
Dallas
35.6%
24.6%
16.0%
52.9%
42.4%
32.2%
St. Louis
21.6%
12.5%
7.0%
34.6%
23.5%
15.1%
Atlanta
52.8%
37.7%
27.0%
75.8%
64.9%
55.2%
<60 ppb
Dallas
75.1%
63.3%
52.7%
92.0%
87.2%
82.1%
St. Louis
47.4%
33.6%
24.4%
69.3%
57.2%
47.3%
3D-166
-------
100%
90%
80%
70%
60%
50%
40% -£
30%
20% -r
10%
0% -t"
Risk Contribution from 7-hr Average Exposures
One Decrement - Current Standard (70 ppb)
E-R Function Risk Approach
>20% FEV1 Reduction - One Day
¦ >15% FEV1 Reduction - One Day
¦ >10% FEV1 Reduction - One Day
ATL
DAL
STL
Exposures
< 30 ppb
Exposures
< 40 ppb
Exposures
< 50 ppb
Exposures
< 60 ppb
100% T
Risk Contribution from 7-hr Average Exposures
Two Decrements - Current Standard (70 ppb)
E-R Function Risk Approach
>20% FEV1 Reduction - Two Days
I £15% FEV1 Reduction - Two Days
I >10% FEV1 Reduction - Two Days
60% :
50% --
40%
20% :
10%
Exposures
< 30 ppb
Exposures
< 40 ppb
Exposures
< 50 ppb
Exposures
< 60 ppb
Figure 3D-15. Estimated lung function risk contribution resulting from selected 7-hr
average O3 exposures in children, using the E-R function risk approach and
air quality adjusted to just meet the current standard, for one decrement (top
panel) and two decrements (bottom panel), 2016.
3D-167
-------
As was done with the E-R function results, we evaluated the influence exposure level has
on risks estimated using the MSS model. New APEX simulations were performed to estimate the
continuous hourly time-series of O3 exposures and FEVi decrements. All simulation conditions
remained the same as done for the main body risk results except that for this evaluation, a single
year of air quality (2016) was used and fewer children were simulated to maintain a tractable
analysis (10,000 rather than the 60,000 done for the main body results). Note, there is little
difference in risks estimated when varying the total number of simulated children (Table 3D-67).
Because the risk estimated using the MSS model is calculated from a cumulative time-series of
O3 exposures (and EVR, along with contributions from other variables used by the MSS model),
we calculated the 7-hr average O3 exposure occurring just prior to the FEVi decrements to allow
for reasonable comparison with the above E-R function risk contribution results.
Table 3D-68 and Figure 3D-16 present the risk contribution resulting from selected 7-hr
average O3 exposures that occur prior to a lung function decrement of interest, estimated using
the MSS model. While the general pattern in the risk contributions across the air quality
scenarios, study areas, and decrements are similar to that described above using the E-R function
approach, there are noteworthy differences between the two risk approaches. First, there is less
variability in the risk contribution values across the study areas and decrements when using the
MSS model risk approach. For example, the overall coefficient of variation (COV; standard
deviation/mean) ranges from 1 to 31% (mean 11%) across study areas when evaluating the MSS
model risk contributions, while the COV ranges from 6 to 49% (mean 26%) for the same
evaluation using the E-R function. Second, the MSS model consistently calculates a greater
percent of lung function decrements that result from low O3 exposures (Table 3D-68) relative to
that estimated when using the E-R function (Table 3D-66). While the majority of risk (84 to
98%), mean 91%>) using the E-R function risk approach was attributed to 7-hr average exposures
>40 ppb, when using the MSS model, between 33 to 75% (mean 54%) of risk is attributed to 7-hr
average exposures >40 ppb when considering the three air quality scenarios and all three
decrements. Based on this evaluation, the MSS model more frequently predicts responses to
occur at lower O3 exposures than does the E-R function approach.
Table 3D-67. MSS model risk estimates from varying the number of simulated children.
Study Area
(2016 AQ)
APEX Simulation, 70 ppb AQ Scenario
(number of simulated children)
% of Children Experiencing at least One Decrement
FEVi >10%
FEVi >15%
FEVi >20%
Atlanta
(worst year)
Sensitivity (n = 10,000)
14.6%
5.1%
2.1%
Main Results, Table 3D-52 (n = 60,000)
15.1%
5.0%
2.1%
Dallas
(best year)
Sensitivity (n = 10,000)
13.3%
4.1%
1.7%
Main Results, Table 3D-52 (n = 60,000)
13.1%
4.0%
1.6%
St. Louis
(worst year)
Sensitivity (n = 10,000)
16.3%
5.8%
2.5%
Main Results, Table 3D-52 (n = 60,000)
16.3%
5.9%
2.7%
3D-168
-------
Table 3D-68. Estimated lung function risk contribution resulting from selected 7-hr
average O3 exposures in children, using the MSS model risk approach, 2016.
Air Quality
Scenario
7-hr
Exposure
Study
Area
Risk Contribution from Indicated 7-
hr Exposure, MSS Model Approach
One Decrement/FEVi Reduction
Two Decrements/FEVi Reduction
>10%
>15%
>20%
>10%
>15%
>20%
65 ppb
< 30 ppb
Atlanta
33.5%
16.2%
10.1%
33.9%
17.4%
10.4%
Dallas
36.0%
19.6%
8.8%
36.9%
20.6%
9.1%
St. Louis
29.6%
13.8%
9.0%
30.3%
15.1%
9.4%
<40 ppb
Atlanta
70.9%
52.9%
41.6%
71.6%
55.4%
44.0%
Dallas
71.7%
57.2%
38.9%
72.6%
60.6%
42.1%
St. Louis
64.9%
46.4%
35.1%
65.5%
49.3%
40.3%
<50 ppb
Atlanta
93.0%
86.8%
81.3%
93.6%
88.9%
84.9%
Dallas
93.7%
89.2%
81.6%
94.0%
90.9%
86.0%
St. Louis
89.7%
82.3%
70.9%
90.1%
84.5%
76.5%
<60 ppb
Atlanta
99.1%
97.9%
96.6%
99.3%
98.0%
96.5%
Dallas
99.5%
98.7%
97.9%
99.6%
99.1%
99.4%
St. Louis
98.4%
97.2%
95.9%
98.5%
97.7%
96.6%
Current
Standard
(70 ppb)
< 30 ppb
Atlanta
28.6%
13.8%
7.3%
29.3%
14.4%
8.9%
Dallas
32.4%
18.4%
8.8%
33.0%
19.4%
9.5%
St. Louis
24.8%
10.5%
5.2%
25.5%
11.1%
5.0%
<40 ppb
Atlanta
62.7%
44.2%
33.4%
63.3%
45.8%
36.2%
Dallas
66.7%
51.0%
37.5%
67.6%
54.0%
41.5%
St. Louis
56.6%
36.1%
25.3%
57.4%
37.9%
27.7%
<50 ppb
Atlanta
87.7%
79.7%
70.9%
88.3%
81.4%
75.1%
Dallas
90.6%
84.0%
77.5%
91.1%
86.7%
81.1%
St. Louis
83.5%
72.0%
62.9%
84.4%
73.5%
65.8%
<60 ppb
Atlanta
97.5%
95.2%
92.5%
97.8%
96.2%
94.6%
Dallas
98.8%
98.0%
96.0%
99.0%
98.6%
97.8%
St. Louis
95.9%
91.9%
87.8%
96.2%
92.8%
89.7%
75 ppb
< 30 ppb
Atlanta
25.1%
11.7%
6.5%
25.6%
12.2%
7.2%
Dallas
29.7%
16.7%
9.0%
30.3%
17.7%
10.2%
St. Louis
21.9%
9.4%
4.9%
22.4%
10.0%
5.0%
<40 ppb
Atlanta
55.9%
38.1%
28.2%
56.6%
39.1%
30.0%
Dallas
62.1%
46.3%
33.0%
63.2%
48.9%
38.0%
St. Louis
51.9%
32.2%
22.0%
52.6%
33.7%
23.1%
<50 ppb
Atlanta
81.4%
70.5%
62.9%
82.2%
72.2%
66.1%
Dallas
87.0%
78.8%
71.2%
87.9%
81.5%
75.8%
St. Louis
78.3%
63.7%
53.2%
79.1%
66.1%
56.3%
<60 ppb
Atlanta
94.6%
90.6%
87.2%
95.0%
92.1%
90.4%
Dallas
97.7%
95.5%
93.0%
98.0%
96.7%
94.7%
St. Louis
93.1%
87.1%
83.0%
93.6%
88.9%
85.6%
3D-169
-------
>20% FEV1 Reduction - One Day
m >15% FEVl Reduction - One Day
¦ >10% FEVl Reduction - One Day
>20% FEVl Reduction - Two Days
¦ >15% FEVl Reduction - Two Days
¦ >10% FEVl Reduction - Two Days
Risk Contribution from 7-hr Average Exposures
Two Decrements - Current Standard (70 ppb)
MSS Model Risk Approach
100%
100%
Risk Contribution from 7-hr Average Exposures
One Decrement - Current Standard (70 ppb)
MSS Model Risk Approach
80%
c
.2 70%
+•>
J 60%
¦£ 50%
O
-------
3D.3.4.2.4 Influence of Ventilation Rate in Lung Function Risk Estimates
A second important variable used to estimate lung function risk in both the E-R function
and MSS model is the ventilation rate. Recall that the E-R function approach uses a threshold
value for EVR to designate whether an individual is at moderate or greater exertion (EVR >17.32
± 1.25 L/min-m2). Technically, while any 7-hr average O3 exposure can potentially lead to a lung
function decrement using the E-R function approach, a lung function decrement is only
calculated when individuals are at or above their designated EVR value and when it occurs
simultaneously with their daily maximum 7-hr average O3 exposure. This is not the case with the
MSS model lung function risk approach; both O3 exposure and ventilation rate are considered
cumulatively over time (among other influential MSS model variables) and neither of which
have a designated level or duration to attain.
Because of this notable difference in the MSS model approach, we first visually
evaluated the relationship between the time-series of O3 exposure and ventilation rate (as
represented by EVR), along with the simultaneous occurrence of lung function decrements
calculated by the MSS model. Of particular interest to this evaluation was whether the pattern of
these variables was correlated, and more importantly, how increases in both exposure and
ventilation rates eventually corresponded to increases in the magnitude of the FEVi decrement.
As was done above to evaluate the risk contribution from selected O3 exposure levels, we used
the same APEX simulation of 10,000 children (and 2016 air quality) which output the hourly
time series of O3 exposure, EVR, and MSS model calculated FEVi decrements for each
simulated individual. The initial goal was to observe how the MSS model functions and see if
there were general patterns in the O3 exposure, EVR, and FEVi reductions.
Figure 3D-17 illustrates an example of the estimated hourly time-series of O3 exposure,
EVR, and FEVi decrement for a child considering 2016 air quality adjusted to just meet the
current standard in the Atlanta study area. As shown here (and among all other visualizations of
children we reviewed that had a lung function decrement of interest), the O3 exposure and EVR
are well correlated with subsequent occurrence of a lung function decrement. With increasing O3
exposures and breathing rates, there is an increase in the magnitude of the FEVi reduction and,
following a continuous episode of high exposure along with elevated breathing rate, a lung
function decrement of interest is attained (Figure 3D-17).
3D-171
-------
EVR,dFEV1
26
12
hour
Exp (ppb)
EVR (l/min/m2) dFEV1 (%)
Figure 3D-17. Example time-series of O3 exposures, EVR, and FEVi reductions estimated
using MSS model for a simulated child in the Atlanta study area, based on a
day in a year (2016) of the current standard air quality scenario.
When considering the influence of EVR in isolation, we can discern how, in many
instances, the MSS model risk estimates are greater than those estimated using the E-R function
approach when both use a generally similar O3 exposure profile (i.e., any level, though using
different averaging times). Recall, that the E-R function risk is only estimated for those attaining
moderate or greater exertion levels EVR >17.32 ± 1.25 L/mm-m2. While there is likely a
minimum EVR in the MSS model, considering both the level and duration, that would lead to
lung function decrements, that minimum is not explicitly defined as it is in the E-R function risk
approach.
Note, the E-R method is a fairly direct translation of the controlled human exposure study
data to exposure-dependent response probabilities, particularly considering the strict adherence
to exertion level needed for a response. As described above (section 3D.3.4.2.2), there is low
statistical uncertainty associated with the risk estimates. We already know that relatively lower
ventilation rates substantially influence MSS model risk estimates based on analyses described in
the 2014 HREA (Chapter 6, Tables 6-9 and 6-10). In that assessment, when restricting the MSS
3D-172
-------
results to when an 8-hr EVR of at least 13 L/min-m2 was not achieved by simulated individuals
(at that time, the threshold for moderate exertion threshold), about 40 to 50% fewer simulated
individuals were estimated to experience a lung function decrement, a result better aligned with
the E-R function risk results.
As a second evaluation of the influence of EVR, a similar evaluation of the degree to
which low-level EVRs influence MSS risk estimates was performed here. We limited the
evaluation to a single year (2016) of air quality adjusted to just meet the current standard in the
three study areas and using the same simulation of 10,000 children described above for
generating the hourly data for the MSS model lung function risks. We identified the days when
children were exercising at moderate or greater exertion, i.e., 7-hr average EVR >17.3 L/min-m2
and calculated the percent of children experiencing one or more lung function decrements of
interest (i.e., >10%, >15%, and >20%). Results for each the main body MSS model approach and
the MSS model restricted to children at moderate or greater exertion are presented, along with
results using the E-R function risk approach (Table 3D-69).
The pattern of risk estimates was consistent across the three study areas. Using the
Atlanta study area results as an example, the E-R function risk approach predicts the percent of
children experiencing one or more FEVi decrements >10% to be 2.5%, while the main body
MSS model risk approach predicts 14.6% of children experience the same decrement (Table 3D-
69). When using the MSS model and restricting the risk results to children at moderate or greater
exertion, 8.5% of children experiencing one or more FEVi decrements >10%. Even with this
adjustment for moderate or greater exertion, this indicates an uncertainty in the MSS model
estimates such that the MSS model is potentially overpredicting risks for children by about a
factor of three or more, particularly when considering the larger lung function decrements.
Note that the MSS model used an age-term that extends information developed for 18-
year olds to estimate lung function risks in the simulated children (ages 5 to 18). The age term at
age 18 is at a maximum value and progressively decreases in value (and hence risk) through age
35 adults (the age range of study subjects in the controlled human exposure studies). Therefore,
use of this extrapolation might also contribute to some of the noted differences in the two risk
approaches because this approach uses the maximum possible observed value. However, the
2013 ISA indicates children's responses to O3 exposure are similar to those for young adults
(2013 ISA, section 8.3.1.1), which lends credence to use of the age-term extrapolation in the
MSS model and, overall, supports the application of E-R risk approach for children.
3D-173
-------
Table 3D-69. Percent of children experiencing one or more FEVi decrements >10,15, 20%,
2016 air quality adjusted to just meet the current standard, considering
influence of moderate or greater exertion level in the MSS model and E-R
function risk approaches.
Study Area
(2016 AQ)
Lung Function
Risk Approach
Exertion Level
(L/min-m2)
% of Children Experiencing at least One Decrement
FEV1 >10%
FEV1 >15%
FEV1 >20%
Atlanta
(worst year)
E-R functionA
>17.32 ±1.25
2.5%
0.6%
0.2%
MSS modelB
Any
14.6%
5.1%
2.1%
MSS modelc
>17.3
8.5%
3.5%
1.6%
Dallas
(best year)
E-R function
>17.32 ±1.25
2.1%
0.5%
0.2%
MSS model
Any
13.3%
4.1%
1.7%
MSS model
>17.3
7.9%
2.9%
1.3%
St. Louis
(worst year)
E-R function
>17.32 ±1.25
2.8%
0.8%
0.3%
MSS model
Any
16.3%
5.8%
2.5%
MSS model
>17.3
9.7%
3.9%
1.9%
A The median (50th percentile) E-R function used to generate the main body results (Table 3D-40).
B Sensitivity results for 10,000 children simulation (Table 3D-67).
c Screened sensitivity results for only those children achieving moderate or greater exertion level.
3D.3.4.2.5 Influence of MSS Model Variability Parameter Settings
In this evaluation, we considered how the values for two MSS model variables, U and vi,
influenced the calculated lung function decrements. These variables are used to account for inter-
and intra-individual variability, respectively, in the estimated lung function decrements. Both of
these variables are in the 2012 MSS model (McDonnell et al., 2012; and used in the 2014 HREA
to estimate lung function risk) and the 2013 MSS model (McDonnell et al. (2013); and used for
the current assessment). However, because the 2013 MSS model adjusted the structure of the
intra-individual variability to now include two explanatory variables, vi and V2, the interpretation
of vi has changed (McDonnell et al. (2013)).85 Each of these variables is discussed in greater
detail below.
The first variable is U, a random variable meant to address inter-individual variability not
accounted for by the other MSS model variables. The impact of the values assigned to U is
apparent simply from its roles in the MSS model calculations, as an exponent to the natural
logarithm used in estimating the base AFEVi (Equation 3D-15) and within the calculation of an
intra-individual variance term s (Equation 3D-16). Based on these roles, it is likely that high
85 Effectively, in McDonnell et al. (2012), intra-personal variability (e) was solely represented by 17. In McDonnell
et al. (2013), the intra-personal variability (e) is represented by 17 + v: x (e,:' x \ lijk) (see Equation 3D-16).
According to McDonnell et al. (2013), this was done such that "individuals experiencing small effects either
because exposure was low, or because of demographics (e.g. older age) or because baseline value of
responsiveness (U,) was small would be expected to exhibit less variability in response than those with larger
mean responses."
3D-174
-------
values for U would likely yield high lung function decrements, particularly for instances of high
O3 exposures that occur simultaneously with high ventilation rates over a few to several hours.
Note that when comparing the variance of U in the 2012 MSS model versus the 2013 MSS
model, its standard error is greater (0.917 versus 1.123) in the most recent model.
For this evaluation, we used the same APEX simulation (as described in the prior section)
of 10,000 children (and 2016 air quality), which output the hourly time series of O3 exposure,
EVR, and FEVi decrements for each simulated individual. We screened the output data for
simulated individuals having experienced each of the three FEVi decrements of interest (i.e.,
>10%, >15%, and >20%) and occurring on separate days. We recognize there are a limited
number of children experiencing lung function decrements on multiple days per year (e.g., Table
3D-57), particularly when considering the highest lung function decrement, but we were
interested in controlling for the influence personal variables might have on the magnitude of each
of the decrements. We identified a few simulated individual children having multiple decrements
at each level of interest, and first visually compared how variation in the value assigned the U
variable appeared to influence the magnitude of FEVi reduction for the subset of these simulated
children that had similar time-series of O3 exposure and ventilation rate.
As an example, Figure 3D-18 illustrates the estimated hourly time-series of O3 exposure,
EVR, and FEVi decrement for two simulated children (top and bottom panels) that differ in the
value they were assigned for the U variable (both runs used the 2016 year for the current
standard air quality scenario for the Atlanta study area). In both cases, the O3 exposure and EVR
are well correlated for each child prior to the occurrence of a lung function decrement, consistent
with the controlled human exposure study data. With increasing magnitude of the FEVi
decrement (Figure 3D-18, from left to right panels) there is also progressively higher exposures
and breathing rates, each occurring as peak events that continue over a few to several hours just
prior to eliciting the indicated FEVi decrement of interest. In general, for each of the three
magnitudes of FEVi decrement, the time-series of O3 exposures appears similar for the two
simulated children - a consistently high exposure maintained across multiple hours for all of the
instances where a lung function decrement occurred, with the highest decrement achieved when
exposures were also highest. There is however a recognizable difference in the EVR time-series
for the two simulated children. For the first child, with the lower value for the U variable (top
panels, Figure 3D-18), the peak of the EVR time-series is broader, that is, longer in duration,
than it is for the peak EVR for the second child (bottom panels, Figure 3D-18). The peak EVR
for the second child (that has the higher value for the U variable) is similar in magnitude to that
for the first child, but it does not persist over as long a duration. The figure illustrates this
difference for three magnitudes of decrement (10%, 15% and 20%) in vertical pairs of panels
from left to right, with the pairs of upper and lower panels differing only by the value of
3D-175
-------
parameter U. Specifically, the lower panel child achieves the same decrement as the upper panel
child but while having a lower average EVR for the event.
The first simulated child (upper panel) has a U value of 0.963, which falls within one
standard deviation of the distribution of U (i.e., 1J has a standard error of 1.123, Table 3D-21).
The second (lower panel) simulated child has a U value of 1.78, within the U variable
parameterization (i.e., within 2 standard deviations), but is nearly twice that of the first child.
Specifically, while the second child has a lower overall "normalized dose" (i.e., C x V^6 in
Equation 3D-12) over a similar exposure duration as the first child, the similar risk result is
likely a result of the second child being assigned a higher value for U. This higher value of U
yielded lung function decrements for the second child similar in magnitude to that predicted for
the first simulated child even though the second child had relatively lower doses than the first
child for each of the three days.
The second variable, vi, a constant, is used on the calculation of the intra-individual
variance term s (Equations 3D-16). In evaluating the MSS model parameters used for this
assessment, McDonnell et al. (2013) notes the estimate of vi is consistent with intra-subject
FEVi variability observed in the forced air trials and below threshold O3 exposures. The variable
vi could be interpreted to represent a separate, non-ozone related contribution to response
variability in the study observations. This suggests the use of non-zero values for vi, as is
provided by McDonnell et al. (2013) in MSS model applications (and as was done for the current
risk analysis), could lead to a greater number of simulated individuals at or above the lung
function decrements (in particular the lowest decrement) and a greater portion of that risk would
be attributed to relatively lower exposure levels and ventilation rates, when compared to
simulation results having vi set as zero.
We evaluated the influence that the value of vi has on risk estimates. A new APEX
simulation was required for this evaluation. All model settings were the same as was done for
generating the main assessment results reported in section 3D.3.3.2, except for varying the value
of vi (the MSS model default vi value is 9.112, a new simulation had vi set as zero) Again, both
simulations were performed for 10,000 children in three study areas (Atlanta, Dallas, and St.
Louis) for a simulated year using 2016 air quality adjusted just meet the current standard. Results
for this evaluation are presented in Table 3D-70.
3D-176
-------
EVR.dFEVI
26
24
22
20
18
16
14
12
6
4 """
2 V
0 2 4 6
8 10 12 14 16 18 20 22 24 0 2
hour
Exp (ppb)
EVR (l/min/m2) dFEV1 (%)
6 8 10 12 14 16 18 20 22 24 0 2
hour
Exp (ppb)
EVR (l/min/m2) dFEV1 (%)
> 8 10 12 14 16 IS 20 22 24
hour
- — E*p (ppb)
- EVR (l/min/ft»2> dFEV1 f%)
EVR.dFEVI
26
20
18
16
0 2 4 6
10 12 14 16
hour
Exp (ppb)
Exp (ppb)
Exp
/»
eo
f\
1 *
' Jr-\
70
60
\\\
50
\L
40
30
20
10
0
• EVR (l/min/m2) dFEVI (%)
- - EVR (l/min/m2) dFEVI (%)
10 12 14 16 18 20 22 24
hour
Exp (ppb)
EVR (l/min/m2) dFEVI (%)
Figure 3D-18. Time-series of O3 exposures, EVR, and FEVi reductions of 10% (left panel), 15% (middle panel), and 20%
(right panel) estimated using MSS model for two simulated children (interpersonal variability parameter U =
0.963, top panel; U= 1.78, bottom panel) in the Atlanta study area on three days in a year (2016) of the current
air quality scenario.
3D-177
-------
For each value of vi, there were small differences in estimated risk across the three study
areas. However, setting the vi to zero (compared to the value reported by McDonnell et al.,
2013) resulted in a decrease in the percent of children experiencing lung function decrements of
>10, >15, and >20% of about 35, 22, and 20% (regardless of study area). This reduction in risk is
similar in magnitude to that resulting from excluding the contribution from low-level exposures
(section 3D.3.4.2.3) and not using ventilation rates below moderate or greater exertion (section
3D.3.4.2.4) when estimating lung function decrements using the MSS model.
Table 3D-70. Percent of children experiencing one or more FEVi decrements >10,15, 20%,
2016 air quality adjusted to just meet the current standard, considering the
setting of variability parameter, vi, in the MSS model.
Study Area
MSS Model Parameter
Setting A
Decrement (FEVi Reduction)
>10%
>15%
>20%
Atlanta
vi = 9.112 (default)
15%
5.1%
2.1%
vi = 0
9.7%
3.9%
1.7%
Dallas
vi= 9.112 (default)
13%
4.1%
1.7%
vi = 0
7.9%
3.2%
1.3%
St. Louis
vj — 9.112 (default)
16%
5.8%
2.5%
vi = 0
11%
4.6%
2.1%
A See Table 3D-21 and Equation 3D-16.
3D.4 REFERENCES
Abt. (2013). Updated Analysis for Estimating Ozone Exposure-Response Function (Revised
v.3). Technical memorandum from Jin Huang and Jacky Haskell to John Langstaff
(WA3-03, Contract No EP-D-08-100).
Adams, WC (2000). Ozone dose-response effects of varied equivalent minute ventilation rates. J
Expo Anal Environ Epidemiol 10(3): 217-226.
Adams, WC (2002). Comparison of chamber and face-mask 6.6-hour exposures to ozone on
pulmonary function and symptoms responses. Inhal Toxicol 14(7): 745-764.
Adams, WC (2003). Comparison of chamber and face mask 6.6-hour exposure to 0.08 ppm
ozone via square-wave and triangular profiles on pulmonary responses. Inhal Toxicol
15(3): 265-281.
Adams, WC (2006). Comparison of chamber 6.6-h exposures to 0.04-0.08 PPM ozone via
square-wave and triangular profiles on pulmonary responses. Inhal Toxicol 18(2): 127-
136.
3D-178
-------
Adams, WC and Ollison, WM (1997). Effects of prolonged simulated ambient ozone dosing
patterns on human pulmonary function and symptomatology Air & Waste Management
Association Pittsburgh, PA.
Ainsworth, BE, Haskell, WL, Herrmann, SD, Meckes, N, Bassett, DR, Jr., Tudor-Locke, C,
Greer, JL, Vezina, J, Whitt-Glover, MC and Leon, AS (2011). 2011 Compendium of
Physical Activities: A Second Update of Codes and MET Values. Med Sci Sports Exerc
43(8): 1575-1581.
Aitken ML, Franklin JL, Pierson DJ, Schoene RB. (1986) Influence of body size and gender on
control of ventilation. J Appl Physiol 60:1894-1899.
American Petroleum Institute (1997). Sensitivity Testing of pNEM/03 Exposure to Changes in
the Model Algorithms. Health and Environmental Sciences Department. API Publication
Number FR2. March 1997.
Arjomandi, M, Balmes, JR, Frampton, MW, Bromberg, P, Rich, DQ, Stark, P, Alexis, NE,
Costantini, M, Hollenbeck-Pringle, D, Dagincourt, N and Hazucha, MJ (2018).
Respiratory Responses to Ozone Exposure. MOSES (The Multicenter Ozone Study in
Older Subjects). Am J Respir Crit Care Med 197(10): 1319-1327.
Bennett, DH, Fisk, W, Apte, MG, Wu, X, Trout, A, Faulkner, D and Sullivan, D (2012).
Ventilation, temperature, and HVAC characteristics in small and medium commercial
buildings in California. Indoor Air 22(4): 309-320.
Box, GEP and Tiao, GC (1973). Bayesian Inference in Statistical Analysis. Wesley Publishing
Co, Wiley Classic Library.
Brown, JS, Bateson, TF and McDonnell, WF (2008). Effects of exposure to 0.06 ppm ozone on
FEV1 in humans: a secondary analysis of existing data. Environ Health Perspect 116(8):
1023-1026.
Burmaster, DE (1998). Lognormal distributions for skin area as a function of body weight. Risk
Anal 18(1): 27-32.
Che, WW, Frey, HC and Lau, AK (2014). Assessment of the effect of population and diary
sampling methods on estimation of school-age children exposure to fine particles. Risk
Anal 34(12): 2066-2079.
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.
3D-179
-------
Davidian, M and Giltinan, DM (2003). Nonlinear models for repeated measurement data: An
overview and update. J Agr Biol Env Stat 8(4): 387.
Downey, N, Emery, C, Jung, J, Sakulyanontvittaya, T, Hebert, L, Blewitt, D and Yarwood, G
(2015). Emission reductions and urban ozone responses under more stringent US
standards. Atmos Environ 101: 209-216.
Esmail, S, Bhambhani, Y and Brintnell, S (1995). Gender differences in work performance on
the Baltimore Therapeutic Equipment work simulator. Am J Occup Ther 49(5): 405-411.
Folinsbee, LJ, Horstman, DH, Kehrl, HR, Harder, S, Abdul-Salaam, S and Ives, PJ (1994).
Respiratory responses to repeated prolonged exposure to 0.12 ppm ozone. Am J Respir
CritCare Med 149(1): 98-105.
Folinsbee, LJ, McDonnell, WF and Horstman, DH (1988). Pulmonary function and symptom
responses after 6.6-hour exposure to 0.12 ppm ozone with moderate exercise. JAPCA
38(1): 28-35.
Ford, ES, Heath, GW, Mannino, DM and Redd, SC (2003). Leisure-time physical activity
patterns among US adults with asthma. Chest 124(2): 432-437.
Frey, HC. (2014). Letter from Dr. H. Christopher Frey, Chair, Clean Air Scientific Advisory
Committee, to Administrator Gina McCarthy. Re: Health Risk and Exposure Assessment
for Ozone (Second External Review Draft - February 2014) EPA-CASAC-14-005. Office
of the Administrator, Science Advisory Board Washington, DC. Available at:
https://nepis. epa.gov/Exe/ZyPURL. cgi?Dockey=P 100JR81. txt.
Glasgow, G and Smith, A (2017). Uncertainty in the estimated risk of lung function decrements
owing to ozone exposure. J Expo Sci Environ Epidemiol 27(5): 535-538.
Glen, G, Smith, L, Isaacs, K, McCurdy, T and Langstaff, J (2008). A new method of longitudinal
diary assembly for human exposure modeling. J Expo Sci Environ Epidemiol 18(3): 299-
311.
Goldstein, B, Tardiff, R, Hoffnagle, G and Kester, R (1992). Valdez Air Health Study: Summary
Report. Prepared for Alyeska Pipeline Service Company. Anchorage, AK.
Graham, SE and McCurdy, T (2004). Developing meaningful cohorts for human exposure
models. J Expo Anal Environ Epidemiol 14(1): 23-43.
Graham, SE and McCurdy, T (2009). Appendix A: Revised ventilation rate (VE) equations for
use in inhalation-oriented exposure models Metabolically Derived Human Ventilation
Rates: A Revised Approach Based Upon Oxygen Consumption Rates. U.S. EPA, Office
of Research Development, National Center for Environmental Assessment. Washington,
DC EPA/600/R-06/129F.
Hartwell, TD, Clayton, CA, Michie, RM, Jr., Whitmore, RW, Zelon, HS, Whitehurst, DA and
Akland, GG (1984). Study of Carbon Monoxide Exposure of Residents of Washington,
3D-180
-------
D.C. and Denver, Colorado. Prepared for the U.S. Environmental Protection Agency.
Research Triangle Park, NC.
Haugen, HA, Melanson, EL, Tran, ZV, Kearney, JT and Hill, JO (2003). Variability of measured
resting metabolic rate. Am J Clin Nutr 78(6): 1141-1145.
Henderson, R. (2006). Letter from Dr. Rogene Henderson, Chair, Clean Air Scientific Advisory
Committee to Honorable Stephen L. Johnson, Administrator, US EPA. Re: CASAC Peer
Review of the Agency's 2nd Draft Ozone Staff Paper October 24, 2006. EPA-CASAC-
07-001. Office of the Administrator, Science Advisory Board U.S. EPA HQ, Washington
DC. Available at: https://nepis.epa.gov/Exe/ZyPURL.cgi?Dockey=P1000W07.txt.
Henry, CJ (2005). Basal metabolic rate studies in humans: measurement and development of
new equations. Public Health Nutr 8(7a): 1133-1152.
Horstman, DH, Folinsbee, LJ, Ives, PJ, Abdul-Salaam, S and McDonnell, WF (1990). Ozone
concentration and pulmonary response relationships for 6.6-hour exposures with five
hours of moderate exercise to 0.08, 0.10, and 0.12 ppm. Am Rev Respir Dis 142(5):
1158-1163.
IOM (2005). Dietary Reference Intakes for Energy, Carbohydrate, Fiber, Fat, Fatty Acids,
Cholesterol, Protein, and Amino Acids. The National Academies Press. Washington, DC.
Available at: https://www.nap.edU/catalog/l0490/dietary-reference-intakes-for-energy-
carbohydrate-fiber-fat-fatty-acids-cholesterol-protein-and-amino-acids.
Isaacs, K, Glen, G, McCurdy, T and Smith, L (2007). Modeling energy expenditure and oxygen
consumption in human exposure models: accounting for fatigue and EPOC. Journal of
Exposure Science and Environmental Epidemiology.
Isaacs, K, McCurdy, T, Glen, G, Nysewander, M, Errickson, A, Forbes, S, Graham, S, McCurdy,
L, Smith, L, Tulve, N and Vallero, D (2013). Statistical properties of longitudinal time-
activity data for use in human exposure modeling. J Expo Sci Environ Epidemiol 23(3):
328-336.
Isaacs, K and Smith, L (2005). New Values for Physiological Parameters for the Exposure
Model Input File Physiology.txt. Memorandum submitted to the U.S. Environmental
Protection Agency under EPA Contract EP-D-05-065. NERL WA 10. Alion Science and
Technology. Available in the 2009 SO2 REA, Appendix B at:
https://www3.epa.gov/ttn/naaqs/standards/so2Zs so2 cr rea.html.
Johnson, T (1984). Study of Personal Exposure to Carbon Monoxide in Denver, Colorado.
Prepared for U.S. Environmental Protection Agency, Environmental Monitoring Systems
Laboratory. Research Triangle Park, NC.
Johnson, T (1989). Human activity patterns in Cincinnati, Ohio [final report], EN-6204. Electric
Power Research Institute. Palo Alto, CA.
3D-181
-------
Johnson, T (2002). A guide to selected algorithms, distributions, and databases used in exposure
models developed by the Office of Air Quality Planning and Standards [revised draft],
U.S. Environmental Protection Agency. Research Triangle Park, NC. Available at:
http://www.epa.gov/ttn/fera/data/human/report052202.pdf.
Johnson, T, Capel, J and Wijnberg, L (1986). Selected Data Analyses Relating to Studies of
Personal Carbon Monoxide Exposure in Denver and Washington, DC. Prepared for U.S.
Environmental Protection Agency. Environmental Monitoring Systems Laboratory,
Research Triangle Park, NC.
Johnson, T, Pakrasi, A, Wisbeth, A, Meiners, G and Ollison, W (1995). Ozone exposures Within
Motor Vehicles - Results of a Field Study in Cincinnati, Ohio Proceedings 88th annual
meeting and exposition of the Air & Waste Management Association San Antonio, TX.
Kim, CS, Alexis, NE, Rappold, AG, Kehrl, H, Hazucha, MJ, Lay, JC, Schmitt, MT, Case, M,
Devlin, RB, Peden, DB and Diaz-Sanchez, D (2011). Lung function and inflammatory
responses in healthy young adults exposed to 0.06 ppm ozone for 6.6 hours. Am J Respir
Crit Care Med 183(9): 1215-1221.
Klepeis, NE, Tsang, AM and Behar, JV (1995). Analysis of the national human activity pattern
survey (NHAPS) respondents from a standpoint of exposure assessment. Final Report.
EPA/600/R-96/074. Prepared for U.S. Environmental Protection Agency, National
Exposure Research Laboratory. Las Vegas, NV. Available at:
http://exposurescience. org/pub/reports/NHAPS Reportl.pdftt... .Local
Knowledge Networks (2009). Field report: National-scale activity survey (NSAS). Research
Triangle Institute. Research Triangle Park, NC.
Lagus Applied Technology (1995). Air change rates in non-residential buildings in California,
California Energy Commission. July 1995. Contract 400-91-034. Sacramento, CA.
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.
Lee, K, Vallarino, J, Dumyahn, T, Ozkaynak, H and Spengler, JD (1999). Ozone decay rates in
residences. J Air Waste Manag Assoc 49(10): 1238-1244.
Leonard, WR, Levy, SB, Tarskaia, LA, Klimova, TM, Fedorova, VI, Baltakhinova, ME,
Krivoshapkin, VG and Snodgrass, JJ (2014). Seasonal variation in basal metabolic rates
among the Yakut (Sakha) of Northeastern Siberia. Am J Hum Biol 26(4): 437-445.
Liu, LJ, Box, M, Kalman, D, Kaufman, J, Koenig, J, Larson, T, Lumley, T, Sheppard, L and
Wallace, L (2003). Exposure assessment of particulate matter for susceptible populations
in Seattle. Environ Health Perspect 111(7): 909-918.
Lunn, D, Jackson, C, Best, N and Thomas, A (2012). The BUGS Book - A Practical Introduction
to Bayesian Analysis. CRC Press: Chapman and Hall.
3D-182
-------
Mansfield, C, Houtven, G, Johnson, FR and Yang, J-C (2009). Environmental Risks and
Behavior: Do children spend less time outdoors when ozone pollution is high? . ASSA
annual meeting, January 5, 2009. Update of Houtven et al. (2003) using the OAB CHAD
data set and related to Mansfield et al. (2006).
McCurdy, T (2000). Conceptual basis for multi-route intake dose modeling using an energy
expenditure approach. J Expo Anal Environ Epidemiol 10(1): 86-97.
McDonnell, WF, 3rd, Chapman, RS, Leigh, MW, Strope, GL and Collier, AM (1985).
Respiratory responses of vigorously exercising children to 0.12 ppm ozone exposure. The
American review of respiratory disease 132(4): 875-879.
McDonnell, WF, Abbey, DE, Nishino, N and Lebowitz, MD (1999). Long-term ambient ozone
concentration and the incidence of asthma in nonsmoking adults: the AHSMOG Study.
Environ Res 80(2 Pt 1): 110-121.
McDonnell, WF, Horstman, DH, Hazucha, MJ, Seal, E, Jr., Haak, ED, Salaam, SA and House,
DE (1983). Pulmonary effects of ozone exposure during exercise: Dose-response
characteristics. J Appl Physiol (1985) 54(5): 1345-1352.
McDonnell, WF, Kehrl, HR, Abdul-Salaam, S, Ives, PJ, Folinsbee, LJ, Devlin, RB, O'Neil, JJ
and Horstman, DH (1991). Respiratory response of humans exposed to low levels of
ozone for 6.6 hours. Arch Environ Health 46(3): 145-150.
McDonnell, WF, Stewart, PW and Smith, MV (2007). The temporal dynamics of ozone-induced
FEV1 changes in humans: an exposure-response model. Inhal Toxicol 19(6-7): 483-494.
McDonnell, WF, Stewart, PW and Smith, MV (2010). Prediction of ozone-induced lung function
responses in humans. Inhal Toxicol 22(2): 160-168.
McDonnell, WF, Stewart, PW and Smith, MV (2013). Ozone exposure-response model for lung
function changes: an alternate variability structure. Inhal Toxicol 25(6): 348-353.
McDonnell, WF, Stewart, PW, Smith, MV, Kim, CS and Schelegle, ES (2012). Prediction of
lung function response for populations exposed to a wide range of ozone conditions.
Inhal Toxicol 24(10): 619-633.
Ogden, CL, Lamb, MM, Carroll, MD and Flegal, KM (2010). Obesity and Socioeconomic Status
in Children and Adolescents: United States, 2005-2008. NCHS Data Brief. Number 51.
Available at: https://files.eric.ed.gov/fulltext/ED530165.pdf.
Roth Associates (1988). LA_partl and LA_part2 (A Study of Activity Patterns Among a Group
of Los Angeles Asthmatics) Electric Power Research Institute.
Samet, JM. (2009). Letter from Jonathan Samet, Chair, Clean Air Scientific Advisory
Committee's (CASAC), to Administrator Lisa P. Jackson. RE: CASAC Review of EPA's
Risk and Exposure Assessment (REA) to Support the Review of the SO2 Primary
National Ambient Air Quality Standards: Second Draft. May 18th, 2009. EPA-CASAC-
3D-183
-------
09-007. U.S. EPA Washington DC. Available at:
https://yosemite.epa.gov/sab/sabproduct.nsf/D21C 10017'5A2BE99852575BB00452632/S
File/EPA-CASA C-09-00 7-unsigned, pdf.
Santuz, P, Baraldi, E, Filippone, M and Zacchello, F (1997). Exercise performance in children
with asthma: is it different from that of healthy controls? Eur Respir J 10(6): 1254-1260.
SAS. (2017). Base SAS® 9.4 Procedures Guide, Seventh Edition. Available at:
http://documentation, sas. com/api/collections/pgmmvacdc/9.4/docsets/proc/content/proc.p
df?locale=en#nameddest=bookinfo. Also used were SAS/ETS 14.2 User's Guide, edited
by A. Baxter, E. Huddleston and SAS/QC 14.2 User's Guide, edited by A. Baxter, V.
Clark, E. Huddleston, S. Prabhu, R. Rodriguez, D. Sawyer, J. Simmons.
Schelegle, ES, Morales, CA, Walby, WF, Marion, S and Allen, RP (2009). 6.6-hour inhalation of
ozone concentrations from 60 to 87 parts per billion in healthy humans. Am J Respir Crit
Care Med 180(3): 265-272.
Schofield, WN (1985). Predicting basal metabolic rate, new standards and review of previous
work. Hum Nutr Clin Nutr 39 Suppl 1: 5-41.
Settergren, SK, Hartwell, TD and Clayton, CA (1984). Study of Carbon Monoxide Exposure of
Residents of Washington, DC: Additional Analyses. Prepared for U.S. Environmental
Protection Agency. Environmental Monitoring Systems Laboratory. Research Triangle
Park, NC. Settergren S.K., Hartwell T.D. and Clayton C.A.
Shendell, DG, Winer, AM, Weker, R and Colome, SD (2004). Evidence of inadequate
ventilation in portable classrooms: results of a pilot study in Los Angeles County. Indoor
Air 14(3): 154-158.
Simon, H, Baker, KR and Phillips, S (2012). Compilation and interpretation of photochemical
model performance statistics published between 2006 and 2012. Atmos Environ 61: 124-
139.
Spier, CE, Little, DE, Trim, SC, Johnson, TR, Linn, WS and Hackney, JD (1992). Activity
patterns in elementary and high school students exposed to oxidant pollution. Journal of
Exposure Science & Environmental Epidemiology 2(3): 277-293.
Tsang, AM and Klepeis, NE (1996). Descriptive Statistics Tables from a Detailed Analysis of
the National Human Activity Pattern Survey (NHAPS) Data. U.S. Environmental
Protection Agency. Washington, DC. Available at: https://nepis.epa.gov.
Turk, BH, Grimsrud, DT, Brown, JT, Geisling-Sobotka, KL, Harrison, J and Prill, RJ (1989).
Commercial building ventilation rates and particle concentrations. American Society of
Heating, Refrigerating and Air-Conditioning Engineers Journal 1: 422-433.
U.S. Bureau of Labor Statistics (2014). American Time Use Survey User's Guide:
Understanding ATUS 2003 to 2013. US Bureau of Labor Statistics. Washington, DC;
December 2014. . Available at: http://www.bls.gov/tus/atususersguide.pdf.
3D-184
-------
U.S. Census Bureau (2012). Technical documentation - 2010 Census Summary File 1—
Technical Documentation/prepared by the U.S. Census Bureau, Revised 2012. Available
at: http://www.census.gov/prod/cen2010/doc/sfl.pdf. Employment Status from the 5-year
American Community Survey (ACS) data, 2010 U.S. Census American FactFinder.
Available at: http://factfinder2.census.gov/. Commuting times file from U.S. Census data
portal (http://dataferrett.census.gov/), Table P31, variables P031001-P031015.
U.S. Census Bureau (2019). American Housing Survey (AHS) Data. 2015, 2017 data. Available
at: https://www.census.gov/programs-surveys/ahs/data/interactive/ahstablecreator.html.
2011 data. Available at https://www.census.gov/programs-surveys/ahs/data/2011/ahs-
2011-summary-tables/ahs-metropolitan-summary-tables. html.
U.S. DHHS (1999). Promoting physical activity: a guide for community action. Department of
Health and Human Services, Centers for Disease Control and Prevention (see Table 2-1).
Available at: http://www.cdc.gov/nccdphp/dnpa/physical/pdf/PAJntensityJable 2 1.pdf.
U.S. DOT (2012). Bureau of Transportation Statistics, Census Transportation Planning Package,
Part 3-The Journey to Work. Available at: http://transtats.bts.gov/.
U.S. EPA (1986). Air quality criteria for ozone and other photochemical oxidants. Research
Triangle Park, NC. U.S. EPA. EPA-600/8-84-020aF - EPA/600/8-84-020eF. Available
at:
https://ntrl.ntis.gov/NTRL/dashboard/searchResults.xhtml? searchQuery=PB87142956&S
tarDB=GRAHIST.
U.S. EPA (1996). Air Quality Criteria for Ozone and Related Photochemical Oxidants. Volumes
1 to 3. Research Triangle Park, NC. U.S. EPA. EPA/600/P-93/004aF, EPA/600/P-
93/004bF, and EPA/600/P-93/004cF.
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://www.epa.gov/ttn/naaqs/standards/ozoneZs o3 cr cd.html.
U.S. EPA (2007a). Ozone Population Exposure Analysis for Selected Urban Areas. Office of Air
Quality Planning and Standards. Research Triangle Park, NC. U.S. EPA. EPA-452-R-07-
010. Available at: http://www.epa.gov/ttn/naaqs/standards/ozoneZs o3 cr td.html.
U.S. EPA (2007b). Ozone Population Health Risk Assessment for Selected Urban Areas.
Research Triangle Park, NC. EPA Office of Air Quality Planning and Standards. EPA-
452/R-07-009. Available at:
https://www 3. epa.gov/ttn/naaqs/standards/ozone/so3crtd. html.
U.S. EPA (2008). Risk and Exposure Assessment to Support the Review of the NO2 Primary
National Ambient Air Quality Standard. EPA-452/R-08-008a. Office of Air Quality
Planning and Standards. Research Triangle Park, NC. Available at:
https://www 3. epa.gov/ttn/naaqs/standards/nox/snoxcrrea. html.
3D-185
-------
U.S. EPA (2009). Risk and Exposure Assessment to Support the Review of the SO2 Primary
National Ambient Air Quality Standard. Office of Air Quality Planning and Standards.
Research Triangle Park, NC. US EPA. EPA-452/R-09-007. Available at:
https://www3.epa.gov/ttn/naaqs/standards/so2/data/200908S02REAFinalReport.pdf.
U.S. EPA (2010). Quantitative Risk and Exposure Assessment for Carbon Monoxide - Amended.
Office of Air Quality Planning and Standards. Research Triangle Park, NC. U.S. EPA.
EPA-452/R-10-006. Available at: https://www.epa.gov/naaqs/carbon-monoxide-co-
standards-risk-and-exposure-assessments-current-review.
U.S. EPA (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=P100KETF.txt.
U.S. EPA (2014). Health Risk and Exposure Assessment for Ozone. (Final Report). Office of Air
Quality Planning and Standards. Research Triangle Park, NC. U.S. EPA. EPA-452/R-14-
004a. August 2014. Available at:
https://nepis. epa.gov/Exe/ZyPURL. cgi?Dockey=P 100KBUF. txt.
U.S. EPA (2018). Risk and Exposure Assessment for the Review of the Primary National
Ambient Air Quality Standard for Sulfur Oxides. Office of Air Quality Planning and
Standards. Research Triangle Park, NC. U.S. EPA. EPA-452/R-18-003. Available at:
\dtps://www .epa.gov/sites/production/files/2018-05/documents/primaryso2 naaqs_-
final rea - m ay 20I8.pdf.
U.S. EPA (2019a). Air Pollutants Exposure Model Documentation (APEX, Version 5) Volume I:
User's Guide. Office of Air Quality Planning and Standards. Research Triangle Park, NC.
EPA-452/R-19-005a. Available at: https://www.epa.gov/fera/apex-user-guides.
U.S. EPA (2019b). Air Pollutants Exposure Model Documentation (APEX, Version 5) Volume
II: Technical Support Document. Office of Air Quality Planning and Standards. Research
Triangle Park, NC. EPA-452/R-19-005b. Available at: https://www .epa.gov/fera/apex-
user-guides.
U.S. EPA (2019c). The Consolidated Human Activity Database (CHAD). Documentation and
User's Guide. Research Triangle Park, NC. US EPA. EPA-452/B-19-001. Available at:
https://www.epa.gov/healthresearch/consolidated-human-activity-database-chad-use-
human-exposure-and-health-studies-and.
U.S. EPA (2019d). 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.
U.S. EPA (2020). 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.
3D-186
-------
University of Michigan (2016). "Panel Study of Income Dynamics." from
http://psidonline. isr. umich. edu/Studies. aspx.
van Gent, R, van der Ent, CK, van Essen-Zandvliet, LE, Rovers, MM, Kimpen, JL, de Meer, G
and Klijn, PH (2007). No differences in physical activity in (un)diagnosed asthma and
healthy controls. Pediatr Pulmonol 42(11): 1018-1023.
van Ooijen, AM, van Marken Lichtenbelt, WD, van Steenhoven, AA and Westerterp, KR (2004).
Seasonal changes in metabolic and temperature responses to cold air in humans. Physiol
Behav 82(2-3): 545-553.
Whitfield, R, Biller, W, Jusko, M and Keisler, J (1996). A Probabilistic Assessment of Health
Risks Associated with Short- and Long-Term Exposure to Tropospheric Ozone. Argonne,
IL: Argonne National Laboratory.
WHO (2008). WHO/IPCS Harmonization Project Document No. 6. Part 1: Guidance Document
on Characterizing and Communicating Uncertainty in Exposure Assessment.
International Programme on Chemical Safety. World Health Organization. Geneva,
Switzerland. Available at:
http://www.who.int/ipcs/methods/harmonization/areas/exposure/en/.
Wiley, J, Robinson, J, Cheng, Y, Piazza, T, Stork, L and Pladsen, K (1991a). Study of Children's
Activity Patterns. Final Report under contract no A733-149. Prepared for California Air
Resources Board. Research Division. Sacramento, CA.
Wiley, JA, Robinson, JP, Piazza, T, Garrett, K, Cirksena, K, Cheng, YT and Martin, G (1991b).
Activity patterns of California residents Final report. California Environmental Protection
Agency. Berkeley, CA.
https://ntrl.ntis.gov/NTRL/dashboard/searchResults.xhtml? searchQuery=PB94108719.
Williams, R, Rea, A, Vette, A, Croghan, C, Whitaker, D, Stevens, C, McDow, S, Fortmann, R,
Sheldon, L, Wilson, H, Thornburg, J, Phillips, M, Lawless, P, Rodes, C and Daughtrey, H
(2009). The design and field implementation of the Detroit Exposure and Aerosol
Research Study. J Expo Sci Environ Epidemiol 19(7): 643-659.
Williams, R, Suggs, J, Creason, J, Rodes, C, Lawless, P, Kwok, R, Zweidinger, R and Sheldon,
L (2000). The 1998 Baltimore particulate matter epidemiology-exposure study: Part 2-
Personal exposure assessment associated with an elderly study population. J Expo Anal
Environ Epidemiol 10(6): 533-543.
Williams, R, Suggs, J, Rea, A, Leovic, K, Vette, A, Croghan, C, Sheldon, L, Rodes, C,
Thornburg, J, Ejire, A, Herbst, M and Sanders, W, Jr. (2003a). The Research Triangle
Park particulate matter panel study: PM mass concentration relationships. Atmos Environ
37(38): 5349-5363.
Williams, R, Suggs, J, Rea, A, Sheldon, L, Rodes, C and Thornburg, J (2003b). The Research
Triangle Park particulate matter panel study: modeling ambient source contribution to
personal and residential PM mass concentrations. Atmos Environ 37(38): 5365-5378.
3D-187
-------
Williams, RW, Wallace, LA, Suggs, JC, Evans, EG, Creason, JP, Highsmith, VR, Sheldon, LS,
Rea, AW, Vette, AF, Zweidinger, RB, Leovic, KW, Norris, GA, Landis, MS,
HowardReed, C, Stevens, C, Conner, TL, Rodes, CE, Lawless, PA, Thornburg, J, Liu,
LS, Kalman, D, Kaufman, J, Koenig, JQ, Larson, TL, Lumley, T, Sheppard, L, Brown,
K, Suh, H, Wheeler, A, Gold, D, Koutrakis, P and Lippmann, M (2001). Preliminary
particulate matter mass concentrations associated with longitudinal panel studies:
assessing human exposures of high risk subpopulations to particulate matter. EPA/600/R-
01/086. National Exposure Research Laboratory, Office of Research Development, U.S.
EPA. Research Triangle Park, NC.
Xue, J, McCurdy, T, Spengler, J and Ozkaynak, H (2004). Understanding variability in time
spent in selected locations for 7-12-year old children. J Expo Anal Environ Epidemiol
14(3): 222-233.
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.
3D-188
-------
APPENDIX 3D, ATTACHMENT 1:
ESTIMATING U.S. CENSUS TRACT LEVEL ASTHMA PREVALENCE (2013-2017)
OVERVIEW
This attachment describes the development of the 2013-2017 census tract-level asthma
prevalence file used by EPA's Air Pollution Exposure Model (APEX) to identify individuals
with asthma during exposure model simulations. The approach used to estimate the APEX file
four basic steps: 1) processing National Health Interview Survey (NHIS) regional asthma
prevalence data, 2) processing U.S. Census poverty/income status data, and 3) combining the
two sets considering variables known to influence asthma (e.g., age, sex, poverty status, U.S.
region) to estimate asthma prevalence stratified by age and sex for all U.S. Census tracts, and 4)
the NHIS regionally derived data were adjusted to account for state level asthma prevelance data
obtained from the Behavioral Risk Factor Surveillance System (BRFSS). Details regarding the
data sets and the processing approaches used are provided below.
GENERAL HISTORY
The current NHIS data processing approach is in part based on work originally performed
by Cohen and Rosenbaum (2005) and then revised and extended by U.S. EPA (2014, 2018).
Briefly, Cohen and Rosenbaum (2005) calculated asthma prevalence for children aged 0 to 17
years for each age, sex, and four U.S. regions using 2003 NHIS survey data.1 The regions
defined by the NHIS were 'Midwest', 'Northeast', 'South', and 'West'. The asthma prevalence
was defined as the probability of a 'Yes' response to the question "EVER been told that [the
child] had asthma?"2 among those persons that responded either 'Yes' or 'No' to this question.3
The responses were weighted to take into account the complex survey design of the NHIS.4
Standard errors and confidence intervals for the prevalence were calculated using a logistic
model (PROC SURVEY LOGISTIC). A scatterplot technique (LOESS smoother) was applied to
smooth the prevalence curves across ages and used to compute the standard errors and
1 The National Health Interview Survey (NHIS) is the principal source of information on the health of the civilian
noninstitutionalized population of the United States and is one of the major data collection programs of the
National Center for Health Statistics (NCHS) which is part of the Centers for Disease Control and Prevention
(CDC). See https://www.cdc.gov/nchs/nhis/data-questionnaires-documentation.htm for data and documentation.
2 The response was recorded as variable "CASHMEV" in the downloaded dataset. Data and documentation are
available at http://www. cdc.gov/nchs/nhis/quest_data_related_l99 7Jorward. htm.
3 If there were another response to this variable other than "yes" or "no" (i.e., refused, not ascertained, don't know,
and missing), the NHIS surveyed individual was excluded from the analysis data set.
4 In the SURVEY LOGISTIC procedure, the variable "WTF SC" was used for weighting, "PSU" was used for
clustering, and "STRATUM" was used to define the stratum.
3D-Attachmentl -1
-------
confidence intervals for the smoothed prevalence estimates. Logistic analysis of the raw and
smoothed prevalence curves showed statistically significant differences in prevalence by gender
and region, supporting their use as stratification variables in the final data set (Cohen and
Rosenbaum, 2005). These smoothed prevalence estimates were then used as an input to APEX to
estimate air pollutant exposure in children with asthma (U.S. EPA 2007; 2008; 2009).
For the 2014 O3 REA (U.S. EPA, 2014), we updated the asthma prevalence database used
by APEX by combining several years of NHIS survey data (2006-2010). Asthma prevalence for
children (by age year) was estimated as described above and, for this update, we also included an
estimate of asthma prevalence for adults. In addition, two sets of asthma prevalence for each
adults and children were estimated. The first data set, as was done previously, was based on
responses to the question "EVER been told that [the child/adult] had asthma". A second data set
was developed using the probability of a 'Yes' response to a question that followed those that
answered 'Yes' to the first question regarding ever having asthma, specifically, do those persons
"STILL have asthma?".5 Further, in addition to the nominal variables region and sex, the asthma
prevalence were stratified by a income/poverty threshold (i.e., whether the family income was
below or at/above the US Census estimate of poverty level for the given year). These 2006-2010
asthma prevalence data were then linked to 2000 U.S. Census tract level income/poverty
threshold probabilities, also stratified by age (section 5C-5 of Appendix 5C, US EPA, 2014).
Staff considered the variability in population exposures to be better represented when accounting
for and modeling these newly refined attributes of this at-risk population. This is was done
because of the 1) significant observed differences in asthma prevalence by age, sex, region, and
poverty status, 2) the variability in the spatial distribution of poverty status across census tracts,
stratified by age, and 3) the potential for spatial variability in local scale ambient concentrations.
And finally, asthma prevalence files used by APEX for the most recent SO2 REA
(Appendix E of U.S. EPA, 2018) were updated in a similar manner using data that reasonably
bounded the exposure assessment period of interest (2011-2015) and, as was done for the 2014
O3 REA, linked the asthma prevalence to the 2010 U.S census tract income to poverty ratio
probabilities. The approach to update the asthma prevalence used for the current O3 REA
analyses follows the same approach used previously, although now employs an adjustment to
account for local more asthma prevalence information at the state level, rather than relying solely
on the regional data. This is described in the fours steps that follow below.
Step 1: NHIS Data Set Description and Processing
5 The response was recorded as variable "CASSTILL" for children and "AASSTILL" for adults in the respective
downloaded datasets. Ultimately, the asthma prevalence used by APEX was based on this variable rather than
those using the data for those individuals responding "Yes" to "Ever" having asthma.
3D-Attachmentl -2
-------
The objective of this processing step is to estimate asthma prevalence for children and
adults considering several influential variables. First, raw 2013-2017 data and associated
documentation were downloaded from the Center for Disease Control (CDC) and Prevention's
NHIS website.6 The 'Sample Child' and 'Sample Adult' files were selected because of the
availability of person-level attributes of interest within these files, i.e., age in years ('age_p'), sex
('sex'), U.S. geographic region ('region'), coupled with the response to questions of whether or
not the surveyed individual ever had and still has asthma. In total, five years of survey data were
used, comprising nearly 60,000 children and 165,000 adults for years 2013-2017 (Table 1).
Information regarding personal and family income and poverty ranking are also provided
by the NHIS in additional survey files. Data files ('INCIMPx.dat') are available for every survey
year, each containing either the actual response for the desired financial variable (where provided
by survey participant) or the imputed value.7 For this current analysis, the ratio of family
income-to-poverty was provided as a continuous variable ('POVRATI3') and used to develop a
nominal variable for this evaluation: either the survey participant was below or above a selected
family income-to-poverty ratio threshold. This was done to be consistent with data generated as
part of the next data set processing step, i.e., developing a database containing the census tract
level family income-to-poverty ratio probabilities, stratified by age (see Step 2 below).
When considering the number of stratification variables used in the development of the
asthma prevalence file (i.e., age years and sex), the level of asthma prevalence (8%, on average),
and the distribution of family income-to-poverty ratios among the surveyed population (12%, on
average), sample size was an important motivation for aggregating the adult data into age groups.
When considering the adult data, there were insufficient numbers of persons available to stratify
the data by single age years (for some ages there were no survey persons). Therefore, the adult
survey data were grouped into the following age groups: ages 18-24, 25-34, 35-44, 45-54, 55-64,
65-74, and, >75.8 To increase the number of persons within the age, sex, and four region
groupings of our characterization of 'below poverty', the family income-to-poverty ratio
threshold was selected as <1.5, thus including persons that were within 50% above the threshold.
For individuals containing the imputed family income information, typically there were 5
estimated values. If the mean of the 5 imputed values were <1.5, the person's family income was
6 Data and documentation are available at http://www.cdc.gov/nchs/nhis/quest_data_related_1997_forward.htm (for
2013-2015, accessed April 11, 2017; for 2016-2017 accessed March 11, 2019).
7 Financial information was not collected from all persons; therefore, the NHIS provides imputed data. Details into
the available variables and imputation method are provided with each year's data set. For example, see "Multiple
Imputation of Family Income and Personal Earnings in the National Health Interview Survey: Methods and
Examples" at https://www.cdc.gov/nchs/data/nhis/tecdocl5.pdf.
8 These same age groupings were used to create the companion file containing the census tract level family income-
to-poverty ratio probabilities (Step 2).
3D-Attachmentl -3
-------
categorized 'below' the poverty threshold; if the mean of the 5 values were >1.5, the person's
family income was categorized 'above' the poverty threshold.
These processed person-level income files were then merged with the ' Sample Adult' and
'Sample Child' files using the 'HHX' (a household identifier), 'FMX' (a family identifier), and
'FPX' (an individual identifier) variables. Note, all persons within the 'Sample Adult' and
' Sample Child' files had corresponding financial survey data.
As was done for previous asthma prevalence data analysis, two asthma survey response
variables were of interest in this analysis and were used to develop the two separate prevalence
data sets for each children and adults. The response to the first question "Have you EVER been
told by a doctor or other health professional that you [or your child] had asthma?" was recorded
as variable name 'CASHMEV' for children and 'AASMEV' for adults. Only persons having
responses of either 'Yes' or 'No' to this question were retained to estimate the asthma
prevalence. This assumes that the exclusion of those responding otherwise, i.e., those that
'refused' to answer, instances where it was "not ascertained', or the person 'does not know',
does not affect the estimated prevalence rate if either 'Yes' or 'No' answers could actually be
given by these persons. There were very few persons providing an unusable response (Table 1),
thus the above assumption is reasonable. A second question was asked as a follow to persons
responding "Yes" to the first question, specifically, "Do you STILL have asthma?" and noted as
variables 'CASSTILL' and 'AASSTILL' for children and adults, respectively. Again, while only
persons responding 'Yes' and 'No' were retained for further analysis, the representativeness of
the screened data set is assumed unchanged from the raw survey data given the few persons in
each survey year having unusable data.
Table 1. Number of total surveyed persons from NHIS (2013-2017) sample adult and child
files and the number of those responding to asthma survey questions.
Children
2013
2014
2015
2016
2017
TOTAL
All Children
12,860
13,380
12,291
11,107
8,845
58,483
Yes/No to Ever Have Asthma
12,851
13,366
12,281
11,098
8,832
58,428
Yes/No to Still Have Asthma
12,844
13,359
12,269
11,087
8,823
58,382
Adults
All Adults
34,557
36,697
33,672
33,028
26,742
164,696
Yes/No to Ever Have Asthma
34,525
36,667
33,641
33,007
26,720
164,560
Yes/No to Still Have Asthma
34,498
36,615
33,614
32,959
26,681
164,367
Logistic Models
As described in the previous section, four person-level analytical data sets were created
from the raw NHIS data files, generally containing similar variables: a 'Yes' or 'No' asthma
3D-Attachmentl -4
-------
response variable (either 'EVER' or 'STILL'), an age (or age group for adults), their sex ('male'
or 'female'), US geographic region ('Midwest', 'Northeast', 'South', and 'West'), and poverty
status ('below' or above'). One approach to calculate prevalence rates and their uncertainties for
a given sex, region, poverty status, and age is to calculate the proportion of 'Yes' responses
among the 'Yes' and 'No' responses for that demographic group, appropriately weighting each
response by the survey weight. This simplified approach was initially used to develop 'raw'
asthma prevalence rates however this approach may not be completely appropriate. The two
main issues with such a simplified approach are that the distributions of the estimated prevalence
rates would not be well approximated by normal distributions and that the estimated confidence
intervals based on a normal approximation would often extend outside the [0, 1] interval. A
better approach for such survey data is to use a logistic transformation and fit the model:
Prob (asthma) = exp(beta) / (1 + exp(beta)),
where beta may depend on the explanatory variables for age, sex, poverty status, or region. This
is equivalent to the model:
Beta = logit {prob (asthma)} = log {prob (asthma) / [1 prob (asthma)]}.
The distribution of the estimated values of beta is more closely approximated by a normal
distribution than the distribution of the corresponding estimates of Prob (asthma). By applying a
logit transformation to the confidence intervals for beta, the corresponding confidence intervals
for Prob (asthma) will always fall within [0, 1], Another advantage of the logistic modeling is
that it can be used to compare alternative statistical models, e.g., as models where the prevalence
probability depends upon age, region, poverty status, and sex, or on age, region, poverty status
but not sex.
In earlier analyses using the NHIS asthma prevalence data, a variety of logistic models
were developed and evaluated for use in estimating asthma prevalence, where the transformed
probability variable beta is a given function of age, gender, poverty status, and region (Cohen
and Rosenbaum, 2005; U.S. EPA, 2014). The SAS procedure SURVEYLOGISTIC was used to
fit the various logistic models, taking into account the NHIS survey weights and survey design
(using both stratification and clustering options), as well as considering various combinations of
the selected explanatory variables.
As an example, Table 2 lists the models fit and their log-likelihood goodness-of-fit
measures using the 'Sample Child' data set and for the "STILL" asthma response variable using
the 2013-2017 NHIS data. A total of 32 logistic models were fit, depending on the inclusion of
selected explanatory variables and how age was considered in the model. The 'Strata' column
lists the eight possible stratifications: no stratification, stratified by sex, by region, by poverty
status, by region and sex, by region and poverty status, by sex and poverty status, and by region,
3D-Attachmentl-5
-------
gender and poverty status. For example, "5. region, sex" indicates that separate prevalence
estimates were made for each combination of region and gender. As another example, "2. sex"
means that separate prevalence estimates were made for each sex, so that for each sex, the
prevalence is assumed to be the same for each region. Note the prevalence estimates are
independently calculated for each stratum. The 'Description' column of Table 2 indicates how
beta depends upon the age:
Linear in age Beta = a + /? x age, where a and ft vary with strata
Quadratic in age Beta = a + /? x age + y x age2 where a /?and yvary with strata
Cubic in age Beta = a + /? x age y x age2 8 x age3 where aft, y and 8
vary with the strata
f(age) Beta = arbitrary function of age, with different functions for
different strata
The category f(age) is equivalent to making age one of the stratification variables, and is
also equivalent to making beta a polynomial of degree 17 in age (since the maximum age for
children is 17), with coefficients that may vary with the strata. The fitted models are listed in
order of complexity, where the simplest model (model 1) is a non-stratified linear model in age
and the most complex model (model 32) has a prevalence that is an arbitrary function of age,
sexr, poverty status, and region. Model 32 is equivalent to calculating independent prevalence
estimates for each of the 288 combinations of age, sex, poverty status, and region.
Table 2 also includes the -2 Log Likelihood statistic, a goodness-of-fit measure, and the
associated degrees of freedom (DF), which is the total number of estimated parameters. Any two
models can be compared using their -2 Log Likelihood values: models having lower values are
preferred. If the first model is a special case of the second model, then the approximate statistical
significance of the first model is estimated by comparing the difference in the -2 Log Likelihood
values with a chi-squared random variable having r degrees of freedom, where r is the difference
in the DF (hence a likelihood ratio test). For all pairs of models from Table 2, all the differences
in the -2 Log Likelihood statistic are at least 50,000 and thus are significant at p-values well
below 1 percent. Based on its having the lowest -2 Log Likelihood value, the last model fit
(model 32: retaining all explanatory variables and using f(agej) was preferred and used to
estimate the asthma prevalence in the prior analyses9 as well as employed for this 2013-2017
NHIS data analysis.
9 Similar results were obtained when estimating prevalence using the 'EVER' have asthma variable as well as when
investigating model fit using the adult data sets. In the Cohen and Rosenbaum (2005) analysis, adult data were not
used and the family income-to-poverty ratio was not a variable in their models. Also, because age was a
3D-Attachmentl -6
-------
Table 2. Logistic models and model fit statistics for estimating child asthma prevalence
using the "STILL" asthma response variable from 2013-2017 NHIS data.
Model
Description
Strata
- 2 Log Likelihood
DF
1
°g
t(prob) = linear in age
1.none
209411405
2
2
°g
t(prob) = linear in age
2. gender
208645067
4
3
°g
t(prob) = linear in age
3. region
209056169.8
8
4
°g
t(prob) = linear in age
4. poverty
208433518.7
4
5
°g
t(prob) = linear in age
5. region, gender
208230032
16
6
°g
t(prob) = linear in age
6. region, poverty
207999872.9
16
7
°g
t(prob) = linear in age
7. gender, poverty
207630301.3
8
8
°g
t(prob) = linear in age
8. region, gender, poverty
207046731.4
32
9
2.
°g
t(prob) = quadratic in age
1.none
207554776.3
3
10
2.
°g
t(prob) = quadratic in age
2. gender
206754508.8
6
11
2.
°g
t(prob) = quadratic in age
3. region
207092990.7
12
12
2.
°g
t(prob) = quadratic in age
4. poverty
206568831.2
6
13
2.
°g
t(prob) = quadratic in age
5. region, gender
206177195.9
24
14
2.
°g
t(prob) = quadratic in age
6. region, poverty
205966568.6
24
15
2.
°g
t(prob) = quadratic in age
7. gender, poverty
205719195.5
12
16
2.
°g
t(prob) = quadratic in age
8. region, gender, poverty
204888997.5
48
17
3.
°g
t(prob) = cubic in age
1.none
207244848.3
4
18
3.
°g
t(prob) = cubic in age
2. gender
206429982.6
8
19
3.
°g
t(prob) = cubic in age
3. region
206770493.7
16
20
3.
°g
t(prob) = cubic in age
4. poverty
206240699
8
21
3.
°g
t(prob) = cubic in age
5. region, gender
205817245.3
32
22
3.
°g
t(prob) = cubic in age
6. region, poverty
205532902.7
32
23
3.
°g
t(prob) = cubic in age
7. gender, poverty
205380882.1
16
24
3.
°g
t(prob) = cubic in age
8. region, gender, poverty
204406907.3
64
25
4.
°g
t(prob) = f(age)
1.none
206929745.9
18
26
4.
°g
t(prob) = f(age)
2. gender
205902376.7
36
27
4.
°g
t(prob) = f(age)
3. region
205961955.1
72
28
4.
°g
t(prob) = f(age)
4. poverty
205783757.8
36
29
4.
»g
t(prob) = f(age)
5. region, gender
204430849.5
144
30
4.
»g
t(prob) = f(age)
6. region, poverty
204133603.6
144
31
4.
°g
t(prob) = f(age)
7. gender, poverty
204565028.6
72
32
4.
»g
t(prob) = f(age)
8. region, gender, poverty
201725493.2
288
categorical variable in the adult data sets in U.S. EPA (2014, 2018) and analyses conducted here, it could only be
evaluated using f(age_group).
3D-Attachmentl -7
-------
The SURVEYLOGISTIC procedure produces estimates of the beta values and their 95%
confidence intervals for each combination of age, region, poverty status, and gender. By
applying the inverse logit transformation,
Prob (asthma) = exp( beta) / (1 + exp(beta) ),
one can convert the beta values and associated 95% confidence intervals into predictions and
95% confidence intervals for the prevalence. The standard error for the prevalence was estimated
as:
StdError {Prob (asthma)} = StdError (beta) x exp(- beta) / (1 + exp(beta) )2,
which follows from the delta method (i.e., a first order Taylor series approximation).
Estimated asthma prevalence using this approach and termed here as 'unsmoothed' are
provided in the supplement at the end of this document. Graphical representation is provided in a
series of figures incorporating the following variables:
• Region
• Gender
• Age (in years) or Age_group (age categories)
• Poverty Status
• Prevalence = predicted prevalence
• SE = standard error of predicted prevalence
• LowerCI = lower bound of 95% confidence interval for predicted prevalence
• UpperCI = upper bound of 95% confidence interval for predicted prevalence
A series of plots are provided per figure that vary by the four regions and two income-to-
poverty ratios. Historically, we have used the prevalence results based on the 'STILL' have
asthma variable. Supplemental Figures S-l through S-4 show the estimated prevalence for
children and adults by age (or age-group), startified by gender. Data used for each figure/plot (as
well as plots for the 'EVER' variable) can be provided upon request.
Loess Smoother
The estimated prevalence curves show that the prevalence is not necessarily a smooth
function of age. The linear, quadratic, and cubic functions of age modeled by
SURVEYLOGISTIC were identified as a potential method for smoothing the curves, but they
did not provide the best fit to the data. One reason for this might be due to the attempt to fit a
global regression curve to all the age groups, which means that the predictions for age A are
3D-Attachmentl-8
-------
affected by data for very different ages. A local regression approach that separately fits a
regression curve to each age A and its neighboring ages was used, giving a regression weight of
1 to the age and lower weights to the neighboring ages using a tri-weight function:
Weight = {1 - [ \age - A \ / q J 3}, where \ age -A\ < = q.
The parameter q defines the number of points in the neighborhood of the age A. Instead
of calling q the smoothing parameter, SAS defines the smoothing parameter as the proportion of
points in each neighborhood. A quadratic function of age to each age neighborhood was fit
separately for each gender and region combination. These local regression curves were fit to the
beta values, the logits of the asthma prevalence estimates, and then converted them back to
estimated prevalence rates by applying the inverse logit function exp(beta) / (1 + exp(beta)). In
addition to the tri-weight variable, each beta value was assigned a weight of
1 / [std error (beta)]2, to account for their uncertainties.
In this application of LOESS, weights of 1 / [std error (beta)]2 were used such that a2 =
1. The LOESS procedure estimates a2 from the weighted sum of squares. Because it is assumed
a2 = 1, the estimated standard errors are multiplied by 1 / estimated a and adjusted the widths of
the confidence intervals by the same factor.
There are several potential values that can be selected for the smoothing parameter; the
optimum value was determined by evaluating three regression diagnostics: the residual standard
error, normal probability plots, and studentized residuals. To generate these statistics, the LOESS
procedure was applied to estimated smoothed curves for beta, the logit of the prevalence, as a
function of age, separately for each region, gender, and poverty classification. For the children
data sets, curves were fit using the choices of 0.4, 0.5, 0.6, 0.7, 0.8, 0.9, and 1.0 for the
smoothing parameter. This selected range of values was bounded using the following
observations. With only 18 points (i.e., the number of single year ages for children), a smoothing
parameter of 0.2 cannot be used because the weight function assigns zero weights to all ages
except age A, and a quadratic model cannot be uniquely fit to a single value. A smoothing
parameter of 0.3 also cannot be used because that choice assigns a neighborhood of 5 points only
(0.3 x 18 = 5, rounded down), of which the two outside ages have assigned weight zero, making
the local quadratic model fit exactly at every point except for the end points (ages 0, 1, 16 and
17). Usually one uses a smoothing parameter below 1 so that not all the data are used for the
local regression at a given x value. Note also that a smoothing parameter of 0 can be used to
generate the raw, unsmoothed, prevalence. The selection of the smoothing parameter used for the
adult curves would follow a similar logic, although the lower bound could effectively be
extended only to 0.9 given the number of age groups. This limits the selection of smoothing
3D-Attachmentl -9
-------
parameter applied to the two adult data sets to a value of 0.9, though values of 0.8 - 1.0 were
nevertheless compared for good measure.
The first regression diagnostic used was the residual standard error, which is the LOESS
estimate of a. As discussed above, the true value of a equals 1, so the best choice of smoothing
parameter should have residual standard errors as close to 1 as possible. For children 'EVER'
having asthma and when considering the best models (of the 112 possible, those having
0.95
-------
Probability Plot for <<(u.*F £e*.*J
Iwiisfe \Utr Abe-', rf
1'«iS-Faa»le-Abc.'ey&',Ht\ljr,el
WrU-M»J*-Ab<*^e%ti;yLrv«l
Figure 2. Studentized residuals versus model predicted betas generated using a logistic
model and the 'STILL' prevalence data, smoothing set to 0.5 and 0.8 for children
(left) and adults (right), respectively.
3D-Attachmentl-11
-------
When considering both children asthma prevalence responses evaluated, the residual
standard error (estimated values for sigma) suggests the choice of smoothing parameter as
varied, ranging from 0.7 to 0.8. The normal probability plots of the studentized residuals suggest
preference for smoothing at or above 0.6. The plots of residuals against smoothed predictions
suggest the choices of 0.4 through 0.6. We therefore chose the final value of 0.6 to use for
smoothing the children's asthma prevalence. For the adults, there were small differences in the
statistical metrics used to evaluate the smoothing. A value of 0.9 was selected for smoothing
based on the above findings and to remain consistent with what was used in the prior analysis
(U.S. EPA, 2014; 2018).
The smoothed asthma prevalence and associated graphical presentation are provided in
Supplemental Figures S-5 through S-8. A similar format to that presented using the non-
smoothed asthma prevalence was followed, and again, only providing the results for children and
adults that reported 'STILL' having asthma.
Step 2: U.S. Census Tract Poverty Ratio Data Set Description and Processing
This section briefly describes the approach used to generate census tract level poverty
ratios for all U.S. census tracts, stratified by age and age groups where available. The following
steps were peformed using data from the 2017 U.S. Census 5-year American Community Survey
(ACS)10 and modified SAS data processing files.11
First, ACS internal point latitudes and longitudes were obtained from the 2017 Gazetteer
files.12 Next, the individual state level ACS sequence files (SF-56) were downloaded,13 retaining
the number of persons across the variable "B17024" for each state considering the appropriate
logical record number.14 The data provided by the B17024 variable is stratified by age or age
groups (ages <5, 5, 6-11, 12-14, 15, 16-17, 18-24, 25-34, 35-44, 45-54, 55-64, 65-74, and >75)
10 https://www.census.gov/newsroom/press-kits/2018/acs-5year. html.
11 ACS file processing code was adapted from ACS 2012 SAS programs and from ACS 2012 SAS Macros available
at http://www2. census.gov/acs2012_5yr/summaryfile/UserTools/SF20125YR_SAS.zip and
http://www2.census.gov/acs2012_5yr/summaryfile/UserTools/SF_All_Macro.sas. These were the same
processing files used for updating the 2011-2015 asthma prevalence data set (US EPA, 2018).
12 Data available at: https://www.census.gov/geographies/reference-files/time-series/geo/gazetteer-files.
13 We used the summary tables (B17024), giving census tract populations by poverty income ratio and age group
downloaded from https://www2.census.gov/programs-surveys/acs/summary Jile/2017/data/5_year_by_state/.
Each state's ACS2017 5-yr table compressed fie was unzipped with the sequence file 56 (SF-56; e20175[state
abbreviationJ0056000.txt) and appropriate geography file (g20175[state abbreviationJ.txt) retained.
14 Variable names (2017 Code Listpdf) are available at https://www.census.gov/programs-surveys/acs/technical-
documentation/summary-file-documentation.html, along with the file for the appropriate logical record number
(A CS_2 017_SF_5 YR_Appendices. xls).
3D-Attachment 1-12
-------
and income/poverty ratios, given in increments of 0.25. We calculated two new variables for
each age using the number of persons from the B17024 stratifications; the fraction of those
persons having poverty ratios <1.5 and > 1.5 by summing the appropriate B17024 variable and
dividing by the total number of persons in that age/age group. Then, the individual state level
geographic data (g20175[xx].txt files) were screened for tract level information using the
"sumlev" variable equal to ' 140'. Also identified was the US Region for each state, consistent
with that used for the NHIS asthma prevalence data.15
Finally, the poverty ratio data were combined with the above described census tract level
geographic data using the "stusab" and "logrecno" variables. Because APEX requires the input
data files to be entirely complete (no missing values), additional processing of the poverty
probability file was needed. For where there was missing tract level poverty information,16 we
substituted an age-specific value using the average for the particular county the tract was located
within, or the state-wide average. The percent of tracts substituted using county averaged values
varied by age group though, on average, was approximately 1.6% of the total tracts (Table 4).
Few tracts in six of the age groups were substituted using state averaged values (in total only 9
tracts had a substitution using state values for one of the age groups). The final output was a
single file containing relevant tract level poverty probabilities (pov_acs2017_5yr.sas7bdat) by
age groups for all U.S. census tracts.
Table 4. Percent of tracts substituted with county average or state average poverty status.
Percent
Substituted
Age Groups (years)
<5
6-11
12-17
18-24
25-34
35- 44
45-54
55-64
65-74
>75
all
Filled using
County Average
1.9%
2.0%
1.9%
1.5%
1.4%
1.4%
1.3%
1.3%
1.6%
1.9%
1.6%
Filled using State
Average
<0.1%
<0.1%
<0.1%
none
<0.1%
none
none
none
none
none
<0.1%
Step 3: Combining Census Tract Poverty Ratios with the NHIS Regional Asthma
Prevalence Data
The two data sets were merged considering the region identifier and stratified by age and
sex. The Census tract-level asthma prevalence data set was calculated using the following
weighting scheme:
15 https://www2.census.gov/geo/pdfs/maps-data/maps/reference/(using file us_regdiv.pdf)
16 Whether there were no data collected by the Census for poverty status or there were no people in an age group is
relatively inconsequential to estimating the exposed people with asthma, particularly considering latter case as no
people in that age group would be modeled by APEX when using the same Census population data set.
3D-Attachmentl -13
-------
Asthmaprevalence=round((pov_prob*prev_belowpov)+((1-pov_prob)*prev abovepov),0.0001);
whereas each U.S. census tract contains a tract-specific poverty-weighted asthma
prevalence, stratified by ages (children 0-17), age groups (adults), and two sexes.
To evaluate the overall accuracy of the Census tract-level estimated asthma prevalence,
we first compared these values with the NHIS national summary data for asthma prevalence
reported for 2013 to 2017.17 According to the CDC, the NHIS are the principal source of
national asthma prevalence data for the US. Note also, the NHIS 2013-2017 raw data was used to
estimate the asthma prevalence for four U.S. regions in step 1 above. The NHIS national
summary data are stratified by two age groups (children and adults) and for the two sexes (male
and female) and were simply averaged across the five years of data available for the comparison.
The Census tract-level estimated asthma prevalence were population-weighed using 2010 U.S.
Census tract population data and aggregated to generate a similar national summary metric (and
also considered data from 2013-2017 in their initial development). Table 5 show reasonable
agreement between the two data sets: where present, the differences between the two data sets
were generally small (< 0.1 percentage points) with the greatest percentage point difference
found for adult females (-0.4 percentage points). The adult asthma prevalence estimated for both
sexes using the Census tract-level was lower than the NHIS reported value, while the children's
asthma prevalence data were generally similar between the two data sets. Overall, this degree of
aggreement was expected given that the 2013-2017 NHIS regional asthma prevalence (stratified
by age, sex, and family income) served as the source for extrapolating asthma prevalence to the
census tract level.
Step 4: Adjusting NHIS Regionally-derived Prevalence Data to Reflect State-level Asthma
Prevalence
We then compared the NHIS Regionally-derived census tract-level estimated asthma
prevalence to the Behavioral Risk Factor Surveillance System (BRFSS),18 an independent source
providing state (and national) data about U.S. residents regarding their chronic health conditions
such as asthma (among other health issues). For this comparison, the BRFSS asthma prevalence
data were available for 2013-2016 and averaged across those four years to obtain a national
summary metric. This BRFSS metric is similar to that calculated using the Census tract-level and
17 Downloaded was Table 4-1, the 2013-2017 NHIS current asthma prevalence percents by age groups and sex
available at https://www.cdc.gOv/asthma/nhis/default.htm#anchor_1524067853614. Accessed 5/7/19.
18 Downloaded was table C2.1 (for each adults and children), the 2013-2016 BRFSS current asthma prevalence
percents by state and sex available at https://www.cdc.gov/asthma/brfss/default.htm. Table CI was also
downloaded to obtain the asthma prevalence for the two age groups not stratified by sex. Accessed 5/3/19.
3D-Attachment 1-14
-------
NHIS asthma prevalence data sets and is provided in Table 5. The asthma prevalence data
reported from BRFSS are consistently greater than that calculated using the Census tract-level
data, particularly when considering adults. Overall, the BRFSS adult asthma prevalence is 1.6
percentage points greater than that estimated using the Census tract-level estimated prevalence,
with the greatest difference observed for the two data sets of 2.8 percentage points observed for
adult females. Asthma prevalence for the two data sets were closer when considering children,
though the Census-tract level estimated data were still consistently lower than the BRFSS
reported values (-0.2 to 0.4 percentage points).
Table 5. Asthma prevalence stratified by two age groups and sex using Census tract-level
estimates, NHIS and BRFSS reported data.
Data Set (years of data)
All Ages,
Both Sexes
Children (<18 years old)
Adults (>18 years old)
all
female
male
all
female
male
NHIS (2013-2017)
7.8%
8.4%
7.2%
9.6%
7.6%
9.6%
5.5%
Census tract-level
estimate
7.6%
8.5%
7.2%
9.7%
7.3%
9.2%
5.3%
BRFSS (2013-2016) A
n/a
8.8%
7.4%
10.1%
8.9%
11.4%
6.3%
A The BRFSS does not have any data for some states, and where represented, not all four years of data were available for those state, n/a
is not available.
It is unlikely that additional data are available for meaningful comparison, certainly not to
the extent to which the NHIS Regionally-derived Census tract-level asthma prevalence is
stratified and also not without inconsistencies in methodology used in their collection and
reporting, if these data do exist at a local level (e.g., county health department data across all US
counties). However, we were concerned with the potential for underestimating asthma
prevalence that is indicated by the comparison of the NHIS Regionally-derived census tract-level
asthma prevalence with the BRFSS data. Note, we used the NHIS 2013-2017 raw data set in Step
1 to serve as the basis for the census tract-level estimated asthma prevalence given its large
sample size for both children and adults and because of the stratification of important influential
variables (i.e., age, sex, family income). Contrary to this, the NHIS data are aggregated to four
US regions and could account for less spatial variability than that provided by the individual
state-level data obtained from BRFSS. With that in mind, we chose to adjust the NHIS-Census
tract-level data (upwards or downwards) based on the percent difference observed between a
population weighted state level aggregate of the census tract level data and the BRFSS state-
level asthma prevalence (Table 6) and was calculated as follows:
State Adjustment Factor = (NHIS CeflSUSregional prevalence — BRFSSstate prevalence)/BRFSSstate prevalence
3D-Attachmentl -15
-------
Table 6. Factors used to adjust NHIS Regionally-derived census tract-level asthma
prevalence and based on BRFSS state level data.
State
Adjustment Factor - Children A
Adjustment Factor - Adults A
male
female
male
female
Alabama
0.510
0.413
0.356
0.399
Alaska
0
0
0.076
0.296
ArizonaB
0.157
0.058
0.199
0.237
Arkansas
0
0
0.343
0.299
California
0.099
0.199
-0.023
0.108
Coloroado
0
0
0.165
0.179
Conneticut
0.114
0.153
0.220
0.365
Delaware
0
0
0.286
0.476
Florida
-0.124
-0.11
0.155
0.136
Georgia
0.234
0.015
0.183
0.320
Hawaii
0.59
1.002
0.277
0.355
Idaho
0
0
0.182
0.171
Illinois
-0.016
-0.151
0.044
0.134
Indiana
-0.107
0.030
0.239
0.388
Iowa
0
0
0.044
0.049
Kansas
0.140
0.035
0.111
0.176
Kentucky
0.076
-0.016
0.701
0.628
Lousiana
-0.051
-0.174
0.250
0.130
Maine
-0.021
-0.104
0.494
0.478
Maryland
0.200
0.218
0.399
0.399
Massachusetts
0.257
0.061
0.328
0.479
Michigan
0.169
0.036
0.414
0.38
Minnesota
-0.228
-0.059
-0.014
0.069
Mississippi
0.127
-0.026
0.151
0.120
Missouri
0.003
0.226
0.264
0.301
Montana
-0.137
0.107
0.154
0.173
Nebraska
-0.180
-0.210
0.030
0
Nevada
-0.143
0.068
-0.070
0.129
New Hampshire
-0.031
0.009
0.276
0.502
New Jersey
-0.094
0.009
0.052
0.078
New Mexico
0.141
0.208
0.340
0.302
New York
-0.040
-0.024
0.285
0.237
North Carolina
0.171
0.416
0.154
0.225
North Dakota
0
0
0.254
0.132
Ohio
0.024
0.016
0.233
0.332
Oklahoma
0.298
0.065
0.549
0.365
Oregon
-0.047
0.237
0.400
0.443
Pennsylvania
0.137
0.003
0.172
0.357
3D-Attachmentl -16
-------
State
Adjustment Factor - Children A
Adjustment Factor - Adults A
male
female
male
female
Puerto Ricoc
0
0
0
0
Rhode Island
0.083
0.136
0.376
0.447
South Carolina
0
0
0.240
0.252
South Dakota
0
0
0.040
-0.043
Tennessee
0.116
-0.111
0.256
0.368
Texas
-0.034
-0.210
0.068
0.111
Utah
-0.079
-0.032
0.239
0.160
Vermont
-0.114
0.131
0.333
0.453
Virginia
0
0
0.218
0.333
Wash DC
0.389
0.436
0.656
0.577
Washington
-0.108
0.091
0.246
0.294
West Virginia
0.041
-0.032
0.561
0.581
Wisconsin
-0.097
0.232
0.279
0.284
Wyoming
0
0
0.146
0.190
A Values of zero indicate there were no BRFSS data were available, therefore no adjustment was made.
B Data reported for Arizona children in the 2013 BRFSS were atypical: prevelance for females were greater than that of male, having rates
almost opposite that expected. These data were not used to calculate the adjustment factor.
cThe NHIS-Census regional data was not used for estimating asthma prevalence for Puerto Rico, therefore only BRFSS data for the two
age groups and sexes were used.
The adjustment factor was applied to the census tract estimated asthma prevalence considering
the state level information as follows:
PrevalenceAdjusted = NHIS/Censusprevaience + (Adjustmen Factor x NHIS/Censiisprevaience)
By design, the adjustment has better aligned the estimated NHIS Regionally-derived census
tract-level asthma prevalence with the BRFSS reported values at the state and national level
(Table 7). These BRFSS-adjusted census tract-level asthma prevalence data are used for the
APEX simulations and are found within the asthmajprev 1317tract 051319 adjusted, txt file.
For brevity, data are shown only for a few states most relevant to the study areas of interest in the
current O3 exposure and risk analysis.
3D-Attachmentl -17
-------
Table 7. Population-weighted state level asthma prevalence stratified by two age groups and sex: Original census tract-level
estimates based on 2013-2017 NHIS regional prevalence and US Census family income data, 2013-2016 BRFSS
reported prevalence, and BRFSS-adjusted census tract-level estimates used for the APEX asthma prevalence file.
State
Related
Study Area A
Sex
Child Asthma Prevalence
Adult Asthma Prevalence
Census tract-
level estimate
BRFSS state
reported data
Adjusted APEX
prevalence file
Census tract-
level estimate
BRFSS state
reported data
Adjusted APEX
prevalence file
Georgia
Atlanta
female
7.9%
8.1%
8.0%
8.7%
11.4%
11.4%
male
10.0%
12.4%
12.3%
4.7%
5.6%
5.5%
Massachusetts2
Boston
female
7.7%
8.2%
8.1%
9.5%
14.0%
14.0%
male
10.9%
13.7%
13.6%
5.8%
7.7%
7.6%
Texas2
Dallas
female
7.9%
6.2%
6.3%
8.6%
9.5%
9.5%
male
10.0%
9.6%
9.6%
4.7%
5.0%
5.0%
Michigan
Detroit
female
7.1%
7.4%
7.4%
9.7%
13.4%
13.4%
male
9.9%
11.6%
11.6%
5.8%
8.2%
8.2%
Pennsylvania
Philadelphia
female
8.0%
8.0%
8.0%
9.6%
13.0%
13.0%
male
11.0%
12.6%
12.5%
5.8%
6.8%
6.8%
Arizona B
Phoenix
female
5.9%
6.2%
6.2%
9.5%
11.7%
11.7%
male
8.6%
9.9%
9.9%
5.6%
6.8%
6.7%
California
Sacramento
female
5.9%
7.1%
7.0%
9.4%
10.4%
10.4%
male
8.6%
9.4%
9.4%
5.6%
5.5%
5.5%
Missouri
St. Louis
female
7.0%
8.5%
8.5%
9.7%
12.6%
12.6%
male
9.7%
9.7%
9.7%
5.8%
7.3%
7.3%
All US States
female
7.2%
7.4%
7.4%
9.2%
11.4%
11.4%
male
9.7%
10.1%
10.1%
5.3%
6.3%
6.3%
both
8.5%
CO
CO
0s-
*-9
0s-
CO
CO
7.3%
8.9%
8.9%
A Each study area is defined by a Consolidated Statistical Area (CSA) may involve counties from more than one US state. This information is added for relevance to the spatial
scale and not meant to be absolute in defining the prevalence for any of the study areas.
B Data for children were only available for the following years in a few states: 2016 (Arizona), 2015 and 2016 (Massachusetts), 2013-2015 (Texas). Adults based on 2013-2016.
3D-Attachmentl -18
-------
The asthma prevalence estimates vary for the different ages and sexes of children and
adults that reside in each census tract of each study area. We evaluated the spatial distribution of
the asthma prevalence using the specific census tracts that comprise the consolidated statistical
area (CSA) that generally define each study area. We first separated data for children from those
for adults and calculated simple descriptive statistics of asthma prevalence for the tracts,
stratified by sex (Table 8). Consistent with broadly defined national asthma prevalence (e.g.,
Table 3-1 of the draft PA), on average, children have higher estimated rates than adults, male
children have higher rates than female children, and adult females have higher rates than adult
males.
By using age, sex, and family income variables to develop the tract level prevalence, we
also observe that there is spatial variability in the estimated prevalence both within and across
the CSAs. Atlanta, Boston, Detroit, and Philadelphia have some of the highest asthma prevalence
for male children considering most of the statistics with rates as high as 25.5% in one or more
census tracts for males of a given year of age. The Dallas study area exhibits some of the lowest
asthma prevalence when considering adults (both sexes) with rates as low as 3.8% in one or
more tracts for males within a given age group. These summary statistics represent the range of
age- and sex-specific values for the census blocks used in each APEX simulation to estimate the
number of individuals that have asthma.
3D-Attachmentl -19
-------
Table 8. Descriptive statistics for non-population weighted asthma prevalence for children
(ages 5-17) and adults (age >17) using all census tracts from 8 consolidated
statistical areas (CSAs) in the APEX asthma prevalence file (2013-2017).
CSA Name - ID#
(# tracts)
and
Population qroup
Sex
Asthma Prevalence across all ages (or age groups) and census tractsA
Mean
Standard
Deviation
Minimum
Median
95th
percentile
99th
percentile
Maximum
Atlanta-122
(1,077)
adult
female
11.1%
1.8%
7.7%
11.1%
14.0%
15.9%
20.9%
male
5.5%
0.8%
4.3%
5.4%
7.1%
7.5%
7.9%
child
female
9.7%
1.7%
6.5%
9.6%
12.9%
13.9%
15.0%
male
14.1%
1.7%
10.6%
14.0%
16.8%
17.6%
18.3%
Boston-148
(1,753)
adult
female
13.8%
1.8%
10.5%
13.5%
17.3%
20.5%
28.9%
male
7.6%
0.9%
5.4%
7.5%
9.1%
10.0%
12.9%
child
female
9.4%
2.0%
5.6%
9.5%
12.4%
13.5%
17.1%
male
15.4%
2.5%
8.7%
15.1%
19.5%
20.8%
23.4%
Dallas-206
(1,422)
adult
female
9.3%
1.5%
6.5%
9.3%
11.8%
13.5%
16.5%
male
4.9%
0.7%
3.8%
4.9%
6.4%
6.8%
9.7%
child
female
7.6%
1.3%
5.0%
7.4%
10.0%
10.9%
13.5%
male
11.0%
1.4%
8.3%
11.0%
13.2%
13.8%
18.1%
Detroit-220
(1,583)
adult
female
13.3%
2.5%
7.8%
13.4%
17.8%
20.6%
25.6%
male
7.9%
2.2%
1.0%
7.6%
12.4%
14.7%
19.0%
child
female
8.6%
1.5%
6.4%
8.2%
11.6%
12.5%
13.2%
male
13.3%
3.0%
7.7%
12.7%
19.9%
23.6%
25.5%
Philadelphia-
428
(1,725)
adult
female
12.1%
2.3%
8.2%
12.0%
16.4%
19.8%
26.5%
male
6.5%
0.9%
4.6%
6.4%
8.1%
9.0%
11.4%
child
female
9.1%
1.9%
5.6%
9.2%
12.0%
13.1%
15.3%
male
13.6%
2.4%
8.2%
13.3%
17.8%
19.2%
21.1%
Phoenix-429
(988)
adult
female
11.6%
1.6%
8.6%
11.7%
14.4%
16.0%
19.7%
male
7.0%
1.5%
5.1%
7.1%
9.1%
11.7%
16.7%
child
female
7.6%
1.5%
4.6%
8.0%
9.5%
9.6%
9.6%
male
11.5%
1.8%
8.5%
11.6%
14.8%
15.9%
17.1%
Sacramento-
472
(539)
adult
female
10.4%
1.4%
7.7%
10.5%
12.7%
14.0%
16.5%
male
5.7%
1.1%
4.2%
5.9%
7.3%
9.0%
13.6%
child
female
8.5%
1.7%
5.2%
9.0%
10.7%
10.9%
10.9%
male
10.8%
1.7%
8.1%
10.9%
13.7%
14.8%
16.2%
St. Louis-476
(638)
adult
female
11.8%
2.1%
6.8%
11.9%
15.0%
17.4%
21.5%
male
6.5%
1.8%
0.9%
6.5%
9.9%
11.8%
14.5%
child
female
9.2%
2.0%
5.3%
9.1%
12.9%
14.2%
15.6%
male
11.1%
2.4%
6.5%
10.7%
15.9%
19.3%
21.9%
A As described in the text, prevalence is based on single year ages (children) or age group (adults) and sex derived from 2013-2017 CDC
NHIS asthma prevalence and considering U.S. census tract level family income/poverty ratio data. Data presented are not population-
weighted and represent the distribution of applied probabilities used by APEX for tracts having a non-zero population. Note also, upper and
lower percentiles could represent prevalence for a single-year age/sex group residing in a single tract within a study area.
3D-Attachmentl -20
-------
Evaluation of Additional Asthma Prevalence Questions and Responses
To estimate asthma prevalence, we used responses to the question of whether an NHIS
study participant responded 'Yes" to the survey question of 'STILL' having asthma rather than
using the responses to the question of 'EVER' having asthma (with the former being a subset of
the latter group). According to the CDC, lifetime asthma is defined by responding 'Yes' to
"Have you ever been told by a doctor {nurse or other health professional} that you have
asthma?", while current asthma is defined as responding 'Yes' to both the aforementioned and
this subsequent question "Do you still have asthma?".19 Because the exposure and risk analyses
in this review reflect a generally current actualized hypothetical single-year scenario that is not
covering the lifetime of the simulated individuals, the prevalence estimate based on those
participants responding as currently ('STILL') having asthma was deemed most appropriate. We
note that the response of survey particpants who stated they do not still have asthma does not
reflect a doctor's/health professional's diagnosis, thus it is possible there may be individuals in
this group that might actually still have asthma and experience asthma-related health effects,
potentially leading to an underestimate in the asthma prevalence used in our exposure and risk
simulations. Because we used the responses to the "STILL" having asthma question to estimate
prevalence in this assessment, we evaluated additional related questions in the NHIS data to
estimate the magnitude of this potential underestimate in asthma prevalence.
There are two additional questions related to asthma prevalence that are asked of NHIS
survey participants who responded 'Yes' to the 'EVER' having asthma question that could
provide insight into the likelihood that people could 'STILL' have asthma but did not respond
'Yes' to that latter question. The first additional asthma question is, "DURING THE PAST 12
MONTHS, have you had an episode of asthma or an asthma attack?" (i.e., variable 'CASHYR'
or ' AASHYR' for children and adults, respectively); the second is, "DURING THE PAST 12
MONTHS, have you had to visit an emergency room or urgent care center because of asthma?"
(i.e., variable 'CASERYR1' or 'AASERYR1'). We evaluated the responses to all four of these
asthma questions using children's 2017 data set as an example, the results of which are presented
in Table 9.
Most survey participants responded either yes or no to the 'EVER' having asthma
question; those not providing a response were removed from the analysis. There were few
individuals not responding to the question (13 of 8,845), thus it was assumed there would be no
bias to the overall conclusions following their removal. Of the remaining children surveyed,
13.2% (i.e., 1,168 of 8,832) had a doctor/health professional diagnose them as having asthma at
some time in their life, with a majority of those 'EVER' having asthma (63.3%) responding
19 https://www.cdc.gov/asthma/brfss/default.htm.
3D-Attachmentl -21
-------
'Yes' to 'STILL' having asthma. Based on these responses to the 'STILL' having asthma
question, the overall asthma prevalence for children would be estimated as 8.4% (i.e., 739 of
8,832). As mentioned above, it is possible that prevalence is underestimated due to the nature of
the diagnosis (i.e., self assessment) and at most, could be underestimated by a factor of 1.6 (i.e.,
13.2/8.4) if assuming 'EVER' having asthma response was appropriate to use in this assessment.
We suggest solely using this 'EVER' having asthma response would likely bias the prevalence
high based on the below analysis of responses to the two additional asthma questions.
Table 9. Chidren's responses to four questions regarding their asthma status, 2017 NHIS.
Diganosed by a
Doctor as EVER
Having Asthma?
Participant
Reported as
Participant Reported in Past 12
Months Did You Have:
Survey
Participants
(n)
STILL Having
Asthma?
Asthma
Attack?
Asthma-related
ER Visit?
Did not respond
-
-
-
13
No
-
-
-
7,664
No
No
396
No
No
Yes
5
(n=420)
Yes
No
15
Yes
Yes
4
1 don't know
No
No
5
Yes
(n=9)
Yes
No
4
(n=1,168)
No
No
336
No
Yes
22
Yes
Yes
No
248
(n=739)
Yes
Yes
131
I don't know
No
1
I don't know
I don't know
1
Sum of EVER (Y/N), all ages
8,832
There were a few participants (6.5%, 28 of 429) who reported they did not or did not
know they 'STILL' have asthma (note also, an unprofessional diagnosis), but also reported they
had an asthma attack and/or had to be treated by a doctor because of asthma. Based on these
data, asthma prevalence estimated using the response for the 'STILL' having asthma question
alone might be underestimated by about 0.3 percentage points (i.e., 28/8832, the number
reporting asthma attack or ER visit but also reporting "no" for still having asthma divided by
total respondents), such that the overall asthma prevalence for children might be 8.7% rather than
8.4%. This would be with the assumption that the individual has accurately self-diagnosed an
asthma attack, a perhaps reasonable assumption given they had been diagnosed with asthma at
some time in their life. When considering the participants that stated they 'STILL' have asthma,
3D-Attachmentl -22
-------
approximately 54% reported they had an asthma attack and/or had to be treated by a doctor
because of asthma (i.e., 401 of 739). This clearly indicates that when survey participants reported
they 'STILL' have asthma, they are more likely to have asthma attacks/ER visits than those who
do not state they 'STILL' have asthma. An alternative hypothesis is also possible, in that they
could have indicated they still have asthma as a result of the asthma attack/ER visit. Regardless,
the health condition and the adverse response appear to be interrelated.
Additionally, we could assume that all participants that 'EVER and 'STILL' have asthma
(100% rather than the 54% estimated above) would have an asthma attack/ER visit at some time
in their life (and perhaps not just within 12 months). Applying that information to survey
participants who stated they did not 'STILL' have asthma and also report they have experienced
an asthma attack/ER visit, implies that the asthma prevalence derived without these individuals
(i.e., 0.3 percentage points) might be underestimated by a factor of about two. Thus, based on
this analysis and including assumptions made using the responses to the additional questions, it
is possible that asthma prevalence estimated using the 'STILL' variable alone (as was done for
this assessment) could be underestimated by about 0.6 percentage points (i.e., an overall
'current' asthma prevalence for children would be about 9.0% rather than the 8.4% used in the
simulations).
REFERENCES
Cohen J and Rosenbaum A. (2005). Analysis of NHIS Asthma Prevalence Data. Memorandum
to John Langstaff by ICF Incorporated. For US EPA Work Assignment 3-08 under EPA
contract 68D01052. Available in US EPA (2007) Appendix G.
U.S. EPA. (2007). Ozone Population Exposure Analysis for Selected Urban Areas (July 2007).
Office of Air Quality Planning and Standards, Research Triangle Park, NC. EPA-452/R-07-
010. Available at http://epa.g0v/ttn/naaqs/standards/0z0ne/s o3 cr td.html.
U.S. EPA. (2008). Risk and Exposure Assessment to Support the Review of the NO2 Primary
National Ambient Air Quality Standard. Report no. EPA-452/R-08-008a. November 2008.
Available at
http://www.epa.gov/ttn/naaqs/standards/nox/data/20081121 N02R/EA final.pelf.
U.S. EPA. (2009). Risk and Exposure Assessment to Support the Review of the SO2 Primary
National Ambient Air Quality Standard. Report no. EPA-452/R-09-007. August 2009.
Available at
http://www.epa.gov/ttn/naaqs/standards/so2/data/200908S02REAFinalReport.pdf
U.S. EPA. (2014). Health Risk and Exposure Assessment for Ozone, Final Report. Chapter 5
Appendices. Report no. EPA-452/R-14-004c. August 2014. Available at
https://nepis.epa.gOv/Exe/ZyPDF.cgi/P 100KC17.PDF?Dockey=P 100KCI7.PDF.
U.S. EPA. (2018). Risk and Exposure Assessment for the Review of the Primary National
Ambient Air Quality Standard for Sulfur Oxides. Office of Air Quality Planning and
Standards, Research Triangle Park, NC, EPA-452/R-18-003, May 2018. Available at
https://www.epa.gov/sites/production/files/2018-05/documents/primary so2 naaqs -
final rea - may 20I8.pdf.
3D-Attachmentl -23
-------
SUPPLEMENTAL FIGURES S-l to S-4, ASTHMA PREVALENCE NON-SMOOTHED
Figure S-l. Non-smoothed asthma prevalence for children that still have asthma. Above (left panels) and below poverty level
(right panels) for Midwest (top panels) and Northeast (bottom panels) regions.
Raw asthma 'STILL' prevalence rates and confidence intervals-2013-2017
region=Midwest pov_rat=Above Poverty Level
Raw asthma 'STILL' prevalence rates and confidence intervals-2013-2017
region=Midwest pov_rat=Below Poverty Level
9 10 11 12 13 14 15 16 17 0
2 3 4 5 6 7
9 10 11 12 13 14 15 16 17
age
gender Female Male
gender Female Male
Raw asthma 'STILL' prevalence rates and confidence intervals-2013-2017
region=Northeast pov_rat=Above Poverty Level
Raw asthma 'STILL' prevalence rates and confidence intervals-2013-2017
region=Northeast pov_rat=Below Poverty Level
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 0
9 10 11 12 13 14 15 16 17
gender Female Male
age
gender Female
3D-Attachmentl-24
-------
Figure S-2. Non-smoothed asthma prevalence for children that still have asthma. Above (left panels) and below poverty level
(right panels) for South (top panels) and West (bottom panels) regions.
Raw asthma 'STILL' prevalence rates and confidence intervals-2013-2017
region=South pov_rat=Above Poverty Level
Raw asthma 'STILL' prevalence rates and confidence intervals-2013-2017
region=South pov_rat=Below Poverty Level
10 11 12 13 14 15 16 17
gender
- Female Male
gender
age
- Female
Raw asthma 'STILL' prevalence rates and confidence intervals-2013-2017
region=West pov_rat=Above Poverty Level
Raw asthma 'STILL' prevalence rates and confidence intervals-2013-2017
region=West pov_rat=Below Poverty Level
9 10 11 12 13 14 15 16 17 0
gender Female Male
gender
age
- Female
3D-Attachmentl-25
-------
Figure S-3. Non-smoothed asthma prevalence for adults that still have asthma. Above (left panels) and below poverty level
(right panels) for Midwest (top panels) and Northeast (bottom panels) regions.
Raw adult asthma 'STILL' prevalence rates and confidence intervals-2013-2017
region=Midwest pov_rat=Above Poverty Level
Raw adult asthma 'STILL' prevalence rates and confidence intervals-2013-2017
region=Midwest pov_rat=Below Poverty Level
75+ 18-24
gender
- Female Male
gender
- Female Male
Raw adult asthma 'STILL' prevalence rates and confidence intervals-2013-2017
region=Northeast pov_rat=Above Poverty Level
Raw adult asthma 'STILL' prevalence rates and confidence intervals-2013-2017
region=Northeast pov_rat=Below Poverty Level
75+ 18-24
gender
- Female Male
gender Female
3D-Attachmentl-26
-------
Figure S-4. Non-smoothed asthma prevalence for adults that still have asthma. Above (left panels) and below poverty level
(right panels) for South (top panels) and West (bottom panels) regions.
Raw adult asthma 'STILL' prevalence rates and confidence intervals-2013-2017
region=South pov_rat=Above Poverty Level
Raw adult asthma 'STILL' prevalence rates and confidence intervals-2013-2017
region=South pov_rat=Below Poverty Level
75+ 18-24
gender
- Female Male
gender
- Female Male
Raw adult asthma 'STILL' prevalence rates and confidence intervals-2013-2017
region=West pov_rat=Above Poverty Level
Raw adult asthma 'STILL' prevalence rates and confidence intervals-2013-2017
region=West pov_rat=Below Poverty Level
75+ 18-24
gender
- Female Male
gender Female
3D-Attachmentl-27
-------
SUPPLEMENTAL FIGURES S-5 to S-8, ASTHMA PREVALENCE SMOOTHED
Figure S-5. Smoothed asthma prevalence for children that still have asthma. Above (left panels) and below poverty level (right
panels) for Midwest (top panels) and Northeast (bottom panels) regions.
Smoothed asthma 'STILL' prevalence rates and confidence intervals-2013-2017
region=Midwest pov_rat=Above Poverty Level
Smoothed asthma 'STILL' prevalence rates and confidence intervals-2013-2017
region=Midwest pov_rat=Below Poverty Level
9 10 11 12 13 14 15 16 17 0
gender Female Male
Smoothed asthma 'STILL' prevalence rates and confidence intervals-2013-2017
region=Northeast pov_rat=Above Poverty Level
2 3 4 5 6 7
9 10 11 12 13 14 15 16 17
age
gender Female Male
Smoothed asthma 'STILL' prevalence rates and confidence intervals-2013-2017
region=Northeast pov_rat=Below Poverty Level
2 3 4 5 6 7
9 10 11 12 13 14 15 16 17 0
2 3 4 5 6 7
9 10 11 12 13 14 15 16 17
gender Female
gender
age
" Female
3D-Attachmentl-28
-------
Figure S-6. Smoothed asthma prevalence for children that still have asthma. Above (left panels) and below poverty level (right
panels) for South (top panels) and West (bottom panels) regions.
Smoothed asthma 'STILL' prevalence rates and confidence intervals-2013-2017
region=South pov_rat=Above Poverty Level
Smoothed asthma 'STILL' prevalence rates and confidence intervals-2013-2017
region=South pov_rat=Below Poverty Level
9 10 11 12 13 14 15 16 17 0
gender
- Female Male
gender
age
- Female
Smoothed asthma 'STILL' prevalence rates and confidence intervals-2013-2017
region=West pov_rat=Above Poverty Level
Smoothed asthma 'STILL' prevalence rates and confidence intervals-2013-2017
region=West pov_rat=Below Poverty Level
2 3 4 5 6 7
9 10 11 12 13 14 15 16 17 0
gender Female Male
gender
age
- Female
3D-Attachmentl-29
-------
Figure S-7. Smoothed asthma prevalence for adults that still have asthma. Above (left panels) and below poverty level (right
panels) for Midwest (top panels) and Northeast (bottom panels) regions.
Smoothed adult asthma 'STILL' prevalence rates and confidence intervals-2013-2017
region=Midwest pov_rat=Above Poverty Level
Smoothed adult asthma 'STILL' prevalence rates and confidence intervals-2013-2017
region=Midwest pov_rat=Below Poverty Level
75+ 18-24
gender
- Female Male
gender
- Female Male
Smoothed adult asthma 'STILL' prevalence rates and confidence intervals-2013-2017
region=Northeast pov_rat=Above Poverty Level
Smoothed adult asthma 'STILL' prevalence rates and confidence intervals-2013-2017
region=Northeast pov_rat=Below Poverty Level
75+ 18-24
gender
- Female Male
gender Female
3D-Attachment 1-30
-------
Figure S-8. Smoothed asthma prevalence for adults that still have asthma. Above (left panels) and below poverty level (right
panels) for South (top panels) and West (bottom panels) regions.
Smoothed adult asthma 'STILL' prevalence rates and confidence intervals-2013-2017
region=South pov_rat=Above Poverty Level
Smoothed adult asthma 'STILL' prevalence rates and confidence intervals-2013-2017
region=South pov_rat=Below Poverty Level
75+ 18-24
gender
- Female Male
gender
- Female Male
Smoothed adult asthma 'STILL' prevalence rates and confidence intervals-2013-2017
region=West pov_rat=Above Poverty Level
Smoothed adult asthma 'STILL' prevalence rates and confidence intervals-2013-2017
region=West pov_rat=Below Poverty Level
75+ 18-24
gender Female Male
gender Female
3D-Attachmentl-31
-------
APPENDIX 3D, ATTACHMENT 2:
ICF TECHNICAL MEMO: IDENTIFICATION OF SIMULATED INDIVIDUALS AT
MODERATE EXERTION
3D-Attachment2-1
-------
MEMORANDUM
To: John Langstaff and Stephen Graham, EPA
From: Jeanne Luh, Graham Glen, and Chris Holder, ICF
Date: March 26, 2019
Re: Identification of Simulated Individuals at Moderate Exertion
1. Introduction
Under Work Assignment 4-55 of U.S. Environmental Protection Agency (EPA) Contract EP-W-
12-010, the EPA Work Assignment Manager (WAM) asked ICF (hereafter "us", "we", etc.) to
evaluate the approach used in the Air Pollutants Exposure Model (APEX; U.S. EPA, 2017a and
2017b) to identify when simulated individuals are at moderate exertion on average during any 8-
hour exposure period. APEX uses the ModEVR8 parameter, where EVR is equivalent
ventilation rate, to define the threshold EVR for moderate exertion. EVR, calculated as
ventilation rate divided by body surface area (Ve/BSA), values at or above ModEVR8 (but below
HeavyEVR8, the threshold for heavy exertion) are classified as moderate exertion. The
ModEVR8 value typically used in regulatory runs of APEX is 13 L/min-m2, which was developed
by Whitfield (1996) using clinical data from McDonnell et al. (1991). In McDonnell et al., study
participants were required to maintain a Ve of 40 L/min while exposed to ozone and performing
activities classified as moderate exertion over a 6.6-hour period. Using this data, Whitfield
(1996) defined the EVR range to be 13-27 L/min-m2 for 8-hour-average exposures at moderate
exertion.
The approach used to define moderate exertion was noted in public comments in the last review
of EPA's Health Risk and Exposure Assessment for Ozone in 2014 (U.S. EPA, 2014). The
bullets below summarize two critiques that some public commenters had about the ModEVR8
value of 13 L/min-m2.
¦ A ModEVR8 value of 13 L/min-m2 was too low and resulted in an overstatement of the
number of exposures. This, in turn, resulted in an overestimation of the lung function
decrement risk when exposure-response functions were used to estimate risk.
¦ The strenuous nature of the exercise performed in the clinical studies to achieve an EVR of
20 L/min-m2 was not comparable to the activities and range of actual 8-hour EVRs in the
populations of interest. They suggested that use of the clinical studies data may not be
reasonable in defining ModEVR8.
Due to the lack of available controlled studies for human exposure to ozone, we focused on
evaluating how ModEVR8 is defined and we performed our analyses using an expanded
dataset of clinical studies provided by the EPA WAM where the target EVR under moderate
exertion was 20 L/min-m2.
3D-Attachment2-2
-------
2. Data Sources
In Table 1 we list the clinical studies with data available on Ve and EVR for individuals
undergoing moderate exertion during 6.6-hour exposure to filtered air and ozone. Adult study
participants were required to maintain an EVR of 20 L/min-m2 while undergoing intermittent
moderate exercise, which consisted of six periods of 50-minute exercise on the treadmill or
cycle ergometer, each followed by a 10-minute break, and with a 35-minute lunch after the third
period.
Table 1. Clinical Studies with 6.6-hour Moderate Exertion
Reference
No. Subjects / Gender
Age Range (years)
O3 Exposure (ppm)
Folinsbee et al. (1998)
10 Males
18-33
0, 0.12
Horstman et al. (1990)
22 Males
18-35
FA, 0.08, 0.10, 0.12
McDonnell et al. (1991)
28 Males
18-30
0, 0.08
McDonnell et al. (1991)
10 Males
18-30
0, 0.08, 0.1
Folinsbee et al. (1994)
17 Males
25±4
FA, 0.12
Schelegle et al. (2009)
15 Males, 16 Females
18-25
Mean: FA, 0.06, 0.07, 0.08, 0.087
Max: n/a, 0.09, 0.09, 0.15, 0.12
Kim et al. (2011)
27 Males, 32 Females
19-35
FA, 0.06
Notes: No. = number; O3 = ozone; ppm = parts per million; FA = filtered air; max = maximum; n/a = not available.
3. Equivalent Ventilation Rates
3.1. Original EVR Threshold
The ModEVR8 of 13 L/min-m2 typically used in regulatory runs of APEX was based on the
range of 13-27 L/min-m2 defined by Whitfield (1996) for 8-hour exposures. However, details
were not available on how this range was obtained from the McDonnell et al. (1991) data. We
analyzed the data to determine
¦ if the mean EVR was calculated based on all data points or based on the person-averaged
EVR values, and
¦ the number of standard deviations away from the mean that would result in the range of
values reported.
The EPA WAM provided a SAS data file with 4,024 individual EVR data points corresponding to
485 experiments. The McDonnell et al. (1991) data were provided as two separate datasets with
Study IDs of "Ozi-2" and "Pokoz", which were identified within the SAS dataset as OZI and POK,
respectively. Using the McDonnell et al. (1991) OZI and POK datasets individually and
combined, we calculated the mean, standard deviation, and upper and lower bounds (defined
as mean ± 3 standard deviations) using (i) all individual EVR data points and (ii) person-
averaged EVR values. The person-averaged EVRs are the average overtime, resulting in one
person-averaged EVR per unique subject and experiment, which is more consistent with how
APEX evaluates whether a profile is at moderate exertion (by calculating the profile's 8-hour-
average EVR). In Table 2 we present the results of this analysis, which suggest that the range
of 13-27 L/min-m2 used by Whitfield (1996) was obtained using individual EVR data from the
OZI dataset and three standard deviations away from the mean (see gray-shaded cells in the
table).
3D-Attachment2-3
-------
Table 2. EVR Metrics for Individual EVR Data Points and Person-averaged EVRs, during
Intermittent Moderate Exercise
McDonnell et al. (1991) Datasets
OZI POK OZI + POK Superset
Individual EVR Data Points (L/min-m2)
Mean
20.29
20.22
20.26
Standard Deviation
2.30
1.95
2.14
Lower Bound
13.37
14.38
13.83
Upper Bound
27.20
26.06
26.69
Person-averaged EVRs (L/min-m2)
Mean
20.29
20.22
20.26
Standard Deviation
2.05
1.61
1.85
Lower Bound
14.15
15.39
14.72
Upper Bound
26.43
25.06
25.80
Notes: EVR = equivalent ventilation rate; L/min-m2 = liters per minute per square meter; lower bound = mean -
3 standard deviations; upper bound = mean + 3 standard deviations.
Cells shaded in gray indicate metrics lining up with the 13-27 L/min-m2 range of moderate-exertion EVRs defined by
Whitfield (1996) for 8-hour exposures based on the McDonnell et al. (1991) data.
3.2. EVR Threshold from All Clinical Studies
ModEVR8 can be re-calculated for the expanded dataset following the original approach of
three standard deviations away from the mean. In Table 3 we present the mean, median,
standard deviation, and upper- and lower-bound EVRs using person-averaged EVR values from
all datasets listed in Table 1.
The EVRs measured in the studies were collected during periods of exertion and represent
exercise-only conditions. However, during the 6.6-hour experiment, only 5 hours were used for
exercise (i.e., six 50-minute periods of treadmill or cycle ergometer), with the remaining 1.6
hours for rest or lunch. During resting times/lunch, EVR values are expected to drop. As
discussed below, we estimated the impact on EVRs from incorporating rest time.
Of the studies in Table 1, only Schelegle et al. (2009) mentioned resting Ve (and, by default,
resting EVR), which was estimated using regression equations derived from the data of Aitken
et al. (1986). For college-age males, this was Ve = 7.61 xBSA, and for college-age females, this
was Ve = 8.05xBSA. These resting EVR values, 7.61 and 8.05 L/min-m2 for college-age males
and females respectively, are consistent with expected resting EVR values. For example,
Adams (2006) reported group-mean-total and exercise-only Ve, which can be used with their
reported BSAs to estimate a resting EVR of 6.38 L/min-m2 for that study. In our analysis, we
used those college-age male and female values to calculate resting EVR for each study, as the
weighted average based on the number of males and females in the study. We then calculated
total (exercise and rest) EVR as a weighted average based on 5 hours of exercise and 1.6
hours of rest/lunch. As expected, the values in Table 3 show that total (exercise and rest) EVRs
are lower than exercise-only EVRs.
3D-Attachment2-4
-------
Table 3. EVR Metrics Derived from All Clinical Studies in Table 1, during Intermittent Moderate
Exercise
Person-averaged EVRs (L/min-m2)
Exercise Only
Exercise + Rest
Mean
20.39
17.32
Standard Deviation
1.65
1.25
Lower Bound
15.44
13.57
Upper Bound
25.34
21.08
Median
20.35
17.31
Notes: EVR = equivalent ventilation rate;
L/min-m2 = liters per minute per square meter; lower bound = mean -
3 standard deviations; upper bound = mean + 3 standard deviations.
3.3. Parameters for Distribution Sampling
An alternative to setting ModEVR8 to a single value is to allow it to be sampled from a
distribution for each person. This introduces variability in ModEVR8 and reflects the variability
across individuals in Ve, and thus EVR, when performing moderate-exertion activities.
We modified the APEX code to allow for sampling ModEVR8 from a distribution. The
distribution parameters are specified in the modified physiology input file, where users can
specify the distribution shape and corresponding parameters. For each profile, the APEX code
samples ModEVR8 from the distribution. EVR values at or above this sampled ModEVR8 (but
below HeavyEVR8) are classified as being at moderate exertion. The sampled ModEVR8
values are then written to the Profile Summary output file.
4. Comparison of Approaches in Defining Moderate
Exertion
4.1. APEX Runs
We conducted four APEX runs, listed in Table 4, to compare how different ModEVR8 values
(including dynamic sampling of values from a distribution) would affect the exposure outcomes.
We used internal version APEX5.04, modified on December 20, 2018 to allow sampling of
ModEVR8 from a distribution. (A more updated version will be provided to the EPA WAM soon
following this memorandum, containing additional model updates unrelated to EVR). The
simulations were for the Los Angeles area, time period of January 1 to December 31, 2007, for
10,000 profiles, and for both children (ages 5 to 18 years) and total population (ages 5 years
and up). We calculated the ModEVR8 values listed in Table 4 from exercise-only data.
3D-Attachment2-5
-------
Table 4. Model Runs
Run Name
ModEVR8
(L/min-m2)
Comments
EVR 13
13
Original ModEVR8 value, calculated as:
¦ Three standard deviations below the mean (see shaded lower-bound value in
Table 2)
¦ Using the OZI group of McDonnell et al. (1991) data
¦ From individual EVR data points (instead of person-averaged EVRs)
EVR 16
15.4
Updated ModEVR8 value, calculated as:
¦ Three standard deviations below the mean (see lower-bound exercise-only
value in Table 3)
¦ Using the data specified in Table 1
¦ From person-averaged EVRs (instead of individual EVR data points)
EVR_Med
20.4
Median value using person-averaged EVRs from the data specified in Table 1 (see
median exercise-only value in Table 3)
DIST20_1
varies
ModEVR8 sampled for each profile from a distribution.
¦ Distribution parameters calculated using person-averaged EVRs from the data
specified in Table 1
¦ Normal distribution; mean = 20.4; standard deviation = 1.7; upper truncation =
25.3; lower truncation = 15.4 (see exercise-only column in Table 3)
Notes: L/min-m2 = liters per minute per square meter; EVR = equivalent ventilation rate; ModEVR8 = the model
parameter for the threshold of moderate-exertion EVR for an 8-hour period.
4.2. Simulated Population Results
Across the test runs, for all profiles and children only, we compared the percent of the profiles
reaching moderate exertion at least once and the person-day counts at moderate exertion.
Results for both metrics and profile groups, presented in Table 5 to Table 8 and graphically in
Figure and Figure 2, show that as the ModEVR8 value increases, the metrics decrease as
expected (EVR 13 > EVR 15 > EVR_Med).
3D-Attachment2-6
-------
Table 5. Percent of Modeled Profiles Reaching Moderate Exertion (Ages 5 Years and Up)
Run Name (see
Table 4)
Level (ppm)
EVR13
EVR15
EVR Med
DIST20 1
0
86.7
66.3
18.5
20.8
0.01
84.8
64.3
17.2
19.4
0.02
83.9
63.3
16.3
18.5
0.03
82.2
61.0
14.4
16.8
0.04
79.8
57.3
12.1
14.3
0.05
76.3
51.7
9.1
11.2
0.06
69.8
43.2
6.0
7.9
0.07
57.3
31.2
3.4
4.8
0.08
38.3
18.4
1.5
2.1
0.09
18.3
7.3
0.46
0.74
0.10
3.5
1.3
0.04
0.06
0.11
0.36
0.07
0
0
0.12
0
0
0
0
0.13
0
0
0
0
0.14
0
0
0
0
0.15
0
0
0
0
0.16
0
0
0
0
Notes: ppm = parts per million.
Shading indicates relative magnitude of values (reds and oranges are higher values; yellows and greens are lower
values).
Table 6. Percent of Modeled Child Profiles (Ages 5 to 18 Years) Reaching Moderate Exertion
Run Name (see
Table 4)
Level (ppm)
EVR13
EVR15
EVR Med
DIST20 1
0
99.4
93.7
41.2
43.3
0.01
98.4
90.7
37.1
39.2
0.02
98.2
89.8
35.4
37.6
0.03
97.6
88.0
31.2
33.7
0.04
97.1
85.8
26.7
29.9
0.05
96.0
82.1
20.8
24.7
0.06
93.4
72.9
13.6
17.6
0.07
86.3
59.4
8.3
11.2
0.08
65.7
38.6
4.0
5.4
0.09
33.8
17.2
1.3
2.1
0.10
5.8
2.6
0
0.1
0.11
0.3
0.1
0
0
0.12
0
0
0
0
0.13
0
0
0
0
0.14
0
0
0
0
0.15
0
0
0
0
0.16
0
0
0
0
Notes: ppm = parts per million.
Shading indicates relative magnitude of values (reds and oranges are higher values; yellows and greens are lower
values).
3D-Attachment2-7
-------
c
o
tr
0)
X
0)
0>
15
s_
CL)
¦o
o
ro
c
.2
ZJ
Q.
O
CL
100
90
80
70
60
50
40
30
20
10
0
fa) Total (5 Years and Up)
o o o o
Levels (ppm)
EVR13
EVR15
EVR Med
¦DIST20 1
c
o
tr
0)
TO
i—
-------
Table 7. Number of Modeled Person-days Reaching Moderate Exertion (Ages 5 Years and Up)
Run Name (see
Table 4)
Level (ppm)
EVR13
EVR15
EVR Med
DIST20 1
0
1.7E+06
7.4E+05
4.9E+04
8.1E+04
0.01
1.5E+06
6.5E+05
4.1E+04
6.9E+04
0.02
1.3E+06
5.3E+05
3.2E+04
5.4E+04
0.03
9.1E+05
3.6E+05
2.2E+04
3.6E+04
0.04
5.0E+05
1.9E+05
1.1E+04
1.8E+04
0.05
2.3E+05
8.6E+04
4.8E+03
8.0E+03
0.06
9.3E+04
3.4E+04
1.9E+03
3.2E+03
0.07
3.4E+04
1.3E+04
6.8E+02
1.1E+03
0.08
1.1E+04
3.9E+03
2.0E+02
3.2E+02
0.09
2.7E+03
9.6E+02
4.8E+01
8.7E+01
0.10
4.0E+02
1.3E+02
4.0E+00
6.0E+00
0.11
3.8E+01
8.0E+00
0
0
0.12
0
0
0
0
0.13
0
0
0
0
0.14
0
0
0
0
0.15
0
0
0
0
0.16
0
0
0
0
Notes: ppm = parts per million.
Shading indicates relative magnitude of values (reds and oranges are higher values; yellows and greens are lower
values).
Table 8. Number of Modeled Person-days Reaching Moderate Exertion (Ages 5 to 18 Years)
Run Name (see
Table 4)
Level (ppm)
EVR13
EVR15
EVR Med
DIST20 1
0
6.4E+05
3.8E+05
3.4E+04
5.2E+04
0.01
5.8E+05
3.4E+05
2.8E+04
4.4E+04
0.02
4.9E+05
2.7E+05
2.1E+04
3.4E+04
0.03
3.7E+05
2.0E+05
1.4E+04
2.3E+04
0.04
2.1E+05
1.1E+05
7.1E+03
1.2E+04
0.05
1.0E+05
4.9E+04
3.1E+03
5.2E+03
0.06
4.3E+04
2.0E+04
1.2E+03
2.1E+03
0.07
1.6E+04
7.3E+03
4.4E+02
7.3E+02
0.08
5.0E+03
2.2E+03
1.2E+02
2.0E+02
0.09
1.2E+03
5.3E+02
2.9E+01
5.6E+01
0.10
1.4E+02
6.0E+01
0
2.0E+00
0.11
7.0E+00
4.0E+00
0
0
0.12
0
0
0
0
0.13
0
0
0
0
0.14
0
0
0
0
0.15
0
0
0
0
0.16
0
0
0
0
Notes: ppm = parts per million.
Shading indicates relative magnitude of values (reds and oranges are higher values; yellows and greens are lower
values).
3D-Attachment2-9
-------
1.80E+06
1.60E+06
1.40E+06
V)
1 1.20E+06
o
^ 1.00E+06
TO
t 8.00E+05
o
% 6.00E+05
CL
4.00E+05
2.00E+05
O.OOE+OO
EVR13 EVR15 EVR Med DIST20 1
7.00E+05
6.00E+05
j2 5.00E+05
c
Z5
8 4.00E+05
_g
c 3.00E+05
o
CO
£ 2.00E+05
1.00E+05
0.00E+00
EVR 13 EVR 15 EVR Med DIST20 1
Notes: ppm = parts per million.
Legend entries are the run names specified in Table 4.
Figure 2. Number of Modeled Person-days Reaching Moderate Exertion for
(a) All Profiles and (b) Children Only
\
(a) Total (5 Years and Up)
^-MCO^lD(DNCOQr
oooooooooc
ooooooooo
Levels (ppm)
CN CO ^ in CD
o o o o o o
(b) Children (5 to 18 Years)
\
\
\ v
CM CO LO CD
OT-fMc0^iD(DS(DO)r
O O O O O O O O O q x— t— ¦¦
OOOOOOOOO oooo oo
Levels (ppm)
3D-Attachment2-10
-------
The alternative method where one ModEVR8 value per person is sampled from a distribution
resulted in higher metrics as compared to setting the ModEVR8 equal to the median of the
distribution (DIST20_1 > EVR_Med). These results are expected because sampling from the
distribution allows the selection of ModEVR8 values lower than the median value. Lower
ModEVR8 values will result in more profiles reaching "moderate exertion" in the modeling.
Specifically, for person-day counts, sampling ModEVR8 from a distribution results in counts that
are more than 50 percent greater than when ModEVR8 is set to the median value. While the
sampling also allows the selection of higher ModEVR8 values (resulting in fewer profiles
reaching "moderate exertion"), profiles reach lower EVRs much more commonly than higher
EVRs, so much so that using lower ModEVR8 values brings many more profiles into the
"moderate exertion" pool than are excluded when higher ModEVR8 values are used.
However, sampling from a distribution still gives metrics that are much lower than when setting
the ModEVR8 value to three standard deviations below the mean (DIST20_1 < EVR15). As an
example, for an exposure level of 0.05 parts per million, DIST20_1 results in 40 percent fewer
profiles overall reaching moderate exertion at least once (11.2 percent with DIST20_1 versus
51.7 percent with EVR15), and 57 percent fewer children (82.1 percent with DIST20_1 versus
24.7 percent with EVR15). When considering person-day counts, in general, DIST20_1 counts
were nearly an order of magnitude lower than EVR15 counts.
5, References
Adams WC. (2006) Comparison of Chamber 6.6-h Exposures to 0.04-0.08 PPM Ozone via
Square-wave and Triangular Profiles on Pulmonary Responses. Inhal Toxicol 18(2): 127-
136. DOI: https://doi.org/10.1080/08958370500306107.
Aitken ML, Franklin JL, Pierson DJ, Schoene RB. (1986) Influence of Body Size and Gender on
Control of Ventilation. J Appi Physiol 60(6): 1894-1899. DOI:
https://doi. org/10.1152/jappl. 1986.60.6.1894.
Folinsbee LJ, McDonnell WF, Horstman DH. (1988) Pulmonary Function and Symptom
Responses after 6.6-Hour Exposure to 0.12 ppm Ozone with Moderate Exercise. JAPCA
38(1):28-35. DOI: https://doi.org/10.1080/08940630.1988.10466349.
Folinsbee LJ, Horstman DH, Kehrl HR, Harder S, Abdul-Salaam S, Ives PJ. (1994) Respiratory
Responses to Repeated Prolonged Exposure to 0.12 ppm Ozone. Am J Respir Crit Care
Med 149(1):98-105. DOI: https://doi.Org/10.1164/ajrccm.149.1.8111607.
Horstman DH, Folinsbee LJ, Ives PJ, Abdul-Salaam S, McDonnell WF. (1990) Ozone
Concentration and Pulmonary Response Relationships for 6.6-Hour Exposures with Five
Hours of Moderate Exercise to 0.08, 0.10, and 0.12 ppm. Am Rev Respir Dis
142(5):1158-1163. DOI: https://doi.org/10.1164/ajrccm/142.5.1158.
Kim CS, Alexis NE, Rappold AG, Kehrl H, Hazucha MJ, Lay JC, Schmitt MT, Case M, Devlin
RB, Peden DB, Diaz-Sanchez D. (2011) Lung Function and Inflammatory Responses in
Healthy Young Adults Exposed to 0.06 ppm Ozone for 6.6 Hours. Am J Respir Crit Care
Med 183(9):1215-1221. DOI: https://doi.org/10.1164/rccm.201011-18130C.
McDonnell WF, Kehrl HR, Abdul-Salaam S, Ives PJ, Folinsbee LJ, Devlin RB, O'Neil JJ,
Horstman DH. (1991) Respiratory Response of Humans Exposed to Low Levels of
3D-Attachment2-11
-------
Ozone for 6.6 Hours. Arch Environ Health 46(3): 145-150. DOI:
https://doi.org/10.1080/00039896.1991.9937441.
Schelegle ES, Morales CA, Walby WF, Marion S, Allen RP. (2009) 6.6-Hour Inhalation of Ozone
Concentrations from 60 to 87 Parts per Billion in Healthy Humans. Am J Respir Crit Care
Med 180(3):265-272. DOI: https://doi.org/10.1164/rccm.200809-14840C.
U.S. EPA, 2014. Health Risk and Exposure Assessment for Ozone: Final Report. Office of Air
Quality Planning and Standards, U.S. Environmental Protection Agency, Research
Triangle Park, NC, 27711. EPA-452/R-14-004a. August 2014. Available at:
https://www3.epa.gov/ttn/naaqs/standards/ozone/data/20140829healthrea.pdf.
U.S. EPA, 2017a. Air Pollutants Exposure Model Documentation (APEX, Version 5) Volume I:
User's Guide. Office of Air Quality Planning and Standards, U.S. Environmental
Protection Agency, Research Triangle Park, NC, 27711. EPA-452/B-17-001a. January
2017. Available at: https://www.epa.gov/fera/apex-user-guides.
U.S. EPA, 2017b. Air Pollutants Exposure Model Documentation (APEX, Version 5) Volume II:
Technical Support Document. Office of Air Quality Planning and Standards, U.S.
Environmental Protection Agency, Research Triangle Park, NC, 27711. EPA-452/B-17-
001b. January 2017. Available at: https://www.epa.gov/fera/apex-user-guides.
Whitfield RG, Biller WG, Jusko MJ, Keisler JM. (1996) A Probabilistic Assessment of Health
Risks Associated with Short- and Long-Term Exposure to Tropospheric Ozone.
Argonne, IL: Argonne National Laboratory.
3D-Attachment2-12
-------
APPENDIX 3D, ATTACHMENT 3:
ICF TECHNICAL MEMO: UPDATES TO THE METEOROLOGY DATA AND
ACTIVITY LOCATIONS WITHIN CHAD
3D-Attachment3-1
-------
MEMORANDUM
To: John Langstaff and Stephen Graham, U.S. EPA-OAQPS
From- ^ader, Graham Glen, Caroline Foster, Samuel Kovach, Delaney Reilly, Chris
Holder, River Williams, Anna Stamatogiannakis, and George Agyeman-Badu, ICF
Date: June 18,2019
Re: Updates to the Meteorology Data and Activity Locations within CHAD
1. Introduction
In the November 1, 2016 version of CHAD, approximately 18 percent (32,723 out of 179,912) of
diary-days are missing values for daily-maximum temperature (Tmax) and thus cannot be used
by APEX. The temperature data currently in CHAD originate from a variety of sources, including
from the original studies and from EPA or contractors who encoded the study data into CHAD.
As discussed in Section 2, we used a methodical process to replace most of these missing
values. As part of this exercise, for diary-days without county-location information, we identified
county locations for over 10,000 diary-days based on respondent zip code and for over
6,000 diary-days based on the metropolitan locations of several of the studies. Some of the
diary-days that received repaired county locations were not missing temperature data;
nonetheless, we made the repairs as part of a "cleaning up" of the diary data. After this process,
only 0.3 percent (565) of diary-days have missing values for Tmax and remain unusable by
APEX.
In the same version of CHAD, six studies have at least 200 minutes per day (on average) of
time spent in locations that are not sufficiently clear (they are ambiguous). Unspecified and
missing location codes are ambiguous, as are those taking place at a residence or a place of
employment without specifying whether they are in the three broad microenvironments (MEs) of
indoors, outdoors, or in-vehicle. If studies have an apparent bias (via ambiguity) in time spent in
the three broad MEs, then the APEX-modeled exposures will also be biased. As discussed in
Section 3, we used paired activity-location information from the other 15 studies in CHAD to
derive frequency distributions of location codes used per each activity code, with different
distributions intended for reassigning unspecified/missing locations, ambiguous residential
locations, and ambiguous workplace locations. For the six targeted studies, for a diary event
with an ambiguous location code, we reassigned the location code based on the activity by
sampling from these frequency distributions. After this process, the time spent per day in
ambiguous locations dropped substantially for the six studies, though one study still had more
than 200 minutes per day spent in ambiguous locations. These location-code reassignments will
substantially reduce bias in APEX exposure estimates, particularly given that one of the
six studies constitutes more than half of all CHAD diary-days.
These modifications do not impact the official EPA CHAD-Master database, which remains
unchanged. Instead, the modifications are specific to the version of the diary data used for
APEX modeling.
3D-Attachment3-2
-------
2. Temperature Data
2.1. Overview and Objectives
The current CHAD questionnaire file includes Tmax and daily-average temperature (Tavg; °F)
as well as daily precipitation (inches) and daily number of hours with precipitation. Only Tmax is
typically used by APEX modelers, and it is used to help select a set of diaries that have similar
temperature values as those experienced by a simulated profile at his/her location on a given
modeling day. Diary-days without values for Tmax cannot be selected for use by any simulated
profile.
As shown in Table 2-1, approximately 18 percent of diary-days are currently unusable by APEX
on the basis of missing Tmax. Less than 1 percent of those are missing all indicators of
respondent location (state, county, and zip code) and are not from studies of a single
metropolitan area; it will not be possible to identify reasonable temperature data for those diary-
days. Most of the remaining diary-days have only state information (no information on county or
zip code).
Table 2-1. Information on Diary-days Missing Daily-maximum Temperature Values
Count
Percent of All
Diary-Days
Percent of Diary-days
Missing Tmax
Missing Tmax
32,723
18%
100%
From the 1980s
14
0.008%
0.04%
From the 1990s
1,230
0.7%
4%
From the 2000s
25,512
14%
78%
From the 2010s
5,967
3%
18%
Missing All Location Information (state,
county, zip code; is not a single-
metropolitan study)
111
0.06%
0.3%
Is a Study of a Single Metropolitan Area
0
0%
0%
Has State Location but not County (and is
not a single-metropolitan study)
30,895
17%
94%
—> Has Zip Code
30
0.02%
0.09%
Notes: Studies limited to one metropolitan area were put into CHAD without county or zip-code information.
Tmax = daily-maximum temperature
The objective of this task is to use historical meteorological records to identify reasonable
temperature values for diary-days currently missing those values. Identifying these values relies
on knowing or estimating the geographic location of each diary-day. Since most of the target
diary-days identify the respondent's state but not county or zip code, in most cases we have
made assumptions about respondent locations within the state.
A structured methodology of identifying appropriate temperature data allows us to identify
reasonable temperature values for nearly all diary-days, not just those currently missing
temperature data. While we will generally not update temperature data in CHAD that are not
already missing (unless we believe the current values are erroneous), we can compare current
and "new" temperatures as part of quality control (QC). With this in mind, as detailed in
Section 2.2, we developed a hierarchy to assign a county location to nearly all diary-days. Then,
as detailed in Section 2.3, we matched county locations to the five closest meteorological
stations from the historical records, thus enabling the assignment of temperature values.
3D-Attachments-3
-------
2.2. Assigning County Locations to Diary-days
Matching diary-days with nearby meteorological stations requires knowing (or estimating) where
the diary-days took place. County is the primary indicator of diary location, though zip codes are
also available for some diaries, and assigning temperature data on a county basis is reasonable
given the typical spatial resolution of counties and typical temperature gradients.
About 43 percent (77,811) of all diary-days already had county designations. For these diary-
days, we "cleaned up" the county names to be more consistent with the names provided by the
U.S. Census Bureau. While the county and state locations of diary-days are not used in APEX,
creating consistent location designations (and use of the more reliable state-county FIPS
designations) made the temperature-assignment process more reliable.
The remaining 57 percent (102,101) of all diary-days had no county locations. As indicated in
Table 2-2, 111 had no location information at all and they were not from studies located in a
single metropolitan area. We could not assign counties to these 111 diary-days, and thus we
could not replace missing temperature data if needed.
Table 2-2. Information on Diary-days Without County Designations
How County Locations Were Determined
(showing counts of diary-days)
Count
Percent of
All Diary-
Days
Metropolitan
Study
Location
Zip Code
State's Population
Distribution
Missing All Location Information
111
0.06%
0
0
0
(state, county, zip code; is not a
single-metropolitan study)
Is a Study of a Single Metropolitan
Area
6,150
2%
6,150
0
0
Has State Location but not County
95,840
55%
0
0
84,141
(and is not a single-metropolitan
study)
(14 from 1980s;
6,139 from 1990s;
64,046 from 2000s;
13,942 from 2010s)
Has Zip Code
11,699
7%
0
11,635
64
(1 from 1980s;
62 from 1990s;
1 from 2000s;
0 from 2010s)
Note: Studies limited to one metropolitan area were put into CHAD without county or zip-code information.
For the other 101,990 diary-days without county designations, a small amount (6,150) were
from studies located within a single metropolitan area. Diary-days from these studies were
originally put into CHAD without county or zip-code information. We made the assumption that
all such respondents lived in the primary county associated with the area, as listed below.
3D-Attachments-4
-------
¦ Hamilton County, Ohio for the Cincinnati Activity Patterns Study (CIN)
¦ Wayne County, Michigan for the Detroit Exposure and Aerosol Research Study (DEA)
¦ Denver County, Colorado for the Denver, Colorado Personal Exposure Study (DEN)
¦ King County, Washington for the Seattle Study (SEA)
¦ District of Columbia for the Washington, DC Study (WAS)
Additionally, a small amount (11,635) of diary-days without county designations had reliable zip
codes that we geocoded to their most likely counties, following the process listed below. Note
that we used geospatial files representing the year 2000 because most of the CHAD diary-days
(129,569 diary-days, which is 72 percent of all diary-days) were from the 2000s, and county
boundaries have remained unchanged through the last few decades for nearly all U.S. counties.
¦ Use GIS software to convert the year-2000 county polygons1 to centroid points (one centroid
per county).
¦ Use GIS software to identify the county centroid (year 2000) closest to each zip-code
centroid (also year 2000; from the zip-code tabulation areas file.2 These centroid-proximity
matches were restricted to within the same state (e.g., a zip-code centroid located in
California could only be matched to a county in California).
¦ A small number of zip codes (145) could not be identified in the Gazetteer files. We
identified the county locations of 85 such zip codes with reasonable confidence using
Internet searches, leaving 60 zip codes unmatched to counties.
For the remaining 84,205 diary-days without county designations (which includes 64 diary-days
that could not be reliably matched to counties via zip code), we assigned them to counties within
the state based on population distributions. We used U.S. Census data to calculate the
population distributions within each state. Since such distributions change over time, we did this
on a decadal basis, covering the decades represented by the CHAD diary-days (the 1980s
through 2010s), as indicated below. The majority of such population-based assignments were
for diary-days in the 2000s decade (as indicated in Table 2-2).
¦ 2000s and 2010s: We queried decadal census data from the U.S. Census Bureau (filtering
by Population Total, the 2010 or 2000 year, and All Counties within United States).3 The SF1
100% datasets were employed.
¦ 1980s and 1990s: We used intercensal data from the U.S. Census Bureau's State and
County Intercensal Datasets websites for 1980 to 1989 and 1990 to 1999.4 The county
1 From the U.S. Census cartographic boundary files available at https://www.census.gov/geographies/mapping-files/time-
series/geo/carto-boundary-file.2000.html.
2 From the U.S. Census Gazetteer files available at https://www.census.gov/geographies/reference-files/time-
series/geo/gazetteer-files.2000.html.
3 The American FactFinder website, used at the time of these analyses were performed, has been decommissioned as of March
30, 2020. Similar data queries can be made at https://data.census.gov/cedsci/.
4 Data available at https://www.census.gov/data/tables/time-series/demo/popest/1980s-county.html and
https://www.census.gov/data/datasets/time-series/demo/popest/intercensal-1990-2000-state-and-county-
characteristics, h tml.
3D-Attachment3-5
-------
populations were partitioned by demographics, which we aggregated to county-total
population values.
2.3. Assigning Temperature Data to Diary-days
The National Centers for Environmental Information (NCEI) distributes several databases of
land-based meteorology station data. We utilized the Global Historical Climatology Network-
Daily (GHCND), as it provided QCed daily temperature data at a relatively high spatial
resolution across the U.S.5 We narrowed the GHCND database based on the criteria listed
below.
¦ Stations must be located 24-50° N and 126-66° W (for contiguous U.S.), 51-72° N and
179.999-129° W (for Alaska; we did not use any stations in the far-western Aleutian
Islands), and 18.5-22.5° N and 160.5-154.5° W (for Hawaii). Note that these boundaries
may extend somewhat into neighboring countries.
¦ Stations must include Tmax and daily-minimum temperature (Tmin) as typically reported
parameters (requiring Tavg was too restrictive; we elected to calculate Tavg as the average
of Tmax and Tmin).
¦ On a decadal basis, stations must report data for the entirety of that decade (or for
2010-2014 for the 2010s).
Some of the GHCND stations were of 'higher quality' than others, as they are part of the U.S.
Historical Climatology Network (HCN), the U.S. Climate Reference Network (CRN) and/or the
Global Climate Observing System Surface Network (GSN). We preferred data from these
stations in our temperature assignments.
In Table 2-3, we indicate the number of meteorological stations per decade, including the
number of higher-quality stations, that meet all the selection criteria listed above. In Figure 2-1
and Figure 2-2, for the 1980s and 2010s respectively, we show examples of the geographic
spread of meteorology stations (with higher-quality stations differentiated) in North and South
Carolina.
Table 2-3. Number of GHCND Meteorological Stations
Meeting Selection Criteria, per Decade and U.S. Region
Year
Number of Meteorological Station Counts (higher-quality Stations)3
Contiguous U.S.
Alaska
Hawaii
1980
6,621 (1,225)
230 (19)
54 (2)
1990
7,207 (1,233)
251 (19)
56 (2)
2000
7,813 (1,151)
341 (21)
72 (2)
2010
8,445 (1,210)
388 (29)
85 (4)
a Note that a small number of stations included here may be across the U.S. border in other countries.
5 https://www.ncdc.noaa.gov/data-access/land-based-station-data/land-based-datasets.
3D-Attachments-6
-------
1980s Meteorology Stations
Legend
High quality met stations
© Other met stations
Figure 2-1. GHCND Meteorological Stations from the 1980s
Meeting Selection Criteria, in the North and South Carolina Region
3 D-Attachments -7
-------
Google E
;-2G«eCco$*
0*1* 5^0. NOM U % NJv, NGA GE0CO
iTagw CUW; 12/I3/201S W 34,438700* ton ^.66i337J «fev 109ft «y»*t
2010s Meteorology Stations
Figure 2-2. GHCND Meteorological Stations from the 2010s
Meeting Selection Criteria, in the North and South Carolina Region
By decade (with county locations fixed at the year-2000 definitions), we used ArcMap's
"Generate Near Table" tool to map each U.S. county to its five closest meteorological stations
from the GHCND dataset. The stations were initially sorted by closest proximity to the county
centroid. Then, we resorted the matches to ensure that the closest higher-quality within 30 miles
of the county centroid was the preferred station of the five stations.
The median distance from county centroid to the preferred meteorological station was 19 km—
only in Alaska were some county centroids more than 100 km from the preferred station, and a
few counties in Arizona, California, Nevada, and Texas were 50-70 km from the preferred
station. The median distance from county centroid to the fifth selected station was 42 km.
Based on the county location and decade of the diary-day, and the five meteorological stations
selected for that county and decade, we identified Tmax and Tmin from the preferred station. If
the preferred station's Tmax and Tmin values were missing, then we used the values from the
second station, and so on until we identified non-missing values. If none of the five stations
supplied non-missing Tmax and Tmin values, then the values were left missing.
Using the method above, 178,893 diary-days (> 99 percent) were matched with new Tmax and
Tavg values, leaving 1,019 diary-days (0.6 percent) without matched values. As a QC check, we
compared the newly matched temperature values ("new" temperatures) to the existing
temperature values where available ("old" temperatures). Using Tmax, there were
146,735 diary-days (82 percent) available for comparison. In Table 2-4, we indicate how many
diary-days were negligibly different (< 5°), 5-10° different, 10-20° different, or > 20° different.
3 D-Attachments -8
-------
Table 2-4. Comparison of Old (in Current CHAD-Master)
and New (Identified Here) Daily-maximum Temperatures
Difference between Old Tmax
and New Tmax
Number of Diary-days
Percent of Diary-days Available
for Comparison
LL
o
LO
VI
101,507
69.2%
5-10 °F
24,604
16.8%
10-20 °F
16,032
10.9%
> 20 °F
4,592
3.1%
During this QC check, we further examined the 4,592 diary-days (3 percent) where the Tmax
values were > 20° different. During this step, we discovered that most of these diary-days were
from the American Time Use Survey by the Bureau of Labor Statistics (BLS). In 2,431 of the
4,592 diary-days with differences over 20°, they were from the BLS study and the old Tmax was
equivalent to the old Tavg. This indicated a systematic error in the old BLS temperatures.
Using a similar approach, we compared the old and new Tavg values. The results are indicated
in Table 2-5. The results comparing the old and new Tavg values were similar to those for
Tmax.
Table 2-5. Comparison of Old and New Average Temperatures
Difference between Old Tmax
Percent of Diary-days Available
and New Tmax
Number of Diary-days
for Comparison
LL
o
LO
VI
109,632
74.7%
5-10 °F
24,430
16.6%
10-20 °F
10,271
7.0%
> 20 °F
2,363
1.6%
We further examined the 2,363 diary-days (1.3%) where differences in Tavg values were > 20°.
For 1,569 of these diary-days, they were from the BLS study and the old Tavg was equivalent to
the old Tmax, again indicating a systematic error in the old BLS temperatures.
As an additional check, we examined the mean Tmax and mean Tavg across all diary-days.
The mean Tmax and mean Tavg for the old values were 68.0° and 58.4°, respectively. For the
new data, the mean Tmax and mean Tavg were 68.4° and 57.8° respectively. The consistency
between the two was expected and provides additional assurance.
At the direction of EPA, and given the errors found in the temperatures of the BLS study, we
developed a diary dataset using a combination of the old and new temperatures. To create this
dataset, we replaced all the old temperatures (maximum and average) of the BLS diary-days.
Next, we replaced all previously missing values where new values were available (across all
studies). Following these rules, we replaced values for 125,581 diary-days, such that the new
diary dataset now has Tmax and Tavg values for 179,347 diary-days. Temperatures remain
missing for 565 diary-days, while 53,766 diary-days retained their old temperatures.
In addition to the new temperature data, we updated the dataset with information that was used
as intermediate to this process, with fields indicated in Table 2-6.
Table 2-6. Updated or Added Fields in the CHAD Dataset
Field Name
Description
county
Values updated to include newly georeferenced data
state
Values updated to include newly georeferenced data
FIPS
Field added to provide a unique ID to every state-county
3D-Attachments-9
-------
Field Name
Description
old avgtemp
Field renamed to identify the temperatures (°F) in the November 2016 CHAD
old maxtemp
Field renamed to identify the temperatures (°F) in the November 2016 CHAD
FlPSfromZip
Field added: TRUE or FALSE—if the county originally was missing, did we identify by
zip code?
FlPSfromStudy
Field added: TRUE or FALSE—if the county originally was missing, did we identify by
study location?
FIPSfromCountyRandom
Field added: TRUE or FALSE—if the county originally was missing, did we identify by
county population distributions in the state?
new avgtemp
Field added to provide new temperatures (°F) queried in this task
new maxtemp
Field added to provide new temperatures (°F) queried in this task
ReplacedMaxT emp
Field added to provide the final temperatures (°F) to use in future applications (either
the old or new value, depending on the study and other criteria as discussed in this
memorandum)
ReplacedAvgT emp
Field added to provide the final temperatures (°F) to use in future applications (either
the old or new value, depending on the study and other criteria as discussed in this
memorandum)
3. CHAD Activity Locations
3.1. Introduction
Each diary-day reports a series of "events" covering 24 hours. Event durations vary, but each
event has one location code and one activity code. To use diaries in APEX, the location codes
are mapped to APEX MEs, each of which has a method for determining its air quality. While the
number of MEs is flexible, generally all APEX runs distinguish between time spent in three basic
MEs: indoor, outdoor, and in-vehicle. Yet six of the location codes are ambiguous, even at that
coarse level of defining MEs (i.e., they do not distinguish between the three basic MEs). CHAD
is composed of 21 originally separate studies, and some of these studies use these ambiguous
codes, but others do not.
These six ambiguous location codes are shown below, and in Table 3-1 we show the average
amount of time spent in ambiguous locations (by study).
¦ Residential:
~ 30000 (Residence, general)
~ 30010 (Your residence)
~ 30020 (Other's residence)
¦ Workplace:
~ 33400 (At work: no specific location, moving among locations)
3D-Attachments -10
-------
¦ Unknown:
~ U (Uncertain)
~ X (Missing)
Table 3-1. Average Amount of Ambiguous Time by Study
Study
Average Ambiguous Time
(minutes per day)
BAL: Baltimore Retirement Home Study
3
BLS: American Time Use Survey (ATUS), Bureau of Labor Statistics
498
CAA: California Adults Activity Pattern Studies
67
CAC: California Children Activity Pattern Studies
0
CAY: California Youth Activity Pattern Studies
101
CIN: Cincinnati Activity Patterns Study
2
DEA: Detroit Exposure and Aerosol Research Study
1,186
DEN: Denver, Colorado Personal Exposure Study
16
EPA: EPA Longitudinal Studies
333
ISR: Population Study of Income Dynamics 1, II, III
58
l_AE: Los Angeles Ozone Exposure Study: Elementary School
34
LAH: Los Angeles Ozone Exposure Study: High School
2
NHA: National Human Activity Pattern Study: Air
18
NHW: National Human Activity Pattern Study: Water
18
NSA: National-scale Activity Study
154
OAB: RTI Ozone Averting Behavior Stud
121
RTP: RTP Particulate Matter Panel Study
1,081
SEA: Seattle Study
1,205
SUP: Study of Use of Products and Exposure-related Behaviors
804
VAL: Valdez Air Health Study
2
WAS: Washington, DC Study
16
Note: Bolded studies have relatively large average amounts of ambiguous time.
APEX assigns MEs based only on the location code (not the activity code), and furthermore,
APEX uses a deterministic mapping (that is, the same location code maps to the same ME
throughout that APEX run). But this rule may lead to an unavoidable bias if applied to certain
diary studies. We examined the CHAD activity code that is paired with each location code (on
the event level), to determine the likely place of occurrence of each event. Since this is not
always a certainty, part of this exercise is to probabilistically assign specific locations to events
with ambiguous location codes, based on the paired activity.
3.2. Methods
The starting point is the November 2016 version of CHAD. It has 179,912 diary-days. Two of
those (EPA002171 and EPA002172) have been deleted because they each contained 24 hours
of missing data.
For our purposes, we divided all location codes into six general MEs and temporarily related
them to the location codes shown as shown below, which are unambiguous. The codes are
typical examples of the categories shown. For example, 31110 is a car; while not all vehicular
travel is in a car, it is reasonable that the air quality in a car would be similar to that found in
other types of vehicles.
¦ IH (indoors at a residence) —~ Code 30120 (Your residence, indoor)
3D-Attachment3-11
-------
¦ 10 (indoors elsewhere)
Code 32000 (Other, indoor general)
¦ OH (outdoors at a residence) —~ Code 30200 (Residence, outdoor)
¦ OV (outdoors near traffic)
Code 35200 (Public garage / parking lot)
¦ O (outdoors elsewhere)
Code 35000 (Other outdoor, general)
¦ V (in an enclosed vehicle)
Code 31110 (Motorized travel by car)
The six ambiguous location codes had more than one mapping option for a location category,
as shown below. They were reassigned location codes based on activity (and occupation where
applicable), as discussed later.
¦ Codes 30000 (residence, general), 30010 (your residence), 30020 (other's residence)
t Could be either IH or OH; occasionally V or OV
¦ Code 33400 (at work; no specific location, moving among locations)
~ Could be any, but depends on occupation
- Occupation TRANS (transportation and material moving)
• V (specifically 31120, travel by truck)
» Occupation FARM (farming, forestry, and fishing)
• O
» Occupation HSHLD (private household)
• IH
- Activity code > 18000 (travel)
• V
» Activity codes 17700-17823 (active-leisure activities; exercise activities)
• OV
- All others
. IO
¦ Codes U (uncertain), X (missing)
~ Could be any
For analysis purposes, we divided CHAD into two parts. The "bad" part consisted of the six
studies with at least 200 minutes per day on average spent in ambiguous locations (see
Table 3-1; the studies were BLS, DEA, EPA [EPA Longitudinal Studies], RTP [RTP Particulate
Matter Panel Study], SEA, and SUP [Study of Use of Products and Exposure-related
Behaviors]). The "good" part consisted of the 15 studies with an average of fewer than
200 minutes per day of ambiguous time.
3D-Attachment3-12
-------
For the purposes of replacing location codes U and X in the "bad" part of CHAD, we analyzed
the "good" part to determine the time fractions in each of the six location categories for each
activity code (except activity codes U and X). We excluded any time in ambiguous locations. For
example, the "eating" code (14400) divided as IH = 76 percent, IO = 21 percent,
OH = 2 percent, 0=1 percent, and OV and V = less than 1 percent. A few activity codes did not
have examples in the "good" part of CHAD, and so we mapped them to similar activities. These
cases occurred extremely rarely in the "bad" part of CHAD, as well. The number of such cases
increased if we stratified CHAD by age group, and for most activities the allocation to the six
location categories was not very different between age groups. Therefore, we did not treat age
groups separately. We linked the time-fraction distributions to the activities in the six studies in
the "bad" part of CHAD. We reassigned U and X locations by activity (excluding activity codes U
and X), following these distributions from the "good" part of CHAD.
For the purposes of replacing ambiguous residential location codes (30000 - Residence,
general; 30010 - Your residence; and 30020 - Other's residence), we made separate time-
fraction determinations (also from the "good" part of CHAD) where we generally restricted time
to three categories: IH, OH, and OV. We used the last of these (OV) for time in the garage or
working on cars. We made an exception for selected travel activity codes over 18000, which
indicate that the person was in a vehicle. For example, we assigned 18031 (drive a motor
vehicle) and similar codes to V. We linked these refined time-fraction determinations to the
activities in the six studies in the "bad" part of CHAD, for all events with location codes 30000,
30010, or 30020. We reassigned these locations by activity (for activities other than U and X),
following these distributions of time spent. We made an exception for the DEA study, where it
was clear that the residential codes up to 30020 were used only for indoor events. Note the
before the location reassignments, the DEA study averaged 83 minutes in OH locations but only
29 minutes in IH locations.
In many cases, the same diary had the same activity code for several consecutive events with
ambiguous location codes. For example, the person might be sleeping for several hours, but the
location is not clear. It would not make sense for them to be relocated part way through, so for
such consecutive events we determined the reassignment (from the activity's distribution across
the six location categories) only for the first of such events, and then subsequent events
received the same new location reassignment.
tissioii
As shown in Table 3-2, five of the six studies where we reassigned location codes now have
fewer than 200 minutes per day of ambiguous location time. The exception is the SUP study, in
which most diaries were shorter than 24 hours and were padded with missing activities and
locations to fill out the day. Many of the SUP diaries were previously rejected by APEX, and
might continue to be, but most of the other diaries will now be acceptable. In particular, the BLS
diaries constitute more than half of CHAD, and they have gone from 498 ambiguous minutes to
just 10 such minutes per diary-day.
3D-Attachment3-13
-------
Table 3-2. Minutes per Day in the Six Location Categories, Before ("Old") and After ("New")
Location Reassignments, For the Six Studies With 200 Minutes per Day or More of Time Spent in
Ambiguous Locations
BLS
DEA
EPA
RTP
SEA
SUP
Location Category
Old
New
Old
New
Old
New
Old
New
Old
New
Old
New
IH
754
1,049
29
1,157
677
903
90
973
0.04
1,121
327
787
IO
79
228
48
95
246
346
131
170
139
145
175
176
OH
22
47
83
83
50
55
36
77
16
73
22
47
O
17
23
19
19
23
23
17
17
24
25
45
45
OV
0.3
1.7
3.3
3.4
24
24
5.8
6.8
1.0
2.1
5.0
5.1
V
70
81
72
72
87
87
80
80
54
54
61
61
Ambiguous
498
10
1,186
10.3
333
2.4
1,081
116
1,205
21
804
317
Indoor Total
833
1,277
78
1,252
923
1,249
220
1,143
139
1,265
503
963
Outdoor Total
39
72
105
106
96
102
59
101
41
99
72
98
Several questions remained, as listed below. We discussed these questions with EPA in
May 2019, with decisions noted below.
1. Should the "good" part of CHAD be defined differently?
a. No, keep it as-is.
2. Should other location codes be deemed ambiguous?
a. Not at this time.
3. Should this method be applied to the ambiguous events in "good" CHAD?
a. No.
The last question is perhaps the most important. The CAY, NSA, and OAB studies average over
100 minutes of ambiguous time per diary, which is significant. The same method could be
applied there, and might significantly reduce the ambiguous time in those studies. One reason
not to apply this method is that the time percentages would then be applied to some of the same
studies used to derive the percentages, and this presents the appearance of circular reasoning.
It is not exactly circular because we excluded ambiguous time when deriving the percentages,
but even so, there may be a correlation between the choice of location code and choice of
activity code within a single study. For example, there may be a reason particular to the given
study for why some eating events were assigned specific location codes, and others were
assigned location X. Hence, it is not clear whether general percentages for all eating events
should apply to those (relatively few) coded with location X. This is less of a concern when most
or all eating events are paired with location X.
4. Diagram of Processing
In Figure 4-1, we indicate the input and output files for the temperature and location-code
updates discussed above, as well as the processing programs and ancillary files. We briefly
discuss these files and programs below the figure.
3D-Attachments -14
-------
Meteoroiogical Data
Population Census Data
£SP5_County.csv
Study_County,csv
Z i p C od e_Co j nty. csv
CHAD Master files
quest_l 10116, sasTbdat
events 110116.sas7odat
CHA D_s-a s_f i 1 e_f orm atte r, R
Convert CHAD Master 5AS files to APEX input format
Current_CHAD,csv (quest) tvents_2016.txt
Cou nty_pop_m«t_st« ti o n_pr a«es.s or.R
CHAD_County_assigrtments.R
ChadCodc_MetAisignment_Top5.R
CHAD_PortProc««lng.R
Updates to the meteoroiogical data
Rnal_CHAD_W<'thTennp_Ftnai_Repiaced.csv
CHAD_aet_cw.csv
CHAD Iocs cwxsv
N«w_locs_5.R
Recod-ng ambiguous CHAD locations
chad_nevv2-csv (events i
Chad2Q19a.SAS
Convert to SAS format, using variable names in the CHAD Waster files
quest_new_Q6(H19.sas7bdat
events new 060419,sas7bdat
YVrKtApsxChadFikt.sai
Fixes nighttime OAB activity and location codes.
Replaces missing activity codes for 12 locations.
Writes several data summaries.
Writes the APEX CHAD Input files.
Summary
tables
APEX CHAD input files
quest_neiv_D60419A.txt
events new 060419A.t»T
Figure 4-1. Files and Processing Programs Used in this Task
Both the temperature and location-code tasks began with the November 2016 version of the
CHAD-master files (quest_1 10116.sas7bdatand events_ 110116.sas7bdat), which we
converted to text or CSV files (Current_CHAD.csv for the questionnaire file; Events_2016.txtfor
the events file) for easier processing in R programs.
We used four different R scripts to modify temperatures and county designations in the
questionnaire file. County_pop_met_station_processor.R reformatted GIS data, outputting the
3D-Attachment3 -15
-------
ranking of up to five meteorology stations for every county, by decade and reorganized based
on distance and station quality. CHAD_County_assignments.Rf\\\ed in missing location data,
based on zip code, study, and random assignment based on population density.
ChadCode_MetAssignment_Top5.R combined the outputs of the previous two scripts to assign
temperatures (and other intermediate details) the questionnaire file. CHAD_PostProcessing.R
cleaned the data of unnecessary fields and reformatted the data for processing back into a SAS
dataset. The resulting updated questionnaire file was
Final_ CHA D_ With Temp_ Final_ Replaced, csv.
The location-code reassignments were made by New_locs_5.R (where 5 is the version number
of the script). The output events file was chad_new2.csv.
The new questionnaire and events files were not directly suitable as input to APEX because
they contains extra variables, including both the old and new location codes, details about
county reassignments and meteorological stations, etc. The program Chad2019a.sas converted
the files to SAS format and utilized field names conforming to those of CHAD-Master, producing
quest_new_060419.sas7bdat and events_new_060419.sas7bdat.
Finally, the EPA WAM's program (WriteApexChadFiles.sas) processed the above-mentioned
SAS datasets in various ways, most importantly producing the APEX-ready diary files
(quest_new_060419A.txtand events_new_060419A.txt).
3D-Attachment3-16
-------
APPENDIX 3D, ATTACHMENT 4:
DETAILED EXPOSURE AND RISK RESULTS
Study Group
Study Area
AQ
Scenario
Year
Benchmark
(PPb)
Number of People with 7-hr Exposure at or above Benchmark
>1 Day
> 2 Days
> 3 Days
> 4 Days
> 5 Days
> 6 Days
All Children
Atlanta
S65
2015
0
782548
702709
652006
613670
581448
553463
All Children
Atlanta
S65
2015
10
737635
650392
594402
553161
518598
488576
All Children
Atlanta
S65
2015
20
638508
531491
470336
424273
390699
360596
All Children
Atlanta
S65
2015
30
498684
383940
321655
279970
246820
222104
All Children
Atlanta
S65
2015
40
321736
210321
154209
118800
93619
74875
All Children
Atlanta
S65
2015
50
110083
37448
14991
6335
2805
1352
All Children
Atlanta
S65
2015
60
11501
585
101
20
0
0
All Children
Atlanta
S65
2015
70
424
20
0
0
0
0
All Children
Atlanta
S65
2015
80
0
0
0
0
0
0
All Children
Atlanta
S65
2015
90
0
0
0
0
0
0
All Children
Atlanta
S65
2015
100
0
0
0
0
0
0
All Children
Atlanta
S65
2016
0
790356
714553
665988
626301
592142
564218
All Children
Atlanta
S65
2016
10
752243
669035
614820
572006
537181
508631
All Children
Atlanta
S65
2016
20
662941
562038
500480
454800
419047
390195
All Children
Atlanta
S65
2016
30
534497
422316
358800
315319
284086
259612
All Children
Atlanta
S65
2016
40
357993
246719
191476
156167
129352
107985
All Children
Atlanta
S65
2016
50
139824
55465
25887
13720
6981
3773
All Children
Atlanta
S65
2016
60
21165
1917
202
61
0
0
All Children
Atlanta
S65
2016
70
1473
20
0
0
0
0
All Children
Atlanta
S65
2016
80
81
0
0
0
0
0
All Children
Atlanta
S65
2016
90
0
0
0
0
0
0
All Children
Atlanta
S65
2016
100
0
0
0
0
0
0
All Children
Atlanta
S65
2017
0
787491
708984
655456
615486
582962
554835
All Children
Atlanta
S65
2017
10
743406
655093
596157
553746
519042
489362
All Children
Atlanta
S65
2017
20
640061
533004
470074
424596
389488
360696
All Children
Atlanta
S65
2017
30
493943
379743
317155
274704
244822
220005
All Children
Atlanta
S65
2017
40
295708
189377
137604
105160
82462
64767
All Children
Atlanta
S65
2017
50
70477
18744
6638
2320
868
504
All Children
Atlanta
S65
2017
60
3955
40
0
0
0
0
All Children
Atlanta
S65
2017
70
40
0
0
0
0
0
All Children
Atlanta
S65
2017
80
0
0
0
0
0
0
All Children
Atlanta
S65
2017
90
0
0
0
0
0
0
All Children
Atlanta
S65
2017
100
0
0
0
0
0
0
All Children
Atlanta
S70
2015
0
782548
702709
652006
613670
581448
553463
All Children
Atlanta
S70
2015
10
741489
655557
599325
558225
524187
494528
All Children
Atlanta
S70
2015
20
653358
549630
486417
441806
406941
376999
All Children
Atlanta
S70
2015
30
530462
416666
353433
311526
277266
250391
All Children
Atlanta
S70
2015
40
382185
267884
208121
169826
141680
119163
All Children
Atlanta
S70
2015
50
191173
94729
54598
32787
19470
11985
All Children
Atlanta
S70
2015
60
38981
6618
1493
363
81
20
All Children
Atlanta
S70
2015
70
4580
182
20
0
0
0
All Children
Atlanta
S70
2015
80
182
0
0
0
0
0
All Children
Atlanta
S70
2015
90
0
0
0
0
0
0
All Children
Atlanta
S70
2015
100
0
0
0
0
0
0
3D-Attachment4-1
-------
AQ
Benchmark
Number of People with 7-hr Exposure at or above Benchmark
Study Group
Study Area
Scenario
Year
(PPb)
>1 Day
> 2 Days
> 3 Days
> 4 Days
> 5 Days
> 6 Days
All Children
Atlanta
S70
2016
0
790356
714553
665988
626301
592142
564218
All Children
Atlanta
S70
2016
10
755774
673776
620167
577675
543436
513857
All Children
Atlanta
S70
2016
20
676823
578381
517468
472071
435531
407264
All Children
Atlanta
S70
2016
30
564036
453105
391244
346311
313302
286891
All Children
Atlanta
S70
2016
40
415839
301135
242926
204631
177513
153927
All Children
Atlanta
S70
2016
50
226825
124368
77559
51148
35349
24131
All Children
Atlanta
S70
2016
60
63455
13801
3914
1190
343
101
All Children
Atlanta
S70
2016
70
9947
282
20
0
0
0
All Children
Atlanta
S70
2016
80
1211
0
0
0
0
0
All Children
Atlanta
S70
2016
90
121
0
0
0
0
0
All Children
Atlanta
S70
2016
100
0
0
0
0
0
0
All Children
Atlanta
S70
2017
0
787491
708984
655456
615486
582962
554835
All Children
Atlanta
S70
2017
10
747340
660036
602573
559436
524914
495012
All Children
Atlanta
S70
2017
20
654730
549529
486094
440414
406356
376495
All Children
Atlanta
S70
2017
30
528767
412591
348328
305534
273877
248333
All Children
Atlanta
S70
2017
40
359082
248031
191112
155198
130058
108005
All Children
Atlanta
S70
2017
50
147228
62003
31597
17029
9887
5770
All Children
Atlanta
S70
2017
60
17291
1675
343
61
0
0
All Children
Atlanta
S70
2017
70
1069
0
0
0
0
0
All Children
Atlanta
S70
2017
80
0
0
0
0
0
0
All Children
Atlanta
S70
2017
90
0
0
0
0
0
0
All Children
Atlanta
S70
2017
100
0
0
0
0
0
0
All Children
Atlanta
S75
2015
0
782548
702709
652006
613670
581448
553463
All Children
Atlanta
S75
2015
10
744495
658725
602654
561595
528021
498018
All Children
Atlanta
S75
2015
20
664192
562724
500016
454921
419471
389953
All Children
Atlanta
S75
2015
30
557055
442896
379057
335698
301862
274260
All Children
Atlanta
S75
2015
40
427219
312717
251642
210442
180197
156651
All Children
Atlanta
S75
2015
50
269115
159395
108792
77417
55808
41201
All Children
Atlanta
S75
2015
60
92348
28126
10573
4136
1897
807
All Children
Atlanta
S75
2015
70
16202
1412
141
40
20
0
All Children
Atlanta
S75
2015
80
2320
40
0
0
0
0
All Children
Atlanta
S75
2015
90
141
0
0
0
0
0
All Children
Atlanta
S75
2015
100
0
0
0
0
0
0
All Children
Atlanta
S75
2016
0
790356
714553
665988
626301
592142
564218
All Children
Atlanta
S75
2016
10
758457
677267
623819
581610
547632
517852
All Children
Atlanta
S75
2016
20
687718
591173
529352
485186
449635
419975
All Children
Atlanta
S75
2016
30
588470
478528
415052
370825
336464
309327
All Children
Atlanta
S75
2016
40
460429
344999
283178
243773
215123
191617
All Children
Atlanta
S75
2016
50
299642
189660
136837
101831
78265
60913
All Children
Atlanta
S75
2016
60
129312
48424
22053
10431
4984
2603
All Children
Atlanta
S75
2016
70
33957
4257
666
182
20
0
All Children
Atlanta
S75
2016
80
6658
121
0
0
0
0
All Children
Atlanta
S75
2016
90
1069
0
0
0
0
0
All Children
Atlanta
S75
2016
100
161
0
0
0
0
0
All Children
Atlanta
S75
2017
0
787491
708984
655456
615486
582962
554835
All Children
Atlanta
S75
2017
10
749862
663668
606407
563350
529070
498906
All Children
Atlanta
S75
2017
20
666311
563209
499612
453852
419027
388580
3D-Attachment4-2
-------
Study Group
Study Area
AQ
Scenario
Year
Benchmark
(PPb)
Number of People with 7-hr Exposure at or above Benchmark
>1 Day
> 2 Days
> 3 Days
> 4 Days
> 5 Days
> 6 Days
All Children
Atlanta
S75
2017
30
553786
439062
374881
331198
297625
270003
All Children
Atlanta
S75
2017
40
408475
292762
233120
194845
167001
144848
All Children
Atlanta
S75
2017
50
222527
120656
76106
51168
34885
24373
All Children
Atlanta
S75
2017
60
57846
12832
3793
1211
424
222
All Children
Atlanta
S75
2017
70
5387
262
20
0
0
0
All Children
Atlanta
S75
2017
80
404
0
0
0
0
0
All Children
Atlanta
S75
2017
90
0
0
0
0
0
0
All Children
Atlanta
S75
2017
100
0
0
0
0
0
0
All Children
Boston
S65
2015
0
862212
765847
702953
654281
613301
578919
All Children
Boston
S65
2015
10
828103
724661
659401
605997
564037
530315
All Children
Boston
S65
2015
20
719792
597213
524695
471381
429331
395632
All Children
Boston
S65
2015
30
564197
430947
358701
309347
271005
242494
All Children
Boston
S65
2015
40
352717
224814
164424
124171
95705
74976
All Children
Boston
S65
2015
50
124717
41686
16383
6667
2822
1456
All Children
Boston
S65
2015
60
22754
1456
114
0
0
0
All Children
Boston
S65
2015
70
3095
0
0
0
0
0
All Children
Boston
S65
2015
80
114
0
0
0
0
0
All Children
Boston
S65
2015
90
0
0
0
0
0
0
All Children
Boston
S65
2015
100
0
0
0
0
0
0
All Children
Boston
S65
2016
0
865716
770989
706503
658309
617829
583447
All Children
Boston
S65
2016
10
832676
731123
663975
612618
571000
535344
All Children
Boston
S65
2016
20
725617
603038
527539
472769
430036
396747
All Children
Boston
S65
2016
30
564310
430924
357654
309278
273554
244315
All Children
Boston
S65
2016
40
359884
232050
168383
128995
100393
79754
All Children
Boston
S65
2016
50
137300
47739
19887
8988
3572
1502
All Children
Boston
S65
2016
60
15314
1160
114
0
0
0
All Children
Boston
S65
2016
70
1251
0
0
0
0
0
All Children
Boston
S65
2016
80
0
0
0
0
0
0
All Children
Boston
S65
2016
90
0
0
0
0
0
0
All Children
Boston
S65
2016
100
0
0
0
0
0
0
All Children
Boston
S65
2017
0
862462
763776
699927
649867
610434
576484
All Children
Boston
S65
2017
10
825941
722226
656693
605155
564288
526902
All Children
Boston
S65
2017
20
718654
597805
519871
464532
421185
386393
All Children
Boston
S65
2017
30
563013
424871
353672
302247
265089
235008
All Children
Boston
S65
2017
40
377155
243518
175369
131475
100142
77706
All Children
Boston
S65
2017
50
173912
69924
29968
12925
5006
2162
All Children
Boston
S65
2017
60
33631
4323
432
46
0
0
All Children
Boston
S65
2017
70
3140
23
0
0
0
0
All Children
Boston
S65
2017
80
0
0
0
0
0
0
All Children
Boston
S65
2017
90
0
0
0
0
0
0
All Children
Boston
S65
2017
100
0
0
0
0
0
0
All Children
Boston
S70
2015
0
862212
765847
702953
654281
613301
578919
All Children
Boston
S70
2015
10
828717
725776
660493
606952
565312
532022
All Children
Boston
S70
2015
20
727164
605655
532318
479391
437682
402185
All Children
Boston
S70
2015
30
585199
452449
378338
328256
289505
259810
All Children
Boston
S70
2015
40
396041
265772
201468
157802
127061
104761
All Children
Boston
S70
2015
50
184288
81643
41686
21867
12287
7031
3D-Attachment4-3
-------
Study Group
Study Area
AQ
Scenario
Year
Benchmark
(PPb)
Number of People with 7-hr Exposure at or above Benchmark
>1 Day
> 2 Days
> 3 Days
> 4 Days
> 5 Days
> 6 Days
All Children
Boston
S70
2015
60
46465
6690
1479
410
114
46
All Children
Boston
S70
2015
70
7554
91
0
0
0
0
All Children
Boston
S70
2015
80
592
0
0
0
0
0
All Children
Boston
S70
2015
90
0
0
0
0
0
0
All Children
Boston
S70
2015
100
0
0
0
0
0
0
All Children
Boston
S70
2016
0
865716
770989
706503
658309
617829
583447
All Children
Boston
S70
2016
10
833541
732397
665204
614120
571842
536391
All Children
Boston
S70
2016
20
735037
612709
537597
482440
438205
405166
All Children
Boston
S70
2016
30
587497
453519
380659
330122
292144
262336
All Children
Boston
S70
2016
40
409466
275238
209227
166836
134365
111315
All Children
Boston
S70
2016
50
208021
97139
52836
30969
18067
10672
All Children
Boston
S70
2016
60
50242
8783
1661
137
46
23
All Children
Boston
S70
2016
70
5438
273
0
0
0
0
All Children
Boston
S70
2016
80
91
0
0
0
0
0
All Children
Boston
S70
2016
90
0
0
0
0
0
0
All Children
Boston
S70
2016
100
0
0
0
0
0
0
All Children
Boston
S70
2017
0
862462
763776
699927
649867
610434
576484
All Children
Boston
S70
2017
10
826737
723250
657877
605951
564652
527835
All Children
Boston
S70
2017
20
726527
605655
528700
473475
429217
393493
All Children
Boston
S70
2017
30
585313
446647
372081
321020
282178
251300
All Children
Boston
S70
2017
40
418909
284977
213687
168428
134274
108038
All Children
Boston
S70
2017
50
238512
116890
63735
35497
20320
10922
All Children
Boston
S70
2017
60
81939
18477
4369
865
68
23
All Children
Boston
S70
2017
70
11923
660
23
0
0
0
All Children
Boston
S70
2017
80
432
0
0
0
0
0
All Children
Boston
S70
2017
90
0
0
0
0
0
0
All Children
Boston
S70
2017
100
0
0
0
0
0
0
All Children
Boston
S75
2015
0
862212
765847
702953
654281
613301
578919
All Children
Boston
S75
2015
10
828581
725457
659970
606952
565107
531203
All Children
Boston
S75
2015
20
729826
609137
535981
482758
440799
404847
All Children
Boston
S75
2015
30
594164
461028
387895
336902
297446
267524
All Children
Boston
S75
2015
40
417408
285637
219558
174731
143012
120235
All Children
Boston
S75
2015
50
218625
107765
60845
34974
21730
12970
All Children
Boston
S75
2015
60
68559
14677
3823
1069
296
91
All Children
Boston
S75
2015
70
12788
341
0
0
0
0
All Children
Boston
S75
2015
80
1047
0
0
0
0
0
All Children
Boston
S75
2015
90
23
0
0
0
0
0
All Children
Boston
S75
2015
100
0
0
0
0
0
0
All Children
Boston
S75
2016
0
865716
770989
706503
658309
617829
583447
All Children
Boston
S75
2016
10
833905
732921
665295
614165
571933
536027
All Children
Boston
S75
2016
20
737950
616577
541215
486103
441982
407760
All Children
Boston
S75
2016
30
598032
463827
389966
338950
301337
270255
All Children
Boston
S75
2016
40
434087
297651
229729
184379
152000
126583
All Children
Boston
S75
2016
50
244815
127971
76705
48649
31629
20684
All Children
Boston
S75
2016
60
82553
20957
5643
1479
387
137
All Children
Boston
S75
2016
70
12356
819
0
0
0
0
All Children
Boston
S75
2016
80
865
0
0
0
0
0
3D-Attachment4-4
-------
Study Group
Study Area
AQ
Scenario
Year
Benchmark
(PPb)
Number of People with 7-hr Exposure at or above Benchmark
>1 Day
> 2 Days
> 3 Days
> 4 Days
> 5 Days
> 6 Days
All Children
Boston
S75
2016
90
0
0
0
0
0
0
All Children
Boston
S75
2016
100
0
0
0
0
0
0
All Children
Boston
S75
2017
0
862462
763776
699927
649867
610434
576484
All Children
Boston
S75
2017
10
826806
723091
658263
605701
564447
527357
All Children
Boston
S75
2017
20
729462
609432
531726
475659
431902
395450
All Children
Boston
S75
2017
30
596212
457569
381296
329712
289983
258604
All Children
Boston
S75
2017
40
440594
303226
233188
184425
150452
124285
All Children
Boston
S75
2017
50
275738
145651
85557
52267
32425
20206
All Children
Boston
S75
2017
60
119711
35724
11400
2890
887
182
All Children
Boston
S75
2017
70
26532
3322
455
23
0
0
All Children
Boston
S75
2017
80
1957
23
0
0
0
0
All Children
Boston
S75
2017
90
0
0
0
0
0
0
All Children
Boston
S75
2017
100
0
0
0
0
0
0
All Children
Dallas
S65
2015
0
931702
848139
790894
748592
711327
679098
All Children
Dallas
S65
2015
10
886185
790965
727382
681320
641596
608800
All Children
Dallas
S65
2015
20
783658
667299
593194
540819
500811
466643
All Children
Dallas
S65
2015
30
636205
503034
429047
377902
341606
310985
All Children
Dallas
S65
2015
40
459431
325149
255276
210066
176632
151780
All Children
Dallas
S65
2015
50
231229
118748
69021
44146
28635
19366
All Children
Dallas
S65
2015
60
39772
7188
1537
213
24
24
All Children
Dallas
S65
2015
70
1017
0
0
0
0
0
All Children
Dallas
S65
2015
80
0
0
0
0
0
0
All Children
Dallas
S65
2015
90
0
0
0
0
0
0
All Children
Dallas
S65
2015
100
0
0
0
0
0
0
All Children
Dallas
S65
2016
0
933499
848896
794440
751618
716860
687137
All Children
Dallas
S65
2016
10
888573
794582
736249
689313
651291
619157
All Children
Dallas
S65
2016
20
783895
670018
601091
547203
505729
473737
All Children
Dallas
S65
2016
30
627078
496176
424507
375443
338036
310016
All Children
Dallas
S65
2016
40
418005
287411
220423
177081
145372
120497
All Children
Dallas
S65
2016
50
152608
62873
29486
16292
7992
4635
All Children
Dallas
S65
2016
60
12343
307
47
24
0
0
All Children
Dallas
S65
2016
70
946
0
0
0
0
0
All Children
Dallas
S65
2016
80
0
0
0
0
0
0
All Children
Dallas
S65
2016
90
0
0
0
0
0
0
All Children
Dallas
S65
2016
100
0
0
0
0
0
0
All Children
Dallas
S65
2017
0
938796
854973
800967
757435
720619
688911
All Children
Dallas
S65
2017
10
899119
806003
744430
698038
658219
624146
All Children
Dallas
S65
2017
20
798578
682053
612749
560587
518214
484850
All Children
Dallas
S65
2017
30
647318
516937
444204
394312
357307
328081
All Children
Dallas
S65
2017
40
455577
319970
251659
205101
171737
144072
All Children
Dallas
S65
2017
50
218200
102243
54645
31094
18278
11208
All Children
Dallas
S65
2017
60
36343
3547
497
24
0
0
All Children
Dallas
S65
2017
70
922
0
0
0
0
0
All Children
Dallas
S65
2017
80
0
0
0
0
0
0
All Children
Dallas
S65
2017
90
0
0
0
0
0
0
All Children
Dallas
S65
2017
100
0
0
0
0
0
0
All Children
Dallas
S70
2015
0
931702
848139
790894
748592
711327
679098
3D-Attachment4-5
-------
AQ
Benchmark
Number of People with 7-hr Exposure at or above Benchmark
Study Group
Study Area
Scenario
Year
(PPb)
>1 Day
> 2 Days
> 3 Days
> 4 Days
> 5 Days
> 6 Days
All Children
Dallas
S70
2015
10
887982
792738
729415
683567
643653
611708
All Children
Dallas
S70
2015
20
793802
677797
603290
551578
510955
477284
All Children
Dallas
S70
2015
30
657793
526466
449311
398261
360357
329169
All Children
Dallas
S70
2015
40
499298
363786
292211
244069
209664
181810
All Children
Dallas
S70
2015
50
304601
181905
122294
87063
63748
47386
All Children
Dallas
S70
2015
60
96261
29273
10759
4422
1773
780
All Children
Dallas
S70
2015
70
9718
757
24
0
0
0
All Children
Dallas
S70
2015
80
236
0
0
0
0
0
All Children
Dallas
S70
2015
90
0
0
0
0
0
0
All Children
Dallas
S70
2015
100
0
0
0
0
0
0
All Children
Dallas
S70
2016
0
933499
848896
794440
751618
716860
687137
All Children
Dallas
S70
2016
10
890677
797491
738992
692008
654436
622774
All Children
Dallas
S70
2016
20
792218
680233
611377
557111
515755
483219
All Children
Dallas
S70
2016
30
648784
518356
444606
395872
358158
328294
All Children
Dallas
S70
2016
40
459006
327111
259414
214512
180108
154405
All Children
Dallas
S70
2016
50
218271
109455
63086
39062
23929
15157
All Children
Dallas
S70
2016
60
34499
5226
1301
260
71
47
All Children
Dallas
S70
2016
70
3168
24
0
0
0
0
All Children
Dallas
S70
2016
80
284
0
0
0
0
0
All Children
Dallas
S70
2016
90
0
0
0
0
0
0
All Children
Dallas
S70
2016
100
0
0
0
0
0
0
All Children
Dallas
S70
2017
0
938796
854973
800967
757435
720619
688911
All Children
Dallas
S70
2017
10
901176
808509
747575
701017
660796
627007
All Children
Dallas
S70
2017
20
807256
691914
622893
571369
528996
494900
All Children
Dallas
S70
2017
30
668552
537958
465106
414576
377145
346217
All Children
Dallas
S70
2017
40
494308
359222
288215
242106
207323
177838
All Children
Dallas
S70
2017
50
279820
153388
95693
62471
40859
28185
All Children
Dallas
S70
2017
60
78621
16907
4469
1277
402
95
All Children
Dallas
S70
2017
70
4705
47
0
0
0
0
All Children
Dallas
S70
2017
80
0
0
0
0
0
0
All Children
Dallas
S70
2017
90
0
0
0
0
0
0
All Children
Dallas
S70
2017
100
0
0
0
0
0
0
All Children
Dallas
S75
2015
0
931702
848139
790894
748592
711327
679098
All Children
Dallas
S75
2015
10
888668
794204
730645
684867
644859
612559
All Children
Dallas
S75
2015
20
799950
685009
610975
559002
517670
483904
All Children
Dallas
S75
2015
30
674652
543042
466099
412732
374260
343048
All Children
Dallas
S75
2015
40
528287
393106
319687
269700
233121
204250
All Children
Dallas
S75
2015
50
362036
231938
164691
123760
95859
75405
All Children
Dallas
S75
2015
60
162894
67602
32725
16670
9293
4942
All Children
Dallas
S75
2015
70
29912
4280
CO
CNJ
CO
95
24
0
All Children
Dallas
S75
2015
80
2081
24
0
0
0
0
All Children
Dallas
S75
2015
90
47
0
0
0
0
0
All Children
Dallas
S75
2015
100
0
0
0
0
0
0
All Children
Dallas
S75
2016
0
933499
848896
794440
751618
716860
687137
All Children
Dallas
S75
2016
10
891623
799217
740623
693663
656138
624311
All Children
Dallas
S75
2016
20
798129
686570
618211
563873
522518
489579
All Children
Dallas
S75
2016
30
664745
534600
459692
409871
370808
341890
3D-Attachment4-6
-------
Study Group
Study Area
AQ
Scenario
Year
Benchmark
(PPb)
Number of People with 7-hr Exposure at or above Benchmark
>1 Day
> 2 Days
> 3 Days
> 4 Days
> 5 Days
> 6 Days
All Children
Dallas
S75
2016
40
489508
355391
285992
241160
206283
178736
All Children
Dallas
S75
2016
50
273507
151662
96852
65734
45399
32347
All Children
Dallas
S75
2016
60
66964
15819
5131
1726
520
189
All Children
Dallas
S75
2016
70
9860
166
24
24
0
0
All Children
Dallas
S75
2016
80
1419
0
0
0
0
0
All Children
Dallas
S75
2016
90
71
0
0
0
0
0
All Children
Dallas
S75
2016
100
0
0
0
0
0
0
All Children
Dallas
S75
2017
0
938796
854973
800967
757435
720619
688911
All Children
Dallas
S75
2017
10
902051
809810
749372
702791
662901
629040
All Children
Dallas
S75
2017
20
813262
699267
631098
577730
535806
502064
All Children
Dallas
S75
2017
30
683118
554250
479837
427557
390008
359151
All Children
Dallas
S75
2017
40
522257
387313
316022
267714
232411
202500
All Children
Dallas
S75
2017
50
328932
197936
133644
94251
67768
49017
All Children
Dallas
S75
2017
60
125037
38755
13998
5344
2317
851
All Children
Dallas
S75
2017
70
16126
804
71
0
0
0
All Children
Dallas
S75
2017
80
71
0
0
0
0
0
All Children
Dallas
S75
2017
90
0
0
0
0
0
0
All Children
Dallas
S75
2017
100
0
0
0
0
0
0
All Children
Detroit
S65
2015
0
658727
585868
537967
501043
471369
444747
All Children
Detroit
S65
2015
10
631377
552916
501598
463617
433093
405847
All Children
Detroit
S65
2015
20
556680
464033
407303
366963
334324
307407
All Children
Detroit
S65
2015
30
448043
347730
291521
253903
225322
202412
All Children
Detroit
S65
2015
40
314483
214118
161863
128912
106001
87409
All Children
Detroit
S65
2015
50
142110
62349
31096
16129
9261
5064
All Children
Detroit
S65
2015
60
14932
1179
121
17
0
0
All Children
Detroit
S65
2015
70
87
0
0
0
0
0
All Children
Detroit
S65
2015
80
0
0
0
0
0
0
All Children
Detroit
S65
2015
90
0
0
0
0
0
0
All Children
Detroit
S65
2015
100
0
0
0
0
0
0
All Children
Detroit
S65
2016
0
666184
595199
547679
510322
479989
454772
All Children
Detroit
S65
2016
10
639771
562646
512264
474179
442545
416842
All Children
Detroit
S65
2016
20
566774
476190
421542
379832
347903
320883
All Children
Detroit
S65
2016
30
464831
363269
306783
267657
238139
215645
All Children
Detroit
S65
2016
40
336769
231982
177941
145318
120205
100521
All Children
Detroit
S65
2016
50
178287
88571
50625
28321
16580
9903
All Children
Detroit
S65
2016
60
38276
5498
884
87
17
0
All Children
Detroit
S65
2016
70
815
17
0
0
0
0
All Children
Detroit
S65
2016
80
0
0
0
0
0
0
All Children
Detroit
S65
2016
90
0
0
0
0
0
0
All Children
Detroit
S65
2016
100
0
0
0
0
0
0
All Children
Detroit
S65
2017
0
661623
588886
542771
505899
475913
448996
All Children
Detroit
S65
2017
10
635990
558553
509212
469409
439267
411813
All Children
Detroit
S65
2017
20
565681
473381
417328
376519
344695
316963
All Children
Detroit
S65
2017
30
458899
357563
301129
261274
232675
209019
All Children
Detroit
S65
2017
40
326727
225721
174316
140133
116008
96948
All Children
Detroit
S65
2017
50
159730
72980
40427
22199
12695
7770
All Children
Detroit
S65
2017
60
17725
1353
139
35
0
0
3D-Attachment4-7
-------
Study Group
Study Area
AQ
Scenario
Year
Benchmark
(PPb)
Number of People with 7-hr Exposure at or above Benchmark
>1 Day
> 2 Days
> 3 Days
> 4 Days
> 5 Days
> 6 Days
All Children
Detroit
S65
2017
70
330
0
0
0
0
0
All Children
Detroit
S65
2017
80
0
0
0
0
0
0
All Children
Detroit
S65
2017
90
0
0
0
0
0
0
All Children
Detroit
S65
2017
100
0
0
0
0
0
0
All Children
Detroit
S70
2015
0
658727
585868
537967
501043
471369
444747
All Children
Detroit
S70
2015
10
632365
553732
502535
464501
433630
406766
All Children
Detroit
S70
2015
20
562663
471733
415056
374352
341677
314708
All Children
Detroit
S70
2015
30
464952
364032
306557
269044
239890
215835
All Children
Detroit
S70
2015
40
348025
246394
192491
158048
132675
112609
All Children
Detroit
S70
2015
50
197330
106868
65470
41398
26882
17499
All Children
Detroit
S70
2015
60
52203
10805
2549
746
191
17
All Children
Detroit
S70
2015
70
1492
69
17
0
0
0
All Children
Detroit
S70
2015
80
0
0
0
0
0
0
All Children
Detroit
S70
2015
90
0
0
0
0
0
0
All Children
Detroit
S70
2015
100
0
0
0
0
0
0
All Children
Detroit
S70
2016
0
666184
595199
547679
510322
479989
454772
All Children
Detroit
S70
2016
10
640326
564016
513287
475132
443204
417831
All Children
Detroit
S70
2016
20
573121
483076
429069
386908
354823
327855
All Children
Detroit
S70
2016
30
481740
380855
323623
283994
253244
228825
All Children
Detroit
S70
2016
40
369669
264361
208569
172027
146203
125478
All Children
Detroit
S70
2016
50
235329
136820
90965
62626
43653
30697
All Children
Detroit
S70
2016
60
95509
28894
9764
3035
1041
399
All Children
Detroit
S70
2016
70
9487
520
0
0
0
0
All Children
Detroit
S70
2016
80
52
0
0
0
0
0
All Children
Detroit
S70
2016
90
0
0
0
0
0
0
All Children
Detroit
S70
2016
100
0
0
0
0
0
0
All Children
Detroit
S70
2017
0
661623
588886
542771
505899
475913
448996
All Children
Detroit
S70
2017
10
636528
559455
510460
470120
440203
412610
All Children
Detroit
S70
2017
20
572202
479763
423866
383092
351372
323467
All Children
Detroit
S70
2017
30
475115
374508
316980
277005
246619
223015
All Children
Detroit
S70
2017
40
359315
255586
202498
166078
140705
120517
All Children
Detroit
S70
2017
50
221142
123379
79327
52602
35710
23934
All Children
Detroit
S70
2017
60
61169
13788
3885
1214
451
156
All Children
Detroit
S70
2017
70
4301
139
0
0
0
0
All Children
Detroit
S70
2017
80
35
0
0
0
0
0
All Children
Detroit
S70
2017
90
0
0
0
0
0
0
All Children
Detroit
S70
2017
100
0
0
0
0
0
0
All Children
Detroit
S75
2015
0
658727
585868
537967
501043
471369
444747
All Children
Detroit
S75
2015
10
631446
552691
501182
463409
432347
405430
All Children
Detroit
S75
2015
20
564398
473589
417102
376173
343221
315784
All Children
Detroit
S75
2015
30
473277
372357
314154
274733
245596
222148
All Children
Detroit
S75
2015
40
366634
263165
208204
172269
146047
125998
All Children
Detroit
S75
2015
50
234982
137375
91242
63424
44329
31738
All Children
Detroit
S75
2015
60
89387
26223
8776
3018
1041
572
All Children
Detroit
S75
2015
70
9296
416
52
0
0
0
All Children
Detroit
S75
2015
80
69
0
0
0
0
0
All Children
Detroit
S75
2015
90
0
0
0
0
0
0
3D-Attachment4-8
-------
Study Group
Study Area
AQ
Scenario
Year
Benchmark
(PPb)
Number of People with 7-hr Exposure at or above Benchmark
>1 Day
> 2 Days
> 3 Days
> 4 Days
> 5 Days
> 6 Days
All Children
Detroit
S75
2015
100
0
0
0
0
0
0
All Children
Detroit
S75
2016
0
666184
595199
547679
510322
479989
454772
All Children
Detroit
S75
2016
10
639615
563132
512056
473970
442059
416513
All Children
Detroit
S75
2016
20
575254
485244
430665
388035
355812
328756
All Children
Detroit
S75
2016
30
489926
389666
332034
291174
260667
234965
All Children
Detroit
S75
2016
40
387931
282450
226328
188052
160615
139526
All Children
Detroit
S75
2016
50
271472
167587
118610
86698
65436
49879
All Children
Detroit
S75
2016
60
144191
60701
28061
12921
5983
2792
All Children
Detroit
S75
2016
70
34981
4561
607
69
0
0
All Children
Detroit
S75
2016
80
1249
17
0
0
0
0
All Children
Detroit
S75
2016
90
0
0
0
0
0
0
All Children
Detroit
S75
2016
100
0
0
0
0
0
0
All Children
Detroit
S75
2017
0
661623
588886
542771
505899
475913
448996
All Children
Detroit
S75
2017
10
635452
558501
509160
468629
438036
410859
All Children
Detroit
S75
2017
20
573607
481029
425965
384601
352482
324178
All Children
Detroit
S75
2017
30
483544
381740
323935
282971
252655
229450
All Children
Detroit
S75
2017
40
376519
271472
216928
179536
153695
132606
All Children
Detroit
S75
2017
50
253262
154180
105880
74298
53070
39248
All Children
Detroit
S75
2017
60
109747
38432
14950
6157
2549
989
All Children
Detroit
S75
2017
70
15106
694
35
0
0
0
All Children
Detroit
S75
2017
80
746
0
0
0
0
0
All Children
Detroit
S75
2017
90
0
0
0
0
0
0
All Children
Detroit
S75
2017
100
0
0
0
0
0
0
All Children
Philadelphia
S65
2015
0
844309
758097
699407
656192
618084
586655
All Children
Philadelphia
S65
2015
10
815062
724070
661496
615312
577248
545077
All Children
Philadelphia
S65
2015
20
730465
621184
550272
501426
463187
430295
All Children
Philadelphia
S65
2015
30
600755
476806
407466
359296
323393
294801
All Children
Philadelphia
S65
2015
40
413773
289279
226901
184974
153894
130169
All Children
Philadelphia
S65
2015
50
163999
72811
37955
21040
12419
7268
All Children
Philadelphia
S65
2015
60
20975
2204
437
44
0
0
All Children
Philadelphia
S65
2015
70
393
0
0
0
0
0
All Children
Philadelphia
S65
2015
80
0
0
0
0
0
0
All Children
Philadelphia
S65
2015
90
0
0
0
0
0
0
All Children
Philadelphia
S65
2015
100
0
0
0
0
0
0
All Children
Philadelphia
S65
2016
0
846469
759581
703030
656803
619197
588532
All Children
Philadelphia
S65
2016
10
817310
724070
664639
616928
580064
548831
All Children
Philadelphia
S65
2016
20
729047
618826
551276
500487
459957
429095
All Children
Philadelphia
S65
2016
30
593814
471022
401660
353730
317609
287009
All Children
Philadelphia
S65
2016
40
401660
274416
211776
169674
139489
116200
All Children
Philadelphia
S65
2016
50
155531
63644
30556
14754
7508
3623
All Children
Philadelphia
S65
2016
60
21171
1702
109
0
0
0
All Children
Philadelphia
S65
2016
70
175
0
0
0
0
0
All Children
Philadelphia
S65
2016
80
0
0
0
0
0
0
All Children
Philadelphia
S65
2016
90
0
0
0
0
0
0
All Children
Philadelphia
S65
2016
100
0
0
0
0
0
0
All Children
Philadelphia
S65
2017
0
843414
754867
696046
650277
613959
581897
All Children
Philadelphia
S65
2017
10
813403
717894
655690
609703
572185
540101
3D-Attachment4-9
-------
Study Group
Study Area
AQ
Scenario
Year
Benchmark
(PPb)
Number of People with 7-hr Exposure at or above Benchmark
>1 Day
> 2 Days
> 3 Days
> 4 Days
> 5 Days
> 6 Days
All Children
Philadelphia
S65
2017
20
721168
607914
538857
488286
449153
417418
All Children
Philadelphia
S65
2017
30
575524
450462
381231
332298
296896
268042
All Children
Philadelphia
S65
2017
40
380860
254227
193049
151602
122268
99875
All Children
Philadelphia
S65
2017
50
139991
52447
23550
10651
5042
2335
All Children
Philadelphia
S65
2017
60
18203
1484
109
22
0
0
All Children
Philadelphia
S65
2017
70
480
0
0
0
0
0
All Children
Philadelphia
S65
2017
80
0
0
0
0
0
0
All Children
Philadelphia
S65
2017
90
0
0
0
0
0
0
All Children
Philadelphia
S65
2017
100
0
0
0
0
0
0
All Children
Philadelphia
S70
2015
0
844309
758097
699407
656192
618084
586655
All Children
Philadelphia
S70
2015
10
816219
725838
663744
617080
578973
547085
All Children
Philadelphia
S70
2015
20
738344
629521
559853
509807
471240
439047
All Children
Philadelphia
S70
2015
30
621140
497715
427087
378590
342272
312021
All Children
Philadelphia
S70
2015
40
460066
332145
266624
223933
192023
165156
All Children
Philadelphia
S70
2015
50
238250
129711
80493
52426
35947
25165
All Children
Philadelphia
S70
2015
60
54674
11764
3405
1157
349
65
All Children
Philadelphia
S70
2015
70
4627
131
0
0
0
0
All Children
Philadelphia
S70
2015
80
44
0
0
0
0
0
All Children
Philadelphia
S70
2015
90
0
0
0
0
0
0
All Children
Philadelphia
S70
2015
100
0
0
0
0
0
0
All Children
Philadelphia
S70
2016
0
846469
759581
703030
656803
619197
588532
All Children
Philadelphia
S70
2016
10
818467
725860
666472
618717
581286
550795
All Children
Philadelphia
S70
2016
20
736598
627993
560268
509698
469429
437323
All Children
Philadelphia
S70
2016
30
614963
490753
422089
372719
336204
305888
All Children
Philadelphia
S70
2016
40
448913
319726
253114
209986
177160
151296
All Children
Philadelphia
S70
2016
50
229476
119234
69777
44743
28068
18115
All Children
Philadelphia
S70
2016
60
53560
9210
1899
480
44
22
All Children
Philadelphia
S70
2016
70
5151
65
0
0
0
0
All Children
Philadelphia
S70
2016
80
0
0
0
0
0
0
All Children
Philadelphia
S70
2016
90
0
0
0
0
0
0
All Children
Philadelphia
S70
2016
100
0
0
0
0
0
0
All Children
Philadelphia
S70
2017
0
843414
754867
696046
650277
613959
581897
All Children
Philadelphia
S70
2017
10
814844
719487
657436
611187
573996
542196
All Children
Philadelphia
S70
2017
20
728807
616753
547303
497584
458232
425974
All Children
Philadelphia
S70
2017
30
597263
471153
401594
352203
314640
285001
All Children
Philadelphia
S70
2017
40
426148
296481
233143
188531
156796
132133
All Children
Philadelphia
S70
2017
50
205599
99984
55787
31080
18705
11589
All Children
Philadelphia
S70
2017
60
51116
9451
2073
349
87
0
All Children
Philadelphia
S70
2017
70
4191
175
0
0
0
0
All Children
Philadelphia
S70
2017
80
87
0
0
0
0
0
All Children
Philadelphia
S70
2017
90
0
0
0
0
0
0
All Children
Philadelphia
S70
2017
100
0
0
0
0
0
0
All Children
Philadelphia
S75
2015
0
844309
758097
699407
656192
618084
586655
All Children
Philadelphia
S75
2015
10
816895
726187
664508
617932
579431
547740
All Children
Philadelphia
S75
2015
20
745307
637357
568758
518013
478247
446403
All Children
Philadelphia
S75
2015
30
640718
517162
446206
396422
359711
328784
All Children
Philadelphia
S75
2015
40
503608
373701
306281
261167
226835
199815
3D-Attachment4-10
-------
Study Group
Study Area
AQ
Scenario
Year
Benchmark
(PPb)
Number of People with 7-hr Exposure at or above Benchmark
>1 Day
> 2 Days
> 3 Days
> 4 Days
> 5 Days
> 6 Days
All Children
Philadelphia
S75
2015
50
316932
197130
138681
102494
78267
60239
All Children
Philadelphia
S75
2015
60
113669
38850
16784
8185
3536
1550
All Children
Philadelphia
S75
2015
70
20036
1964
393
87
0
0
All Children
Philadelphia
S75
2015
80
1550
22
0
0
0
0
All Children
Philadelphia
S75
2015
90
22
0
0
0
0
0
All Children
Philadelphia
S75
2015
100
0
0
0
0
0
0
All Children
Philadelphia
S75
2016
0
846469
759581
703030
656803
619197
588532
All Children
Philadelphia
S75
2016
10
819165
726668
666821
620201
581963
551101
All Children
Philadelphia
S75
2016
20
742906
637357
568496
518100
477985
445028
All Children
Philadelphia
S75
2016
30
635960
512098
442671
391424
354189
323305
All Children
Philadelphia
S75
2016
40
491582
362504
293993
246675
212605
186654
All Children
Philadelphia
S75
2016
50
310210
189142
129994
94506
69100
52076
All Children
Philadelphia
S75
2016
60
115153
36493
13139
4845
1986
829
All Children
Philadelphia
S75
2016
70
18377
1113
87
0
0
0
All Children
Philadelphia
S75
2016
80
240
0
0
0
0
0
All Children
Philadelphia
S75
2016
90
0
0
0
0
0
0
All Children
Philadelphia
S75
2016
100
0
0
0
0
0
0
All Children
Philadelphia
S75
2017
0
843414
754867
696046
650277
613959
581897
All Children
Philadelphia
S75
2017
10
815477
720513
658811
612170
574171
542982
All Children
Philadelphia
S75
2017
20
736511
624305
554484
505180
465871
433176
All Children
Philadelphia
S75
2017
30
617211
492259
421696
371388
333476
302069
All Children
Philadelphia
S75
2017
40
467094
337296
271076
225482
191303
164588
All Children
Philadelphia
S75
2017
50
281509
162624
105091
70846
48104
33830
All Children
Philadelphia
S75
2017
60
107950
33306
11677
4060
1353
371
All Children
Philadelphia
S75
2017
70
16173
1310
109
0
0
0
All Children
Philadelphia
S75
2017
80
633
0
0
0
0
0
All Children
Philadelphia
S75
2017
90
0
0
0
0
0
0
All Children
Philadelphia
S75
2017
100
0
0
0
0
0
0
All Children
Phoenix
S65
2015
0
573408
527665
497306
475283
455327
438640
All Children
Phoenix
S65
2015
10
554797
505302
472721
448448
427020
409314
All Children
Phoenix
S65
2015
20
507326
445844
408479
379182
355687
336241
All Children
Phoenix
S65
2015
30
435880
362750
321153
291530
268772
249707
All Children
Phoenix
S65
2015
40
332632
254859
213489
185041
164306
147209
All Children
Phoenix
S65
2015
50
173265
100206
66323
46720
33572
25094
All Children
Phoenix
S65
2015
60
11323
2066
510
142
71
14
All Children
Phoenix
S65
2015
70
0
0
0
0
0
0
All Children
Phoenix
S65
2015
80
0
0
0
0
0
0
All Children
Phoenix
S65
2015
90
0
0
0
0
0
0
All Children
Phoenix
S65
2015
100
0
0
0
0
0
0
All Children
Phoenix
S65
2016
0
573705
529561
500023
476840
457549
440310
All Children
Phoenix
S65
2016
10
555165
506718
475722
451067
429483
411494
All Children
Phoenix
S65
2016
20
508459
447599
410036
382565
359905
340812
All Children
Phoenix
S65
2016
30
435611
365510
323602
294276
271418
253543
All Children
Phoenix
S65
2016
40
327140
251561
210064
182904
161574
145383
All Children
Phoenix
S65
2016
50
145100
77178
47230
30996
21683
15526
All Children
Phoenix
S65
2016
60
7982
722
142
42
42
28
All Children
Phoenix
S65
2016
70
0
0
0
0
0
0
3D-Attachment4-11
-------
Study Group
Study Area
AQ
Scenario
Year
Benchmark
(PPb)
Number of People with 7-hr Exposure at or above Benchmark
>1 Day
> 2 Days
> 3 Days
> 4 Days
> 5 Days
> 6 Days
All Children
Phoenix
S65
2016
80
0
0
0
0
0
0
All Children
Phoenix
S65
2016
90
0
0
0
0
0
0
All Children
Phoenix
S65
2016
100
0
0
0
0
0
0
All Children
Phoenix
S65
2017
0
575177
529830
499174
476614
457889
440961
All Children
Phoenix
S65
2017
10
557655
508204
475637
451548
432045
414169
All Children
Phoenix
S65
2017
20
512337
452510
414792
387080
363727
343572
All Children
Phoenix
S65
2017
30
443282
373436
332504
303263
280420
261327
All Children
Phoenix
S65
2017
40
347408
270767
230416
202747
180993
164080
All Children
Phoenix
S65
2017
50
187234
113524
80164
59147
44456
34378
All Children
Phoenix
S65
2017
60
25334
4388
1076
354
71
14
All Children
Phoenix
S65
2017
70
0
0
0
0
0
0
All Children
Phoenix
S65
2017
80
0
0
0
0
0
0
All Children
Phoenix
S65
2017
90
0
0
0
0
0
0
All Children
Phoenix
S65
2017
100
0
0
0
0
0
0
All Children
Phoenix
S70
2015
0
573408
527665
497306
475283
455327
438640
All Children
Phoenix
S70
2015
10
555674
506336
473953
449439
428379
410532
All Children
Phoenix
S70
2015
20
511742
452100
414608
385636
362071
342256
All Children
Phoenix
S70
2015
30
449213
376988
334967
305627
282005
262318
All Children
Phoenix
S70
2015
40
361405
283817
240508
212003
190192
173803
All Children
Phoenix
S70
2015
50
236842
159409
121322
95521
77334
63633
All Children
Phoenix
S70
2015
60
68177
24061
10912
5803
3284
1713
All Children
Phoenix
S70
2015
70
807
0
0
0
0
0
All Children
Phoenix
S70
2015
80
0
0
0
0
0
0
All Children
Phoenix
S70
2015
90
0
0
0
0
0
0
All Children
Phoenix
S70
2015
100
0
0
0
0
0
0
All Children
Phoenix
S70
2016
0
573705
529561
500023
476840
457549
440310
All Children
Phoenix
S70
2016
10
556084
507807
476911
452737
430714
412626
All Children
Phoenix
S70
2016
20
512988
453869
416518
388339
366189
347903
All Children
Phoenix
S70
2016
30
448279
379621
337543
308231
284595
265955
All Children
Phoenix
S70
2016
40
357598
281694
239913
212144
190277
173067
All Children
Phoenix
S70
2016
50
218768
142326
103645
79598
62388
49961
All Children
Phoenix
S70
2016
60
50754
14153
5576
2633
1189
594
All Children
Phoenix
S70
2016
70
269
14
0
0
0
0
All Children
Phoenix
S70
2016
80
0
0
0
0
0
0
All Children
Phoenix
S70
2016
90
0
0
0
0
0
0
All Children
Phoenix
S70
2016
100
0
0
0
0
0
0
All Children
Phoenix
S70
2017
0
575177
529830
499174
476614
457889
440961
All Children
Phoenix
S70
2017
10
558575
509647
476911
452963
433616
415655
All Children
Phoenix
S70
2017
20
517517
458568
421076
393434
371256
350437
All Children
Phoenix
S70
2017
30
456587
386556
346176
316992
293823
274858
All Children
Phoenix
S70
2017
40
375955
298579
258256
230586
208026
190546
All Children
Phoenix
S70
2017
50
254335
175954
136834
111104
93058
78240
All Children
Phoenix
S70
2017
60
89775
36643
18074
9554
5392
3326
All Children
Phoenix
S70
2017
70
4784
269
42
14
0
0
All Children
Phoenix
S70
2017
80
0
0
0
0
0
0
All Children
Phoenix
S70
2017
90
0
0
0
0
0
0
All Children
Phoenix
S70
2017
100
0
0
0
0
0
0
3D-Attachment4-12
-------
Study Group
Study Area
AQ
Scenario
Year
Benchmark
(PPb)
Number of People with 7-hr Exposure at or above Benchmark
>1 Day
> 2 Days
> 3 Days
> 4 Days
> 5 Days
> 6 Days
All Children
Phoenix
S75
2015
0
573408
527665
497306
475283
455327
438640
All Children
Phoenix
S75
2015
10
555575
506251
473372
449198
428167
410065
All Children
Phoenix
S75
2015
20
513992
455016
417170
388849
365198
344931
All Children
Phoenix
S75
2015
30
456728
385763
343332
313567
289308
270527
All Children
Phoenix
S75
2015
40
379408
301664
258553
229539
207007
189471
All Children
Phoenix
S75
2015
50
276655
198147
157470
130678
110382
94134
All Children
Phoenix
S75
2015
60
136778
69054
41186
25844
16899
11535
All Children
Phoenix
S75
2015
70
13191
1613
226
28
0
0
All Children
Phoenix
S75
2015
80
198
0
0
0
0
0
All Children
Phoenix
S75
2015
90
0
0
0
0
0
0
All Children
Phoenix
S75
2015
100
0
0
0
0
0
0
All Children
Phoenix
S75
2016
0
573705
529561
500023
476840
457549
440310
All Children
Phoenix
S75
2016
10
555830
507850
476882
452241
430403
412173
All Children
Phoenix
S75
2016
20
515309
456742
419590
391241
368765
350026
All Children
Phoenix
S75
2016
30
456303
387915
346219
316171
292153
273230
All Children
Phoenix
S75
2016
40
376564
301268
258058
229256
207134
189584
All Children
Phoenix
S75
2016
50
264469
186329
145666
117175
97856
82726
All Children
Phoenix
S75
2016
60
111769
50527
27599
16574
9794
6171
All Children
Phoenix
S75
2016
70
8025
863
113
28
0
0
All Children
Phoenix
S75
2016
80
0
0
0
0
0
0
All Children
Phoenix
S75
2016
90
0
0
0
0
0
0
All Children
Phoenix
S75
2016
100
0
0
0
0
0
0
All Children
Phoenix
S75
2017
0
575177
529830
499174
476614
457889
440961
All Children
Phoenix
S75
2017
10
558717
509322
476755
452836
433899
415358
All Children
Phoenix
S75
2017
20
519710
461611
424289
396152
374469
353734
All Children
Phoenix
S75
2017
30
464880
396180
355121
325612
302244
282543
All Children
Phoenix
S75
2017
40
393434
317233
276132
247315
224868
206228
All Children
Phoenix
S75
2017
50
294871
215216
174185
147237
127295
111274
All Children
Phoenix
S75
2017
60
151653
83660
53358
34902
24400
16956
All Children
Phoenix
S75
2017
70
29255
5775
1302
368
85
14
All Children
Phoenix
S75
2017
80
1062
14
0
0
0
0
All Children
Phoenix
S75
2017
90
0
0
0
0
0
0
All Children
Phoenix
S75
2017
100
0
0
0
0
0
0
All Children
Sacramento
S65
2015
0
311348
285237
266859
252643
241284
230547
All Children
Sacramento
S65
2015
10
297194
266813
246867
231750
219397
208582
All Children
Sacramento
S65
2015
20
261494
224025
200368
183411
170259
159218
All Children
Sacramento
S65
2015
30
207977
166113
141516
125033
113239
103596
All Children
Sacramento
S65
2015
40
127797
85902
64349
50218
40886
33494
All Children
Sacramento
S65
2015
50
32718
9930
3439
1359
551
225
All Children
Sacramento
S65
2015
60
1599
78
8
0
0
0
All Children
Sacramento
S65
2015
70
0
0
0
0
0
0
All Children
Sacramento
S65
2015
80
0
0
0
0
0
0
All Children
Sacramento
S65
2015
90
0
0
0
0
0
0
All Children
Sacramento
S65
2015
100
0
0
0
0
0
0
All Children
Sacramento
S65
2016
0
311681
285283
268039
253940
242146
232146
All Children
Sacramento
S65
2016
10
297411
267612
247962
233093
220857
210849
All Children
Sacramento
S65
2016
20
260516
223901
201493
184847
171633
160949
3D-Attachment4-13
-------
Study Group
Study Area
AQ
Scenario
Year
Benchmark
(PPb)
Number of People with 7-hr Exposure at or above Benchmark
>1 Day
> 2 Days
> 3 Days
> 4 Days
> 5 Days
> 6 Days
All Children
Sacramento
S65
2016
30
207891
165950
141718
125763
114039
104101
All Children
Sacramento
S65
2016
40
132176
88394
66608
52679
42400
34744
All Children
Sacramento
S65
2016
50
41716
16374
7430
3867
1840
908
All Children
Sacramento
S65
2016
60
2632
163
8
0
0
0
All Children
Sacramento
S65
2016
70
0
0
0
0
0
0
All Children
Sacramento
S65
2016
80
0
0
0
0
0
0
All Children
Sacramento
S65
2016
90
0
0
0
0
0
0
All Children
Sacramento
S65
2016
100
0
0
0
0
0
0
All Children
Sacramento
S65
2017
0
311363
284484
266269
252263
240694
230686
All Children
Sacramento
S65
2017
10
297395
267574
247418
231797
219157
209382
All Children
Sacramento
S65
2017
20
262069
225112
201657
185616
172409
161159
All Children
Sacramento
S65
2017
30
210352
167906
144746
128286
115755
105964
All Children
Sacramento
S65
2017
40
132277
88759
66965
52881
43230
35396
All Children
Sacramento
S65
2017
50
33564
9643
3408
1413
590
272
All Children
Sacramento
S65
2017
60
1266
8
0
0
0
0
All Children
Sacramento
S65
2017
70
0
0
0
0
0
0
All Children
Sacramento
S65
2017
80
0
0
0
0
0
0
All Children
Sacramento
S65
2017
90
0
0
0
0
0
0
All Children
Sacramento
S65
2017
100
0
0
0
0
0
0
All Children
Sacramento
S70
2015
0
311348
285237
266859
252643
241284
230547
All Children
Sacramento
S70
2015
10
298141
268063
248078
233218
220803
209863
All Children
Sacramento
S70
2015
20
266067
229289
206012
189405
175631
164645
All Children
Sacramento
S70
2015
30
220384
178318
154109
137067
124357
114676
All Children
Sacramento
S70
2015
40
156811
112890
90684
74869
63207
54558
All Children
Sacramento
S70
2015
50
71810
36313
21398
13354
8735
5544
All Children
Sacramento
S70
2015
60
10645
1281
186
8
0
0
All Children
Sacramento
S70
2015
70
707
31
0
0
0
0
All Children
Sacramento
S70
2015
80
0
0
0
0
0
0
All Children
Sacramento
S70
2015
90
0
0
0
0
0
0
All Children
Sacramento
S70
2015
100
0
0
0
0
0
0
All Children
Sacramento
S70
2016
0
311681
285283
268039
253940
242146
232146
All Children
Sacramento
S70
2016
10
298125
268816
249002
234320
222169
212060
All Children
Sacramento
S70
2016
20
265446
229320
207006
190546
177068
166136
All Children
Sacramento
S70
2016
30
219522
178846
154567
136772
124311
114396
All Children
Sacramento
S70
2016
40
159016
115079
91375
75630
63984
54892
All Children
Sacramento
S70
2016
50
81841
44434
28331
18960
13137
9340
All Children
Sacramento
S70
2016
60
18378
4278
1219
411
179
47
All Children
Sacramento
S70
2016
70
1203
16
0
0
0
0
All Children
Sacramento
S70
2016
80
0
0
0
0
0
0
All Children
Sacramento
S70
2016
90
0
0
0
0
0
0
All Children
Sacramento
S70
2016
100
0
0
0
0
0
0
All Children
Sacramento
S70
2017
0
311363
284484
266269
252263
240694
230686
All Children
Sacramento
S70
2017
10
298413
268730
248971
233155
220616
210655
All Children
Sacramento
S70
2017
20
266782
230213
207480
191090
177681
166788
All Children
Sacramento
S70
2017
30
222519
180049
156175
139909
127587
117144
All Children
Sacramento
S70
2017
40
161167
117308
93736
78184
66849
57881
All Children
Sacramento
S70
2017
50
76748
39247
23634
15381
10482
7050
3D-Attachment4-14
-------
Study Group
Study Area
AQ
Scenario
Year
Benchmark
(PPb)
Number of People with 7-hr Exposure at or above Benchmark
>1 Day
> 2 Days
> 3 Days
> 4 Days
> 5 Days
> 6 Days
All Children
Sacramento
S70
2017
60
15761
2244
435
54
8
0
All Children
Sacramento
S70
2017
70
272
0
0
0
0
0
All Children
Sacramento
S70
2017
80
0
0
0
0
0
0
All Children
Sacramento
S70
2017
90
0
0
0
0
0
0
All Children
Sacramento
S70
2017
100
0
0
0
0
0
0
All Children
Sacramento
S75
2015
0
311348
285237
266859
252643
241284
230547
All Children
Sacramento
S75
2015
10
298661
268715
248854
233823
221564
210624
All Children
Sacramento
S75
2015
20
268893
232806
209677
193077
179544
167999
All Children
Sacramento
S75
2015
30
228280
186167
162091
144901
131679
121694
All Children
Sacramento
S75
2015
40
173395
128783
105413
89838
77462
68425
All Children
Sacramento
S75
2015
50
99838
59015
39775
28479
20878
15699
All Children
Sacramento
S75
2015
60
29457
7919
2562
901
326
93
All Children
Sacramento
S75
2015
70
3603
202
23
0
0
0
All Children
Sacramento
S75
2015
80
116
0
0
0
0
0
All Children
Sacramento
S75
2015
90
0
0
0
0
0
0
All Children
Sacramento
S75
2015
100
0
0
0
0
0
0
All Children
Sacramento
S75
2016
0
311681
285283
268039
253940
242146
232146
All Children
Sacramento
S75
2016
10
298878
269383
249646
234996
222922
212767
All Children
Sacramento
S75
2016
20
268404
232798
210329
193955
180352
169389
All Children
Sacramento
S75
2016
30
226626
186734
162487
144334
131298
120731
All Children
Sacramento
S75
2016
40
173760
130157
106562
90226
78029
68565
All Children
Sacramento
S75
2016
50
109147
66243
46724
35094
26763
20342
All Children
Sacramento
S75
2016
60
41445
15878
7438
3766
1964
1040
All Children
Sacramento
S75
2016
70
6499
761
116
8
0
0
All Children
Sacramento
S75
2016
80
217
0
0
0
0
0
All Children
Sacramento
S75
2016
90
0
0
0
0
0
0
All Children
Sacramento
S75
2016
100
0
0
0
0
0
0
All Children
Sacramento
S75
2017
0
311363
284484
266269
252263
240694
230686
All Children
Sacramento
S75
2017
10
298739
269421
249670
233885
221331
211385
All Children
Sacramento
S75
2017
20
269841
233528
211059
194863
180996
170189
All Children
Sacramento
S75
2017
30
230050
187821
164086
147417
134466
123930
All Children
Sacramento
S75
2017
40
176990
133340
109505
93076
81010
71406
All Children
Sacramento
S75
2017
50
106570
64162
44465
32632
24713
19123
All Children
Sacramento
S75
2017
60
33851
9464
3230
1328
575
248
All Children
Sacramento
S75
2017
70
4643
O
CO
CNJ
16
0
0
0
All Children
Sacramento
S75
2017
80
54
0
0
0
0
0
All Children
Sacramento
S75
2017
90
0
0
0
0
0
0
All Children
Sacramento
S75
2017
100
0
0
0
0
0
0
All Children
St. Louis
S65
2015
0
355693
320478
297229
278469
263525
250485
All Children
St. Louis
S65
2015
10
338755
300106
275045
255985
239703
226653
All Children
St. Louis
S65
2015
20
297047
249820
221043
199980
183269
170010
All Children
St. Louis
S65
2015
30
235887
183552
155822
136744
121463
109297
All Children
St. Louis
S65
2015
40
159583
108140
82360
64911
52636
43356
All Children
St. Louis
S65
2015
50
65585
26864
12348
5828
2978
1439
All Children
St. Louis
S65
2015
60
4034
200
9
0
0
0
All Children
St. Louis
S65
2015
70
0
0
0
0
0
0
All Children
St. Louis
S65
2015
80
0
0
0
0
0
0
3D-Attachment4-15
-------
Study Group
Study Area
AQ
Scenario
Year
Benchmark
(PPb)
Number of People with 7-hr Exposure at or above Benchmark
>1 Day
> 2 Days
> 3 Days
> 4 Days
> 5 Days
> 6 Days
All Children
St. Louis
S65
2015
90
0
0
0
0
0
0
All Children
St. Louis
S65
2015
100
0
0
0
0
0
0
All Children
St. Louis
S65
2016
0
359080
325268
301645
283232
267933
254646
All Children
St. Louis
S65
2016
10
344483
306763
281083
260903
244975
231552
All Children
St. Louis
S65
2016
20
305898
258790
228693
206910
190855
177541
All Children
St. Louis
S65
2016
30
248399
195764
166413
146461
131635
119815
All Children
St. Louis
S65
2016
40
174308
122456
95655
77469
64037
54257
All Children
St. Louis
S65
2016
50
87059
42300
23049
13141
7522
4435
All Children
St. Louis
S65
2016
60
16847
2204
392
109
18
0
All Children
St. Louis
S65
2016
70
55
0
0
0
0
0
All Children
St. Louis
S65
2016
80
0
0
0
0
0
0
All Children
St. Louis
S65
2016
90
0
0
0
0
0
0
All Children
St. Louis
S65
2016
100
0
0
0
0
0
0
All Children
St. Louis
S65
2017
0
355702
320669
297356
279526
264764
252179
All Children
St. Louis
S65
2017
10
342115
304487
279471
260520
244875
231352
All Children
St. Louis
S65
2017
20
305215
259682
231033
210498
194106
180801
All Children
St. Louis
S65
2017
30
251869
200053
171804
151988
137309
125425
All Children
St. Louis
S65
2017
40
179900
128548
101711
83753
70503
60222
All Children
St. Louis
S65
2017
50
77387
35270
19178
11028
6693
3943
All Children
St. Louis
S65
2017
60
5764
464
27
0
0
0
All Children
St. Louis
S65
2017
70
146
0
0
0
0
0
All Children
St. Louis
S65
2017
80
0
0
0
0
0
0
All Children
St. Louis
S65
2017
90
0
0
0
0
0
0
All Children
St. Louis
S65
2017
100
0
0
0
0
0
0
All Children
St. Louis
S70
2015
0
355693
320478
297229
278469
263525
250485
All Children
St. Louis
S70
2015
10
339492
301063
276193
256868
240896
227636
All Children
St. Louis
S70
2015
20
301664
255257
225533
205134
188351
175019
All Children
St. Louis
S70
2015
30
246742
194862
166531
146433
131271
119023
All Children
St. Louis
S70
2015
40
181020
128612
101556
83106
69893
59484
All Children
St. Louis
S70
2015
50
102157
54403
32711
20781
13660
8751
All Children
St. Louis
S70
2015
60
22320
4071
883
209
36
18
All Children
St. Louis
S70
2015
70
446
9
0
0
0
0
All Children
St. Louis
S70
2015
80
0
0
0
0
0
0
All Children
St. Louis
S70
2015
90
0
0
0
0
0
0
All Children
St. Louis
S70
2015
100
0
0
0
0
0
0
All Children
St. Louis
S70
2016
0
359080
325268
301645
283232
267933
254646
All Children
St. Louis
S70
2016
10
345411
307656
282367
262223
246469
232745
All Children
St. Louis
S70
2016
20
310497
264108
234567
212611
196091
183033
All Children
St. Louis
S70
2016
30
259664
206792
177141
156888
141279
129195
All Children
St. Louis
S70
2016
40
195627
143046
114843
96156
82004
71568
All Children
St. Louis
S70
2016
50
120880
70330
46462
32274
22429
16064
All Children
St. Louis
S70
2016
60
47609
14325
5036
1803
519
182
All Children
St. Louis
S70
2016
70
4863
155
9
0
0
0
All Children
St. Louis
S70
2016
80
0
0
0
0
0
0
All Children
St. Louis
S70
2016
90
0
0
0
0
0
0
All Children
St. Louis
S70
2016
100
0
0
0
0
0
0
All Children
St. Louis
S70
2017
0
355702
320669
297356
279526
264764
252179
3D-Attachment4-16
-------
Study Group
Study Area
AQ
Scenario
Year
Benchmark
(PPb)
Number of People with 7-hr Exposure at or above Benchmark
>1 Day
> 2 Days
> 3 Days
> 4 Days
> 5 Days
> 6 Days
All Children
St. Louis
S70
2017
10
343053
305771
280855
261850
246314
232782
All Children
St. Louis
S70
2017
20
309878
264991
236688
216326
199579
186092
All Children
St. Louis
S70
2017
30
262851
211363
182614
162597
147025
134686
All Children
St. Louis
S70
2017
40
202211
149748
122019
103132
89763
78763
All Children
St. Louis
S70
2017
50
119123
69201
45870
32410
23859
17685
All Children
St. Louis
S70
2017
60
28595
6520
1949
574
219
46
All Children
St. Louis
S70
2017
70
1685
36
0
0
0
0
All Children
St. Louis
S70
2017
80
36
0
0
0
0
0
All Children
St. Louis
S70
2017
90
0
0
0
0
0
0
All Children
St. Louis
S70
2017
100
0
0
0
0
0
0
All Children
St. Louis
S75
2015
0
355693
320478
297229
278469
263525
250485
All Children
St. Louis
S75
2015
10
339574
301336
276420
257105
241196
227846
All Children
St. Louis
S75
2015
20
303931
257825
228420
207748
190864
177468
All Children
St. Louis
S75
2015
30
253353
201200
172569
152708
137054
124186
All Children
St. Louis
S75
2015
40
193851
140742
113149
94144
80119
69246
All Children
St. Louis
S75
2015
50
123958
73544
49458
34760
24742
18186
All Children
St. Louis
S75
2015
60
46635
14662
5245
2031
729
291
All Children
St. Louis
S75
2015
70
4162
255
27
0
0
0
All Children
St. Louis
S75
2015
80
0
0
0
0
0
0
All Children
St. Louis
S75
2015
90
0
0
0
0
0
0
All Children
St. Louis
S75
2015
100
0
0
0
0
0
0
All Children
St. Louis
S75
2016
0
359080
325268
301645
283232
267933
254646
All Children
St. Louis
S75
2016
10
345867
308029
282950
262851
246778
233301
All Children
St. Louis
S75
2016
20
313065
266849
237608
215935
199151
185846
All Children
St. Louis
S75
2016
30
266685
213503
183834
163372
147171
134914
All Children
St. Louis
S75
2016
40
208531
155148
126526
107166
92805
81522
All Children
St. Louis
S75
2016
50
141707
89590
63491
46871
35661
27447
All Children
St. Louis
S75
2016
60
70922
29259
13223
6675
3215
1539
All Children
St. Louis
S75
2016
70
17730
2113
310
73
9
0
All Children
St. Louis
S75
2016
80
555
9
0
0
0
0
All Children
St. Louis
S75
2016
90
0
0
0
0
0
0
All Children
St. Louis
S75
2016
100
0
0
0
0
0
0
All Children
St. Louis
S75
2017
0
355702
320669
297356
279526
264764
252179
All Children
St. Louis
S75
2017
10
343545
306290
281420
262369
246815
233392
All Children
St. Louis
S75
2017
20
312455
267942
240076
219768
202757
189143
All Children
St. Louis
S75
2017
30
268880
218348
189107
168644
152963
140642
All Children
St. Louis
S75
2017
40
215106
161833
133420
114788
100673
88780
All Children
St. Louis
S75
2017
50
143109
91931
66141
50159
39022
30835
All Children
St. Louis
S75
2017
60
55623
19743
8351
3843
1730
829
All Children
St. Louis
S75
2017
70
6484
565
64
9
0
0
All Children
St. Louis
S75
2017
80
492
0
0
0
0
0
All Children
St. Louis
S75
2017
90
0
0
0
0
0
0
All Children
St. Louis
S75
2017
100
0
0
0
0
0
0
Asthma Children
Atlanta
S65
2015
0
96464
87526
81735
77135
73342
70214
Asthma Children
Atlanta
S65
2015
10
91420
81271
74714
69912
66119
62689
Asthma Children
Atlanta
S65
2015
20
79899
67854
60409
54961
50825
47072
Asthma Children
Atlanta
S65
2015
30
63011
49392
41705
36277
32061
28953
3D-Attachment4-17
-------
AQ
Benchmark
Number of People with 7-hr Exposure at or above Benchmark
Study Group
Study Area
Scenario
Year
(PPb)
>1 Day
> 2 Days
> 3 Days
> 4 Days
> 5 Days
> 6 Days
Asthma Children
Atlanta
S65
2015
40
41221
27259
20358
15617
12570
10330
Asthma Children
Atlanta
S65
2015
50
14164
5206
2219
928
565
323
Asthma Children
Atlanta
S65
2015
60
1654
101
20
0
0
0
Asthma Children
Atlanta
S65
2015
70
40
0
0
0
0
0
Asthma Children
Atlanta
S65
2015
80
0
0
0
0
0
0
Asthma Children
Atlanta
S65
2015
90
0
0
0
0
0
0
Asthma Children
Atlanta
S65
2015
100
0
0
0
0
0
0
Asthma Children
Atlanta
S65
2016
0
99390
90068
83935
78951
74875
71627
Asthma Children
Atlanta
S65
2016
10
94850
84600
77680
72595
68156
64948
Asthma Children
Atlanta
S65
2016
20
84278
71586
63899
58209
53549
50139
Asthma Children
Atlanta
S65
2016
30
67733
53549
45619
40313
37105
33755
Asthma Children
Atlanta
S65
2016
40
46043
31637
24333
20035
16747
14124
Asthma Children
Atlanta
S65
2016
50
18260
6961
3087
1735
908
484
Asthma Children
Atlanta
S65
2016
60
2724
222
40
20
0
0
Asthma Children
Atlanta
S65
2016
70
222
0
0
0
0
0
Asthma Children
Atlanta
S65
2016
80
20
0
0
0
0
0
Asthma Children
Atlanta
S65
2016
90
0
0
0
0
0
0
Asthma Children
Atlanta
S65
2016
100
0
0
0
0
0
0
Asthma Children
Atlanta
S65
2017
0
96827
88757
82361
77922
73786
70416
Asthma Children
Atlanta
S65
2017
10
91904
82502
75521
70860
66562
62810
Asthma Children
Atlanta
S65
2017
20
80605
68621
60792
54921
50462
46850
Asthma Children
Atlanta
S65
2017
30
63092
49190
41059
35914
32182
29175
Asthma Children
Atlanta
S65
2017
40
38073
25261
18623
13982
11178
9180
Asthma Children
Atlanta
S65
2017
50
9100
2219
807
282
101
61
Asthma Children
Atlanta
S65
2017
60
424
0
0
0
0
0
Asthma Children
Atlanta
S65
2017
70
20
0
0
0
0
0
Asthma Children
Atlanta
S65
2017
80
0
0
0
0
0
0
Asthma Children
Atlanta
S65
2017
90
0
0
0
0
0
0
Asthma Children
Atlanta
S65
2017
100
0
0
0
0
0
0
Asthma Children
Atlanta
S70
2015
0
96464
87526
81735
77135
73342
70214
Asthma Children
Atlanta
S70
2015
10
91985
82119
75218
70376
66764
63254
Asthma Children
Atlanta
S70
2015
20
81574
69952
62325
56938
52963
49049
Asthma Children
Atlanta
S70
2015
30
67047
53085
45660
40333
36257
32747
Asthma Children
Atlanta
S70
2015
40
49049
34744
27480
22396
18724
15879
Asthma Children
Atlanta
S70
2015
50
24515
12913
7627
4620
2805
1634
Asthma Children
Atlanta
S70
2015
60
5044
1090
303
101
20
0
Asthma Children
Atlanta
S70
2015
70
585
40
0
0
0
0
Asthma Children
Atlanta
S70
2015
80
0
0
0
0
0
0
Asthma Children
Atlanta
S70
2015
90
0
0
0
0
0
0
Asthma Children
Atlanta
S70
2015
100
0
0
0
0
0
0
Asthma Children
Atlanta
S70
2016
0
99390
90068
83935
78951
74875
71627
Asthma Children
Atlanta
S70
2016
10
95254
85004
78346
73281
69125
65675
Asthma Children
Atlanta
S70
2016
20
85871
73665
65977
60489
56010
52197
Asthma Children
Atlanta
S70
2016
30
71385
57423
49453
44187
40373
37569
Asthma Children
Atlanta
S70
2016
40
52802
38396
31254
26411
23021
19934
Asthma Children
Atlanta
S70
2016
50
28691
15919
10028
6557
4540
2926
Asthma Children
Atlanta
S70
2016
60
8333
1715
504
101
20
0
3D-Attachment4-18
-------
Study Group
Study Area
AQ
Scenario
Year
Benchmark
(PPb)
Number of People with 7-hr Exposure at or above Benchmark
>1 Day
> 2 Days
> 3 Days
> 4 Days
> 5 Days
> 6 Days
Asthma Children
Atlanta
S70
2016
70
1271
20
0
0
0
0
Asthma Children
Atlanta
S70
2016
80
202
0
0
0
0
0
Asthma Children
Atlanta
S70
2016
90
20
0
0
0
0
0
Asthma Children
Atlanta
S70
2016
100
0
0
0
0
0
0
Asthma Children
Atlanta
S70
2017
0
96827
88757
82361
77922
73786
70416
Asthma Children
Atlanta
S70
2017
10
92187
83067
76207
71385
67349
63738
Asthma Children
Atlanta
S70
2017
20
82381
71022
62830
57100
52742
49090
Asthma Children
Atlanta
S70
2017
30
66966
53125
44731
39425
35813
32726
Asthma Children
Atlanta
S70
2017
40
45821
32424
24898
21105
17554
14608
Asthma Children
Atlanta
S70
2017
50
19127
7990
4156
1997
1211
686
Asthma Children
Atlanta
S70
2017
60
2078
202
81
0
0
0
Asthma Children
Atlanta
S70
2017
70
141
0
0
0
0
0
Asthma Children
Atlanta
S70
2017
80
0
0
0
0
0
0
Asthma Children
Atlanta
S70
2017
90
0
0
0
0
0
0
Asthma Children
Atlanta
S70
2017
100
0
0
0
0
0
0
Asthma Children
Atlanta
S75
2015
0
96464
87526
81735
77135
73342
70214
Asthma Children
Atlanta
S75
2015
10
92368
82522
75884
70719
67208
63818
Asthma Children
Atlanta
S75
2015
20
82704
71264
63778
58653
54335
50542
Asthma Children
Atlanta
S75
2015
30
70537
56575
48626
43400
39385
35874
Asthma Children
Atlanta
S75
2015
40
54941
40494
32747
27541
23486
20802
Asthma Children
Atlanta
S75
2015
50
34522
21185
14628
10270
7808
6134
Asthma Children
Atlanta
S75
2015
60
12025
3834
1614
807
404
161
Asthma Children
Atlanta
S75
2015
70
2119
202
20
20
0
0
Asthma Children
Atlanta
S75
2015
80
282
0
0
0
0
0
Asthma Children
Atlanta
S75
2015
90
0
0
0
0
0
0
Asthma Children
Atlanta
S75
2015
100
0
0
0
0
0
0
Asthma Children
Atlanta
S75
2016
0
99390
90068
83935
78951
74875
71627
Asthma Children
Atlanta
S75
2016
10
95718
85488
78608
73523
69710
66260
Asthma Children
Atlanta
S75
2016
20
87344
75299
67430
62144
57584
54093
Asthma Children
Atlanta
S75
2016
30
74774
60610
52540
47274
43158
39829
Asthma Children
Atlanta
S75
2016
40
58209
43783
36298
31112
27541
24979
Asthma Children
Atlanta
S75
2016
50
37912
23808
17412
13135
10088
7788
Asthma Children
Atlanta
S75
2016
60
17009
5972
2764
1311
605
242
Asthma Children
Atlanta
S75
2016
70
4439
565
101
20
20
0
Asthma Children
Atlanta
S75
2016
80
888
20
0
0
0
0
Asthma Children
Atlanta
S75
2016
90
182
0
0
0
0
0
Asthma Children
Atlanta
S75
2016
100
20
0
0
0
0
0
Asthma Children
Atlanta
S75
2017
0
96827
88757
82361
77922
73786
70416
Asthma Children
Atlanta
S75
2017
10
92368
83672
76711
71808
67934
64141
Asthma Children
Atlanta
S75
2017
20
83773
72373
64726
58835
54356
50805
Asthma Children
Atlanta
S75
2017
30
70114
56837
48323
42794
38739
35269
Asthma Children
Atlanta
S75
2017
40
52277
38073
30547
26129
22376
19491
Asthma Children
Atlanta
S75
2017
50
29054
15899
9685
6477
4459
3067
Asthma Children
Atlanta
S75
2017
60
7425
1614
464
182
61
40
Asthma Children
Atlanta
S75
2017
70
545
61
20
0
0
0
Asthma Children
Atlanta
S75
2017
80
61
0
0
0
0
0
Asthma Children
Atlanta
S75
2017
90
0
0
0
0
0
0
3D-Attachment4-19
-------
Study Group
Study Area
AQ
Scenario
Year
Benchmark
(PPb)
Number of People with 7-hr Exposure at or above Benchmark
>1 Day
> 2 Days
> 3 Days
> 4 Days
> 5 Days
> 6 Days
Asthma Children
Atlanta
S75
2017
100
0
0
0
0
0
0
Asthma Children
Boston
S65
2015
0
110791
99209
91427
85625
79800
75522
Asthma Children
Boston
S65
2015
10
106673
94181
86421
79072
73474
69310
Asthma Children
Boston
S65
2015
20
93157
78526
69333
62484
57000
52677
Asthma Children
Boston
S65
2015
30
73633
57273
47921
41595
36748
33153
Asthma Children
Boston
S65
2015
40
47307
30218
22572
16861
13129
10331
Asthma Children
Boston
S65
2015
50
17066
5689
2275
910
319
182
Asthma Children
Boston
S65
2015
60
3163
319
0
0
0
0
Asthma Children
Boston
S65
2015
70
455
0
0
0
0
0
Asthma Children
Boston
S65
2015
80
68
0
0
0
0
0
Asthma Children
Boston
S65
2015
90
0
0
0
0
0
0
Asthma Children
Boston
S65
2015
100
0
0
0
0
0
0
Asthma Children
Boston
S65
2016
0
115661
104420
96593
90403
85056
80756
Asthma Children
Boston
S65
2016
10
112270
99710
91609
84647
79049
74430
Asthma Children
Boston
S65
2016
20
99209
83486
73815
65988
59935
55498
Asthma Children
Boston
S65
2016
30
78526
60845
50924
43939
39274
35474
Asthma Children
Boston
S65
2016
40
51220
33859
24916
18954
14972
11946
Asthma Children
Boston
S65
2016
50
19705
7031
3163
1456
683
250
Asthma Children
Boston
S65
2016
60
2298
205
23
0
0
0
Asthma Children
Boston
S65
2016
70
91
0
0
0
0
0
Asthma Children
Boston
S65
2016
80
0
0
0
0
0
0
Asthma Children
Boston
S65
2016
90
0
0
0
0
0
0
Asthma Children
Boston
S65
2016
100
0
0
0
0
0
0
Asthma Children
Boston
S65
2017
0
112976
101257
93612
86990
82075
77684
Asthma Children
Boston
S65
2017
10
108607
96661
88469
81962
76364
71540
Asthma Children
Boston
S65
2017
20
96843
81529
71244
63235
57819
52927
Asthma Children
Boston
S65
2017
30
76341
57592
48763
42073
37340
33176
Asthma Children
Boston
S65
2017
40
50697
33563
24734
18067
13903
11286
Asthma Children
Boston
S65
2017
50
23915
9716
4460
1752
592
250
Asthma Children
Boston
S65
2017
60
5165
523
46
0
0
0
Asthma Children
Boston
S65
2017
70
387
0
0
0
0
0
Asthma Children
Boston
S65
2017
80
0
0
0
0
0
0
Asthma Children
Boston
S65
2017
90
0
0
0
0
0
0
Asthma Children
Boston
S65
2017
100
0
0
0
0
0
0
Asthma Children
Boston
S70
2015
0
110791
99209
91427
85625
79800
75522
Asthma Children
Boston
S70
2015
10
106764
94203
86763
79140
73770
69492
Asthma Children
Boston
S70
2015
20
94249
79754
70380
63235
57956
53632
Asthma Children
Boston
S70
2015
30
76341
59867
50378
43893
39069
35269
Asthma Children
Boston
S70
2015
40
52904
35429
27669
21617
17453
14495
Asthma Children
Boston
S70
2015
50
25804
11286
5552
3004
1547
865
Asthma Children
Boston
S70
2015
60
6166
1047
250
91
23
23
Asthma Children
Boston
S70
2015
70
1024
0
0
0
0
0
Asthma Children
Boston
S70
2015
80
114
0
0
0
0
0
Asthma Children
Boston
S70
2015
90
0
0
0
0
0
0
Asthma Children
Boston
S70
2015
100
0
0
0
0
0
0
Asthma Children
Boston
S70
2016
0
115661
104420
96593
90403
85056
80756
Asthma Children
Boston
S70
2016
10
112452
99892
91541
84783
79185
74612
3D-Attachment4-20
-------
Study Group
Study Area
AQ
Scenario
Year
Benchmark
(PPb)
Number of People with 7-hr Exposure at or above Benchmark
>1 Day
> 2 Days
> 3 Days
> 4 Days
> 5 Days
> 6 Days
Asthma Children
Boston
S70
2016
20
100370
84738
75363
67603
61346
56954
Asthma Children
Boston
S70
2016
30
81893
63781
54269
46851
41345
37545
Asthma Children
Boston
S70
2016
40
57887
39456
30559
24552
19842
16747
Asthma Children
Boston
S70
2016
50
29968
14176
8146
4892
2822
1889
Asthma Children
Boston
S70
2016
60
7782
1343
250
0
0
0
Asthma Children
Boston
S70
2016
70
796
68
0
0
0
0
Asthma Children
Boston
S70
2016
80
0
0
0
0
0
0
Asthma Children
Boston
S70
2016
90
0
0
0
0
0
0
Asthma Children
Boston
S70
2016
100
0
0
0
0
0
0
Asthma Children
Boston
S70
2017
0
112976
101257
93612
86990
82075
77684
Asthma Children
Boston
S70
2017
10
108744
96775
88674
81939
76318
71813
Asthma Children
Boston
S70
2017
20
97344
82417
72200
64486
58843
53678
Asthma Children
Boston
S70
2017
30
79140
60504
51015
44144
39297
35019
Asthma Children
Boston
S70
2017
40
56386
39365
29945
23665
18590
15382
Asthma Children
Boston
S70
2017
50
32812
16247
8624
4778
2731
1502
Asthma Children
Boston
S70
2017
60
11605
2617
455
137
0
0
Asthma Children
Boston
S70
2017
70
1616
23
0
0
0
0
Asthma Children
Boston
S70
2017
80
68
0
0
0
0
0
Asthma Children
Boston
S70
2017
90
0
0
0
0
0
0
Asthma Children
Boston
S70
2017
100
0
0
0
0
0
0
Asthma Children
Boston
S75
2015
0
110791
99209
91427
85625
79800
75522
Asthma Children
Boston
S75
2015
10
106809
94021
86603
79094
73588
69333
Asthma Children
Boston
S75
2015
20
94454
80232
70880
63417
58251
53928
Asthma Children
Boston
S75
2015
30
77274
61050
51584
45008
40071
36384
Asthma Children
Boston
S75
2015
40
55430
38318
29695
23938
20183
16725
Asthma Children
Boston
S75
2015
50
29968
14654
8374
5074
3254
1684
Asthma Children
Boston
S75
2015
60
9375
2184
614
182
68
23
Asthma Children
Boston
S75
2015
70
1752
68
0
0
0
0
Asthma Children
Boston
S75
2015
80
137
0
0
0
0
0
Asthma Children
Boston
S75
2015
90
0
0
0
0
0
0
Asthma Children
Boston
S75
2015
100
0
0
0
0
0
0
Asthma Children
Boston
S75
2016
0
115661
104420
96593
90403
85056
80756
Asthma Children
Boston
S75
2016
10
112430
100006
91609
84715
79185
74430
Asthma Children
Boston
S75
2016
20
100666
85147
75704
67968
61915
57341
Asthma Children
Boston
S75
2016
30
82940
65283
55566
48239
42596
38546
Asthma Children
Boston
S75
2016
40
60777
42824
33130
27146
22618
18954
Asthma Children
Boston
S75
2016
50
35383
18795
11013
7486
5097
3459
Asthma Children
Boston
S75
2016
60
12674
3163
933
205
46
0
Asthma Children
Boston
S75
2016
70
1684
137
0
0
0
0
Asthma Children
Boston
S75
2016
80
137
0
0
0
0
0
Asthma Children
Boston
S75
2016
90
0
0
0
0
0
0
Asthma Children
Boston
S75
2016
100
0
0
0
0
0
0
Asthma Children
Boston
S75
2017
0
112976
101257
93612
86990
82075
77684
Asthma Children
Boston
S75
2017
10
108789
96547
88674
81848
76364
71677
Asthma Children
Boston
S75
2017
20
97935
82894
72632
64691
59184
53792
Asthma Children
Boston
S75
2017
30
80528
62120
52335
45145
40412
36271
Asthma Children
Boston
S75
2017
40
59435
41572
32653
26168
21002
17703
3D-Attachment4-21
-------
AQ
Benchmark
Number of People with 7-hr Exposure at or above Benchmark
Study Group
Study Area
Scenario
Year
(PPb)
>1 Day
> 2 Days
> 3 Days
> 4 Days
> 5 Days
> 6 Days
Asthma Children
Boston
S75
2017
50
37932
20274
11696
7122
4505
2958
Asthma Children
Boston
S75
2017
60
16315
4915
1343
319
114
46
Asthma Children
Boston
S75
2017
70
3641
387
68
0
0
0
Asthma Children
Boston
S75
2017
80
319
0
0
0
0
0
Asthma Children
Boston
S75
2017
90
0
0
0
0
0
0
Asthma Children
Boston
S75
2017
100
0
0
0
0
0
0
Asthma Children
Dallas
S65
2015
0
90113
82594
76635
73254
69872
66775
Asthma Children
Dallas
S65
2015
10
86330
77250
71267
67248
63488
60249
Asthma Children
Dallas
S65
2015
20
77155
66373
59350
54408
50247
46463
Asthma Children
Dallas
S65
2015
30
63181
50649
43177
38258
35161
32229
Asthma Children
Dallas
S65
2015
40
46014
32844
26246
21967
18373
15984
Asthma Children
Dallas
S65
2015
50
23196
11846
6834
4043
2601
1702
Asthma Children
Dallas
S65
2015
60
3878
709
95
0
0
0
Asthma Children
Dallas
S65
2015
70
71
0
0
0
0
0
Asthma Children
Dallas
S65
2015
80
0
0
0
0
0
0
Asthma Children
Dallas
S65
2015
90
0
0
0
0
0
0
Asthma Children
Dallas
S65
2015
100
0
0
0
0
0
0
Asthma Children
Dallas
S65
2016
0
90420
83208
78266
74318
71315
68501
Asthma Children
Dallas
S65
2016
10
86164
78361
73325
68879
65427
62069
Asthma Children
Dallas
S65
2016
20
77273
67106
60367
55212
51145
48142
Asthma Children
Dallas
S65
2016
30
62022
49466
42846
37880
33955
31212
Asthma Children
Dallas
S65
2016
40
41427
28895
22038
17592
14802
12059
Asthma Children
Dallas
S65
2016
50
14991
5817
2743
1584
662
355
Asthma Children
Dallas
S65
2016
60
1395
47
24
0
0
0
Asthma Children
Dallas
S65
2016
70
47
0
0
0
0
0
Asthma Children
Dallas
S65
2016
80
0
0
0
0
0
0
Asthma Children
Dallas
S65
2016
90
0
0
0
0
0
0
Asthma Children
Dallas
S65
2016
100
0
0
0
0
0
0
Asthma Children
Dallas
S65
2017
0
91035
83563
78645
74341
70724
68028
Asthma Children
Dallas
S65
2017
10
87819
79023
72592
68241
64623
61218
Asthma Children
Dallas
S65
2017
20
78196
67153
60509
55661
51263
48119
Asthma Children
Dallas
S65
2017
30
63252
51192
43933
39133
35634
32583
Asthma Children
Dallas
S65
2017
40
44974
31638
25301
20524
17427
14282
Asthma Children
Dallas
S65
2017
50
21588
10735
5675
3003
1608
1040
Asthma Children
Dallas
S65
2017
60
3476
402
47
0
0
0
Asthma Children
Dallas
S65
2017
70
118
0
0
0
0
0
Asthma Children
Dallas
S65
2017
80
0
0
0
0
0
0
Asthma Children
Dallas
S65
2017
90
0
0
0
0
0
0
Asthma Children
Dallas
S65
2017
100
0
0
0
0
0
0
Asthma Children
Dallas
S70
2015
0
90113
82594
76635
73254
69872
66775
Asthma Children
Dallas
S70
2015
10
86401
77368
71362
67461
63606
60532
Asthma Children
Dallas
S70
2015
20
77935
67082
60130
55401
51074
47740
Asthma Children
Dallas
S70
2015
30
65285
52706
44903
40150
36556
33931
Asthma Children
Dallas
S70
2015
40
50176
37052
29912
25395
21659
18869
Asthma Children
Dallas
S70
2015
50
30668
18467
12698
8654
6006
4564
Asthma Children
Dallas
S70
2015
60
9813
2861
1064
473
213
71
Asthma Children
Dallas
S70
2015
70
946
118
0
0
0
0
3D-Attachment4-22
-------
Study Group
Study Area
AQ
Scenario
Year
Benchmark
(PPb)
Number of People with 7-hr Exposure at or above Benchmark
>1 Day
> 2 Days
> 3 Days
> 4 Days
> 5 Days
> 6 Days
Asthma Children
Dallas
S70
2015
80
24
0
0
0
0
0
Asthma Children
Dallas
S70
2015
90
0
0
0
0
0
0
Asthma Children
Dallas
S70
2015
100
0
0
0
0
0
0
Asthma Children
Dallas
S70
2016
0
90420
83208
78266
74318
71315
68501
Asthma Children
Dallas
S70
2016
10
86259
78598
73727
69116
65640
62471
Asthma Children
Dallas
S70
2016
20
77817
67933
61384
56087
52375
49041
Asthma Children
Dallas
S70
2016
30
64245
51665
44761
40079
35894
32702
Asthma Children
Dallas
S70
2016
40
45707
32867
25608
21446
18373
15677
Asthma Children
Dallas
S70
2016
50
21494
10144
5746
3689
2270
1513
Asthma Children
Dallas
S70
2016
60
2908
378
118
24
24
0
Asthma Children
Dallas
S70
2016
70
426
0
0
0
0
0
Asthma Children
Dallas
S70
2016
80
0
0
0
0
0
0
Asthma Children
Dallas
S70
2016
90
0
0
0
0
0
0
Asthma Children
Dallas
S70
2016
100
0
0
0
0
0
0
Asthma Children
Dallas
S70
2017
0
91035
83563
78645
74341
70724
68028
Asthma Children
Dallas
S70
2017
10
88008
79236
72899
68525
64883
61502
Asthma Children
Dallas
S70
2017
20
79047
68052
61407
56773
52446
49183
Asthma Children
Dallas
S70
2017
30
65238
52942
46085
41261
37525
34570
Asthma Children
Dallas
S70
2017
40
49159
35137
28469
24378
20974
18089
Asthma Children
Dallas
S70
2017
50
28114
15913
9907
6503
4209
2767
Asthma Children
Dallas
S70
2017
60
8134
1561
402
142
24
0
Asthma Children
Dallas
S70
2017
70
355
0
0
0
0
0
Asthma Children
Dallas
S70
2017
80
0
0
0
0
0
0
Asthma Children
Dallas
S70
2017
90
0
0
0
0
0
0
Asthma Children
Dallas
S70
2017
100
0
0
0
0
0
0
Asthma Children
Dallas
S75
2015
0
90113
82594
76635
73254
69872
66775
Asthma Children
Dallas
S75
2015
10
86519
77604
71504
67650
63748
60627
Asthma Children
Dallas
S75
2015
20
78621
67815
60840
56134
52115
48568
Asthma Children
Dallas
S75
2015
30
66798
54077
46487
41782
37951
35043
Asthma Children
Dallas
S75
2015
40
52824
40055
32536
28044
24095
21210
Asthma Children
Dallas
S75
2015
50
36580
23622
16859
12958
10357
7425
Asthma Children
Dallas
S75
2015
60
15724
6313
2956
1490
969
520
Asthma Children
Dallas
S75
2015
70
2956
355
47
0
0
0
Asthma Children
Dallas
S75
2015
80
166
0
0
0
0
0
Asthma Children
Dallas
S75
2015
90
0
0
0
0
0
0
Asthma Children
Dallas
S75
2015
100
0
0
0
0
0
0
Asthma Children
Dallas
S75
2016
0
90420
83208
78266
74318
71315
68501
Asthma Children
Dallas
S75
2016
10
86282
78739
73774
69352
65876
62708
Asthma Children
Dallas
S75
2016
20
78408
68525
62046
56655
53131
49916
Asthma Children
Dallas
S75
2016
30
65971
53250
46345
41592
37431
34286
Asthma Children
Dallas
S75
2016
40
48781
35847
28611
23764
20572
18136
Asthma Children
Dallas
S75
2016
50
27311
15015
9104
6313
4327
3121
Asthma Children
Dallas
S75
2016
60
6006
1348
331
142
24
0
Asthma Children
Dallas
S75
2016
70
1111
0
0
0
0
0
Asthma Children
Dallas
S75
2016
80
47
0
0
0
0
0
Asthma Children
Dallas
S75
2016
90
0
0
0
0
0
0
Asthma Children
Dallas
S75
2016
100
0
0
0
0
0
0
3D-Attachment4-23
-------
Study Group
Study Area
AQ
Scenario
Year
Benchmark
(PPb)
Number of People with 7-hr Exposure at or above Benchmark
>1 Day
> 2 Days
> 3 Days
> 4 Days
> 5 Days
> 6 Days
Asthma Children
Dallas
S75
2017
0
91035
83563
78645
74341
70724
68028
Asthma Children
Dallas
S75
2017
10
88127
79401
73277
68595
65214
61715
Asthma Children
Dallas
S75
2017
20
79685
68737
61951
57151
52989
49679
Asthma Children
Dallas
S75
2017
30
66822
54763
47598
42373
38826
35657
Asthma Children
Dallas
S75
2017
40
52091
38164
31236
26578
23196
20666
Asthma Children
Dallas
S75
2017
50
32560
19508
13596
10144
7236
5249
Asthma Children
Dallas
S75
2017
60
13100
3831
1064
473
213
47
Asthma Children
Dallas
S75
2017
70
1371
95
0
0
0
0
Asthma Children
Dallas
S75
2017
80
24
0
0
0
0
0
Asthma Children
Dallas
S75
2017
90
0
0
0
0
0
0
Asthma Children
Dallas
S75
2017
100
0
0
0
0
0
0
Asthma Children
Detroit
S65
2015
0
77853
70222
64551
60129
56729
53504
Asthma Children
Detroit
S65
2015
10
74853
66806
60319
55550
51995
48960
Asthma Children
Detroit
S65
2015
20
66285
55741
49168
44242
39872
36542
Asthma Children
Detroit
S65
2015
30
53486
41467
34652
30246
26656
23847
Asthma Children
Detroit
S65
2015
40
38155
25911
19632
15366
12435
10024
Asthma Children
Detroit
S65
2015
50
16927
7267
3711
1890
1041
468
Asthma Children
Detroit
S65
2015
60
1717
173
0
0
0
0
Asthma Children
Detroit
S65
2015
70
0
0
0
0
0
0
Asthma Children
Detroit
S65
2015
80
0
0
0
0
0
0
Asthma Children
Detroit
S65
2015
90
0
0
0
0
0
0
Asthma Children
Detroit
S65
2015
100
0
0
0
0
0
0
Asthma Children
Detroit
S65
2016
0
80871
72806
67187
62747
58741
55706
Asthma Children
Detroit
S65
2016
10
78079
68627
62956
58204
54180
51249
Asthma Children
Detroit
S65
2016
20
69008
58065
51492
46393
42508
38987
Asthma Children
Detroit
S65
2016
30
56244
44034
37496
32796
29136
26223
Asthma Children
Detroit
S65
2016
40
40323
28287
21488
17777
13996
11655
Asthma Children
Detroit
S65
2016
50
21904
10857
6053
3018
1682
850
Asthma Children
Detroit
S65
2016
60
5064
486
104
17
0
0
Asthma Children
Detroit
S65
2016
70
52
0
0
0
0
0
Asthma Children
Detroit
S65
2016
80
0
0
0
0
0
0
Asthma Children
Detroit
S65
2016
90
0
0
0
0
0
0
Asthma Children
Detroit
S65
2016
100
0
0
0
0
0
0
Asthma Children
Detroit
S65
2017
0
79917
71558
66060
62140
58516
54718
Asthma Children
Detroit
S65
2017
10
76847
67968
62123
57215
53885
50451
Asthma Children
Detroit
S65
2017
20
68609
57649
50937
45959
41762
38033
Asthma Children
Detroit
S65
2017
30
56122
43236
36663
31998
28443
25442
Asthma Children
Detroit
S65
2017
40
39820
28079
21540
17222
14169
11377
Asthma Children
Detroit
S65
2017
50
19806
8914
5238
2636
1422
954
Asthma Children
Detroit
S65
2017
60
2012
173
35
0
0
0
Asthma Children
Detroit
S65
2017
70
35
0
0
0
0
0
Asthma Children
Detroit
S65
2017
80
0
0
0
0
0
0
Asthma Children
Detroit
S65
2017
90
0
0
0
0
0
0
Asthma Children
Detroit
S65
2017
100
0
0
0
0
0
0
Asthma Children
Detroit
S70
2015
0
77853
70222
64551
60129
56729
53504
Asthma Children
Detroit
S70
2015
10
74905
66910
60406
55723
51977
49064
Asthma Children
Detroit
S70
2015
20
67205
56469
50139
45005
40756
37635
3D-Attachment4-24
-------
Study Group
Study Area
AQ
Scenario
Year
Benchmark
(PPb)
Number of People with 7-hr Exposure at or above Benchmark
>1 Day
> 2 Days
> 3 Days
> 4 Days
> 5 Days
> 6 Days
Asthma Children
Detroit
S70
2015
30
55793
43288
36369
31825
28200
25668
Asthma Children
Detroit
S70
2015
40
41762
29518
23396
18904
15730
12938
Asthma Children
Detroit
S70
2015
50
24107
12366
7544
4613
3035
1960
Asthma Children
Detroit
S70
2015
60
6209
1301
243
52
0
0
Asthma Children
Detroit
S70
2015
70
121
0
0
0
0
0
Asthma Children
Detroit
S70
2015
80
0
0
0
0
0
0
Asthma Children
Detroit
S70
2015
90
0
0
0
0
0
0
Asthma Children
Detroit
S70
2015
100
0
0
0
0
0
0
Asthma Children
Detroit
S70
2016
0
80871
72806
67187
62747
58741
55706
Asthma Children
Detroit
S70
2016
10
78218
68852
63112
58446
54215
51336
Asthma Children
Detroit
S70
2016
20
69771
58897
52498
47555
43583
40115
Asthma Children
Detroit
S70
2016
30
58325
45994
39334
34998
31148
28096
Asthma Children
Detroit
S70
2016
40
44624
32137
25390
21020
17690
14672
Asthma Children
Detroit
S70
2016
50
28477
16372
10753
7267
4839
3313
Asthma Children
Detroit
S70
2016
60
11776
3469
1110
243
139
69
Asthma Children
Detroit
S70
2016
70
1110
17
0
0
0
0
Asthma Children
Detroit
S70
2016
80
0
0
0
0
0
0
Asthma Children
Detroit
S70
2016
90
0
0
0
0
0
0
Asthma Children
Detroit
S70
2016
100
0
0
0
0
0
0
Asthma Children
Detroit
S70
2017
0
79917
71558
66060
62140
58516
54718
Asthma Children
Detroit
S70
2017
10
76986
68193
62297
57371
53954
50538
Asthma Children
Detroit
S70
2017
20
69321
58533
51544
46792
42681
38918
Asthma Children
Detroit
S70
2017
30
57770
45526
38242
33993
30229
27125
Asthma Children
Detroit
S70
2017
40
43878
31270
25113
20500
17222
14655
Asthma Children
Detroit
S70
2017
50
27055
15314
9712
6348
4492
2792
Asthma Children
Detroit
S70
2017
60
7648
1769
572
191
139
35
Asthma Children
Detroit
S70
2017
70
503
17
0
0
0
0
Asthma Children
Detroit
S70
2017
80
0
0
0
0
0
0
Asthma Children
Detroit
S70
2017
90
0
0
0
0
0
0
Asthma Children
Detroit
S70
2017
100
0
0
0
0
0
0
Asthma Children
Detroit
S75
2015
0
77853
70222
64551
60129
56729
53504
Asthma Children
Detroit
S75
2015
10
74732
66823
60146
55602
51821
48873
Asthma Children
Detroit
S75
2015
20
67465
56885
50399
45318
41086
37947
Asthma Children
Detroit
S75
2015
30
56573
44346
37652
32605
28980
26258
Asthma Children
Detroit
S75
2015
40
43878
31235
24957
20621
17308
14499
Asthma Children
Detroit
S75
2015
50
28911
16615
10631
7249
4891
3208
Asthma Children
Detroit
S75
2015
60
11030
3243
798
277
52
35
Asthma Children
Detroit
S75
2015
70
1214
35
0
0
0
0
Asthma Children
Detroit
S75
2015
80
0
0
0
0
0
0
Asthma Children
Detroit
S75
2015
90
0
0
0
0
0
0
Asthma Children
Detroit
S75
2015
100
0
0
0
0
0
0
Asthma Children
Detroit
S75
2016
0
80871
72806
67187
62747
58741
55706
Asthma Children
Detroit
S75
2016
10
78061
68748
62817
58412
54163
51232
Asthma Children
Detroit
S75
2016
20
69962
59140
52671
47763
43722
40132
Asthma Children
Detroit
S75
2016
30
59504
47277
40097
35536
31825
28894
Asthma Children
Detroit
S75
2016
40
46861
34010
27680
23084
19442
16597
Asthma Children
Detroit
S75
2016
50
32744
20066
14169
10094
7353
5654
3D-Attachment4-25
-------
Study Group
Study Area
AQ
Scenario
Year
Benchmark
(PPb)
Number of People with 7-hr Exposure at or above Benchmark
>1 Day
> 2 Days
> 3 Days
> 4 Days
> 5 Days
> 6 Days
Asthma Children
Detroit
S75
2016
60
17673
7492
3261
1231
520
225
Asthma Children
Detroit
S75
2016
70
4544
520
104
17
0
0
Asthma Children
Detroit
S75
2016
80
69
0
0
0
0
0
Asthma Children
Detroit
S75
2016
90
0
0
0
0
0
0
Asthma Children
Detroit
S75
2016
100
0
0
0
0
0
0
Asthma Children
Detroit
S75
2017
0
79917
71558
66060
62140
58516
54718
Asthma Children
Detroit
S75
2017
10
76899
68037
62175
57284
53781
50243
Asthma Children
Detroit
S75
2017
20
69598
58689
51787
47121
42890
38883
Asthma Children
Detroit
S75
2017
30
58828
46306
38970
34548
30992
27801
Asthma Children
Detroit
S75
2017
40
45925
33229
26726
22130
18696
16233
Asthma Children
Detroit
S75
2017
50
30853
18991
12955
8828
6452
4717
Asthma Children
Detroit
S75
2017
60
13458
4683
1821
763
364
208
Asthma Children
Detroit
S75
2017
70
1682
87
0
0
0
0
Asthma Children
Detroit
S75
2017
80
52
0
0
0
0
0
Asthma Children
Detroit
S75
2017
90
0
0
0
0
0
0
Asthma Children
Detroit
S75
2017
100
0
0
0
0
0
0
Asthma Children
Philadelphia
S65
2015
0
100049
90817
84248
78616
73924
70432
Asthma Children
Philadelphia
S65
2015
10
97016
87172
79315
73575
69493
66198
Asthma Children
Philadelphia
S65
2015
20
87631
75146
66612
60872
56703
52709
Asthma Children
Philadelphia
S65
2015
30
73073
58428
50243
43848
39505
35620
Asthma Children
Philadelphia
S65
2015
40
50352
35314
27675
22284
18552
15605
Asthma Children
Philadelphia
S65
2015
50
20102
9058
4954
2794
1441
808
Asthma Children
Philadelphia
S65
2015
60
2248
218
44
0
0
0
Asthma Children
Philadelphia
S65
2015
70
44
0
0
0
0
0
Asthma Children
Philadelphia
S65
2015
80
0
0
0
0
0
0
Asthma Children
Philadelphia
S65
2015
90
0
0
0
0
0
0
Asthma Children
Philadelphia
S65
2015
100
0
0
0
0
0
0
Asthma Children
Philadelphia
S65
2016
0
98522
89180
82632
77241
72745
69144
Asthma Children
Philadelphia
S65
2016
10
95400
85666
78442
72331
67856
64233
Asthma Children
Philadelphia
S65
2016
20
85775
72854
64975
58973
54237
50767
Asthma Children
Philadelphia
S65
2016
30
69581
55590
47711
42036
38348
35030
Asthma Children
Philadelphia
S65
2016
40
47515
33110
25711
21040
16937
14230
Asthma Children
Philadelphia
S65
2016
50
18421
7923
3667
1724
851
327
Asthma Children
Philadelphia
S65
2016
60
2750
262
0
0
0
0
Asthma Children
Philadelphia
S65
2016
70
22
0
0
0
0
0
Asthma Children
Philadelphia
S65
2016
80
0
0
0
0
0
0
Asthma Children
Philadelphia
S65
2016
90
0
0
0
0
0
0
Asthma Children
Philadelphia
S65
2016
100
0
0
0
0
0
0
Asthma Children
Philadelphia
S65
2017
0
98958
89180
82501
76958
73051
69515
Asthma Children
Philadelphia
S65
2017
10
95728
85426
77743
72505
68140
64561
Asthma Children
Philadelphia
S65
2017
20
85688
72593
64408
58820
54477
50614
Asthma Children
Philadelphia
S65
2017
30
68620
53910
45812
40094
35947
32761
Asthma Children
Philadelphia
S65
2017
40
46052
30185
23528
18465
15191
12506
Asthma Children
Philadelphia
S65
2017
50
16937
6308
3012
1179
655
306
Asthma Children
Philadelphia
S65
2017
60
1964
218
0
0
0
0
Asthma Children
Philadelphia
S65
2017
70
65
0
0
0
0
0
Asthma Children
Philadelphia
S65
2017
80
0
0
0
0
0
0
3D-Attachment4-26
-------
Study Group
Study Area
AQ
Scenario
Year
Benchmark
(PPb)
Number of People with 7-hr Exposure at or above Benchmark
>1 Day
> 2 Days
> 3 Days
> 4 Days
> 5 Days
> 6 Days
Asthma Children
Philadelphia
S65
2017
90
0
0
0
0
0
0
Asthma Children
Philadelphia
S65
2017
100
0
0
0
0
0
0
Asthma Children
Philadelphia
S70
2015
0
100049
90817
84248
78616
73924
70432
Asthma Children
Philadelphia
S70
2015
10
97125
87325
79533
73858
69668
66394
Asthma Children
Philadelphia
S70
2015
20
88635
75910
67791
61767
57402
53691
Asthma Children
Philadelphia
S70
2015
30
75212
60741
52491
46423
41578
37562
Asthma Children
Philadelphia
S70
2015
40
56114
40574
33001
27282
23397
19840
Asthma Children
Philadelphia
S70
2015
50
28897
15562
9953
6766
4845
3274
Asthma Children
Philadelphia
S70
2015
60
6264
1375
437
109
22
22
Asthma Children
Philadelphia
S70
2015
70
611
0
0
0
0
0
Asthma Children
Philadelphia
S70
2015
80
0
0
0
0
0
0
Asthma Children
Philadelphia
S70
2015
90
0
0
0
0
0
0
Asthma Children
Philadelphia
S70
2015
100
0
0
0
0
0
0
Asthma Children
Philadelphia
S70
2016
0
98522
89180
82632
77241
72745
69144
Asthma Children
Philadelphia
S70
2016
10
95510
85841
78682
72593
68009
64473
Asthma Children
Philadelphia
S70
2016
20
86561
73880
66285
60021
55437
51487
Asthma Children
Philadelphia
S70
2016
30
72374
57860
49806
44241
40290
37017
Asthma Children
Philadelphia
S70
2016
40
53189
38413
30578
25776
21717
18399
Asthma Children
Philadelphia
S70
2016
50
27326
14885
8425
5020
3318
2030
Asthma Children
Philadelphia
S70
2016
60
6504
1113
175
44
0
0
Asthma Children
Philadelphia
S70
2016
70
655
44
0
0
0
0
Asthma Children
Philadelphia
S70
2016
80
0
0
0
0
0
0
Asthma Children
Philadelphia
S70
2016
90
0
0
0
0
0
0
Asthma Children
Philadelphia
S70
2016
100
0
0
0
0
0
0
Asthma Children
Philadelphia
S70
2017
0
98958
89180
82501
76958
73051
69515
Asthma Children
Philadelphia
S70
2017
10
95946
85557
77896
72636
68424
64779
Asthma Children
Philadelphia
S70
2017
20
86517
73553
65368
59672
55481
51858
Asthma Children
Philadelphia
S70
2017
30
71567
56420
48584
42429
37890
34485
Asthma Children
Philadelphia
S70
2017
40
51400
35532
28177
22786
19490
16326
Asthma Children
Philadelphia
S70
2017
50
24248
12463
7224
3907
2401
1288
Asthma Children
Philadelphia
S70
2017
60
6024
1375
284
22
0
0
Asthma Children
Philadelphia
S70
2017
70
524
0
0
0
0
0
Asthma Children
Philadelphia
S70
2017
80
0
0
0
0
0
0
Asthma Children
Philadelphia
S70
2017
90
0
0
0
0
0
0
Asthma Children
Philadelphia
S70
2017
100
0
0
0
0
0
0
Asthma Children
Philadelphia
S75
2015
0
100049
90817
84248
78616
73924
70432
Asthma Children
Philadelphia
S75
2015
10
97147
87434
79642
73989
69733
66547
Asthma Children
Philadelphia
S75
2015
20
89638
76783
68642
62575
58078
54477
Asthma Children
Philadelphia
S75
2015
30
77569
62946
54739
48519
44154
39963
Asthma Children
Philadelphia
S75
2015
40
61461
46074
37824
31778
27588
24248
Asthma Children
Philadelphia
S75
2015
50
38261
23768
16850
12572
9800
7552
Asthma Children
Philadelphia
S75
2015
60
13859
4474
2095
939
393
196
Asthma Children
Philadelphia
S75
2015
70
2183
175
44
0
0
0
Asthma Children
Philadelphia
S75
2015
80
218
0
0
0
0
0
Asthma Children
Philadelphia
S75
2015
90
0
0
0
0
0
0
Asthma Children
Philadelphia
S75
2015
100
0
0
0
0
0
0
Asthma Children
Philadelphia
S75
2016
0
98522
89180
82632
77241
72745
69144
3D-Attachment4-27
-------
Study Group
Study Area
AQ
Scenario
Year
Benchmark
(PPb)
Number of People with 7-hr Exposure at or above Benchmark
>1 Day
> 2 Days
> 3 Days
> 4 Days
> 5 Days
> 6 Days
Asthma Children
Philadelphia
S75
2016
10
95619
85884
78529
72702
68184
64473
Asthma Children
Philadelphia
S75
2016
20
87085
75059
67027
61265
56441
52731
Asthma Children
Philadelphia
S75
2016
30
74950
60392
52120
46598
42386
38981
Asthma Children
Philadelphia
S75
2016
40
58275
43564
35380
30098
26082
23157
Asthma Children
Philadelphia
S75
2016
50
36798
23354
15976
11524
8272
6068
Asthma Children
Philadelphia
S75
2016
60
14078
4562
1375
480
196
87
Asthma Children
Philadelphia
S75
2016
70
2270
196
0
0
0
0
Asthma Children
Philadelphia
S75
2016
80
65
0
0
0
0
0
Asthma Children
Philadelphia
S75
2016
90
0
0
0
0
0
0
Asthma Children
Philadelphia
S75
2016
100
0
0
0
0
0
0
Asthma Children
Philadelphia
S75
2017
0
98958
89180
82501
76958
73051
69515
Asthma Children
Philadelphia
S75
2017
10
95946
85710
78049
72767
68577
64779
Asthma Children
Philadelphia
S75
2017
20
87434
74208
65914
60501
56398
52578
Asthma Children
Philadelphia
S75
2017
30
74077
58886
50876
44852
40050
36624
Asthma Children
Philadelphia
S75
2017
40
55961
40552
32957
27304
23463
20342
Asthma Children
Philadelphia
S75
2017
50
33088
19512
13205
9080
6155
4212
Asthma Children
Philadelphia
S75
2017
60
13292
4103
1724
567
262
44
Asthma Children
Philadelphia
S75
2017
70
1942
131
22
0
0
0
Asthma Children
Philadelphia
S75
2017
80
87
0
0
0
0
0
Asthma Children
Philadelphia
S75
2017
90
0
0
0
0
0
0
Asthma Children
Philadelphia
S75
2017
100
0
0
0
0
0
0
Asthma Children
Phoenix
S65
2015
0
58411
53655
50924
48659
46819
44993
Asthma Children
Phoenix
S65
2015
10
56727
51745
48532
46196
44158
42488
Asthma Children
Phoenix
S65
2015
20
51886
46253
42502
39658
36742
34548
Asthma Children
Phoenix
S65
2015
30
44923
37520
33260
30274
27641
25844
Asthma Children
Phoenix
S65
2015
40
34619
26778
22306
19107
17126
15243
Asthma Children
Phoenix
S65
2015
50
18541
10643
7034
4897
3468
2633
Asthma Children
Phoenix
S65
2015
60
1500
226
14
0
0
0
Asthma Children
Phoenix
S65
2015
70
0
0
0
0
0
0
Asthma Children
Phoenix
S65
2015
80
0
0
0
0
0
0
Asthma Children
Phoenix
S65
2015
90
0
0
0
0
0
0
Asthma Children
Phoenix
S65
2015
100
0
0
0
0
0
0
Asthma Children
Phoenix
S65
2016
0
56995
52580
49848
47584
45701
44130
Asthma Children
Phoenix
S65
2016
10
55127
50329
47456
45163
43026
41512
Asthma Children
Phoenix
S65
2016
20
50442
44654
41186
38384
36360
34548
Asthma Children
Phoenix
S65
2016
30
43380
37266
33133
30161
27910
26014
Asthma Children
Phoenix
S65
2016
40
33289
25518
21414
18725
16630
15073
Asthma Children
Phoenix
S65
2016
50
15059
7982
5081
3354
2364
1571
Asthma Children
Phoenix
S65
2016
60
778
71
0
0
0
0
Asthma Children
Phoenix
S65
2016
70
0
0
0
0
0
0
Asthma Children
Phoenix
S65
2016
80
0
0
0
0
0
0
Asthma Children
Phoenix
S65
2016
90
0
0
0
0
0
0
Asthma Children
Phoenix
S65
2016
100
0
0
0
0
0
0
Asthma Children
Phoenix
S65
2017
0
58340
53684
50457
48121
46437
44894
Asthma Children
Phoenix
S65
2017
10
56925
51518
48150
45645
44031
42375
Asthma Children
Phoenix
S65
2017
20
52325
46338
42460
39488
37252
35341
Asthma Children
Phoenix
S65
2017
30
45772
38058
33855
30939
28901
27160
3D-Attachment4-28
-------
Study Group
Study Area
AQ
Scenario
Year
Benchmark
(PPb)
Number of People with 7-hr Exposure at or above Benchmark
>1 Day
> 2 Days
> 3 Days
> 4 Days
> 5 Days
> 6 Days
Asthma Children
Phoenix
S65
2017
40
35907
27953
24146
21103
18697
16956
Asthma Children
Phoenix
S65
2017
50
19348
12002
8464
6199
4671
3637
Asthma Children
Phoenix
S65
2017
60
2788
481
127
42
0
0
Asthma Children
Phoenix
S65
2017
70
0
0
0
0
0
0
Asthma Children
Phoenix
S65
2017
80
0
0
0
0
0
0
Asthma Children
Phoenix
S65
2017
90
0
0
0
0
0
0
Asthma Children
Phoenix
S65
2017
100
0
0
0
0
0
0
Asthma Children
Phoenix
S70
2015
0
58411
53655
50924
48659
46819
44993
Asthma Children
Phoenix
S70
2015
10
56826
51801
48659
46281
44257
42573
Asthma Children
Phoenix
S70
2015
20
52339
46819
42984
40210
37520
35256
Asthma Children
Phoenix
S70
2015
30
46310
39148
34789
31788
29085
27203
Asthma Children
Phoenix
S70
2015
40
37351
29566
24896
22192
19716
17946
Asthma Children
Phoenix
S70
2015
50
24995
16588
12738
10063
8209
6610
Asthma Children
Phoenix
S70
2015
60
7034
2406
1203
665
354
170
Asthma Children
Phoenix
S70
2015
70
127
0
0
0
0
0
Asthma Children
Phoenix
S70
2015
80
0
0
0
0
0
0
Asthma Children
Phoenix
S70
2015
90
0
0
0
0
0
0
Asthma Children
Phoenix
S70
2015
100
0
0
0
0
0
0
Asthma Children
Phoenix
S70
2016
0
56995
52580
49848
47584
45701
44130
Asthma Children
Phoenix
S70
2016
10
55269
50414
47612
45347
43224
41639
Asthma Children
Phoenix
S70
2016
20
50782
45305
41823
38922
36983
35100
Asthma Children
Phoenix
S70
2016
30
44427
38440
34435
31519
29283
27089
Asthma Children
Phoenix
S70
2016
40
35978
28717
24542
21584
19475
17777
Asthma Children
Phoenix
S70
2016
50
22192
14734
10714
8591
6680
5308
Asthma Children
Phoenix
S70
2016
60
5336
1444
623
212
127
85
Asthma Children
Phoenix
S70
2016
70
0
0
0
0
0
0
Asthma Children
Phoenix
S70
2016
80
0
0
0
0
0
0
Asthma Children
Phoenix
S70
2016
90
0
0
0
0
0
0
Asthma Children
Phoenix
S70
2016
100
0
0
0
0
0
0
Asthma Children
Phoenix
S70
2017
0
58340
53684
50457
48121
46437
44894
Asthma Children
Phoenix
S70
2017
10
57010
51674
48362
45871
44173
42559
Asthma Children
Phoenix
S70
2017
20
52806
46947
43040
40210
38030
35949
Asthma Children
Phoenix
S70
2017
30
46961
39445
35270
32425
30331
28505
Asthma Children
Phoenix
S70
2017
40
38624
30868
26877
23877
21640
19857
Asthma Children
Phoenix
S70
2017
50
26453
18343
14351
11507
9822
8308
Asthma Children
Phoenix
S70
2017
60
9143
3977
2109
1033
609
CNJ
CO
CO
Asthma Children
Phoenix
S70
2017
70
552
71
0
0
0
0
Asthma Children
Phoenix
S70
2017
80
0
0
0
0
0
0
Asthma Children
Phoenix
S70
2017
90
0
0
0
0
0
0
Asthma Children
Phoenix
S70
2017
100
0
0
0
0
0
0
Asthma Children
Phoenix
S75
2015
0
58411
53655
50924
48659
46819
44993
Asthma Children
Phoenix
S75
2015
10
56769
51730
48518
46267
44243
42602
Asthma Children
Phoenix
S75
2015
20
52495
47046
43295
40549
37818
35511
Asthma Children
Phoenix
S75
2015
30
47145
40139
35681
32737
29977
27995
Asthma Children
Phoenix
S75
2015
40
39191
31307
26637
23877
21626
19659
Asthma Children
Phoenix
S75
2015
50
28844
20480
16404
13814
11450
9780
Asthma Children
Phoenix
S75
2015
60
14451
7402
4161
2774
1840
1161
3D-Attachment4-29
-------
Study Group
Study Area
AQ
Scenario
Year
Benchmark
(PPb)
Number of People with 7-hr Exposure at or above Benchmark
>1 Day
> 2 Days
> 3 Days
> 4 Days
> 5 Days
> 6 Days
Asthma Children
Phoenix
S75
2015
70
1302
170
28
0
0
0
Asthma Children
Phoenix
S75
2015
80
14
0
0
0
0
0
Asthma Children
Phoenix
S75
2015
90
0
0
0
0
0
0
Asthma Children
Phoenix
S75
2015
100
0
0
0
0
0
0
Asthma Children
Phoenix
S75
2016
0
56995
52580
49848
47584
45701
44130
Asthma Children
Phoenix
S75
2016
10
55269
50428
47555
45234
43054
41526
Asthma Children
Phoenix
S75
2016
20
51108
45489
42078
39191
37195
35270
Asthma Children
Phoenix
S75
2016
30
45078
39077
35185
32524
30090
27981
Asthma Children
Phoenix
S75
2016
40
37605
30628
26495
23466
21202
19447
Asthma Children
Phoenix
S75
2016
50
26835
19135
14847
12172
10091
8733
Asthma Children
Phoenix
S75
2016
60
11634
5364
2802
1755
1019
580
Asthma Children
Phoenix
S75
2016
70
807
142
14
0
0
0
Asthma Children
Phoenix
S75
2016
80
0
0
0
0
0
0
Asthma Children
Phoenix
S75
2016
90
0
0
0
0
0
0
Asthma Children
Phoenix
S75
2016
100
0
0
0
0
0
0
Asthma Children
Phoenix
S75
2017
0
58340
53684
50457
48121
46437
44894
Asthma Children
Phoenix
S75
2017
10
56953
51773
48291
45942
44215
42403
Asthma Children
Phoenix
S75
2017
20
53033
47272
43338
40323
38172
36204
Asthma Children
Phoenix
S75
2017
30
47470
40479
36233
33331
30996
29184
Asthma Children
Phoenix
S75
2017
40
40507
32425
28618
25787
23636
21810
Asthma Children
Phoenix
S75
2017
50
30670
22504
18145
15187
13064
11620
Asthma Children
Phoenix
S75
2017
60
15668
8931
5789
3595
2562
1897
Asthma Children
Phoenix
S75
2017
70
3128
637
113
99
14
0
Asthma Children
Phoenix
S75
2017
80
156
0
0
0
0
0
Asthma Children
Phoenix
S75
2017
90
0
0
0
0
0
0
Asthma Children
Phoenix
S75
2017
100
0
0
0
0
0
0
Asthma Children
Sacramento
S65
2015
0
30691
27958
26095
24744
23595
22454
Asthma Children
Sacramento
S65
2015
10
29286
26297
24239
22733
21382
20334
Asthma Children
Sacramento
S65
2015
20
25536
21941
19620
17935
16600
15637
Asthma Children
Sacramento
S65
2015
30
20691
16569
14006
12384
11258
10202
Asthma Children
Sacramento
S65
2015
40
12904
8735
6491
5124
4185
3540
Asthma Children
Sacramento
S65
2015
50
3362
1017
435
163
70
31
Asthma Children
Sacramento
S65
2015
60
217
16
8
0
0
0
Asthma Children
Sacramento
S65
2015
70
0
0
0
0
0
0
Asthma Children
Sacramento
S65
2015
80
0
0
0
0
0
0
Asthma Children
Sacramento
S65
2015
90
0
0
0
0
0
0
Asthma Children
Sacramento
S65
2015
100
0
0
0
0
0
0
Asthma Children
Sacramento
S65
2016
0
31786
28968
27198
25738
24597
23463
Asthma Children
Sacramento
S65
2016
10
30280
27190
25101
23587
22267
21351
Asthma Children
Sacramento
S65
2016
20
26281
22461
20233
18548
17166
16126
Asthma Children
Sacramento
S65
2016
30
20668
16763
14185
12741
11421
10489
Asthma Children
Sacramento
S65
2016
40
13137
8851
6755
5443
4270
3463
Asthma Children
Sacramento
S65
2016
50
4262
1770
784
342
155
78
Asthma Children
Sacramento
S65
2016
60
295
23
0
0
0
0
Asthma Children
Sacramento
S65
2016
70
0
0
0
0
0
0
Asthma Children
Sacramento
S65
2016
80
0
0
0
0
0
0
Asthma Children
Sacramento
S65
2016
90
0
0
0
0
0
0
3D-Attachment4-30
-------
Study Group
Study Area
AQ
Scenario
Year
Benchmark
(PPb)
Number of People with 7-hr Exposure at or above Benchmark
>1 Day
> 2 Days
> 3 Days
> 4 Days
> 5 Days
> 6 Days
Asthma Children
Sacramento
S65
2016
100
0
0
0
0
0
0
Asthma Children
Sacramento
S65
2017
0
30598
27990
26118
24698
23448
22500
Asthma Children
Sacramento
S65
2017
10
29340
26281
24302
22632
21320
20412
Asthma Children
Sacramento
S65
2017
20
25893
22190
19721
18215
16918
15885
Asthma Children
Sacramento
S65
2017
30
20497
16219
13773
12423
11328
10350
Asthma Children
Sacramento
S65
2017
40
12702
8595
6530
5132
4177
3432
Asthma Children
Sacramento
S65
2017
50
3253
1002
318
148
85
47
Asthma Children
Sacramento
S65
2017
60
124
0
0
0
0
0
Asthma Children
Sacramento
S65
2017
70
0
0
0
0
0
0
Asthma Children
Sacramento
S65
2017
80
0
0
0
0
0
0
Asthma Children
Sacramento
S65
2017
90
0
0
0
0
0
0
Asthma Children
Sacramento
S65
2017
100
0
0
0
0
0
0
Asthma Children
Sacramento
S70
2015
0
30691
27958
26095
24744
23595
22454
Asthma Children
Sacramento
S70
2015
10
29379
26390
24356
22920
21561
20458
Asthma Children
Sacramento
S70
2015
20
26080
22454
20125
18463
17112
16126
Asthma Children
Sacramento
S70
2015
30
21910
17834
15249
13502
12399
11452
Asthma Children
Sacramento
S70
2015
40
15753
11320
9092
7648
6475
5582
Asthma Children
Sacramento
S70
2015
50
7221
3781
2337
1374
893
598
Asthma Children
Sacramento
S70
2015
60
1157
179
39
0
0
0
Asthma Children
Sacramento
S70
2015
70
70
8
0
0
0
0
Asthma Children
Sacramento
S70
2015
80
0
0
0
0
0
0
Asthma Children
Sacramento
S70
2015
90
0
0
0
0
0
0
Asthma Children
Sacramento
S70
2015
100
0
0
0
0
0
0
Asthma Children
Sacramento
S70
2016
0
31786
28968
27198
25738
24597
23463
Asthma Children
Sacramento
S70
2016
10
30389
27291
25194
23727
22430
21491
Asthma Children
Sacramento
S70
2016
20
26786
22989
20839
19146
17741
16693
Asthma Children
Sacramento
S70
2016
30
21864
17966
15668
13874
12539
11553
Asthma Children
Sacramento
S70
2016
40
15994
11592
9200
7710
6467
5427
Asthma Children
Sacramento
S70
2016
50
8269
4705
2958
1918
1343
924
Asthma Children
Sacramento
S70
2016
60
1871
396
132
70
23
0
Asthma Children
Sacramento
S70
2016
70
155
8
0
0
0
0
Asthma Children
Sacramento
S70
2016
80
0
0
0
0
0
0
Asthma Children
Sacramento
S70
2016
90
0
0
0
0
0
0
Asthma Children
Sacramento
S70
2016
100
0
0
0
0
0
0
Asthma Children
Sacramento
S70
2017
0
30598
27990
26118
24698
23448
22500
Asthma Children
Sacramento
S70
2017
10
29449
26367
24434
22811
21483
20513
Asthma Children
Sacramento
S70
2017
20
26421
22609
20280
18688
17547
16491
Asthma Children
Sacramento
S70
2017
30
21763
17539
14860
13533
12492
11514
Asthma Children
Sacramento
S70
2017
40
15621
11196
8975
7469
6328
5551
Asthma Children
Sacramento
S70
2017
50
7283
3750
2182
1312
939
668
Asthma Children
Sacramento
S70
2017
60
1522
272
31
0
0
0
Asthma Children
Sacramento
S70
2017
70
54
0
0
0
0
0
Asthma Children
Sacramento
S70
2017
80
0
0
0
0
0
0
Asthma Children
Sacramento
S70
2017
90
0
0
0
0
0
0
Asthma Children
Sacramento
S70
2017
100
0
0
0
0
0
0
Asthma Children
Sacramento
S75
2015
0
30691
27958
26095
24744
23595
22454
Asthma Children
Sacramento
S75
2015
10
29426
26468
24418
22958
21654
20536
3D-Attachment4-31
-------
Study Group
Study Area
AQ
Scenario
Year
Benchmark
(PPb)
Number of People with 7-hr Exposure at or above Benchmark
>1 Day
> 2 Days
> 3 Days
> 4 Days
> 5 Days
> 6 Days
Asthma Children
Sacramento
S75
2015
20
26359
22811
20497
18805
17485
16336
Asthma Children
Sacramento
S75
2015
30
22578
18587
16041
14286
13121
12190
Asthma Children
Sacramento
S75
2015
40
17306
13044
10575
9006
7811
7050
Asthma Children
Sacramento
S75
2015
50
10000
5940
4162
2943
2174
1685
Asthma Children
Sacramento
S75
2015
60
3059
846
318
93
39
16
Asthma Children
Sacramento
S75
2015
70
427
23
8
0
0
0
Asthma Children
Sacramento
S75
2015
80
23
0
0
0
0
0
Asthma Children
Sacramento
S75
2015
90
0
0
0
0
0
0
Asthma Children
Sacramento
S75
2015
100
0
0
0
0
0
0
Asthma Children
Sacramento
S75
2016
0
31786
28968
27198
25738
24597
23463
Asthma Children
Sacramento
S75
2016
10
30435
27337
25280
23836
22531
21553
Asthma Children
Sacramento
S75
2016
20
27066
23370
21103
19534
18044
16965
Asthma Children
Sacramento
S75
2016
30
22586
18696
16413
14527
13207
12151
Asthma Children
Sacramento
S75
2016
40
17345
13207
10808
9270
7950
6863
Asthma Children
Sacramento
S75
2016
50
10947
6778
4860
3540
2717
2065
Asthma Children
Sacramento
S75
2016
60
4301
1739
769
349
163
78
Asthma Children
Sacramento
S75
2016
70
675
78
23
0
0
0
Asthma Children
Sacramento
S75
2016
80
23
0
0
0
0
0
Asthma Children
Sacramento
S75
2016
90
0
0
0
0
0
0
Asthma Children
Sacramento
S75
2016
100
0
0
0
0
0
0
Asthma Children
Sacramento
S75
2017
0
30598
27990
26118
24698
23448
22500
Asthma Children
Sacramento
S75
2017
10
29496
26452
24503
22873
21522
20606
Asthma Children
Sacramento
S75
2017
20
26646
22920
20629
19030
17850
16801
Asthma Children
Sacramento
S75
2017
30
22586
18455
15668
14247
13082
12065
Asthma Children
Sacramento
S75
2017
40
17260
12772
10536
9061
7756
6724
Asthma Children
Sacramento
S75
2017
50
10388
6242
4363
3106
2236
1724
Asthma Children
Sacramento
S75
2017
60
3253
978
357
171
70
16
Asthma Children
Sacramento
S75
2017
70
404
8
0
0
0
0
Asthma Children
Sacramento
S75
2017
80
23
0
0
0
0
0
Asthma Children
Sacramento
S75
2017
90
0
0
0
0
0
0
Asthma Children
Sacramento
S75
2017
100
0
0
0
0
0
0
Asthma Children
St. Louis
S65
2015
0
37965
34049
31527
29587
27775
26391
Asthma Children
St. Louis
S65
2015
10
36016
31764
29150
27101
25161
23586
Asthma Children
St. Louis
S65
2015
20
31490
26200
23049
20690
18942
17557
Asthma Children
St. Louis
S65
2015
30
24642
19333
16292
14288
12594
11247
Asthma Children
St. Louis
S65
2015
40
16647
11256
8360
6584
5400
4471
Asthma Children
St. Louis
S65
2015
50
6793
2477
1093
474
228
91
Asthma Children
St. Louis
S65
2015
60
CO
CNJ
CO
9
0
0
0
0
Asthma Children
St. Louis
S65
2015
70
0
0
0
0
0
0
Asthma Children
St. Louis
S65
2015
80
0
0
0
0
0
0
Asthma Children
St. Louis
S65
2015
90
0
0
0
0
0
0
Asthma Children
St. Louis
S65
2015
100
0
0
0
0
0
0
Asthma Children
St. Louis
S65
2016
0
36882
32765
30461
28595
27037
25826
Asthma Children
St. Louis
S65
2016
10
35133
30880
28367
26327
24697
23440
Asthma Children
St. Louis
S65
2016
20
30871
26027
23194
21045
19361
17913
Asthma Children
St. Louis
S65
2016
30
25025
19743
16665
14525
13050
11702
Asthma Children
St. Louis
S65
2016
40
17785
12230
9243
7549
6211
5409
3D-Attachment4-32
-------
Study Group
Study Area
AQ
Scenario
Year
Benchmark
(PPb)
Number of People with 7-hr Exposure at or above Benchmark
>1 Day
> 2 Days
> 3 Days
> 4 Days
> 5 Days
> 6 Days
Asthma Children
St. Louis
S65
2016
50
8779
4216
2176
1166
729
483
Asthma Children
St. Louis
S65
2016
60
1657
173
36
0
0
0
Asthma Children
St. Louis
S65
2016
70
27
0
0
0
0
0
Asthma Children
St. Louis
S65
2016
80
0
0
0
0
0
0
Asthma Children
St. Louis
S65
2016
90
0
0
0
0
0
0
Asthma Children
St. Louis
S65
2016
100
0
0
0
0
0
0
Asthma Children
St. Louis
S65
2017
0
37656
33967
31427
29505
27966
26864
Asthma Children
St. Louis
S65
2017
10
36344
32419
29596
27647
25926
24533
Asthma Children
St. Louis
S65
2017
20
32501
27529
24497
22329
20490
19087
Asthma Children
St. Louis
S65
2017
30
26737
21073
18013
16292
14552
13259
Asthma Children
St. Louis
S65
2017
40
19370
13897
10682
8597
7249
6101
Asthma Children
St. Louis
S65
2017
50
8251
3643
1849
965
483
310
Asthma Children
St. Louis
S65
2017
60
510
36
9
0
0
0
Asthma Children
St. Louis
S65
2017
70
9
0
0
0
0
0
Asthma Children
St. Louis
S65
2017
80
0
0
0
0
0
0
Asthma Children
St. Louis
S65
2017
90
0
0
0
0
0
0
Asthma Children
St. Louis
S65
2017
100
0
0
0
0
0
0
Asthma Children
St. Louis
S70
2015
0
37965
34049
31527
29587
27775
26391
Asthma Children
St. Louis
S70
2015
10
36135
31873
29305
27165
25316
23723
Asthma Children
St. Louis
S70
2015
20
32000
26855
23577
21291
19442
17913
Asthma Children
St. Louis
S70
2015
30
25808
20399
17230
15317
13696
12230
Asthma Children
St. Louis
S70
2015
40
18741
13268
10409
8469
7121
6065
Asthma Children
St. Louis
S70
2015
50
10582
5309
3114
2013
1229
738
Asthma Children
St. Louis
S70
2015
60
2195
337
82
0
0
0
Asthma Children
St. Louis
S70
2015
70
18
0
0
0
0
0
Asthma Children
St. Louis
S70
2015
80
0
0
0
0
0
0
Asthma Children
St. Louis
S70
2015
90
0
0
0
0
0
0
Asthma Children
St. Louis
S70
2015
100
0
0
0
0
0
0
Asthma Children
St. Louis
S70
2016
0
36882
32765
30461
28595
27037
25826
Asthma Children
St. Louis
S70
2016
10
35242
30962
28504
26445
24879
23531
Asthma Children
St. Louis
S70
2016
20
31390
26582
23795
21519
19861
18368
Asthma Children
St. Louis
S70
2016
30
26136
20936
17785
15691
13951
12676
Asthma Children
St. Louis
S70
2016
40
19697
14234
11128
9407
8014
6939
Asthma Children
St. Louis
S70
2016
50
12121
6967
4562
3096
2067
1457
Asthma Children
St. Louis
S70
2016
60
4927
1211
455
155
64
9
Asthma Children
St. Louis
S70
2016
70
437
0
0
0
0
0
Asthma Children
St. Louis
S70
2016
80
0
0
0
0
0
0
Asthma Children
St. Louis
S70
2016
90
0
0
0
0
0
0
Asthma Children
St. Louis
S70
2016
100
0
0
0
0
0
0
Asthma Children
St. Louis
S70
2017
0
37656
33967
31427
29505
27966
26864
Asthma Children
St. Louis
S70
2017
10
36454
32556
29769
27748
26063
24688
Asthma Children
St. Louis
S70
2017
20
32993
28230
25180
22912
21073
19643
Asthma Children
St. Louis
S70
2017
30
27839
22184
19087
17129
15572
14243
Asthma Children
St. Louis
S70
2017
40
21482
16000
12959
10855
9307
8050
Asthma Children
St. Louis
S70
2017
50
12631
7130
4617
3269
2340
1748
Asthma Children
St. Louis
S70
2017
60
2969
592
173
64
36
9
Asthma Children
St. Louis
S70
2017
70
118
9
0
0
0
0
3D-Attachment4-33
-------
Study Group
Study Area
AQ
Scenario
Year
Benchmark
(PPb)
Number of People with 7-hr Exposure at or above Benchmark
>1 Day
> 2 Days
> 3 Days
> 4 Days
> 5 Days
> 6 Days
Asthma Children
St. Louis
S70
2017
80
9
0
0
0
0
0
Asthma Children
St. Louis
S70
2017
90
0
0
0
0
0
0
Asthma Children
St. Louis
S70
2017
100
0
0
0
0
0
0
Asthma Children
St. Louis
S75
2015
0
37965
34049
31527
29587
27775
26391
Asthma Children
St. Louis
S75
2015
10
36153
31946
29350
27147
25371
23750
Asthma Children
St. Louis
S75
2015
20
32201
27165
23914
21619
19634
18213
Asthma Children
St. Louis
S75
2015
30
26436
20927
17822
15909
14370
12767
Asthma Children
St. Louis
S75
2015
40
19980
14716
11629
9535
8241
7012
Asthma Children
St. Louis
S75
2015
50
12968
7495
4954
3506
2359
1712
Asthma Children
St. Louis
S75
2015
60
4808
1402
492
155
46
9
Asthma Children
St. Louis
S75
2015
70
364
0
0
0
0
0
Asthma Children
St. Louis
S75
2015
80
0
0
0
0
0
0
Asthma Children
St. Louis
S75
2015
90
0
0
0
0
0
0
Asthma Children
St. Louis
S75
2015
100
0
0
0
0
0
0
Asthma Children
St. Louis
S75
2016
0
36882
32765
30461
28595
27037
25826
Asthma Children
St. Louis
S75
2016
10
35315
30980
28522
26500
24888
23622
Asthma Children
St. Louis
S75
2016
20
31645
26919
24069
21892
20135
18723
Asthma Children
St. Louis
S75
2016
30
26855
21637
18477
16355
14416
13177
Asthma Children
St. Louis
S75
2016
40
21027
15426
12303
10345
9125
8050
Asthma Children
St. Louis
S75
2016
50
14142
8788
6138
4526
3379
2586
Asthma Children
St. Louis
S75
2016
60
7212
2705
1102
619
346
137
Asthma Children
St. Louis
S75
2016
70
1767
191
27
0
0
0
Asthma Children
St. Louis
S75
2016
80
82
0
0
0
0
0
Asthma Children
St. Louis
S75
2016
90
0
0
0
0
0
0
Asthma Children
St. Louis
S75
2016
100
0
0
0
0
0
0
Asthma Children
St. Louis
S75
2017
0
37656
33967
31427
29505
27966
26864
Asthma Children
St. Louis
S75
2017
10
36454
32592
29806
27729
26145
24715
Asthma Children
St. Louis
S75
2017
20
33266
28494
25517
23203
21400
19898
Asthma Children
St. Louis
S75
2017
30
28540
22967
19616
17621
16182
14853
Asthma Children
St. Louis
S75
2017
40
22812
17239
14088
12157
10491
9234
Asthma Children
St. Louis
S75
2017
50
15181
9607
6812
5081
3943
2969
Asthma Children
St. Louis
S75
2017
60
5719
1931
856
301
155
100
Asthma Children
St. Louis
S75
2017
70
610
55
18
0
0
0
Asthma Children
St. Louis
S75
2017
80
18
0
0
0
0
0
Asthma Children
St. Louis
S75
2017
90
0
0
0
0
0
0
Asthma Children
St. Louis
S75
2017
100
0
0
0
0
0
0
All Adults
Atlanta
S65
2015
0
1444098
1258787
1143840
1058756
992690
936765
All Adults
Atlanta
S65
2015
10
1250476
1029174
893026
791391
713280
645875
All Adults
Atlanta
S65
2015
20
845906
585443
443590
360197
299694
252363
All Adults
Atlanta
S65
2015
30
523391
319838
230388
178478
139388
111848
All Adults
Atlanta
S65
2015
40
282368
140022
81703
49867
31836
20074
All Adults
Atlanta
S65
2015
50
72265
15073
3663
986
141
0
All Adults
Atlanta
S65
2015
60
5564
0
0
0
0
0
All Adults
Atlanta
S65
2015
70
211
0
0
0
0
0
All Adults
Atlanta
S65
2015
80
0
0
0
0
0
0
All Adults
Atlanta
S65
2015
90
0
0
0
0
0
0
All Adults
Atlanta
S65
2016
0
1444309
1253434
1141445
1056925
990647
933878
3D-Attachment4-34
-------
Study Group
Study Area
AQ
Scenario
Year
Benchmark
(PPb)
Number of People with 7-hr Exposure at or above Benchmark
>1 Day
> 2 Days
> 3 Days
> 4 Days
> 5 Days
> 6 Days
All Adults
Atlanta
S65
2016
10
1277311
1066293
940498
843089
769556
704335
All Adults
Atlanta
S65
2016
20
912325
650946
510502
413304
346181
293637
All Adults
Atlanta
S65
2016
30
591148
379707
281170
222218
182282
151291
All Adults
Atlanta
S65
2016
40
336390
181437
113398
76843
54234
37682
All Adults
Atlanta
S65
2016
50
93254
20919
5916
1197
493
141
All Adults
Atlanta
S65
2016
60
10706
352
0
0
0
0
All Adults
Atlanta
S65
2016
70
704
0
0
0
0
0
All Adults
Atlanta
S65
2016
80
0
0
0
0
0
0
All Adults
Atlanta
S65
2016
90
0
0
0
0
0
0
All Adults
Atlanta
S65
2017
0
1447972
1257872
1140248
1056854
990154
935568
All Adults
Atlanta
S65
2017
10
1256111
1034034
900210
806111
722577
650665
All Adults
Atlanta
S65
2017
20
844568
581569
446055
357379
296877
251729
All Adults
Atlanta
S65
2017
30
521560
320120
236868
184254
147910
121216
All Adults
Atlanta
S65
2017
40
269972
128189
76491
46345
29934
19017
All Adults
Atlanta
S65
2017
50
45218
7536
1761
352
70
0
All Adults
Atlanta
S65
2017
60
1057
0
0
0
0
0
All Adults
Atlanta
S65
2017
70
0
0
0
0
0
0
All Adults
Atlanta
S65
2017
80
0
0
0
0
0
0
All Adults
Atlanta
S65
2017
90
0
0
0
0
0
0
All Adults
Atlanta
S70
2015
0
1444098
1258787
1143840
1058756
992690
936765
All Adults
Atlanta
S70
2015
10
1262379
1040162
907888
807590
731240
663554
All Adults
Atlanta
S70
2015
20
894224
627281
478596
389356
325051
274831
All Adults
Atlanta
S70
2015
30
567835
353224
254194
200665
160166
128893
All Adults
Atlanta
S70
2015
40
348998
185944
119737
81421
56347
39795
All Adults
Atlanta
S70
2015
50
140022
43950
17397
5705
2043
493
All Adults
Atlanta
S70
2015
60
20003
1972
211
0
0
0
All Adults
Atlanta
S70
2015
70
2254
0
0
0
0
0
All Adults
Atlanta
S70
2015
80
0
0
0
0
0
0
All Adults
Atlanta
S70
2015
90
0
0
0
0
0
0
All Adults
Atlanta
S70
2016
0
1444309
1253434
1141445
1056925
990647
933878
All Adults
Atlanta
S70
2016
10
1290834
1078759
955078
861120
785615
722718
All Adults
Atlanta
S70
2016
20
956275
697925
549804
451760
379143
323783
All Adults
Atlanta
S70
2016
30
638761
415205
308851
245320
201792
168406
All Adults
Atlanta
S70
2016
40
405767
231656
155869
111708
85084
64940
All Adults
Atlanta
S70
2016
50
168547
60784
25779
10917
3663
1479
All Adults
Atlanta
S70
2016
60
34160
3592
282
141
0
0
All Adults
Atlanta
S70
2016
70
5001
0
0
0
0
0
All Adults
Atlanta
S70
2016
80
352
0
0
0
0
0
All Adults
Atlanta
S70
2016
90
0
0
0
0
0
0
All Adults
Atlanta
S70
2017
0
1447972
1257872
1140248
1056854
990154
935568
All Adults
Atlanta
S70
2017
10
1267098
1048332
915494
824847
741242
672499
All Adults
Atlanta
S70
2017
20
885560
621998
481061
386821
322163
272507
All Adults
Atlanta
S70
2017
30
565299
351956
261308
204187
164533
138261
All Adults
Atlanta
S70
2017
40
331178
174464
112694
76209
53389
37752
All Adults
Atlanta
S70
2017
50
106566
26624
9649
3522
1127
423
All Adults
Atlanta
S70
2017
60
9790
211
70
0
0
0
All Adults
Atlanta
S70
2017
70
282
0
0
0
0
0
3D-Attachment4-35
-------
Study Group
Study Area
AQ
Scenario
Year
Benchmark
(PPb)
Number of People with 7-hr Exposure at or above Benchmark
>1 Day
> 2 Days
> 3 Days
> 4 Days
> 5 Days
> 6 Days
All Adults
Atlanta
S70
2017
80
0
0
0
0
0
0
All Adults
Atlanta
S70
2017
90
0
0
0
0
0
0
All Adults
Atlanta
S75
2015
0
1444098
1258787
1143840
1058756
992690
936765
All Adults
Atlanta
S75
2015
10
1269634
1047980
918030
816395
740608
674189
All Adults
Atlanta
S75
2015
20
929088
662286
510079
414853
345124
292088
All Adults
Atlanta
S75
2015
30
610517
383651
277156
217217
173971
143332
All Adults
Atlanta
S75
2015
40
397386
221654
148544
104312
77688
57333
All Adults
Atlanta
S75
2015
50
207286
85577
40147
20285
9931
4649
All Adults
Atlanta
S75
2015
60
55220
9368
1479
352
141
0
All Adults
Atlanta
S75
2015
70
7396
282
0
0
0
0
All Adults
Atlanta
S75
2015
80
1057
0
0
0
0
0
All Adults
Atlanta
S75
2015
90
0
0
0
0
0
0
All Adults
Atlanta
S75
2016
0
1444309
1253434
1141445
1056925
990647
933878
All Adults
Atlanta
S75
2016
10
1297737
1086648
961699
870487
796039
733987
All Adults
Atlanta
S75
2016
20
989943
732720
588049
485287
408796
348716
All Adults
Atlanta
S75
2016
30
680951
443661
334629
263633
217499
183197
All Adults
Atlanta
S75
2016
40
455775
271521
189677
140867
109947
84731
All Adults
Atlanta
S75
2016
50
240530
107341
55502
29652
15777
8029
All Adults
Atlanta
S75
2016
60
75153
15214
3522
563
141
70
All Adults
Atlanta
S75
2016
70
15495
916
0
0
0
0
All Adults
Atlanta
S75
2016
80
2817
0
0
0
0
0
All Adults
Atlanta
S75
2016
90
211
0
0
0
0
0
All Adults
Atlanta
S75
2017
0
1447972
1257872
1140248
1056854
990154
935568
All Adults
Atlanta
S75
2017
10
1274353
1055516
925073
834566
750046
684332
All Adults
Atlanta
S75
2017
20
917607
655524
510150
414219
343011
291735
All Adults
Atlanta
S75
2017
30
604812
380059
280818
219119
179183
149601
All Adults
Atlanta
S75
2017
40
382595
212920
145868
100931
75434
55995
All Adults
Atlanta
S75
2017
50
167984
60291
25004
12608
7184
3240
All Adults
Atlanta
S75
2017
60
33456
2888
211
0
0
0
All Adults
Atlanta
S75
2017
70
3099
70
0
0
0
0
All Adults
Atlanta
S75
2017
80
0
0
0
0
0
0
All Adults
Atlanta
S75
2017
90
0
0
0
0
0
0
All Adults
Boston
S65
2015
0
1850655
1592858
1438181
1324300
1230183
1160719
All Adults
Boston
S65
2015
10
1661832
1370968
1193299
1069439
961624
873475
All Adults
Boston
S65
2015
20
1148099
795108
600318
478709
398190
333325
All Adults
Boston
S65
2015
30
699132
430672
313269
244295
194399
159765
All Adults
Boston
S65
2015
40
359447
173462
97151
56745
36199
23481
All Adults
Boston
S65
2015
50
76214
14675
2152
1076
391
98
All Adults
Boston
S65
2015
60
8707
98
0
0
0
0
All Adults
Boston
S65
2015
70
1174
0
0
0
0
0
All Adults
Boston
S65
2015
80
196
0
0
0
0
0
All Adults
Boston
S65
2015
90
0
0
0
0
0
0
All Adults
Boston
S65
2016
0
1865134
1607632
1447964
1340345
1253076
1178330
All Adults
Boston
S65
2016
10
1689128
1394253
1216877
1087441
986279
895977
All Adults
Boston
S65
2016
20
1166883
813501
616461
491036
404647
341446
All Adults
Boston
S65
2016
30
701969
434389
315226
246447
203302
169060
All Adults
Boston
S65
2016
40
378819
186670
113098
72007
44711
30231
3D-Attachment4-36
-------
Study Group
Study Area
AQ
Scenario
Year
Benchmark
(PPb)
Number of People with 7-hr Exposure at or above Benchmark
>1 Day
> 2 Days
> 3 Days
> 4 Days
> 5 Days
> 6 Days
All Adults
Boston
S65
2016
50
101553
19469
4500
1174
294
0
All Adults
Boston
S65
2016
60
7533
98
98
0
0
0
All Adults
Boston
S65
2016
70
685
0
0
0
0
0
All Adults
Boston
S65
2016
80
0
0
0
0
0
0
All Adults
Boston
S65
2016
90
0
0
0
0
0
0
All Adults
Boston
S65
2017
0
1856329
1605871
1448453
1339367
1248478
1172460
All Adults
Boston
S65
2017
10
1682867
1396698
1219323
1092430
987062
896075
All Adults
Boston
S65
2017
20
1182635
811251
617341
494460
407582
335086
All Adults
Boston
S65
2017
30
715373
438694
318846
247915
195573
155754
All Adults
Boston
S65
2017
40
392614
189312
110848
67213
42852
27981
All Adults
Boston
S65
2017
50
135502
26024
6457
1663
587
98
All Adults
Boston
S65
2017
60
15360
978
98
0
0
0
All Adults
Boston
S65
2017
70
1468
0
0
0
0
0
All Adults
Boston
S65
2017
80
0
0
0
0
0
0
All Adults
Boston
S65
2017
90
0
0
0
0
0
0
All Adults
Boston
S70
2015
0
1850655
1592858
1438181
1324300
1230183
1160719
All Adults
Boston
S70
2015
10
1662713
1371555
1191929
1069830
964559
875725
All Adults
Boston
S70
2015
20
1173144
816143
621744
497493
411300
345848
All Adults
Boston
S70
2015
30
730244
448967
324129
249969
200367
164559
All Adults
Boston
S70
2015
40
408463
206726
124251
79149
51755
35612
All Adults
Boston
S70
2015
50
122490
31601
7925
2152
881
196
All Adults
Boston
S70
2015
60
19274
1076
196
98
0
0
All Adults
Boston
S70
2015
70
2152
0
0
0
0
0
All Adults
Boston
S70
2015
80
294
0
0
0
0
0
All Adults
Boston
S70
2015
90
0
0
0
0
0
0
All Adults
Boston
S70
2016
0
1865134
1607632
1447964
1340345
1253076
1178330
All Adults
Boston
S70
2016
10
1689911
1397188
1219714
1091061
985594
896662
All Adults
Boston
S70
2016
20
1196429
838352
642290
512071
423236
355045
All Adults
Boston
S70
2016
30
738364
450826
327846
257307
209074
174930
All Adults
Boston
S70
2016
40
425388
219738
140100
92552
63202
45004
All Adults
Boston
S70
2016
50
159570
44026
15654
6261
2837
1370
All Adults
Boston
S70
2016
60
23383
1468
98
0
0
0
All Adults
Boston
S70
2016
70
2739
98
98
0
0
0
All Adults
Boston
S70
2016
80
98
0
0
0
0
0
All Adults
Boston
S70
2016
90
0
0
0
0
0
0
All Adults
Boston
S70
2017
0
1856329
1605871
1448453
1339367
1248478
1172460
All Adults
Boston
S70
2017
10
1684237
1396601
1221182
1093702
987649
896662
All Adults
Boston
S70
2017
20
1209833
837080
639550
511484
421768
347805
All Adults
Boston
S70
2017
30
748441
455913
332347
254568
202813
161135
All Adults
Boston
S70
2017
40
440846
226782
137557
89226
60267
40308
All Adults
Boston
S70
2017
50
188235
48331
14675
4794
1859
391
All Adults
Boston
S70
2017
60
48429
4794
98
0
0
0
All Adults
Boston
S70
2017
70
5283
196
0
0
0
0
All Adults
Boston
S70
2017
80
489
0
0
0
0
0
All Adults
Boston
S70
2017
90
0
0
0
0
0
0
All Adults
Boston
S75
2015
0
1850655
1592858
1438181
1324300
1230183
1160719
All Adults
Boston
S75
2015
10
1659582
1368130
1187820
1064351
961429
871029
3D-Attachment4-37
-------
Study Group
Study Area
AQ
Scenario
Year
Benchmark
(PPb)
Number of People with 7-hr Exposure at or above Benchmark
>1 Day
> 2 Days
> 3 Days
> 4 Days
> 5 Days
> 6 Days
All Adults
Boston
S75
2015
20
1183319
827101
630451
503755
415996
350642
All Adults
Boston
S75
2015
30
742082
455815
328140
251339
201443
164364
All Adults
Boston
S75
2015
40
426954
220619
135209
88150
59288
40993
All Adults
Boston
S75
2015
50
152819
44124
14480
4598
2152
587
All Adults
Boston
S75
2015
60
31014
3131
489
391
0
0
All Adults
Boston
S75
2015
70
4305
98
0
0
0
0
All Adults
Boston
S75
2015
80
489
0
0
0
0
0
All Adults
Boston
S75
2015
90
0
0
0
0
0
0
All Adults
Boston
S75
2016
0
1865134
1607632
1447964
1340345
1253076
1178330
All Adults
Boston
S75
2016
10
1686878
1394253
1216290
1088223
982659
893433
All Adults
Boston
S75
2016
20
1209148
849994
649334
521463
430476
362480
All Adults
Boston
S75
2016
30
756072
459044
332347
257992
210542
175615
All Adults
Boston
S75
2016
40
444271
234414
149982
102825
70931
49896
All Adults
Boston
S75
2016
50
188627
58897
23089
10958
5087
2348
All Adults
Boston
S75
2016
60
44124
4403
489
98
98
0
All Adults
Boston
S75
2016
70
5283
196
98
0
0
0
All Adults
Boston
S75
2016
80
391
0
0
0
0
0
All Adults
Boston
S75
2016
90
0
0
0
0
0
0
All Adults
Boston
S75
2017
0
1856329
1605871
1448453
1339367
1248478
1172460
All Adults
Boston
S75
2017
10
1682182
1393470
1218638
1091256
984909
892357
All Adults
Boston
S75
2017
20
1218834
847646
647279
517843
424801
352403
All Adults
Boston
S75
2017
30
765758
465697
336652
256231
204085
161918
All Adults
Boston
S75
2017
40
462468
240773
148318
98031
68289
47255
All Adults
Boston
S75
2017
50
213379
63495
21817
7925
3620
1565
All Adults
Boston
S75
2017
60
78366
10762
1272
196
0
0
All Adults
Boston
S75
2017
70
10664
783
98
0
0
0
All Adults
Boston
S75
2017
80
685
0
0
0
0
0
All Adults
Boston
S75
2017
90
0
0
0
0
0
0
All Adults
Dallas
S65
2015
0
1659225
1460056
1340038
1250572
1177202
1118443
All Adults
Dallas
S65
2015
10
1471229
1236976
1093362
983580
895364
818634
All Adults
Dallas
S65
2015
20
1054684
754797
588992
480070
401621
341925
All Adults
Dallas
S65
2015
30
663456
419670
307232
241598
194638
158304
All Adults
Dallas
S65
2015
40
400527
210734
133613
87903
59462
42506
All Adults
Dallas
S65
2015
50
162836
50867
18206
7345
3047
1250
All Adults
Dallas
S65
2015
60
20472
1016
234
0
0
0
All Adults
Dallas
S65
2015
70
547
0
0
0
0
0
All Adults
Dallas
S65
2015
80
0
0
0
0
0
0
All Adults
Dallas
S65
2015
90
0
0
0
0
0
0
All Adults
Dallas
S65
2016
0
1670711
1470213
1353634
1261824
1191110
1126335
All Adults
Dallas
S65
2016
10
1483731
1252057
1103051
994754
906147
830667
All Adults
Dallas
S65
2016
20
1043745
748390
588210
481789
400918
344581
All Adults
Dallas
S65
2016
30
658142
424437
315827
246676
200654
166899
All Adults
Dallas
S65
2016
40
367866
191590
120096
76574
50085
34927
All Adults
Dallas
S65
2016
50
97749
21253
4766
938
313
156
All Adults
Dallas
S65
2016
60
4923
78
78
0
0
0
All Adults
Dallas
S65
2016
70
234
0
0
0
0
0
All Adults
Dallas
S65
2016
80
0
0
0
0
0
0
3D-Attachment4-3 8
-------
Study Group
Study Area
AQ
Scenario
Year
Benchmark
(PPb)
Number of People with 7-hr Exposure at or above Benchmark
>1 Day
> 2 Days
> 3 Days
> 4 Days
> 5 Days
> 6 Days
All Adults
Dallas
S65
2016
90
0
0
0
0
0
0
All Adults
Dallas
S65
2017
0
1672743
1467478
1349961
1259792
1182828
1124460
All Adults
Dallas
S65
2017
10
1490841
1258698
1109380
1003349
920446
848014
All Adults
Dallas
S65
2017
20
1069061
774097
611886
502182
418420
359662
All Adults
Dallas
S65
2017
30
683693
438345
329735
264101
214875
179323
All Adults
Dallas
S65
2017
40
396464
214328
135254
92748
62509
45553
All Adults
Dallas
S65
2017
50
141974
37662
11720
3438
1172
469
All Adults
Dallas
S65
2017
60
17659
547
0
0
0
0
All Adults
Dallas
S65
2017
70
156
0
0
0
0
0
All Adults
Dallas
S65
2017
80
0
0
0
0
0
0
All Adults
Dallas
S65
2017
90
0
0
0
0
0
0
All Adults
Dallas
S70
2015
0
1659225
1460056
1340038
1250572
1177202
1118443
All Adults
Dallas
S70
2015
10
1476308
1241664
1099691
990769
902162
825432
All Adults
Dallas
S70
2015
20
1082423
783317
617668
505151
420686
357474
All Adults
Dallas
S70
2015
30
703071
445221
324188
255115
207452
170025
All Adults
Dallas
S70
2015
40
440533
241598
159007
110563
77433
54539
All Adults
Dallas
S70
2015
50
226674
87044
39928
18753
9064
4610
All Adults
Dallas
S70
2015
60
54461
6798
1328
234
78
0
All Adults
Dallas
S70
2015
70
4141
0
0
0
0
0
All Adults
Dallas
S70
2015
80
0
0
0
0
0
0
All Adults
Dallas
S70
2015
90
0
0
0
0
0
0
All Adults
Dallas
S70
2016
0
1670711
1470213
1353634
1261824
1191110
1126335
All Adults
Dallas
S70
2016
10
1485997
1258698
1110317
1001317
912867
839028
All Adults
Dallas
S70
2016
20
1070468
772846
610088
501635
418108
357864
All Adults
Dallas
S70
2016
30
687756
445143
331611
261757
212453
177838
All Adults
Dallas
S70
2016
40
410606
221438
144865
99546
69229
47819
All Adults
Dallas
S70
2016
50
148459
43131
14924
5470
1953
625
All Adults
Dallas
S70
2016
60
14611
1250
313
0
0
0
All Adults
Dallas
S70
2016
70
1328
78
0
0
0
0
All Adults
Dallas
S70
2016
80
78
0
0
0
0
0
All Adults
Dallas
S70
2016
90
0
0
0
0
0
0
All Adults
Dallas
S70
2017
0
1672743
1467478
1349961
1259792
1182828
1124460
All Adults
Dallas
S70
2017
10
1496155
1263386
1120084
1012491
929197
856530
All Adults
Dallas
S70
2017
20
1095940
801601
638061
525467
440923
377867
All Adults
Dallas
S70
2017
30
715026
461551
345441
276915
227377
191200
All Adults
Dallas
S70
2017
40
439595
247380
161430
114470
83684
61571
All Adults
Dallas
S70
2017
50
195810
65088
27348
10314
4297
1406
All Adults
Dallas
S70
2017
60
40865
3125
313
78
78
0
All Adults
Dallas
S70
2017
70
1485
0
0
0
0
0
All Adults
Dallas
S70
2017
80
0
0
0
0
0
0
All Adults
Dallas
S70
2017
90
0
0
0
0
0
0
All Adults
Dallas
S75
2015
0
1659225
1460056
1340038
1250572
1177202
1118443
All Adults
Dallas
S75
2015
10
1476152
1240414
1099534
991238
902084
827307
All Adults
Dallas
S75
2015
20
1104535
803085
635561
520388
434204
368647
All Adults
Dallas
S75
2015
30
733153
466396
339346
265117
214484
176588
All Adults
Dallas
S75
2015
40
471787
265429
174010
124315
90169
63916
All Adults
Dallas
S75
2015
50
276290
115720
58055
32505
16721
10001
3D-Attachment4-39
-------
Study Group
Study Area
AQ
Scenario
Year
Benchmark
(PPb)
Number of People with 7-hr Exposure at or above Benchmark
>1 Day
> 2 Days
> 3 Days
> 4 Days
> 5 Days
> 6 Days
All Adults
Dallas
S75
2015
60
98139
21097
5157
1406
313
156
All Adults
Dallas
S75
2015
70
12971
625
234
0
0
0
All Adults
Dallas
S75
2015
80
781
0
0
0
0
0
All Adults
Dallas
S75
2015
90
0
0
0
0
0
0
All Adults
Dallas
S75
2016
0
1670711
1470213
1353634
1261824
1191110
1126335
All Adults
Dallas
S75
2016
10
1487481
1258464
1111177
1002567
913648
840903
All Adults
Dallas
S75
2016
20
1087970
792381
626653
515856
432172
367397
All Adults
Dallas
S75
2016
30
709243
463427
341768
268398
220579
183855
All Adults
Dallas
S75
2016
40
440455
241676
161742
115095
82278
59618
All Adults
Dallas
S75
2016
50
192059
67588
27895
11252
4923
2969
All Adults
Dallas
S75
2016
60
31489
3125
625
0
0
0
All Adults
Dallas
S75
2016
70
3751
78
0
0
0
0
All Adults
Dallas
S75
2016
80
313
0
0
0
0
0
All Adults
Dallas
S75
2016
90
78
0
0
0
0
0
All Adults
Dallas
S75
2017
0
1672743
1467478
1349961
1259792
1182828
1124460
All Adults
Dallas
S75
2017
10
1498186
1264793
1122819
1014053
932635
861062
All Adults
Dallas
S75
2017
20
1111724
821525
655720
540469
454519
391150
All Adults
Dallas
S75
2017
30
737294
478429
357474
285276
234018
196904
All Adults
Dallas
S75
2017
40
470224
270821
181667
129863
97670
74620
All Adults
Dallas
S75
2017
50
237847
92045
43600
20472
10158
5391
All Adults
Dallas
S75
2017
60
66416
8282
1485
234
78
78
All Adults
Dallas
S75
2017
70
6329
156
0
0
0
0
All Adults
Dallas
S75
2017
80
0
0
0
0
0
0
All Adults
Dallas
S75
2017
90
0
0
0
0
0
0
All Adults
Detroit
S65
2015
0
1188134
1019169
920988
847385
790298
745861
All Adults
Detroit
S65
2015
10
1064065
871635
763557
679533
612353
558740
All Adults
Detroit
S65
2015
20
746779
519153
395805
320497
267343
226708
All Adults
Detroit
S65
2015
30
467441
292380
217008
170604
138358
114697
All Adults
Detroit
S65
2015
40
278682
147009
93396
62789
43519
31198
All Adults
Detroit
S65
2015
50
98771
27790
10159
3474
1311
655
All Adults
Detroit
S65
2015
60
9700
393
0
0
0
0
All Adults
Detroit
S65
2015
70
0
0
0
0
0
0
All Adults
Detroit
S65
2015
80
0
0
0
0
0
0
All Adults
Detroit
S65
2015
90
0
0
0
0
0
0
All Adults
Detroit
S65
2016
0
1181777
1016416
916465
845615
790167
742060
All Adults
Detroit
S65
2016
10
1065703
876551
765851
685891
623364
564311
All Adults
Detroit
S65
2016
20
773126
542814
420383
344027
285367
241782
All Adults
Detroit
S65
2016
30
494838
303588
224414
180763
148189
121841
All Adults
Detroit
S65
2016
40
295526
155005
99951
67835
47780
32836
All Adults
Detroit
S65
2016
50
128068
39128
14026
5374
1901
655
All Adults
Detroit
S65
2016
60
24578
2228
131
0
0
0
All Adults
Detroit
S65
2016
70
786
0
0
0
0
0
All Adults
Detroit
S65
2016
80
0
0
0
0
0
0
All Adults
Detroit
S65
2016
90
0
0
0
0
0
0
All Adults
Detroit
S65
2017
0
1196917
1026247
926100
854791
800588
755102
All Adults
Detroit
S65
2017
10
1074813
885268
774830
694477
630377
572832
All Adults
Detroit
S65
2017
20
764016
539471
418089
337079
279075
236342
3D-Attachment4-40
-------
Study Group
Study Area
AQ
Scenario
Year
Benchmark
(PPb)
Number of People with 7-hr Exposure at or above Benchmark
>1 Day
> 2 Days
> 3 Days
> 4 Days
> 5 Days
> 6 Days
All Adults
Detroit
S65
2017
30
489267
307913
228805
179846
145109
121514
All Adults
Detroit
S65
2017
40
293888
156185
102638
70654
49287
36048
All Adults
Detroit
S65
2017
50
108864
28707
10356
3670
1114
721
All Adults
Detroit
S65
2017
60
8520
328
0
0
0
0
All Adults
Detroit
S65
2017
70
262
0
0
0
0
0
All Adults
Detroit
S65
2017
80
0
0
0
0
0
0
All Adults
Detroit
S65
2017
90
0
0
0
0
0
0
All Adults
Detroit
S70
2015
0
1188134
1019169
920988
847385
790298
745861
All Adults
Detroit
S70
2015
10
1064917
873274
763688
680910
614516
561559
All Adults
Detroit
S70
2015
20
764737
538750
413566
334458
279469
237456
All Adults
Detroit
S70
2015
30
487497
307586
225725
176437
142880
119679
All Adults
Detroit
S70
2015
40
311912
170998
114173
78781
58332
42471
All Adults
Detroit
S70
2015
50
147337
52630
23923
11076
5374
2884
All Adults
Detroit
S70
2015
60
30215
3212
590
0
0
0
All Adults
Detroit
S70
2015
70
1049
0
0
0
0
0
All Adults
Detroit
S70
2015
80
0
0
0
0
0
0
All Adults
Detroit
S70
2015
90
0
0
0
0
0
0
All Adults
Detroit
S70
2016
0
1181777
1016416
916465
845615
790167
742060
All Adults
Detroit
S70
2016
10
1065769
876944
765524
687791
626183
566736
All Adults
Detroit
S70
2016
20
791806
564246
439455
359495
300573
254104
All Adults
Detroit
S70
2016
30
521906
323054
236736
189087
154219
127740
All Adults
Detroit
S70
2016
40
326789
179387
119154
86121
61806
45355
All Adults
Detroit
S70
2016
50
176175
67377
31919
15861
7341
3867
All Adults
Detroit
S70
2016
60
62264
9438
1966
262
66
66
All Adults
Detroit
S70
2016
70
7668
131
0
0
0
0
All Adults
Detroit
S70
2016
80
0
0
0
0
0
0
All Adults
Detroit
S70
2016
90
0
0
0
0
0
0
All Adults
Detroit
S70
2017
0
1196917
1026247
926100
854791
800588
755102
All Adults
Detroit
S70
2017
10
1075469
885923
775420
694083
629787
574143
All Adults
Detroit
S70
2017
20
781843
559068
435260
350843
290545
246042
All Adults
Detroit
S70
2017
30
514500
322136
239488
188038
151597
126560
All Adults
Detroit
S70
2017
40
325216
179059
121055
84483
62854
46534
All Adults
Detroit
S70
2017
50
160576
57283
25299
11732
5964
2425
All Adults
Detroit
S70
2017
60
30280
2884
262
66
0
0
All Adults
Detroit
S70
2017
70
2359
0
0
0
0
0
All Adults
Detroit
S70
2017
80
0
0
0
0
0
0
All Adults
Detroit
S70
2017
90
0
0
0
0
0
0
All Adults
Detroit
S75
2015
0
1188134
1019169
920988
847385
790298
745861
All Adults
Detroit
S75
2015
10
1059804
866064
754840
673110
606127
552842
All Adults
Detroit
S75
2015
20
769259
539865
416122
335703
279600
235621
All Adults
Detroit
S75
2015
30
496607
310994
226577
176503
141832
117778
All Adults
Detroit
S75
2015
40
326462
181091
121710
86318
63379
47845
All Adults
Detroit
S75
2015
50
174733
70785
35065
18155
9307
5768
All Adults
Detroit
S75
2015
60
51450
9372
1966
524
0
0
All Adults
Detroit
S75
2015
70
5964
197
0
0
0
0
All Adults
Detroit
S75
2015
80
0
0
0
0
0
0
All Adults
Detroit
S75
2015
90
0
0
0
0
0
0
3D-Attachment4-41
-------
Study Group
Study Area
AQ
Scenario
Year
Benchmark
(PPb)
Number of People with 7-hr Exposure at or above Benchmark
>1 Day
> 2 Days
> 3 Days
> 4 Days
> 5 Days
> 6 Days
All Adults
Detroit
S75
2016
0
1181777
1016416
916465
845615
790167
742060
All Adults
Detroit
S75
2016
10
1060591
870783
760739
679992
618580
560510
All Adults
Detroit
S75
2016
20
798950
570931
446665
364541
303326
256529
All Adults
Detroit
S75
2016
30
536129
331574
239685
189546
154022
127806
All Adults
Detroit
S75
2016
40
344879
189939
126495
92544
67573
51581
All Adults
Detroit
S75
2016
50
203309
86711
44109
24119
13764
7275
All Adults
Detroit
S75
2016
60
91627
20252
5702
1573
131
131
All Adults
Detroit
S75
2016
70
21170
1180
0
0
0
0
All Adults
Detroit
S75
2016
80
1049
0
0
0
0
0
All Adults
Detroit
S75
2016
90
0
0
0
0
0
0
All Adults
Detroit
S75
2017
0
1196917
1026247
926100
854791
800588
755102
All Adults
Detroit
S75
2017
10
1069242
878517
765655
685432
621398
565688
All Adults
Detroit
S75
2017
20
785776
561690
438144
353006
290545
243290
All Adults
Detroit
S75
2017
30
526560
327445
241192
188235
151597
124922
All Adults
Detroit
S75
2017
40
338194
187186
126560
89398
67377
49746
All Adults
Detroit
S75
2017
50
189087
73931
37948
19597
10290
5505
All Adults
Detroit
S75
2017
60
62526
10093
1966
328
197
0
All Adults
Detroit
S75
2017
70
6161
131
0
0
0
0
All Adults
Detroit
S75
2017
80
328
0
0
0
0
0
All Adults
Detroit
S75
2017
90
0
0
0
0
0
0
All Adults
Philadelphia
S65
2015
0
1659016
1438982
1297637
1196813
1118472
1057385
All Adults
Philadelphia
S65
2015
10
1497106
1244393
1091719
979829
887632
810337
All Adults
Philadelphia
S65
2015
20
1079694
764064
597709
487039
408175
348918
All Adults
Philadelphia
S65
2015
30
694176
443903
330444
262821
212627
181779
All Adults
Philadelphia
S65
2015
40
406170
219512
142303
96641
67622
48800
All Adults
Philadelphia
S65
2015
50
121128
29280
9150
3137
523
174
All Adults
Philadelphia
S65
2015
60
8714
261
0
0
0
0
All Adults
Philadelphia
S65
2015
70
174
0
0
0
0
0
All Adults
Philadelphia
S65
2015
80
0
0
0
0
0
0
All Adults
Philadelphia
S65
2015
90
0
0
0
0
0
0
All Adults
Philadelphia
S65
2016
0
1662589
1436019
1305567
1206137
1125879
1063572
All Adults
Philadelphia
S65
2016
10
1504338
1256506
1104791
992726
899222
826981
All Adults
Philadelphia
S65
2016
20
1083702
774870
603025
492964
413490
351794
All Adults
Philadelphia
S65
2016
30
688512
441115
329311
263431
218814
182563
All Adults
Philadelphia
S65
2016
40
388568
206440
126356
85138
59693
41480
All Adults
Philadelphia
S65
2016
50
104571
23267
6884
1656
784
174
All Adults
Philadelphia
S65
2016
60
8714
174
0
0
0
0
All Adults
Philadelphia
S65
2016
70
87
0
0
0
0
0
All Adults
Philadelphia
S65
2016
80
0
0
0
0
0
0
All Adults
Philadelphia
S65
2016
90
0
0
0
0
0
0
All Adults
Philadelphia
S65
2017
0
1653788
1437500
1306351
1209797
1131369
1064182
All Adults
Philadelphia
S65
2017
10
1503554
1249883
1099649
985493
896259
815042
All Adults
Philadelphia
S65
2017
20
1059477
755437
585248
474490
396933
338722
All Adults
Philadelphia
S65
2017
30
675353
425952
311708
245654
199382
168272
All Adults
Philadelphia
S65
2017
40
371575
191713
116771
74681
51065
33898
All Adults
Philadelphia
S65
2017
50
88624
16121
3747
523
174
87
All Adults
Philadelphia
S65
2017
60
8976
349
0
0
0
0
3D-Attachment4-42
-------
Study Group
Study Area
AQ
Scenario
Year
Benchmark
(PPb)
Number of People with 7-hr Exposure at or above Benchmark
>1 Day
> 2 Days
> 3 Days
> 4 Days
> 5 Days
> 6 Days
All Adults
Philadelphia
S65
2017
70
174
0
0
0
0
0
All Adults
Philadelphia
S65
2017
80
0
0
0
0
0
0
All Adults
Philadelphia
S65
2017
90
0
0
0
0
0
0
All Adults
Philadelphia
S70
2015
0
1659016
1438982
1297637
1196813
1118472
1057385
All Adults
Philadelphia
S70
2015
10
1501463
1249447
1097906
984621
896172
818528
All Adults
Philadelphia
S70
2015
20
1111849
794390
626466
511351
430135
367828
All Adults
Philadelphia
S70
2015
30
728074
468303
349092
274586
222736
189535
All Adults
Philadelphia
S70
2015
40
455667
260120
175592
123916
90802
68232
All Adults
Philadelphia
S70
2015
50
190842
64224
24836
10719
5316
2440
All Adults
Philadelphia
S70
2015
60
27973
2266
349
87
0
0
All Adults
Philadelphia
S70
2015
70
1481
0
0
0
0
0
All Adults
Philadelphia
S70
2015
80
87
0
0
0
0
0
All Adults
Philadelphia
S70
2015
90
0
0
0
0
0
0
All Adults
Philadelphia
S70
2016
0
1662589
1436019
1305567
1206137
1125879
1063572
All Adults
Philadelphia
S70
2016
10
1509393
1263564
1111675
1000830
907588
835957
All Adults
Philadelphia
S70
2016
20
1115074
809465
632218
516144
433620
369309
All Adults
Philadelphia
S70
2016
30
724763
464730
344648
275806
230056
193717
All Adults
Philadelphia
S70
2016
40
442160
246961
162608
113808
81740
60128
All Adults
Philadelphia
S70
2016
50
176550
54638
21350
8191
4183
1743
All Adults
Philadelphia
S70
2016
60
25968
1830
87
0
0
0
All Adults
Philadelphia
S70
2016
70
1481
0
0
0
0
0
All Adults
Philadelphia
S70
2016
80
0
0
0
0
0
0
All Adults
Philadelphia
S70
2016
90
0
0
0
0
0
0
All Adults
Philadelphia
S70
2017
0
1653788
1437500
1306351
1209797
1131369
1064182
All Adults
Philadelphia
S70
2017
10
1509218
1258423
1106098
991854
905845
823931
All Adults
Philadelphia
S70
2017
20
1093114
783584
613569
497321
418283
355367
All Adults
Philadelphia
S70
2017
30
708119
447040
326610
258028
209926
176812
All Adults
Philadelphia
S70
2017
40
424035
234413
146661
102654
71544
50455
All Adults
Philadelphia
S70
2017
50
145005
40434
12548
4444
2091
959
All Adults
Philadelphia
S70
2017
60
25184
1656
261
0
0
0
All Adults
Philadelphia
S70
2017
70
1830
87
0
0
0
0
All Adults
Philadelphia
S70
2017
80
0
0
0
0
0
0
All Adults
Philadelphia
S70
2017
90
0
0
0
0
0
0
All Adults
Philadelphia
S75
2015
0
1659016
1438982
1297637
1196813
1118472
1057385
All Adults
Philadelphia
S75
2015
10
1500853
1248053
1099475
984447
898612
821665
All Adults
Philadelphia
S75
2015
20
1137295
822362
651999
533660
450177
384733
All Adults
Philadelphia
S75
2015
30
763628
493487
365649
287047
232932
197639
All Adults
Philadelphia
S75
2015
40
501417
291578
203565
148142
112588
84615
All Adults
Philadelphia
S75
2015
50
268573
113546
58734
31458
17254
9934
All Adults
Philadelphia
S75
2015
60
71195
12461
3224
523
0
0
All Adults
Philadelphia
S75
2015
70
8017
349
0
0
0
0
All Adults
Philadelphia
S75
2015
80
523
0
0
0
0
0
All Adults
Philadelphia
S75
2015
90
87
0
0
0
0
0
All Adults
Philadelphia
S75
2016
0
1662589
1436019
1305567
1206137
1125879
1063572
All Adults
Philadelphia
S75
2016
10
1509654
1262344
1111762
1003008
909679
838048
All Adults
Philadelphia
S75
2016
20
1142349
839878
659232
540893
453750
386215
All Adults
Philadelphia
S75
2016
30
763106
488259
362512
287395
239293
199643
3D-Attachment4-43
-------
Study Group
Study Area
AQ
Scenario
Year
Benchmark
(PPb)
Number of People with 7-hr Exposure at or above Benchmark
>1 Day
> 2 Days
> 3 Days
> 4 Days
> 5 Days
> 6 Days
All Adults
Philadelphia
S75
2016
40
492877
281906
190232
136378
102392
77731
All Adults
Philadelphia
S75
2016
50
259946
103699
50891
25968
13768
7756
All Adults
Philadelphia
S75
2016
60
62742
8889
1830
436
261
0
All Adults
Philadelphia
S75
2016
70
6884
87
0
0
0
0
All Adults
Philadelphia
S75
2016
80
87
0
0
0
0
0
All Adults
Philadelphia
S75
2016
90
0
0
0
0
0
0
All Adults
Philadelphia
S75
2017
0
1653788
1437500
1306351
1209797
1131369
1064182
All Adults
Philadelphia
S75
2017
10
1509480
1260253
1106621
995078
904712
825848
All Adults
Philadelphia
S75
2017
20
1119343
810947
639538
519107
436932
370529
All Adults
Philadelphia
S75
2017
30
742191
470394
339245
269008
218727
183086
All Adults
Philadelphia
S75
2017
40
469174
266481
173239
125485
91848
68407
All Adults
Philadelphia
S75
2017
50
215329
78341
32766
14814
7407
4270
All Adults
Philadelphia
S75
2017
60
57340
7494
871
87
0
0
All Adults
Philadelphia
S75
2017
70
8627
261
0
0
0
0
All Adults
Philadelphia
S75
2017
80
261
0
0
0
0
0
All Adults
Philadelphia
S75
2017
90
0
0
0
0
0
0
All Adults
Phoenix
S65
2015
0
1055538
935690
867248
816686
771836
738559
All Adults
Phoenix
S65
2015
10
968570
838738
758227
701705
654670
614787
All Adults
Phoenix
S65
2015
20
749734
578529
479840
411100
357756
315539
All Adults
Phoenix
S65
2015
30
520915
360886
287079
241633
206816
181635
All Adults
Phoenix
S65
2015
40
347724
219531
162215
123623
98988
80611
All Adults
Phoenix
S65
2015
50
144881
57168
27516
14106
7251
3725
All Adults
Phoenix
S65
2015
60
7947
546
50
0
0
0
All Adults
Phoenix
S65
2015
70
0
0
0
0
0
0
All Adults
Phoenix
S65
2015
80
0
0
0
0
0
0
All Adults
Phoenix
S65
2015
90
0
0
0
0
0
0
All Adults
Phoenix
S65
2016
0
1041681
923074
850410
799650
758078
722913
All Adults
Phoenix
S65
2016
10
962858
825875
749386
691921
644041
606989
All Adults
Phoenix
S65
2016
20
741439
571079
474873
405735
356217
315638
All Adults
Phoenix
S65
2016
30
515302
360339
288321
239498
206369
180194
All Adults
Phoenix
S65
2016
40
345141
218985
161668
124616
100974
82051
All Adults
Phoenix
S65
2016
50
115179
41522
17930
8096
3775
1738
All Adults
Phoenix
S65
2016
60
4818
199
0
0
0
0
All Adults
Phoenix
S65
2016
70
0
0
0
0
0
0
All Adults
Phoenix
S65
2016
80
0
0
0
0
0
0
All Adults
Phoenix
S65
2016
90
0
0
0
0
0
0
All Adults
Phoenix
S65
2017
0
1040042
922379
855427
801637
762201
726489
All Adults
Phoenix
S65
2017
10
957742
828358
756290
699172
655067
616724
All Adults
Phoenix
S65
2017
20
755098
587271
492555
421728
371713
326366
All Adults
Phoenix
S65
2017
30
528911
371862
296119
245358
210939
185906
All Adults
Phoenix
S65
2017
40
361978
234630
176668
138821
114136
94418
All Adults
Phoenix
S65
2017
50
148457
63277
31887
17483
10579
6457
All Adults
Phoenix
S65
2017
60
12814
745
99
0
0
0
All Adults
Phoenix
S65
2017
70
0
0
0
0
0
0
All Adults
Phoenix
S65
2017
80
0
0
0
0
0
0
All Adults
Phoenix
S65
2017
90
0
0
0
0
0
0
All Adults
Phoenix
S70
2015
0
1055538
935690
867248
816686
771836
738559
3D-Attachment4-44
-------
Study Group
Study Area
AQ
Scenario
Year
Benchmark
(PPb)
Number of People with 7-hr Exposure at or above Benchmark
>1 Day
> 2 Days
> 3 Days
> 4 Days
> 5 Days
> 6 Days
All Adults
Phoenix
S70
2015
10
970308
839533
760810
701954
656359
616773
All Adults
Phoenix
S70
2015
20
767267
601079
501197
431116
378220
334115
All Adults
Phoenix
S70
2015
30
547338
382243
301086
250872
214862
186850
All Adults
Phoenix
S70
2015
40
383137
247196
185757
146222
118060
97994
All Adults
Phoenix
S70
2015
50
216402
106637
64717
40827
25678
17036
All Adults
Phoenix
S70
2015
60
45893
9040
2285
894
497
199
All Adults
Phoenix
S70
2015
70
298
50
0
0
0
0
All Adults
Phoenix
S70
2015
80
0
0
0
0
0
0
All Adults
Phoenix
S70
2015
90
0
0
0
0
0
0
All Adults
Phoenix
S70
2016
0
1041681
923074
850410
799650
758078
722913
All Adults
Phoenix
S70
2016
10
963603
827961
750876
693361
645730
608926
All Adults
Phoenix
S70
2016
20
761456
592983
498167
428036
375786
333121
All Adults
Phoenix
S70
2016
30
542669
378170
300639
249233
215856
188340
All Adults
Phoenix
S70
2016
40
380107
248835
186502
149649
121736
101272
All Adults
Phoenix
S70
2016
50
189482
91736
53244
31042
19619
12020
All Adults
Phoenix
S70
2016
60
33178
4718
1043
199
0
0
All Adults
Phoenix
S70
2016
70
149
0
0
0
0
0
All Adults
Phoenix
S70
2016
80
0
0
0
0
0
0
All Adults
Phoenix
S70
2016
90
0
0
0
0
0
0
All Adults
Phoenix
S70
2017
0
1040042
922379
855427
801637
762201
726489
All Adults
Phoenix
S70
2017
10
959928
829848
757333
702152
657054
619207
All Adults
Phoenix
S70
2017
20
775710
611856
515501
446314
393070
347227
All Adults
Phoenix
S70
2017
30
560301
393865
313503
258024
221319
194846
All Adults
Phoenix
S70
2017
40
397640
264630
201552
161370
135642
113441
All Adults
Phoenix
S70
2017
50
221071
118159
75942
50711
33923
23691
All Adults
Phoenix
S70
2017
60
54585
11324
2881
844
348
199
All Adults
Phoenix
S70
2017
70
1788
50
0
0
0
0
All Adults
Phoenix
S70
2017
80
0
0
0
0
0
0
All Adults
Phoenix
S70
2017
90
0
0
0
0
0
0
All Adults
Phoenix
S75
2015
0
1055538
935690
867248
816686
771836
738559
All Adults
Phoenix
S75
2015
10
965242
833772
753906
695993
648859
610168
All Adults
Phoenix
S75
2015
20
771439
609274
508597
439609
385272
340472
All Adults
Phoenix
S75
2015
30
561891
391729
309181
256534
217495
188886
All Adults
Phoenix
S75
2015
40
401315
261500
198472
154417
126404
105296
All Adults
Phoenix
S75
2015
50
257726
139815
91786
60843
43310
30595
All Adults
Phoenix
S75
2015
60
101769
30148
11175
4619
2334
993
All Adults
Phoenix
S75
2015
70
6854
397
50
0
0
0
All Adults
Phoenix
S75
2015
80
50
0
0
0
0
0
All Adults
Phoenix
S75
2015
90
0
0
0
0
0
0
All Adults
Phoenix
S75
2016
0
1041681
923074
850410
799650
758078
722913
All Adults
Phoenix
S75
2016
10
959977
823391
745810
687053
640117
602171
All Adults
Phoenix
S75
2016
20
769353
602022
505319
436132
381299
338734
All Adults
Phoenix
S75
2016
30
557967
387656
308287
255044
218637
189780
All Adults
Phoenix
S75
2016
40
398335
261252
197926
158937
131123
110014
All Adults
Phoenix
S75
2016
50
237064
127050
81852
55081
38244
26175
All Adults
Phoenix
S75
2016
60
78276
19470
6308
2086
894
199
All Adults
Phoenix
S75
2016
70
4718
99
50
0
0
0
3D-Attachment4-45
-------
Study Group
Study Area
AQ
Scenario
Year
Benchmark
(PPb)
Number of People with 7-hr Exposure at or above Benchmark
>1 Day
> 2 Days
> 3 Days
> 4 Days
> 5 Days
> 6 Days
All Adults
Phoenix
S75
2016
80
0
0
0
0
0
0
All Adults
Phoenix
S75
2016
90
0
0
0
0
0
0
All Adults
Phoenix
S75
2017
0
1040042
922379
855427
801637
762201
726489
All Adults
Phoenix
S75
2017
10
956451
826818
752615
698080
652137
614141
All Adults
Phoenix
S75
2017
20
783309
621244
524590
455651
400073
356713
All Adults
Phoenix
S75
2017
30
576294
406331
322890
264431
225640
197975
All Adults
Phoenix
S75
2017
40
418649
278139
212280
171503
143490
121438
All Adults
Phoenix
S75
2017
50
268007
153970
103358
74502
55777
40728
All Adults
Phoenix
S75
2017
60
105991
35711
13708
6010
3179
1192
All Adults
Phoenix
S75
2017
70
12318
993
99
50
0
0
All Adults
Phoenix
S75
2017
80
397
0
0
0
0
0
All Adults
Phoenix
S75
2017
90
0
0
0
0
0
0
All Adults
Sacramento
S65
2015
0
580206
508974
467127
437713
414074
393207
All Adults
Sacramento
S65
2015
10
515920
432254
384489
350559
324776
301709
All Adults
Sacramento
S65
2015
20
358935
262777
211525
178367
153584
134604
All Adults
Sacramento
S65
2015
30
232591
155328
120626
96844
81094
68917
All Adults
Sacramento
S65
2015
40
115252
58341
34759
22239
14607
10205
All Adults
Sacramento
S65
2015
50
19323
2658
657
57
29
0
All Adults
Sacramento
S65
2015
60
715
29
0
0
0
0
All Adults
Sacramento
S65
2015
70
0
0
0
0
0
0
All Adults
Sacramento
S65
2015
80
0
0
0
0
0
0
All Adults
Sacramento
S65
2015
90
0
0
0
0
0
0
All Adults
Sacramento
S65
2016
0
579921
510346
467641
437456
414474
392607
All Adults
Sacramento
S65
2016
10
512176
432311
385690
351274
324490
300336
All Adults
Sacramento
S65
2016
20
360993
263348
210753
176452
152241
132975
All Adults
Sacramento
S65
2016
30
232992
155528
118797
95415
78750
66516
All Adults
Sacramento
S65
2016
40
112337
55111
31614
20266
12949
8918
All Adults
Sacramento
S65
2016
50
24668
5031
1286
457
114
57
All Adults
Sacramento
S65
2016
60
1172
86
0
0
0
0
All Adults
Sacramento
S65
2016
70
0
0
0
0
0
0
All Adults
Sacramento
S65
2016
80
0
0
0
0
0
0
All Adults
Sacramento
S65
2016
90
0
0
0
0
0
0
All Adults
Sacramento
S65
2017
0
578434
512118
468527
437942
412988
393893
All Adults
Sacramento
S65
2017
10
513891
434740
389434
355247
328549
305825
All Adults
Sacramento
S65
2017
20
359678
263720
211039
176537
151040
131860
All Adults
Sacramento
S65
2017
30
232791
157100
120941
98673
83381
71976
All Adults
Sacramento
S65
2017
40
114824
57426
35216
21981
14549
9747
All Adults
Sacramento
S65
2017
50
19094
2801
515
114
29
0
All Adults
Sacramento
S65
2017
60
1315
29
0
0
0
0
All Adults
Sacramento
S65
2017
70
0
0
0
0
0
0
All Adults
Sacramento
S65
2017
80
0
0
0
0
0
0
All Adults
Sacramento
S65
2017
90
0
0
0
0
0
0
All Adults
Sacramento
S70
2015
0
580206
508974
467127
437713
414074
393207
All Adults
Sacramento
S70
2015
10
518121
435255
388176
354504
328406
305567
All Adults
Sacramento
S70
2015
20
375142
279470
226074
191287
163760
144094
All Adults
Sacramento
S70
2015
30
249713
167419
130173
105505
89241
75692
All Adults
Sacramento
S70
2015
40
147410
83810
56368
39132
28785
21181
3D-Attachment4-46
-------
Study Group
Study Area
AQ
Scenario
Year
Benchmark
(PPb)
Number of People with 7-hr Exposure at or above Benchmark
>1 Day
> 2 Days
> 3 Days
> 4 Days
> 5 Days
> 6 Days
All Adults
Sacramento
S70
2015
50
49280
14292
5202
2115
715
343
All Adults
Sacramento
S70
2015
60
4688
257
0
0
0
0
All Adults
Sacramento
S70
2015
70
372
0
0
0
0
0
All Adults
Sacramento
S70
2015
80
0
0
0
0
0
0
All Adults
Sacramento
S70
2015
90
0
0
0
0
0
0
All Adults
Sacramento
S70
2016
0
579921
510346
467641
437456
414474
392607
All Adults
Sacramento
S70
2016
10
514834
435369
388920
354304
328206
303538
All Adults
Sacramento
S70
2016
20
376342
277155
225017
187885
162645
141836
All Adults
Sacramento
S70
2016
30
250628
168248
128115
104390
86011
72662
All Adults
Sacramento
S70
2016
40
143837
78493
49851
34530
24926
18237
All Adults
Sacramento
S70
2016
50
56654
18551
8147
3544
1744
772
All Adults
Sacramento
S70
2016
60
9176
972
143
0
0
0
All Adults
Sacramento
S70
2016
70
600
29
0
0
0
0
All Adults
Sacramento
S70
2016
80
0
0
0
0
0
0
All Adults
Sacramento
S70
2016
90
0
0
0
0
0
0
All Adults
Sacramento
S70
2017
0
578434
512118
468527
437942
412988
393893
All Adults
Sacramento
S70
2017
10
516921
437570
392864
359449
332151
310198
All Adults
Sacramento
S70
2017
20
375885
279556
227303
188771
160644
140864
All Adults
Sacramento
S70
2017
30
249085
168105
130745
106105
89755
77550
All Adults
Sacramento
S70
2017
40
146438
83609
56140
40161
28785
21781
All Adults
Sacramento
S70
2017
50
54253
16150
6374
3087
1601
772
All Adults
Sacramento
S70
2017
60
8089
486
0
0
0
0
All Adults
Sacramento
S70
2017
70
229
0
0
0
0
0
All Adults
Sacramento
S70
2017
80
0
0
0
0
0
0
All Adults
Sacramento
S70
2017
90
0
0
0
0
0
0
All Adults
Sacramento
S75
2015
0
580206
508974
467127
437713
414074
393207
All Adults
Sacramento
S75
2015
10
519550
436884
389891
356533
330064
307368
All Adults
Sacramento
S75
2015
20
385747
289703
236507
200491
172364
151555
All Adults
Sacramento
S75
2015
30
264034
177166
136891
110622
94128
80036
All Adults
Sacramento
S75
2015
40
167848
98588
68946
50566
38818
30299
All Adults
Sacramento
S75
2015
50
75006
28384
12806
6632
3487
1601
All Adults
Sacramento
S75
2015
60
15779
1887
314
57
0
0
All Adults
Sacramento
S75
2015
70
1515
57
0
0
0
0
All Adults
Sacramento
S75
2015
80
114
0
0
0
0
0
All Adults
Sacramento
S75
2015
90
0
0
0
0
0
0
All Adults
Sacramento
S75
2016
0
579921
510346
467641
437456
414474
392607
All Adults
Sacramento
S75
2016
10
516206
436941
390206
355504
329578
305453
All Adults
Sacramento
S75
2016
20
386347
288731
233849
196203
169477
148325
All Adults
Sacramento
S75
2016
30
264663
176480
134947
109821
90270
77264
All Adults
Sacramento
S75
2016
40
165447
93385
62714
44677
33444
25554
All Adults
Sacramento
S75
2016
50
81351
32872
16865
9804
5202
2830
All Adults
Sacramento
S75
2016
60
22382
4259
943
372
86
57
All Adults
Sacramento
S75
2016
70
2687
114
29
0
0
0
All Adults
Sacramento
S75
2016
80
200
0
0
0
0
0
All Adults
Sacramento
S75
2016
90
0
0
0
0
0
0
All Adults
Sacramento
S75
2017
0
578434
512118
468527
437942
412988
393893
All Adults
Sacramento
S75
2017
10
517978
439200
394779
361679
334238
312056
3D-Attachment4-47
-------
Study Group
Study Area
AQ
Scenario
Year
Benchmark
(PPb)
Number of People with 7-hr Exposure at or above Benchmark
>1 Day
> 2 Days
> 3 Days
> 4 Days
> 5 Days
> 6 Days
All Adults
Sacramento
S75
2017
20
387948
290532
236850
198519
169220
147467
All Adults
Sacramento
S75
2017
30
262005
177138
136891
111622
94128
81294
All Adults
Sacramento
S75
2017
40
165447
99245
69403
51795
39446
30585
All Adults
Sacramento
S75
2017
50
80837
32186
15579
8918
5174
3402
All Adults
Sacramento
S75
2017
60
18008
2172
457
57
0
0
All Adults
Sacramento
S75
2017
70
2944
114
0
0
0
0
All Adults
Sacramento
S75
2017
80
0
0
0
0
0
0
All Adults
Sacramento
S75
2017
90
0
0
0
0
0
0
All Adults
St. Louis
S65
2015
0
677754
588086
534292
493660
463115
438972
All Adults
St. Louis
S65
2015
10
598923
495985
435109
388898
350841
319652
All Adults
St. Louis
S65
2015
20
414328
288034
224154
179588
148935
125865
All Adults
St. Louis
S65
2015
30
257882
161167
117531
90849
73538
60983
All Adults
St. Louis
S65
2015
40
143999
72608
44244
27577
18778
13305
All Adults
St. Louis
S65
2015
50
42849
9121
2361
715
179
0
All Adults
St. Louis
S65
2015
60
1931
0
0
0
0
0
All Adults
St. Louis
S65
2015
70
0
0
0
0
0
0
All Adults
St. Louis
S65
2015
80
0
0
0
0
0
0
All Adults
St. Louis
S65
2015
90
0
0
0
0
0
0
All Adults
St. Louis
S65
2016
0
676395
585725
533433
493839
461291
436754
All Adults
St. Louis
S65
2016
10
603394
501100
442298
398018
361643
330025
All Adults
St. Louis
S65
2016
20
433285
304523
237852
194216
161847
138992
All Adults
St. Louis
S65
2016
30
280737
177871
130622
104011
85913
70748
All Adults
St. Louis
S65
2016
40
168965
89061
56298
38629
27434
19600
All Adults
St. Louis
S65
2016
50
65740
20352
6367
2146
1001
465
All Adults
St. Louis
S65
2016
60
9049
429
72
0
0
0
All Adults
St. Louis
S65
2016
70
107
0
0
0
0
0
All Adults
St. Louis
S65
2016
80
0
0
0
0
0
0
All Adults
St. Louis
S65
2016
90
0
0
0
0
0
0
All Adults
St. Louis
S65
2017
0
675465
586905
532217
492587
461505
436397
All Adults
St. Louis
S65
2017
10
608080
508754
448021
403062
368117
337643
All Adults
St. Louis
S65
2017
20
442334
318293
249978
202979
169895
145823
All Adults
St. Louis
S65
2017
30
288499
185096
138527
110378
90777
76399
All Adults
St. Louis
S65
2017
40
180196
101400
67743
45997
32942
23893
All Adults
St. Louis
S65
2017
50
60554
18134
6688
2253
1288
501
All Adults
St. Louis
S65
2017
60
2897
179
0
0
0
0
All Adults
St. Louis
S65
2017
70
36
0
0
0
0
0
All Adults
St. Louis
S65
2017
80
0
0
0
0
0
0
All Adults
St. Louis
S65
2017
90
0
0
0
0
0
0
All Adults
St. Louis
S70
2015
0
677754
588086
534292
493660
463115
438972
All Adults
St. Louis
S70
2015
10
600676
498131
438686
391652
355062
322943
All Adults
St. Louis
S70
2015
20
429887
303736
235921
191248
158556
133662
All Adults
St. Louis
S70
2015
30
273799
170431
124470
96607
78616
64488
All Adults
St. Louis
S70
2015
40
168249
89847
59016
39308
27076
19922
All Adults
St. Louis
S70
2015
50
72965
23106
8906
4149
1896
465
All Adults
St. Louis
S70
2015
60
11016
858
179
0
0
0
All Adults
St. Louis
S70
2015
70
72
0
0
0
0
0
All Adults
St. Louis
S70
2015
80
0
0
0
0
0
0
3D-Attachment4-48
-------
Study Group
Study Area
AQ
Scenario
Year
Benchmark
(PPb)
Number of People with 7-hr Exposure at or above Benchmark
>1 Day
> 2 Days
> 3 Days
> 4 Days
> 5 Days
> 6 Days
All Adults
St. Louis
S70
2015
90
0
0
0
0
0
0
All Adults
St. Louis
S70
2016
0
676395
585725
533433
493839
461291
436754
All Adults
St. Louis
S70
2016
10
605755
504140
444909
402167
365184
333530
All Adults
St. Louis
S70
2016
20
448164
321262
251086
208094
173757
148649
All Adults
St. Louis
S70
2016
30
298478
189638
139099
110664
92029
75684
All Adults
St. Louis
S70
2016
40
192857
106944
71785
51898
38164
28471
All Adults
St. Louis
S70
2016
50
96715
37591
17490
8119
4006
2289
All Adults
St. Louis
S70
2016
60
28185
4435
572
143
0
0
All Adults
St. Louis
S70
2016
70
2075
0
0
0
0
0
All Adults
St. Louis
S70
2016
80
0
0
0
0
0
0
All Adults
St. Louis
S70
2016
90
0
0
0
0
0
0
All Adults
St. Louis
S70
2017
0
675465
586905
532217
492587
461505
436397
All Adults
St. Louis
S70
2017
10
610583
511901
452420
406996
372302
342901
All Adults
St. Louis
S70
2017
20
457177
335819
266037
216356
182055
155910
All Adults
St. Louis
S70
2017
30
307134
197900
147504
117281
96464
80655
All Adults
St. Louis
S70
2017
40
203945
120107
82658
60018
44781
33872
All Adults
St. Louis
S70
2017
50
100220
39952
19279
9764
4936
2611
All Adults
St. Louis
S70
2017
60
17741
2468
393
36
36
0
All Adults
St. Louis
S70
2017
70
680
0
0
0
0
0
All Adults
St. Louis
S70
2017
80
0
0
0
0
0
0
All Adults
St. Louis
S70
2017
90
0
0
0
0
0
0
All Adults
St. Louis
S75
2015
0
677754
588086
534292
493660
463115
438972
All Adults
St. Louis
S75
2015
10
600712
498989
439294
391866
355026
323944
All Adults
St. Louis
S75
2015
20
438292
312928
243683
196291
164208
137668
All Adults
St. Louis
S75
2015
30
285852
177120
129370
100041
81514
66956
All Adults
St. Louis
S75
2015
40
181912
100184
66313
45889
32942
24429
All Adults
St. Louis
S75
2015
50
93102
35445
17347
8548
4650
2253
All Adults
St. Louis
S75
2015
60
25967
3577
644
107
0
0
All Adults
St. Louis
S75
2015
70
1753
36
0
0
0
0
All Adults
St. Louis
S75
2015
80
0
0
0
0
0
0
All Adults
St. Louis
S75
2015
90
0
0
0
0
0
0
All Adults
St. Louis
S75
2016
0
676395
585725
533433
493839
461291
436754
All Adults
St. Louis
S75
2016
10
606434
504891
446304
401774
366257
334424
All Adults
St. Louis
S75
2016
20
457893
329739
260314
215641
180625
153871
All Adults
St. Louis
S75
2016
30
311605
197292
144321
115528
95177
78795
All Adults
St. Louis
S75
2016
40
206556
116530
80119
58623
44351
33836
All Adults
St. Louis
S75
2016
50
117424
50968
27398
14879
8370
4900
All Adults
St. Louis
S75
2016
60
45961
10802
2361
644
143
0
All Adults
St. Louis
S75
2016
70
8226
179
36
0
0
0
All Adults
St. Louis
S75
2016
80
250
0
0
0
0
0
All Adults
St. Louis
S75
2016
90
0
0
0
0
0
0
All Adults
St. Louis
S75
2017
0
675465
586905
532217
492587
461505
436397
All Adults
St. Louis
S75
2017
10
611585
513117
453386
408892
373053
345154
All Adults
St. Louis
S75
2017
20
467443
346549
276481
226800
190461
163170
All Adults
St. Louis
S75
2017
30
320403
207915
153084
121716
100220
83552
All Adults
St. Louis
S75
2017
40
217930
129585
91528
67529
51111
40024
All Adults
St. Louis
S75
2017
50
124148
56584
30116
17276
10194
6295
3D-Attachment4-49
-------
Study Group
Study Area
AQ
Scenario
Year
Benchmark
(PPb)
Number of People with 7-hr Exposure at or above Benchmark
>1 Day
> 2 Days
> 3 Days
> 4 Days
> 5 Days
> 6 Days
All Adults
St. Louis
S75
2017
60
37270
7583
1538
429
250
107
All Adults
St. Louis
S75
2017
70
3291
215
0
0
0
0
All Adults
St. Louis
S75
2017
80
179
0
0
0
0
0
All Adults
St. Louis
S75
2017
90
0
0
0
0
0
0
Asthma Adults
Atlanta
S65
2015
0
99029
84802
77054
70433
65362
60221
Asthma Adults
Atlanta
S65
2015
10
84309
67898
58037
49867
44091
39936
Asthma Adults
Atlanta
S65
2015
20
55149
38738
29371
23525
20144
16974
Asthma Adults
Atlanta
S65
2015
30
35146
22257
15284
11692
9297
7325
Asthma Adults
Atlanta
S65
2015
40
18665
9579
5353
3592
2254
1761
Asthma Adults
Atlanta
S65
2015
50
5423
1197
282
70
0
0
Asthma Adults
Atlanta
S65
2015
60
423
0
0
0
0
0
Asthma Adults
Atlanta
S65
2015
70
0
0
0
0
0
0
Asthma Adults
Atlanta
S65
2015
80
0
0
0
0
0
0
Asthma Adults
Atlanta
S65
2016
0
99029
86211
75998
69236
64306
60643
Asthma Adults
Atlanta
S65
2016
10
86985
72476
62474
55079
49867
45570
Asthma Adults
Atlanta
S65
2016
20
61700
41767
33033
25567
20637
17538
Asthma Adults
Atlanta
S65
2016
30
37964
23595
17397
13101
10635
8241
Asthma Adults
Atlanta
S65
2016
40
21905
11762
6269
4226
2747
1902
Asthma Adults
Atlanta
S65
2016
50
4860
845
70
0
0
0
Asthma Adults
Atlanta
S65
2016
60
211
0
0
0
0
0
Asthma Adults
Atlanta
S65
2016
70
0
0
0
0
0
0
Asthma Adults
Atlanta
S65
2016
80
0
0
0
0
0
0
Asthma Adults
Atlanta
S65
2017
0
103255
88535
79519
72194
67194
62615
Asthma Adults
Atlanta
S65
2017
10
86704
69940
59939
52543
47050
43035
Asthma Adults
Atlanta
S65
2017
20
57262
37752
28807
22680
19087
16622
Asthma Adults
Atlanta
S65
2017
30
34372
20496
15707
11833
9297
7889
Asthma Adults
Atlanta
S65
2017
40
16411
8100
5212
3240
1690
1338
Asthma Adults
Atlanta
S65
2017
50
2676
352
70
0
0
0
Asthma Adults
Atlanta
S65
2017
60
70
0
0
0
0
0
Asthma Adults
Atlanta
S65
2017
70
0
0
0
0
0
0
Asthma Adults
Atlanta
S65
2017
80
0
0
0
0
0
0
Asthma Adults
Atlanta
S70
2015
0
99029
84802
77054
70433
65362
60221
Asthma Adults
Atlanta
S70
2015
10
85365
69166
58953
50994
44796
40851
Asthma Adults
Atlanta
S70
2015
20
57967
41344
31061
25286
21553
18031
Asthma Adults
Atlanta
S70
2015
30
38245
24159
16693
13030
10706
8170
Asthma Adults
Atlanta
S70
2015
40
22891
12326
7748
4930
3451
2606
Asthma Adults
Atlanta
S70
2015
50
8804
3310
1479
634
282
70
Asthma Adults
Atlanta
S70
2015
60
1268
70
0
0
0
0
Asthma Adults
Atlanta
S70
2015
70
70
0
0
0
0
0
Asthma Adults
Atlanta
S70
2015
80
0
0
0
0
0
0
Asthma Adults
Atlanta
S70
2016
0
99029
86211
75998
69236
64306
60643
Asthma Adults
Atlanta
S70
2016
10
88394
73040
63601
56347
50853
46627
Asthma Adults
Atlanta
S70
2016
20
64588
45782
35710
28526
23313
19651
Asthma Adults
Atlanta
S70
2016
30
40922
25567
18947
14861
11762
9720
Asthma Adults
Atlanta
S70
2016
40
25708
14580
8804
6550
4719
3310
Asthma Adults
Atlanta
S70
2016
50
10072
3029
845
211
70
0
Asthma Adults
Atlanta
S70
2016
60
2113
211
0
0
0
0
3D-Attachment4-50
-------
Study Group
Study Area
AQ
Scenario
Year
Benchmark
(PPb)
Number of People with 7-hr Exposure at or above Benchmark
>1 Day
> 2 Days
> 3 Days
> 4 Days
> 5 Days
> 6 Days
Asthma Adults
Atlanta
S70
2016
70
141
0
0
0
0
0
Asthma Adults
Atlanta
S70
2016
80
0
0
0
0
0
0
Asthma Adults
Atlanta
S70
2017
0
103255
88535
79519
72194
67194
62615
Asthma Adults
Atlanta
S70
2017
10
88042
71208
60855
54304
48599
44303
Asthma Adults
Atlanta
S70
2017
20
59516
40570
30779
25004
20637
17890
Asthma Adults
Atlanta
S70
2017
30
37259
22609
16622
13594
11058
8875
Asthma Adults
Atlanta
S70
2017
40
21130
11340
7466
5212
3944
2747
Asthma Adults
Atlanta
S70
2017
50
6480
1338
563
70
0
0
Asthma Adults
Atlanta
S70
2017
60
775
0
0
0
0
0
Asthma Adults
Atlanta
S70
2017
70
0
0
0
0
0
0
Asthma Adults
Atlanta
S70
2017
80
0
0
0
0
0
0
Asthma Adults
Atlanta
S75
2015
0
99029
84802
77054
70433
65362
60221
Asthma Adults
Atlanta
S75
2015
10
86281
69729
59657
51909
46204
41344
Asthma Adults
Atlanta
S75
2015
20
59868
42964
33526
26624
22539
19228
Asthma Adults
Atlanta
S75
2015
30
41415
26060
18594
14157
11692
9368
Asthma Adults
Atlanta
S75
2015
40
26835
15284
9931
6902
4789
3451
Asthma Adults
Atlanta
S75
2015
50
12889
5564
2888
1690
916
352
Asthma Adults
Atlanta
S75
2015
60
3310
775
70
0
0
0
Asthma Adults
Atlanta
S75
2015
70
423
0
0
0
0
0
Asthma Adults
Atlanta
S75
2015
80
0
0
0
0
0
0
Asthma Adults
Atlanta
S75
2016
0
99029
86211
75998
69236
64306
60643
Asthma Adults
Atlanta
S75
2016
10
89028
73744
63954
56629
51487
47754
Asthma Adults
Atlanta
S75
2016
20
66630
48529
38457
30498
25074
21553
Asthma Adults
Atlanta
S75
2016
30
44655
27187
20567
15848
12889
11128
Asthma Adults
Atlanta
S75
2016
40
29159
16904
10635
7959
6339
4930
Asthma Adults
Atlanta
S75
2016
50
15566
5987
2817
986
282
141
Asthma Adults
Atlanta
S75
2016
60
4156
704
70
0
0
0
Asthma Adults
Atlanta
S75
2016
70
1057
70
0
0
0
0
Asthma Adults
Atlanta
S75
2016
80
70
0
0
0
0
0
Asthma Adults
Atlanta
S75
2017
0
103255
88535
79519
72194
67194
62615
Asthma Adults
Atlanta
S75
2017
10
88605
71842
61348
54727
49163
45077
Asthma Adults
Atlanta
S75
2017
20
61911
42894
32681
26624
22327
19017
Asthma Adults
Atlanta
S75
2017
30
40147
24088
17890
14791
12115
9931
Asthma Adults
Atlanta
S75
2017
40
24863
13735
9861
6550
5423
3874
Asthma Adults
Atlanta
S75
2017
50
10072
3310
1479
704
563
141
Asthma Adults
Atlanta
S75
2017
60
2254
70
0
0
0
0
Asthma Adults
Atlanta
S75
2017
70
211
0
0
0
0
0
Asthma Adults
Atlanta
S75
2017
80
0
0
0
0
0
0
Asthma Adults
Boston
S65
2015
0
171995
144601
128654
117011
105564
99107
Asthma Adults
Boston
S65
2015
10
152525
122783
105564
90498
80225
72692
Asthma Adults
Boston
S65
2015
20
100673
67506
50679
38449
32384
25535
Asthma Adults
Boston
S65
2015
30
58310
33949
24459
19078
13990
11251
Asthma Adults
Boston
S65
2015
40
27296
11642
6164
3424
2152
1468
Asthma Adults
Boston
S65
2015
50
4990
1076
0
0
0
0
Asthma Adults
Boston
S65
2015
60
489
0
0
0
0
0
Asthma Adults
Boston
S65
2015
70
196
0
0
0
0
0
Asthma Adults
Boston
S65
2015
80
0
0
0
0
0
0
3D-Attachment4-51
-------
Study Group
Study Area
AQ
Scenario
Year
Benchmark
(PPb)
Number of People with 7-hr Exposure at or above Benchmark
>1 Day
> 2 Days
> 3 Days
> 4 Days
> 5 Days
> 6 Days
Asthma Adults
Boston
S65
2016
0
177180
150275
133448
122588
112902
105369
Asthma Adults
Boston
S65
2016
10
157711
127382
109674
97444
86878
79247
Asthma Adults
Boston
S65
2016
20
104097
69952
52048
42265
35123
29546
Asthma Adults
Boston
S65
2016
30
62125
39428
28764
21915
17610
14675
Asthma Adults
Boston
S65
2016
40
34145
16632
9588
6457
4011
2739
Asthma Adults
Boston
S65
2016
50
8414
1565
294
196
98
0
Asthma Adults
Boston
S65
2016
60
196
0
0
0
0
0
Asthma Adults
Boston
S65
2016
70
0
0
0
0
0
0
Asthma Adults
Boston
S65
2016
80
0
0
0
0
0
0
Asthma Adults
Boston
S65
2017
0
174636
148416
131589
120925
110750
103216
Asthma Adults
Boston
S65
2017
10
156145
127284
109086
97151
87171
77486
Asthma Adults
Boston
S65
2017
20
105369
71029
54103
44613
36199
29448
Asthma Adults
Boston
S65
2017
30
62713
39526
28372
21719
15947
12621
Asthma Adults
Boston
S65
2017
40
33949
17415
9392
5479
3913
2250
Asthma Adults
Boston
S65
2017
50
11251
2348
489
294
0
0
Asthma Adults
Boston
S65
2017
60
1565
98
0
0
0
0
Asthma Adults
Boston
S65
2017
70
98
0
0
0
0
0
Asthma Adults
Boston
S65
2017
80
0
0
0
0
0
0
Asthma Adults
Boston
S70
2015
0
171995
144601
128654
117011
105564
99107
Asthma Adults
Boston
S70
2015
10
152428
122979
105662
90987
80421
72594
Asthma Adults
Boston
S70
2015
20
103314
69757
51951
39526
33362
26318
Asthma Adults
Boston
S70
2015
30
61343
35319
25731
19763
14675
11447
Asthma Adults
Boston
S70
2015
40
31014
14871
8120
5185
3131
1859
Asthma Adults
Boston
S70
2015
50
8512
1957
685
0
0
0
Asthma Adults
Boston
S70
2015
60
1370
0
0
0
0
0
Asthma Adults
Boston
S70
2015
70
196
0
0
0
0
0
Asthma Adults
Boston
S70
2015
80
98
0
0
0
0
0
Asthma Adults
Boston
S70
2016
0
177180
150275
133448
122588
112902
105369
Asthma Adults
Boston
S70
2016
10
158200
128751
110163
97346
86780
79345
Asthma Adults
Boston
S70
2016
20
107228
71713
53516
43830
36395
30622
Asthma Adults
Boston
S70
2016
30
65550
40895
29448
23187
18491
14969
Asthma Adults
Boston
S70
2016
40
39036
20154
11740
8218
5283
3718
Asthma Adults
Boston
S70
2016
50
12621
3816
1663
587
196
98
Asthma Adults
Boston
S70
2016
60
2055
489
0
0
0
0
Asthma Adults
Boston
S70
2016
70
0
0
0
0
0
0
Asthma Adults
Boston
S70
2016
80
0
0
0
0
0
0
Asthma Adults
Boston
S70
2017
0
174636
148416
131589
120925
110750
103216
Asthma Adults
Boston
S70
2017
10
155950
126990
109282
97444
87269
77583
Asthma Adults
Boston
S70
2017
20
108010
73474
56255
45493
36982
30916
Asthma Adults
Boston
S70
2017
30
65843
40700
29351
22698
16730
13697
Asthma Adults
Boston
S70
2017
40
38743
20643
11447
7631
5381
3522
Asthma Adults
Boston
S70
2017
50
16241
4403
783
294
98
0
Asthma Adults
Boston
S70
2017
60
4207
294
0
0
0
0
Asthma Adults
Boston
S70
2017
70
685
98
0
0
0
0
Asthma Adults
Boston
S70
2017
80
98
0
0
0
0
0
Asthma Adults
Boston
S75
2015
0
171995
144601
128654
117011
105564
99107
Asthma Adults
Boston
S75
2015
10
152428
122392
105662
90400
79834
71909
3D-Attachment4-52
-------
Study Group
Study Area
AQ
Scenario
Year
Benchmark
(PPb)
Number of People with 7-hr Exposure at or above Benchmark
>1 Day
> 2 Days
> 3 Days
> 4 Days
> 5 Days
> 6 Days
Asthma Adults
Boston
S75
2015
20
104488
70833
52538
41091
33851
26611
Asthma Adults
Boston
S75
2015
30
62419
35514
26024
19665
15164
11447
Asthma Adults
Boston
S75
2015
40
32775
16241
9392
6261
3620
2446
Asthma Adults
Boston
S75
2015
50
10664
3033
1076
98
0
0
Asthma Adults
Boston
S75
2015
60
2250
196
0
0
0
0
Asthma Adults
Boston
S75
2015
70
391
0
0
0
0
0
Asthma Adults
Boston
S75
2015
80
196
0
0
0
0
0
Asthma Adults
Boston
S75
2016
0
177180
150275
133448
122588
112902
105369
Asthma Adults
Boston
S75
2016
10
158004
128458
109674
96661
86584
79149
Asthma Adults
Boston
S75
2016
20
108402
72203
54005
44319
37373
31307
Asthma Adults
Boston
S75
2016
30
66724
41287
29448
23285
18687
15262
Asthma Adults
Boston
S75
2016
40
40895
20937
12719
8903
6164
4109
Asthma Adults
Boston
S75
2016
50
16045
5283
2837
1468
391
196
Asthma Adults
Boston
S75
2016
60
3033
587
98
98
98
0
Asthma Adults
Boston
S75
2016
70
294
0
0
0
0
0
Asthma Adults
Boston
S75
2016
80
0
0
0
0
0
0
Asthma Adults
Boston
S75
2017
0
174636
148416
131589
120925
110750
103216
Asthma Adults
Boston
S75
2017
10
155558
126599
109282
97542
86878
77290
Asthma Adults
Boston
S75
2017
20
108499
73768
56745
45591
37569
31992
Asthma Adults
Boston
S75
2017
30
67702
41776
29644
22600
17317
13599
Asthma Adults
Boston
S75
2017
40
41580
21817
12621
8414
5674
3913
Asthma Adults
Boston
S75
2017
50
18784
6066
1174
391
196
98
Asthma Adults
Boston
S75
2017
60
6164
881
0
0
0
0
Asthma Adults
Boston
S75
2017
70
783
98
0
0
0
0
Asthma Adults
Boston
S75
2017
80
98
0
0
0
0
0
Asthma Adults
Dallas
S65
2015
0
102827
89701
81887
75558
70401
66650
Asthma Adults
Dallas
S65
2015
10
90091
74933
64853
57430
51492
46726
Asthma Adults
Dallas
S65
2015
20
63759
44538
34536
27660
22972
18440
Asthma Adults
Dallas
S65
2015
30
39693
24457
17034
12814
10001
7814
Asthma Adults
Dallas
S65
2015
40
23675
11564
7110
4610
3360
2422
Asthma Adults
Dallas
S65
2015
50
9064
1875
703
234
0
0
Asthma Adults
Dallas
S65
2015
60
938
0
0
0
0
0
Asthma Adults
Dallas
S65
2015
70
0
0
0
0
0
0
Asthma Adults
Dallas
S65
2015
80
0
0
0
0
0
0
Asthma Adults
Dallas
S65
2016
0
106500
93217
84700
77668
72432
67510
Asthma Adults
Dallas
S65
2016
10
94154
78605
66807
59696
54539
50007
Asthma Adults
Dallas
S65
2016
20
64853
46726
36255
28754
23753
20784
Asthma Adults
Dallas
S65
2016
30
42115
26098
19143
14143
11486
9454
Asthma Adults
Dallas
S65
2016
40
21956
10158
6720
4376
2422
1641
Asthma Adults
Dallas
S65
2016
50
5079
1328
313
78
78
78
Asthma Adults
Dallas
S65
2016
60
156
0
0
0
0
0
Asthma Adults
Dallas
S65
2016
70
78
0
0
0
0
0
Asthma Adults
Dallas
S65
2016
80
0
0
0
0
0
0
Asthma Adults
Dallas
S65
2017
0
102046
89701
82668
75870
70870
66103
Asthma Adults
Dallas
S65
2017
10
91732
77042
65400
59618
54227
50007
Asthma Adults
Dallas
S65
2017
20
61884
45241
35161
29614
25160
21800
Asthma Adults
Dallas
S65
2017
30
39381
25551
19534
16096
13049
10548
3D-Attachment4-53
-------
Study Group
Study Area
AQ
Scenario
Year
Benchmark
(PPb)
Number of People with 7-hr Exposure at or above Benchmark
>1 Day
> 2 Days
> 3 Days
> 4 Days
> 5 Days
> 6 Days
Asthma Adults
Dallas
S65
2017
40
23285
12814
7970
5626
3829
2813
Asthma Adults
Dallas
S65
2017
50
8908
2578
469
78
0
0
Asthma Adults
Dallas
S65
2017
60
859
0
0
0
0
0
Asthma Adults
Dallas
S65
2017
70
0
0
0
0
0
0
Asthma Adults
Dallas
S65
2017
80
0
0
0
0
0
0
Asthma Adults
Dallas
S70
2015
0
102827
89701
81887
75558
70401
66650
Asthma Adults
Dallas
S70
2015
10
90638
75402
65400
57899
52351
47429
Asthma Adults
Dallas
S70
2015
20
65400
45944
35943
29301
23753
19065
Asthma Adults
Dallas
S70
2015
30
41959
25941
18362
13830
11252
8361
Asthma Adults
Dallas
S70
2015
40
26254
13596
8204
5782
3751
2969
Asthma Adults
Dallas
S70
2015
50
13439
3907
1953
703
313
234
Asthma Adults
Dallas
S70
2015
60
3047
156
0
0
0
0
Asthma Adults
Dallas
S70
2015
70
234
0
0
0
0
0
Asthma Adults
Dallas
S70
2015
80
0
0
0
0
0
0
Asthma Adults
Dallas
S70
2016
0
106500
93217
84700
77668
72432
67510
Asthma Adults
Dallas
S70
2016
10
94389
79074
67432
59931
54930
50398
Asthma Adults
Dallas
S70
2016
20
66963
47585
37974
30161
25082
21566
Asthma Adults
Dallas
S70
2016
30
44147
27738
20315
15002
11642
9923
Asthma Adults
Dallas
S70
2016
40
25472
12189
7970
5782
3516
2266
Asthma Adults
Dallas
S70
2016
50
8048
2344
859
313
156
78
Asthma Adults
Dallas
S70
2016
60
781
78
0
0
0
0
Asthma Adults
Dallas
S70
2016
70
78
0
0
0
0
0
Asthma Adults
Dallas
S70
2016
80
0
0
0
0
0
0
Asthma Adults
Dallas
S70
2017
0
102046
89701
82668
75870
70870
66103
Asthma Adults
Dallas
S70
2017
10
92123
77199
66494
59852
55242
50242
Asthma Adults
Dallas
S70
2017
20
64462
47194
36412
30708
26332
22972
Asthma Adults
Dallas
S70
2017
30
41334
26723
20315
16956
13908
11330
Asthma Adults
Dallas
S70
2017
40
25082
15237
9689
7267
5235
3751
Asthma Adults
Dallas
S70
2017
50
11720
4610
1563
469
156
0
Asthma Adults
Dallas
S70
2017
60
2500
78
0
0
0
0
Asthma Adults
Dallas
S70
2017
70
0
0
0
0
0
0
Asthma Adults
Dallas
S70
2017
80
0
0
0
0
0
0
Asthma Adults
Dallas
S75
2015
0
102827
89701
81887
75558
70401
66650
Asthma Adults
Dallas
S75
2015
10
90404
75402
65322
57899
52508
47507
Asthma Adults
Dallas
S75
2015
20
66728
47819
36802
29848
24613
19534
Asthma Adults
Dallas
S75
2015
30
44381
27191
19456
14768
11564
8986
Asthma Adults
Dallas
S75
2015
40
27504
15002
9064
6642
4610
3204
Asthma Adults
Dallas
S75
2015
50
16565
6251
2813
1485
703
313
Asthma Adults
Dallas
S75
2015
60
5313
859
156
0
0
0
Asthma Adults
Dallas
S75
2015
70
469
0
0
0
0
0
Asthma Adults
Dallas
S75
2015
80
0
0
0
0
0
0
Asthma Adults
Dallas
S75
2016
0
106500
93217
84700
77668
72432
67510
Asthma Adults
Dallas
S75
2016
10
94389
78996
67432
60399
55086
50554
Asthma Adults
Dallas
S75
2016
20
68057
48679
38756
30864
25629
21566
Asthma Adults
Dallas
S75
2016
30
45163
29223
21019
15315
12033
10158
Asthma Adults
Dallas
S75
2016
40
27348
13518
8517
6798
4376
2969
Asthma Adults
Dallas
S75
2016
50
11095
3907
1406
547
234
156
3D-Attachment4-54
-------
Study Group
Study Area
AQ
Scenario
Year
Benchmark
(PPb)
Number of People with 7-hr Exposure at or above Benchmark
>1 Day
> 2 Days
> 3 Days
> 4 Days
> 5 Days
> 6 Days
Asthma Adults
Dallas
S75
2016
60
1719
78
0
0
0
0
Asthma Adults
Dallas
S75
2016
70
234
0
0
0
0
0
Asthma Adults
Dallas
S75
2016
80
78
0
0
0
0
0
Asthma Adults
Dallas
S75
2017
0
102046
89701
82668
75870
70870
66103
Asthma Adults
Dallas
S75
2017
10
92123
77433
66416
60087
55242
50476
Asthma Adults
Dallas
S75
2017
20
65400
48288
37427
32114
27035
23128
Asthma Adults
Dallas
S75
2017
30
42741
27817
21175
17659
14299
11799
Asthma Adults
Dallas
S75
2017
40
26566
16799
10705
7814
5938
4219
Asthma Adults
Dallas
S75
2017
50
14221
6485
3047
1328
547
234
Asthma Adults
Dallas
S75
2017
60
4454
469
0
0
0
0
Asthma Adults
Dallas
S75
2017
70
156
0
0
0
0
0
Asthma Adults
Dallas
S75
2017
80
0
0
0
0
0
0
Asthma Adults
Detroit
S65
2015
0
117319
100672
87825
81337
75569
71178
Asthma Adults
Detroit
S65
2015
10
104145
83762
72751
64034
56497
50729
Asthma Adults
Detroit
S65
2015
20
70850
47452
35458
27921
23202
19859
Asthma Adults
Detroit
S65
2015
30
42799
25823
19138
14812
12125
9831
Asthma Adults
Detroit
S65
2015
40
24709
12781
7931
4916
3408
2294
Asthma Adults
Detroit
S65
2015
50
9045
2884
1049
197
131
66
Asthma Adults
Detroit
S65
2015
60
655
0
0
0
0
0
Asthma Adults
Detroit
S65
2015
70
0
0
0
0
0
0
Asthma Adults
Detroit
S65
2015
80
0
0
0
0
0
0
Asthma Adults
Detroit
S65
2016
0
116467
99557
89923
82189
74717
69867
Asthma Adults
Detroit
S65
2016
10
103883
84745
73210
64296
57611
51450
Asthma Adults
Detroit
S65
2016
20
73406
49877
38735
31001
26020
21563
Asthma Adults
Detroit
S65
2016
30
46141
27658
20383
15861
13108
10093
Asthma Adults
Detroit
S65
2016
40
26741
13567
7996
5374
4195
3015
Asthma Adults
Detroit
S65
2016
50
10552
3408
1311
459
131
66
Asthma Adults
Detroit
S65
2016
60
2032
262
0
0
0
0
Asthma Adults
Detroit
S65
2016
70
66
0
0
0
0
0
Asthma Adults
Detroit
S65
2016
80
0
0
0
0
0
0
Asthma Adults
Detroit
S65
2017
0
116205
99885
89464
81534
75504
70195
Asthma Adults
Detroit
S65
2017
10
104670
85269
73406
64689
57087
52302
Asthma Adults
Detroit
S65
2017
20
72882
51188
39325
32115
25889
21498
Asthma Adults
Detroit
S65
2017
30
46403
29100
22153
17762
14091
12256
Asthma Adults
Detroit
S65
2017
40
28117
15861
10093
7078
4785
3343
Asthma Adults
Detroit
S65
2017
50
10749
2949
1442
328
197
131
Asthma Adults
Detroit
S65
2017
60
721
0
0
0
0
0
Asthma Adults
Detroit
S65
2017
70
66
0
0
0
0
0
Asthma Adults
Detroit
S65
2017
80
0
0
0
0
0
0
Asthma Adults
Detroit
S70
2015
0
117319
100672
87825
81337
75569
71178
Asthma Adults
Detroit
S70
2015
10
104407
84286
73079
64099
56431
50991
Asthma Adults
Detroit
S70
2015
20
72161
49484
36900
29231
24119
20842
Asthma Adults
Detroit
S70
2015
30
44634
27527
19990
15402
12846
10356
Asthma Adults
Detroit
S70
2015
40
28248
15009
9569
6685
4588
3539
Asthma Adults
Detroit
S70
2015
50
13108
4653
2294
655
393
131
Asthma Adults
Detroit
S70
2015
60
2425
197
66
0
0
0
Asthma Adults
Detroit
S70
2015
70
66
0
0
0
0
0
3D-Attachment4-55
-------
Study Group
Study Area
AQ
Scenario
Year
Benchmark
(PPb)
Number of People with 7-hr Exposure at or above Benchmark
>1 Day
> 2 Days
> 3 Days
> 4 Days
> 5 Days
> 6 Days
Asthma Adults
Detroit
S70
2015
80
0
0
0
0
0
0
Asthma Adults
Detroit
S70
2016
0
116467
99557
89923
82189
74717
69867
Asthma Adults
Detroit
S70
2016
10
103818
85204
72948
64427
58004
52040
Asthma Adults
Detroit
S70
2016
20
74914
52105
40374
32574
26741
22612
Asthma Adults
Detroit
S70
2016
30
48763
29494
21301
16582
13698
11208
Asthma Adults
Detroit
S70
2016
40
29756
15664
9897
7013
5112
3998
Asthma Adults
Detroit
S70
2016
50
15206
5637
2622
1507
524
197
Asthma Adults
Detroit
S70
2016
60
5047
590
66
0
0
0
Asthma Adults
Detroit
S70
2016
70
655
0
0
0
0
0
Asthma Adults
Detroit
S70
2016
80
0
0
0
0
0
0
Asthma Adults
Detroit
S70
2017
0
116205
99885
89464
81534
75504
70195
Asthma Adults
Detroit
S70
2017
10
104473
85204
73144
64362
57349
52433
Asthma Adults
Detroit
S70
2017
20
74324
53089
40898
33098
27134
22743
Asthma Adults
Detroit
S70
2017
30
48566
31067
23005
18221
14747
12453
Asthma Adults
Detroit
S70
2017
40
31919
18286
12387
8586
6226
4457
Asthma Adults
Detroit
S70
2017
50
14878
5571
2818
1114
590
262
Asthma Adults
Detroit
S70
2017
60
2425
197
0
0
0
0
Asthma Adults
Detroit
S70
2017
70
197
0
0
0
0
0
Asthma Adults
Detroit
S70
2017
80
0
0
0
0
0
0
Asthma Adults
Detroit
S75
2015
0
117319
100672
87825
81337
75569
71178
Asthma Adults
Detroit
S75
2015
10
104670
83696
71244
62854
55579
50008
Asthma Adults
Detroit
S75
2015
20
73013
49353
37096
30018
23923
20646
Asthma Adults
Detroit
S75
2015
30
45617
28052
20646
15468
12912
9831
Asthma Adults
Detroit
S75
2015
40
29559
15795
10356
7210
4981
3801
Asthma Adults
Detroit
S75
2015
50
15730
6030
3080
1245
524
459
Asthma Adults
Detroit
S75
2015
60
4719
721
197
0
0
0
Asthma Adults
Detroit
S75
2015
70
524
66
0
0
0
0
Asthma Adults
Detroit
S75
2015
80
0
0
0
0
0
0
Asthma Adults
Detroit
S75
2016
0
116467
99557
89923
82189
74717
69867
Asthma Adults
Detroit
S75
2016
10
103031
84286
72489
63510
57349
51253
Asthma Adults
Detroit
S75
2016
20
75438
53351
41029
32902
26938
22808
Asthma Adults
Detroit
S75
2016
30
50467
30018
21629
16844
13764
11273
Asthma Adults
Detroit
S75
2016
40
31329
16975
10945
7799
5899
4653
Asthma Adults
Detroit
S75
2016
50
17631
7144
3867
2228
1245
590
Asthma Adults
Detroit
S75
2016
60
7472
1966
393
131
0
0
Asthma Adults
Detroit
S75
2016
70
1835
66
0
0
0
0
Asthma Adults
Detroit
S75
2016
80
0
0
0
0
0
0
Asthma Adults
Detroit
S75
2017
0
116205
99885
89464
81534
75504
70195
Asthma Adults
Detroit
S75
2017
10
104211
84155
72161
63706
56693
52105
Asthma Adults
Detroit
S75
2017
20
74848
53351
41291
33033
26872
21956
Asthma Adults
Detroit
S75
2017
30
49943
31788
23333
18548
14550
12584
Asthma Adults
Detroit
S75
2017
40
33295
18614
12256
9241
7013
4850
Asthma Adults
Detroit
S75
2017
50
17893
7013
3998
2228
1180
655
Asthma Adults
Detroit
S75
2017
60
6620
786
262
0
0
0
Asthma Adults
Detroit
S75
2017
70
459
0
0
0
0
0
Asthma Adults
Detroit
S75
2017
80
131
0
0
0
0
0
Asthma Adults
Philadelphia
S65
2015
0
137336
115725
101608
91238
85487
80868
3D-Attachment4-56
-------
Study Group
Study Area
AQ
Scenario
Year
Benchmark
(PPb)
Number of People with 7-hr Exposure at or above Benchmark
>1 Day
> 2 Days
> 3 Days
> 4 Days
> 5 Days
> 6 Days
Asthma Adults
Philadelphia
S65
2015
10
120866
97077
84005
75117
67187
60738
Asthma Adults
Philadelphia
S65
2015
20
84964
58995
46708
37297
30761
25533
Asthma Adults
Philadelphia
S65
2015
30
55945
36426
25794
20217
16557
14378
Asthma Adults
Philadelphia
S65
2015
40
33114
17603
12113
7843
5403
3747
Asthma Adults
Philadelphia
S65
2015
50
10370
2876
610
87
0
0
Asthma Adults
Philadelphia
S65
2015
60
871
0
0
0
0
0
Asthma Adults
Philadelphia
S65
2015
70
0
0
0
0
0
0
Asthma Adults
Philadelphia
S65
2015
80
0
0
0
0
0
0
Asthma Adults
Philadelphia
S65
2016
0
131672
112152
101434
93242
85748
79474
Asthma Adults
Philadelphia
S65
2016
10
119298
97077
84790
75291
67448
61958
Asthma Adults
Philadelphia
S65
2016
20
85487
61087
45053
36600
31110
26927
Asthma Adults
Philadelphia
S65
2016
30
55161
33898
24661
19258
16034
13246
Asthma Adults
Philadelphia
S65
2016
40
30587
14814
8279
4531
2701
1830
Asthma Adults
Philadelphia
S65
2016
50
6536
1394
436
87
87
0
Asthma Adults
Philadelphia
S65
2016
60
436
0
0
0
0
0
Asthma Adults
Philadelphia
S65
2016
70
0
0
0
0
0
0
Asthma Adults
Philadelphia
S65
2016
80
0
0
0
0
0
0
Asthma Adults
Philadelphia
S65
2017
0
136726
115725
102654
95247
88972
82088
Asthma Adults
Philadelphia
S65
2017
10
121999
98122
85835
77121
69627
63004
Asthma Adults
Philadelphia
S65
2017
20
81740
58908
45837
37645
32504
28234
Asthma Adults
Philadelphia
S65
2017
30
54900
33986
24923
20043
15773
12636
Asthma Adults
Philadelphia
S65
2017
40
29367
14988
7930
4009
2701
2091
Asthma Adults
Philadelphia
S65
2017
50
7581
697
87
0
0
0
Asthma Adults
Philadelphia
S65
2017
60
871
0
0
0
0
0
Asthma Adults
Philadelphia
S65
2017
70
0
0
0
0
0
0
Asthma Adults
Philadelphia
S65
2017
80
0
0
0
0
0
0
Asthma Adults
Philadelphia
S70
2015
0
137336
115725
101608
91238
85487
80868
Asthma Adults
Philadelphia
S70
2015
10
121041
97599
84702
75378
68058
61610
Asthma Adults
Philadelphia
S70
2015
20
87752
61348
48625
38778
32766
27276
Asthma Adults
Philadelphia
S70
2015
30
58473
37907
27101
21698
17603
14814
Asthma Adults
Philadelphia
S70
2015
40
36251
20827
14466
10283
7146
5664
Asthma Adults
Philadelphia
S70
2015
50
16470
5141
1830
174
87
87
Asthma Adults
Philadelphia
S70
2015
60
2614
174
0
0
0
0
Asthma Adults
Philadelphia
S70
2015
70
0
0
0
0
0
0
Asthma Adults
Philadelphia
S70
2015
80
0
0
0
0
0
0
Asthma Adults
Philadelphia
S70
2016
0
131672
112152
101434
93242
85748
79474
Asthma Adults
Philadelphia
S70
2016
10
119995
98209
85487
75727
67971
63091
Asthma Adults
Philadelphia
S70
2016
20
88101
63265
47493
38517
32853
28757
Asthma Adults
Philadelphia
S70
2016
30
57688
35467
25620
20653
17167
14117
Asthma Adults
Philadelphia
S70
2016
40
35293
17777
12026
6971
4793
3224
Asthma Adults
Philadelphia
S70
2016
50
12113
3573
1220
523
261
174
Asthma Adults
Philadelphia
S70
2016
60
1569
0
0
0
0
0
Asthma Adults
Philadelphia
S70
2016
70
0
0
0
0
0
0
Asthma Adults
Philadelphia
S70
2016
80
0
0
0
0
0
0
Asthma Adults
Philadelphia
S70
2017
0
136726
115725
102654
95247
88972
82088
Asthma Adults
Philadelphia
S70
2017
10
122958
98994
86097
77731
70585
63527
Asthma Adults
Philadelphia
S70
2017
20
84702
61435
48190
39563
33986
29193
3D-Attachment4-57
-------
Study Group
Study Area
AQ
Scenario
Year
Benchmark
(PPb)
Number of People with 7-hr Exposure at or above Benchmark
>1 Day
> 2 Days
> 3 Days
> 4 Days
> 5 Days
> 6 Days
Asthma Adults
Philadelphia
S70
2017
30
56817
35380
26753
20914
16383
14030
Asthma Adults
Philadelphia
S70
2017
40
33811
18387
11503
6884
4357
2963
Asthma Adults
Philadelphia
S70
2017
50
12461
3137
523
0
0
0
Asthma Adults
Philadelphia
S70
2017
60
2353
0
0
0
0
0
Asthma Adults
Philadelphia
S70
2017
70
261
0
0
0
0
0
Asthma Adults
Philadelphia
S70
2017
80
0
0
0
0
0
0
Asthma Adults
Philadelphia
S75
2015
0
137336
115725
101608
91238
85487
80868
Asthma Adults
Philadelphia
S75
2015
10
121041
98558
84964
75030
68145
62307
Asthma Adults
Philadelphia
S75
2015
20
89757
63091
49235
40783
33898
29106
Asthma Adults
Philadelphia
S75
2015
30
59954
40347
28408
22134
18300
15511
Asthma Adults
Philadelphia
S75
2015
40
39563
23877
16034
12200
8627
6449
Asthma Adults
Philadelphia
S75
2015
50
22657
9150
4270
1656
610
349
Asthma Adults
Philadelphia
S75
2015
60
5926
784
87
0
0
0
Asthma Adults
Philadelphia
S75
2015
70
697
0
0
0
0
0
Asthma Adults
Philadelphia
S75
2015
80
0
0
0
0
0
0
Asthma Adults
Philadelphia
S75
2016
0
131672
112152
101434
93242
85748
79474
Asthma Adults
Philadelphia
S75
2016
10
120431
98122
85574
76075
67797
63265
Asthma Adults
Philadelphia
S75
2016
20
90367
65270
49933
40085
33550
29803
Asthma Adults
Philadelphia
S75
2016
30
60477
37297
27276
21350
17603
14378
Asthma Adults
Philadelphia
S75
2016
40
40173
20740
13943
8540
6361
4793
Asthma Adults
Philadelphia
S75
2016
50
19346
6884
3050
1481
871
349
Asthma Adults
Philadelphia
S75
2016
60
3573
436
0
0
0
0
Asthma Adults
Philadelphia
S75
2016
70
523
0
0
0
0
0
Asthma Adults
Philadelphia
S75
2016
80
0
0
0
0
0
0
Asthma Adults
Philadelphia
S75
2017
0
136726
115725
102654
95247
88972
82088
Asthma Adults
Philadelphia
S75
2017
10
123481
99517
86009
77818
70062
63614
Asthma Adults
Philadelphia
S75
2017
20
87578
63004
49845
40783
35206
30238
Asthma Adults
Philadelphia
S75
2017
30
58821
37123
27711
21960
17254
14291
Asthma Adults
Philadelphia
S75
2017
40
38081
21176
13681
9324
6361
3921
Asthma Adults
Philadelphia
S75
2017
50
17603
5839
2091
436
174
174
Asthma Adults
Philadelphia
S75
2017
60
5141
174
0
0
0
0
Asthma Adults
Philadelphia
S75
2017
70
610
0
0
0
0
0
Asthma Adults
Philadelphia
S75
2017
80
0
0
0
0
0
0
Asthma Adults
Phoenix
S65
2015
0
81554
71968
65512
60297
56472
53790
Asthma Adults
Phoenix
S65
2015
10
74055
63475
55826
49817
46588
42565
Asthma Adults
Phoenix
S65
2015
20
55578
42615
34420
28907
25132
22152
Asthma Adults
Phoenix
S65
2015
30
39386
26076
20066
17036
14205
12367
Asthma Adults
Phoenix
S65
2015
40
25231
14950
10877
8245
6357
5215
Asthma Adults
Phoenix
S65
2015
50
10579
3775
1490
944
447
447
Asthma Adults
Phoenix
S65
2015
60
596
50
0
0
0
0
Asthma Adults
Phoenix
S65
2015
70
0
0
0
0
0
0
Asthma Adults
Phoenix
S65
2015
80
0
0
0
0
0
0
Asthma Adults
Phoenix
S65
2016
0
78177
68243
61687
58012
54734
51853
Asthma Adults
Phoenix
S65
2016
10
71621
60247
53095
48327
44751
42118
Asthma Adults
Phoenix
S65
2016
20
52697
39237
31837
26970
23493
21109
Asthma Adults
Phoenix
S65
2016
30
35413
23840
19519
16589
14155
12516
Asthma Adults
Phoenix
S65
2016
40
23940
15347
10977
8146
6755
5612
3D-Attachment4-5 8
-------
Study Group
Study Area
AQ
Scenario
Year
Benchmark
(PPb)
Number of People with 7-hr Exposure at or above Benchmark
>1 Day
> 2 Days
> 3 Days
> 4 Days
> 5 Days
> 6 Days
Asthma Adults
Phoenix
S65
2016
50
8593
2930
1341
447
199
149
Asthma Adults
Phoenix
S65
2016
60
248
0
0
0
0
0
Asthma Adults
Phoenix
S65
2016
70
0
0
0
0
0
0
Asthma Adults
Phoenix
S65
2016
80
0
0
0
0
0
0
Asthma Adults
Phoenix
S65
2017
0
77233
66604
60942
56323
52797
50214
Asthma Adults
Phoenix
S65
2017
10
69882
58757
52598
48277
45098
43012
Asthma Adults
Phoenix
S65
2017
20
53293
39833
33377
28807
23989
21506
Asthma Adults
Phoenix
S65
2017
30
35661
24983
20562
16539
14702
13410
Asthma Adults
Phoenix
S65
2017
40
24884
16539
12715
10281
8195
6804
Asthma Adults
Phoenix
S65
2017
50
10430
4818
2384
1242
497
248
Asthma Adults
Phoenix
S65
2017
60
944
0
0
0
0
0
Asthma Adults
Phoenix
S65
2017
70
0
0
0
0
0
0
Asthma Adults
Phoenix
S65
2017
80
0
0
0
0
0
0
Asthma Adults
Phoenix
S70
2015
0
81554
71968
65512
60297
56472
53790
Asthma Adults
Phoenix
S70
2015
10
74104
63277
55926
49717
46787
43360
Asthma Adults
Phoenix
S70
2015
20
56969
44204
36108
30049
26324
23443
Asthma Adults
Phoenix
S70
2015
30
40728
28162
21158
17533
14553
12914
Asthma Adults
Phoenix
S70
2015
40
27814
17036
12665
10033
7996
6606
Asthma Adults
Phoenix
S70
2015
50
15397
7301
4073
2334
1440
993
Asthma Adults
Phoenix
S70
2015
60
3328
695
199
149
50
50
Asthma Adults
Phoenix
S70
2015
70
50
50
0
0
0
0
Asthma Adults
Phoenix
S70
2015
80
0
0
0
0
0
0
Asthma Adults
Phoenix
S70
2016
0
78177
68243
61687
58012
54734
51853
Asthma Adults
Phoenix
S70
2016
10
71819
60396
53095
48674
45098
42218
Asthma Adults
Phoenix
S70
2016
20
54535
41026
33675
28559
24784
22400
Asthma Adults
Phoenix
S70
2016
30
36853
25082
20016
17284
15049
13063
Asthma Adults
Phoenix
S70
2016
40
26771
17135
13212
10480
8146
6804
Asthma Adults
Phoenix
S70
2016
50
14006
6258
3824
2334
1341
695
Asthma Adults
Phoenix
S70
2016
60
2831
397
99
50
0
0
Asthma Adults
Phoenix
S70
2016
70
0
0
0
0
0
0
Asthma Adults
Phoenix
S70
2016
80
0
0
0
0
0
0
Asthma Adults
Phoenix
S70
2017
0
77233
66604
60942
56323
52797
50214
Asthma Adults
Phoenix
S70
2017
10
70329
58906
52598
48625
45049
42963
Asthma Adults
Phoenix
S70
2017
20
55081
41522
34817
30347
26125
22897
Asthma Adults
Phoenix
S70
2017
30
37499
26175
21655
17682
15149
13559
Asthma Adults
Phoenix
S70
2017
40
27218
18526
14056
11920
10083
8344
Asthma Adults
Phoenix
S70
2017
50
15198
8593
5712
3924
2583
1788
Asthma Adults
Phoenix
S70
2017
60
3973
1142
99
50
0
0
Asthma Adults
Phoenix
S70
2017
70
99
0
0
0
0
0
Asthma Adults
Phoenix
S70
2017
80
0
0
0
0
0
0
Asthma Adults
Phoenix
S75
2015
0
81554
71968
65512
60297
56472
53790
Asthma Adults
Phoenix
S75
2015
10
73856
62879
55032
49419
46290
42814
Asthma Adults
Phoenix
S75
2015
20
57168
44651
36456
30943
26970
23940
Asthma Adults
Phoenix
S75
2015
30
41622
28311
21953
18029
14801
13112
Asthma Adults
Phoenix
S75
2015
40
29403
18029
13659
10381
8394
6953
Asthma Adults
Phoenix
S75
2015
50
18725
9338
6159
3973
2682
1689
Asthma Adults
Phoenix
S75
2015
60
7599
1887
646
298
248
50
3D-Attachment4-59
-------
Study Group
Study Area
AQ
Scenario
Year
Benchmark
(PPb)
Number of People with 7-hr Exposure at or above Benchmark
>1 Day
> 2 Days
> 3 Days
> 4 Days
> 5 Days
> 6 Days
Asthma Adults
Phoenix
S75
2015
70
298
50
50
0
0
0
Asthma Adults
Phoenix
S75
2015
80
0
0
0
0
0
0
Asthma Adults
Phoenix
S75
2016
0
78177
68243
61687
58012
54734
51853
Asthma Adults
Phoenix
S75
2016
10
71770
59850
52946
48128
44999
41622
Asthma Adults
Phoenix
S75
2016
20
55131
41622
33824
29304
25678
22648
Asthma Adults
Phoenix
S75
2016
30
37896
25579
20761
17433
15347
13112
Asthma Adults
Phoenix
S75
2016
40
27913
18129
14106
11424
8791
7450
Asthma Adults
Phoenix
S75
2016
50
17135
8791
5861
3675
2483
1738
Asthma Adults
Phoenix
S75
2016
60
6159
1589
447
149
99
0
Asthma Adults
Phoenix
S75
2016
70
447
0
0
0
0
0
Asthma Adults
Phoenix
S75
2016
80
0
0
0
0
0
0
Asthma Adults
Phoenix
S75
2017
0
77233
66604
60942
56323
52797
50214
Asthma Adults
Phoenix
S75
2017
10
69982
58509
52201
48128
45049
42764
Asthma Adults
Phoenix
S75
2017
20
55677
42317
35810
30744
26870
23890
Asthma Adults
Phoenix
S75
2017
30
38592
27168
21854
17880
15546
13808
Asthma Adults
Phoenix
S75
2017
40
28509
19768
14751
12566
10877
9238
Asthma Adults
Phoenix
S75
2017
50
18178
11175
7251
5364
4222
3129
Asthma Adults
Phoenix
S75
2017
60
7947
2930
993
397
99
50
Asthma Adults
Phoenix
S75
2017
70
1043
50
0
0
0
0
Asthma Adults
Phoenix
S75
2017
80
0
0
0
0
0
0
Asthma Adults
Sacramento
S65
2015
0
36274
31357
28213
26012
24583
23096
Asthma Adults
Sacramento
S65
2015
10
31643
25726
22439
20324
18694
16979
Asthma Adults
Sacramento
S65
2015
20
21467
14978
12063
10233
8833
7546
Asthma Adults
Sacramento
S65
2015
30
13921
9004
7003
5431
4745
4059
Asthma Adults
Sacramento
S65
2015
40
7289
3373
2344
1486
800
572
Asthma Adults
Sacramento
S65
2015
50
1343
200
29
0
0
0
Asthma Adults
Sacramento
S65
2015
60
57
0
0
0
0
0
Asthma Adults
Sacramento
S65
2015
70
0
0
0
0
0
0
Asthma Adults
Sacramento
S65
2015
80
0
0
0
0
0
0
Asthma Adults
Sacramento
S65
2016
0
36960
31586
28127
26040
24383
22896
Asthma Adults
Sacramento
S65
2016
10
31443
25526
22667
20409
18894
17236
Asthma Adults
Sacramento
S65
2016
20
21181
15350
12406
10319
9033
7775
Asthma Adults
Sacramento
S65
2016
30
13806
8947
6717
5460
4602
4030
Asthma Adults
Sacramento
S65
2016
40
6660
3373
1972
1229
772
543
Asthma Adults
Sacramento
S65
2016
50
1658
286
0
0
0
0
Asthma Adults
Sacramento
S65
2016
60
57
0
0
0
0
0
Asthma Adults
Sacramento
S65
2016
70
0
0
0
0
0
0
Asthma Adults
Sacramento
S65
2016
80
0
0
0
0
0
0
Asthma Adults
Sacramento
S65
2017
0
36588
31614
28413
25983
23725
22267
Asthma Adults
Sacramento
S65
2017
10
31814
25926
22582
19723
17665
16265
Asthma Adults
Sacramento
S65
2017
20
21067
14578
11177
9290
7804
6689
Asthma Adults
Sacramento
S65
2017
30
12663
8204
6632
5374
4316
3802
Asthma Adults
Sacramento
S65
2017
40
6117
3144
1801
1201
829
629
Asthma Adults
Sacramento
S65
2017
50
943
143
0
0
0
0
Asthma Adults
Sacramento
S65
2017
60
86
0
0
0
0
0
Asthma Adults
Sacramento
S65
2017
70
0
0
0
0
0
0
Asthma Adults
Sacramento
S65
2017
80
0
0
0
0
0
0
3D-Attachment4-60
-------
Study Group
Study Area
AQ
Scenario
Year
Benchmark
(PPb)
Number of People with 7-hr Exposure at or above Benchmark
>1 Day
> 2 Days
> 3 Days
> 4 Days
> 5 Days
> 6 Days
Asthma Adults
Sacramento
S70
2015
0
36274
31357
28213
26012
24583
23096
Asthma Adults
Sacramento
S70
2015
10
31700
25955
22696
20524
18780
17494
Asthma Adults
Sacramento
S70
2015
20
22296
16207
13006
11005
9376
8032
Asthma Adults
Sacramento
S70
2015
30
14864
9747
7746
6060
5059
4402
Asthma Adults
Sacramento
S70
2015
40
9347
5117
3201
2458
1829
1401
Asthma Adults
Sacramento
S70
2015
50
3259
972
429
57
0
0
Asthma Adults
Sacramento
S70
2015
60
257
0
0
0
0
0
Asthma Adults
Sacramento
S70
2015
70
57
0
0
0
0
0
Asthma Adults
Sacramento
S70
2015
80
0
0
0
0
0
0
Asthma Adults
Sacramento
S70
2016
0
36960
31586
28127
26040
24383
22896
Asthma Adults
Sacramento
S70
2016
10
31529
25755
22839
20724
19123
17465
Asthma Adults
Sacramento
S70
2016
20
21953
16122
13235
11119
9433
8261
Asthma Adults
Sacramento
S70
2016
30
14892
9776
7318
5860
5088
4230
Asthma Adults
Sacramento
S70
2016
40
8347
4688
3030
2230
1629
1172
Asthma Adults
Sacramento
S70
2016
50
3716
1201
343
143
57
29
Asthma Adults
Sacramento
S70
2016
60
286
86
0
0
0
0
Asthma Adults
Sacramento
S70
2016
70
29
0
0
0
0
0
Asthma Adults
Sacramento
S70
2016
80
0
0
0
0
0
0
Asthma Adults
Sacramento
S70
2017
0
36588
31614
28413
25983
23725
22267
Asthma Adults
Sacramento
S70
2017
10
31929
26298
22868
20266
18094
16608
Asthma Adults
Sacramento
S70
2017
20
22124
15436
11948
9804
8318
7175
Asthma Adults
Sacramento
S70
2017
30
14035
9090
7175
5688
4716
4059
Asthma Adults
Sacramento
S70
2017
40
7918
4602
3287
2144
1629
1372
Asthma Adults
Sacramento
S70
2017
50
3030
772
143
57
29
0
Asthma Adults
Sacramento
S70
2017
60
343
29
0
0
0
0
Asthma Adults
Sacramento
S70
2017
70
29
0
0
0
0
0
Asthma Adults
Sacramento
S70
2017
80
0
0
0
0
0
0
Asthma Adults
Sacramento
S75
2015
0
36274
31357
28213
26012
24583
23096
Asthma Adults
Sacramento
S75
2015
10
31729
26098
22982
20781
18894
17637
Asthma Adults
Sacramento
S75
2015
20
23039
16893
13578
11405
9719
8432
Asthma Adults
Sacramento
S75
2015
30
15836
10262
8089
6317
5260
4574
Asthma Adults
Sacramento
S75
2015
40
10490
6088
4202
3116
2315
1915
Asthma Adults
Sacramento
S75
2015
50
4888
1829
915
400
257
114
Asthma Adults
Sacramento
S75
2015
60
972
143
0
0
0
0
Asthma Adults
Sacramento
S75
2015
70
86
0
0
0
0
0
Asthma Adults
Sacramento
S75
2015
80
29
0
0
0
0
0
Asthma Adults
Sacramento
S75
2016
0
36960
31586
28127
26040
24383
22896
Asthma Adults
Sacramento
S75
2016
10
31643
25869
22896
20867
19180
17551
Asthma Adults
Sacramento
S75
2016
20
22953
16722
13663
11777
9976
8575
Asthma Adults
Sacramento
S75
2016
30
15693
10147
7661
6146
5260
4431
Asthma Adults
Sacramento
S75
2016
40
9576
5517
3830
2858
2230
1658
Asthma Adults
Sacramento
S75
2016
50
5145
2001
972
515
286
114
Asthma Adults
Sacramento
S75
2016
60
1629
286
0
0
0
0
Asthma Adults
Sacramento
S75
2016
70
114
0
0
0
0
0
Asthma Adults
Sacramento
S75
2016
80
0
0
0
0
0
0
Asthma Adults
Sacramento
S75
2017
0
36588
31614
28413
25983
23725
22267
Asthma Adults
Sacramento
S75
2017
10
31986
26526
23153
20495
18208
16665
3D-Attachment4-61
-------
Study Group
Study Area
AQ
Scenario
Year
Benchmark
(PPb)
Number of People with 7-hr Exposure at or above Benchmark
>1 Day
> 2 Days
> 3 Days
> 4 Days
> 5 Days
> 6 Days
Asthma Adults
Sacramento
S75
2017
20
23296
16236
12491
10090
8718
7603
Asthma Adults
Sacramento
S75
2017
30
14950
9519
7575
5888
4888
4202
Asthma Adults
Sacramento
S75
2017
40
9004
5460
3887
2773
2144
1744
Asthma Adults
Sacramento
S75
2017
50
4345
1744
829
429
200
114
Asthma Adults
Sacramento
S75
2017
60
972
143
0
0
0
0
Asthma Adults
Sacramento
S75
2017
70
114
0
0
0
0
0
Asthma Adults
Sacramento
S75
2017
80
0
0
0
0
0
0
Asthma Adults
St. Louis
S65
2015
0
59231
50503
45496
41812
38843
36447
Asthma Adults
St. Louis
S65
2015
10
51719
41562
35767
31439
28256
25431
Asthma Adults
St. Louis
S65
2015
20
34873
23463
18062
14378
11696
9693
Asthma Adults
St. Louis
S65
2015
30
20602
12805
9013
6975
5544
4793
Asthma Adults
St. Louis
S65
2015
40
11231
5687
3434
2075
1538
1037
Asthma Adults
St. Louis
S65
2015
50
3505
715
215
72
36
0
Asthma Adults
St. Louis
S65
2015
60
143
0
0
0
0
0
Asthma Adults
St. Louis
S65
2015
70
0
0
0
0
0
0
Asthma Adults
St. Louis
S65
2015
80
0
0
0
0
0
0
Asthma Adults
St. Louis
S65
2016
0
56262
47570
43028
39094
36053
33764
Asthma Adults
St. Louis
S65
2016
10
49824
40524
34945
30402
27398
25180
Asthma Adults
St. Louis
S65
2016
20
34337
23249
17669
14307
11624
9657
Asthma Adults
St. Louis
S65
2016
30
21496
12876
9693
7762
6259
5437
Asthma Adults
St. Louis
S65
2016
40
13019
6903
4149
3076
1967
1431
Asthma Adults
St. Louis
S65
2016
50
5294
1645
572
72
36
0
Asthma Adults
St. Louis
S65
2016
60
930
72
36
0
0
0
Asthma Adults
St. Louis
S65
2016
70
36
0
0
0
0
0
Asthma Adults
St. Louis
S65
2016
80
0
0
0
0
0
0
Asthma Adults
St. Louis
S65
2017
0
58694
50897
45746
42062
39094
36769
Asthma Adults
St. Louis
S65
2017
10
52435
43243
37556
34158
30760
27577
Asthma Adults
St. Louis
S65
2017
20
37198
26217
20495
16524
13556
11338
Asthma Adults
St. Louis
S65
2017
30
23249
15165
10694
8405
6832
5437
Asthma Adults
St. Louis
S65
2017
40
14271
7368
4686
3112
2253
1717
Asthma Adults
St. Louis
S65
2017
50
4328
1180
644
286
179
107
Asthma Adults
St. Louis
S65
2017
60
72
0
0
0
0
0
Asthma Adults
St. Louis
S65
2017
70
0
0
0
0
0
0
Asthma Adults
St. Louis
S65
2017
80
0
0
0
0
0
0
Asthma Adults
St. Louis
S70
2015
0
59231
50503
45496
41812
38843
36447
Asthma Adults
St. Louis
S70
2015
10
52006
42134
36125
31547
28578
25896
Asthma Adults
St. Louis
S70
2015
20
36232
24679
18849
15487
12411
10337
Asthma Adults
St. Louis
S70
2015
30
22211
13949
9621
7475
5830
5007
Asthma Adults
St. Louis
S70
2015
40
13341
7153
4542
3040
2003
1610
Asthma Adults
St. Louis
S70
2015
50
5794
1717
644
215
107
0
Asthma Adults
St. Louis
S70
2015
60
787
72
72
0
0
0
Asthma Adults
St. Louis
S70
2015
70
0
0
0
0
0
0
Asthma Adults
St. Louis
S70
2015
80
0
0
0
0
0
0
Asthma Adults
St. Louis
S70
2016
0
56262
47570
43028
39094
36053
33764
Asthma Adults
St. Louis
S70
2016
10
50432
40632
35195
30796
27791
25287
Asthma Adults
St. Louis
S70
2016
20
35588
24358
18635
15559
12197
10194
Asthma Adults
St. Louis
S70
2016
30
22784
13913
9979
8012
6545
5615
3D-Attachment4-62
-------
Study Group
Study Area
AQ
Scenario
Year
Benchmark
(PPb)
Number of People with 7-hr Exposure at or above Benchmark
>1 Day
> 2 Days
> 3 Days
> 4 Days
> 5 Days
> 6 Days
Asthma Adults
St. Louis
S70
2016
40
14665
8262
5580
3899
2683
1860
Asthma Adults
St. Louis
S70
2016
50
7618
2826
1431
501
286
143
Asthma Adults
St. Louis
S70
2016
60
2325
CO
LO
CO
36
0
0
0
Asthma Adults
St. Louis
S70
2016
70
179
0
0
0
0
0
Asthma Adults
St. Louis
S70
2016
80
0
0
0
0
0
0
Asthma Adults
St. Louis
S70
2017
0
58694
50897
45746
42062
39094
36769
Asthma Adults
St. Louis
S70
2017
10
52685
43672
37985
34372
31082
28256
Asthma Adults
St. Louis
S70
2017
20
38307
27684
21997
17383
14378
12447
Asthma Adults
St. Louis
S70
2017
30
24894
16203
11660
8835
7511
5973
Asthma Adults
St. Louis
S70
2017
40
16453
9478
6009
4221
2790
2325
Asthma Adults
St. Louis
S70
2017
50
7332
2754
1431
CO
CNJ
CO
501
CO
LO
CO
Asthma Adults
St. Louis
S70
2017
60
1073
143
36
0
0
0
Asthma Adults
St. Louis
S70
2017
70
36
0
0
0
0
0
Asthma Adults
St. Louis
S70
2017
80
0
0
0
0
0
0
Asthma Adults
St. Louis
S75
2015
0
59231
50503
45496
41812
38843
36447
Asthma Adults
St. Louis
S75
2015
10
52041
42420
36518
31726
28757
25896
Asthma Adults
St. Louis
S75
2015
20
36876
25538
19600
15809
12912
10802
Asthma Adults
St. Louis
S75
2015
30
23499
14522
10015
7762
6009
5115
Asthma Adults
St. Louis
S75
2015
40
14593
8048
5222
3434
2718
1896
Asthma Adults
St. Louis
S75
2015
50
7618
2861
1180
537
215
72
Asthma Adults
St. Louis
S75
2015
60
2039
250
72
36
0
0
Asthma Adults
St. Louis
S75
2015
70
36
0
0
0
0
0
Asthma Adults
St. Louis
S75
2015
80
0
0
0
0
0
0
Asthma Adults
St. Louis
S75
2016
0
56262
47570
43028
39094
36053
33764
Asthma Adults
St. Louis
S75
2016
10
50360
40846
35481
30831
27720
25395
Asthma Adults
St. Louis
S75
2016
20
36626
24787
19493
15809
12805
10444
Asthma Adults
St. Louis
S75
2016
30
24071
14450
10373
8298
6653
5723
Asthma Adults
St. Louis
S75
2016
40
15416
8870
6080
4614
3291
2432
Asthma Adults
St. Louis
S75
2016
50
9264
3863
2218
1109
I—.
CO
LO
322
Asthma Adults
St. Louis
S75
2016
60
3577
787
143
0
0
0
Asthma Adults
St. Louis
S75
2016
70
CO
LO
CO
72
36
0
0
0
Asthma Adults
St. Louis
S75
2016
80
36
0
0
0
0
0
Asthma Adults
St. Louis
S75
2017
0
58694
50897
45746
42062
39094
36769
Asthma Adults
St. Louis
S75
2017
10
52685
43815
38343
34623
31082
28328
Asthma Adults
St. Louis
S75
2017
20
39237
28542
22855
18456
14986
13019
Asthma Adults
St. Louis
S75
2017
30
26003
16846
12089
9156
7690
6224
Asthma Adults
St. Louis
S75
2017
40
17204
10122
6617
4757
3398
2861
Asthma Adults
St. Louis
S75
2017
50
9550
4256
2182
1252
930
680
Asthma Adults
St. Louis
S75
2017
60
2539
644
107
107
107
36
Asthma Adults
St. Louis
S75
2017
70
107
0
0
0
0
0
Asthma Adults
St. Louis
S75
2017
80
36
0
0
0
0
0
3D-Attachment4-63
-------
Study Group
Study Area
AQ
Scenario
Year
FEV,
(percent)
Number of People at or above FEVt Decrement
>1 Day
> 2 Days
> 3 Days
> 4 Days
> 5 Days
> 6 Days
All Children
Atlanta
S65
2015
10
119728
64262
44045
33433
25947
20943
All Children
Atlanta
S65
2015
15
31879
15718
9806
6739
4802
3612
All Children
Atlanta
S65
2015
20
12812
4903
2663
1897
1291
1069
All Children
Atlanta
S65
2016
10
141680
82320
57846
44510
36156
29619
All Children
Atlanta
S65
2016
15
40636
19955
12994
9261
7122
5649
All Children
Atlanta
S65
2016
20
15213
6457
3914
2583
1957
1392
All Children
Atlanta
S65
2017
10
111112
60772
41019
31072
24373
19713
All Children
Atlanta
S65
2017
15
28368
13357
8535
6134
4600
3470
All Children
Atlanta
S65
2017
20
9443
3975
2119
1372
989
646
All Children
Atlanta
S70
2015
10
154048
87284
61922
47536
38457
31536
All Children
Atlanta
S70
2015
15
48303
24232
16444
11561
8777
6941
All Children
Atlanta
S70
2015
20
20217
8979
4903
3369
2441
1856
All Children
Atlanta
S70
2016
10
182558
110689
80424
63455
51955
43561
All Children
Atlanta
S70
2016
15
60328
32040
21125
15536
12187
9765
All Children
Atlanta
S70
2016
20
25685
11904
7607
5206
3914
3087
All Children
Atlanta
S70
2017
10
141680
80585
57443
44106
35228
28711
All Children
Atlanta
S70
2017
15
40676
20338
13720
9967
7647
5972
All Children
Atlanta
S70
2017
20
15233
7042
4358
2764
1957
1372
All Children
Atlanta
S75
2015
10
192081
114563
83491
65170
53266
44631
All Children
Atlanta
S75
2015
15
68560
36136
24736
18482
14668
11682
All Children
Atlanta
S75
2015
20
29680
14890
9624
6335
4822
3632
All Children
Atlanta
S75
2016
10
226744
142527
107238
85791
70739
60489
All Children
Atlanta
S75
2016
15
85367
47516
32484
25039
19309
15798
All Children
Atlanta
S75
2016
20
39385
19370
12590
9019
6840
5306
All Children
Atlanta
S75
2017
10
175435
105584
75824
59057
47839
40252
All Children
Atlanta
S75
2017
15
56918
30103
20580
15294
11803
9221
All Children
Atlanta
S75
2017
20
23728
11561
7445
5145
3632
2744
All Children
Boston
S65
2015
10
142102
77729
54110
40116
31196
25713
All Children
Boston
S65
2015
15
41823
18909
11855
7850
5529
4210
All Children
Boston
S65
2015
20
15746
5825
3368
2207
1365
956
All Children
Boston
S65
2016
10
144536
79345
54884
40594
32061
25804
All Children
Boston
S65
2016
15
41686
20138
12606
8851
6508
4960
All Children
Boston
S65
2016
20
16474
6713
3868
2435
1752
1206
All Children
Boston
S65
2017
10
155777
83827
56295
41117
32152
26259
All Children
Boston
S65
2017
15
51380
23141
13994
9352
6508
4505
All Children
Boston
S65
2017
20
21435
7668
3868
2480
1661
1115
All Children
Boston
S70
2015
10
168747
94590
66762
51152
40685
33267
All Children
Boston
S70
2015
15
55225
25849
16998
11832
8738
6553
All Children
Boston
S70
2015
20
22368
9079
5234
3390
2480
1616
All Children
Boston
S70
2016
10
177849
99960
70061
52972
41709
34086
All Children
Boston
S70
2016
15
56932
28853
18454
13425
9830
7691
All Children
Boston
S70
2016
20
24461
10672
6030
4050
2890
2230
All Children
Boston
S70
2017
10
192821
107742
73019
54929
42187
34223
All Children
Boston
S70
2017
15
68218
32471
20502
14358
10331
7850
All Children
Boston
S70
2017
20
31924
12697
6485
4323
2981
2048
3D-Attachment4-64
-------
Study Group
Study Area
AQ
Scenario
Year
FEV,
(percent)
Number of People at or above FEVt Decrement
>1 Day
> 2 Days
> 3 Days
> 4 Days
> 5 Days
> 6 Days
All Children
Boston
S75
2015
10
186109
106309
75021
57409
45736
38023
All Children
Boston
S75
2015
15
64054
30969
20502
14654
10786
8328
All Children
Boston
S75
2015
20
26827
11491
7031
4278
3095
2321
All Children
Boston
S75
2016
10
199238
114159
80437
62506
48990
39661
All Children
Boston
S75
2016
15
68673
35156
23050
16656
12265
9602
All Children
Boston
S75
2016
20
30514
13857
8214
5484
3891
2822
All Children
Boston
S75
2017
10
217328
123124
84783
63849
49354
39547
All Children
Boston
S75
2017
15
80960
39820
25417
17384
13107
9989
All Children
Boston
S75
2017
20
39684
16702
9375
5893
3937
2662
All Children
Dallas
S65
2015
10
184931
109857
77865
59799
48355
40103
All Children
Dallas
S65
2015
15
58735
30621
20146
14258
10853
8489
All Children
Dallas
S65
2015
20
23362
10451
6242
4091
2885
2270
All Children
Dallas
S65
2016
10
158708
92406
66018
50956
41261
33600
All Children
Dallas
S65
2016
15
43957
20832
14400
10428
8087
6573
All Children
Dallas
S65
2016
20
16717
7236
3902
2790
1844
1395
All Children
Dallas
S65
2017
10
183891
109739
78621
61100
50128
41238
All Children
Dallas
S65
2017
15
57482
29084
19247
13738
10617
8796
All Children
Dallas
S65
2017
20
23196
10499
6857
4469
3074
2365
All Children
Dallas
S70
2015
10
222622
134637
98578
77108
62660
52209
All Children
Dallas
S70
2015
15
76942
42869
28587
21352
16812
13123
All Children
Dallas
S70
2015
20
34853
16410
10215
7094
5013
4067
All Children
Dallas
S70
2016
10
185830
109975
80300
62140
50625
42207
All Children
Dallas
S70
2016
15
57317
28658
19484
14282
11232
9174
All Children
Dallas
S70
2016
20
22747
10475
6597
4540
3310
2246
All Children
Dallas
S70
2017
10
213211
130192
95173
75878
61833
51760
All Children
Dallas
S70
2017
15
72473
38968
25986
19271
14707
11752
All Children
Dallas
S70
2017
20
31165
14329
9198
6668
4753
3760
All Children
Dallas
S75
2015
10
257783
161049
118582
94062
77628
65380
All Children
Dallas
S75
2015
15
96899
56182
38637
28753
22794
18183
All Children
Dallas
S75
2015
20
45612
23527
15393
10428
7921
6195
All Children
Dallas
S75
2016
10
211343
127142
93305
72828
60367
50861
All Children
Dallas
S75
2016
15
70274
37005
25230
18562
14873
11941
All Children
Dallas
S75
2016
20
29770
13785
8796
6171
4635
3570
All Children
Dallas
S75
2017
10
240285
148494
111110
88718
73230
61927
All Children
Dallas
S75
2017
15
86992
48000
32252
24449
19035
15393
All Children
Dallas
S75
2017
20
39819
19295
12154
8938
7023
5391
All Children
Detroit
S65
2015
10
124524
70240
49393
37392
29483
24523
All Children
Detroit
S65
2015
15
37045
18592
11707
7943
6018
4527
All Children
Detroit
S65
2015
20
14117
6174
3399
2012
1474
1075
All Children
Detroit
S65
2016
10
144399
83143
58585
44364
35710
29778
All Children
Detroit
S65
2016
15
46514
23205
14828
10510
7995
6174
All Children
Detroit
S65
2016
20
18696
8186
4891
3278
2255
1544
All Children
Detroit
S65
2017
10
133039
77108
54561
42525
34322
28079
All Children
Detroit
S65
2017
15
40548
20829
13649
9920
7718
5931
All Children
Detroit
S65
2017
20
16979
7683
4475
2896
2168
1665
3D-Attachment4-65
-------
Study Group
Study Area
AQ
Scenario
Year
FEV,
(percent)
Number of People at or above FEVt Decrement
>1 Day
> 2 Days
> 3 Days
> 4 Days
> 5 Days
> 6 Days
All Children
Detroit
S70
2015
10
149480
87982
63337
48613
38883
32067
All Children
Detroit
S70
2015
15
50434
26015
17326
12574
9348
7458
All Children
Detroit
S70
2015
20
21037
9729
5689
3885
2706
1908
All Children
Detroit
S70
2016
10
176362
107267
76709
59001
47659
40028
All Children
Detroit
S70
2016
15
63632
33819
22928
16025
12504
9816
All Children
Detroit
S70
2016
20
28547
13302
8394
5584
4041
3052
All Children
Detroit
S70
2017
10
162228
97798
69650
54561
44121
36715
All Children
Detroit
S70
2017
15
56018
30004
19910
14759
11533
9053
All Children
Detroit
S70
2017
20
24541
11897
7458
5047
3659
2567
All Children
Detroit
S75
2015
10
168471
100070
73101
56157
45734
37531
All Children
Detroit
S75
2015
15
61048
31929
21922
16129
12123
9747
All Children
Detroit
S75
2015
20
26587
12574
7562
5134
3746
2740
All Children
Detroit
S75
2016
10
202568
126674
91797
71749
58082
48387
All Children
Detroit
S75
2016
15
80143
44017
29986
22130
17100
13649
All Children
Detroit
S75
2016
20
38415
18991
12106
8221
6001
4665
All Children
Detroit
S75
2017
10
185519
113355
82189
64621
52342
43601
All Children
Detroit
S75
2017
15
68679
38727
25703
18522
14516
11811
All Children
Detroit
S75
2017
20
32119
16112
10389
7197
5238
3885
All Children
Philadelphia
S65
2015
10
164741
96099
67878
51749
41556
34419
All Children
Philadelphia
S65
2015
15
46620
23834
15693
11197
8447
6526
All Children
Philadelphia
S65
2015
20
17351
7595
4496
2968
2030
1484
All Children
Philadelphia
S65
2016
10
162035
94266
65084
49872
40749
33612
All Children
Philadelphia
S65
2016
15
45921
23244
15060
10760
8294
6810
All Children
Philadelphia
S65
2016
20
17526
7923
4365
2837
1899
1462
All Children
Philadelphia
S65
2017
10
150445
85099
60021
45267
36493
29356
All Children
Philadelphia
S65
2017
15
42648
20691
12877
9123
6701
5347
All Children
Philadelphia
S65
2017
20
16348
6831
3841
2532
1724
1331
All Children
Philadelphia
S70
2015
10
196192
118929
85841
66547
53888
44568
All Children
Philadelphia
S70
2015
15
62400
32913
22437
16544
12550
10258
All Children
Philadelphia
S70
2015
20
25449
11677
7159
5282
3754
2619
All Children
Philadelphia
S70
2016
10
193355
116397
82458
63404
52076
43586
All Children
Philadelphia
S70
2016
15
61221
32302
21935
15955
12681
9800
All Children
Philadelphia
S70
2016
20
25012
11677
7312
5020
3579
2488
All Children
Philadelphia
S70
2017
10
178688
104546
74775
56943
46358
37955
All Children
Philadelphia
S70
2017
15
56856
29508
19119
13576
10280
7857
All Children
Philadelphia
S70
2017
20
23615
10324
5653
4038
2903
2248
All Children
Philadelphia
S75
2015
10
237072
147695
109936
87085
71720
59824
All Children
Philadelphia
S75
2015
15
84204
46162
32499
24423
18923
15584
All Children
Philadelphia
S75
2015
20
37278
18814
11873
8599
6482
5085
All Children
Philadelphia
S75
2016
10
232750
145032
106575
83855
67987
57904
All Children
Philadelphia
S75
2016
15
83440
46074
32084
24161
18552
15016
All Children
Philadelphia
S75
2016
20
37628
18574
11742
8425
6242
4889
All Children
Philadelphia
S75
2017
10
216861
129929
94615
73793
59759
49828
All Children
Philadelphia
S75
2017
15
77525
42058
27042
19927
15453
12070
All Children
Philadelphia
S75
2017
20
35358
16326
9647
6548
4933
3776
3D-Attachment4-66
-------
Study Group
Study Area
AQ
Scenario
Year
FEV,
(percent)
Number of People at or above FEVt Decrement
>1 Day
> 2 Days
> 3 Days
> 4 Days
> 5 Days
> 6 Days
All Children
Phoenix
S65
2015
10
141180
90695
69549
55764
47541
40592
All Children
Phoenix
S65
2015
15
43974
25150
18145
14196
11436
9398
All Children
Phoenix
S65
2015
20
17423
9115
6015
4529
3482
2760
All Children
Phoenix
S65
2016
10
135773
87001
66676
54349
45390
39106
All Children
Phoenix
S65
2016
15
39856
23424
16970
13035
10601
9143
All Children
Phoenix
S65
2016
20
15484
8166
5463
4161
3298
2831
All Children
Phoenix
S65
2017
10
153889
101649
78027
63633
54306
47187
All Children
Phoenix
S65
2017
15
48588
29524
21230
16927
13899
11860
All Children
Phoenix
S65
2017
20
20041
11054
7459
5308
4232
3453
All Children
Phoenix
S70
2015
10
170378
112646
87680
72125
61256
52933
All Children
Phoenix
S70
2015
15
59798
36034
26070
20409
16956
14535
All Children
Phoenix
S70
2015
20
25844
14493
9851
7501
5959
4869
All Children
Phoenix
S70
2016
10
164589
108372
83873
69479
58906
50980
All Children
Phoenix
S70
2016
15
54688
32468
23990
18895
15781
13417
All Children
Phoenix
S70
2016
20
22773
12596
8860
6893
5746
4572
All Children
Phoenix
S70
2017
10
185182
125399
98889
81962
70399
61921
All Children
Phoenix
S70
2017
15
65827
41342
31251
24698
21060
17932
All Children
Phoenix
S70
2017
20
30302
17210
12384
9469
7530
6128
All Children
Phoenix
S75
2015
10
197312
132787
104721
87482
74135
65219
All Children
Phoenix
S75
2015
15
76117
47484
34930
27712
22872
19645
All Children
Phoenix
S75
2015
20
36105
20607
14351
11153
8747
7218
All Children
Phoenix
S75
2016
10
190306
128668
99908
83137
71389
62869
All Children
Phoenix
S75
2016
15
69167
42771
31619
25901
21456
18399
All Children
Phoenix
S75
2016
20
31576
18130
12894
10006
8152
6949
All Children
Phoenix
S75
2017
10
213036
148242
118449
99271
85684
75848
All Children
Phoenix
S75
2017
15
84439
53528
40889
33374
28080
24188
All Children
Phoenix
S75
2017
20
40549
24499
17762
14012
11337
9568
All Children
Sacramento
S65
2015
10
48758
27826
19938
15808
12896
10645
All Children
Sacramento
S65
2015
15
12919
6250
3804
2873
2205
1755
All Children
Sacramento
S65
2015
20
4193
1716
1040
714
528
373
All Children
Sacramento
S65
2016
10
51701
30303
21444
16406
13269
10986
All Children
Sacramento
S65
2016
15
13828
6972
4441
3253
2601
2112
All Children
Sacramento
S65
2016
20
4985
2244
1328
916
637
458
All Children
Sacramento
S65
2017
10
50614
29030
20132
15559
12640
10536
All Children
Sacramento
S65
2017
15
12896
6367
3898
2896
2073
1623
All Children
Sacramento
S65
2017
20
4255
1677
1017
637
435
334
All Children
Sacramento
S70
2015
10
64364
38712
28059
22182
18649
15862
All Children
Sacramento
S70
2015
15
20024
10559
7267
5311
4169
3269
All Children
Sacramento
S70
2015
20
7904
3540
2213
1599
1188
916
All Children
Sacramento
S70
2016
10
68293
41406
30466
24045
19767
16638
All Children
Sacramento
S70
2016
15
21871
11848
7896
5955
4705
3797
All Children
Sacramento
S70
2016
20
9286
4371
2756
1988
1467
1157
All Children
Sacramento
S70
2017
10
67066
40909
29030
23036
18975
15901
All Children
Sacramento
S70
2017
15
20101
10567
7275
5326
4169
3292
All Children
Sacramento
S70
2017
20
8230
3556
2135
1522
1149
815
3D-Attachment4-67
-------
Study Group
Study Area
AQ
Scenario
Year
FEV,
(percent)
Number of People at or above FEVt Decrement
>1 Day
> 2 Days
> 3 Days
> 4 Days
> 5 Days
> 6 Days
All Children
Sacramento
S75
2015
10
77253
48300
35521
28432
23649
20520
All Children
Sacramento
S75
2015
15
26693
14992
10497
7958
6165
5008
All Children
Sacramento
S75
2015
20
11755
5769
3758
2694
2073
1568
All Children
Sacramento
S75
2016
10
82439
51996
38867
31118
25746
21623
All Children
Sacramento
S75
2016
15
29504
16685
11871
8929
6995
5629
All Children
Sacramento
S75
2016
20
13354
6980
4410
3284
2547
2065
All Children
Sacramento
S75
2017
10
81057
50839
37586
29667
24511
20947
All Children
Sacramento
S75
2017
15
27539
15311
10567
8036
6250
5194
All Children
Sacramento
S75
2017
20
11957
5893
3711
2702
2026
1514
All Children
St. Louis
S65
2015
10
61187
34086
23668
18195
14543
11702
All Children
St. Louis
S65
2015
15
17066
8542
5555
4062
2951
2249
All Children
St. Louis
S65
2015
20
6338
2832
1776
1157
892
583
All Children
St. Louis
S65
2016
10
71851
42218
30070
23149
18696
15508
All Children
St. Louis
S65
2016
15
22839
11584
7750
5418
3980
2941
All Children
St. Louis
S65
2016
20
9243
3779
2195
1357
1020
719
All Children
St. Louis
S65
2017
10
69365
40633
28822
21901
17621
14680
All Children
St. Louis
S65
2017
15
19716
10081
6548
4908
3816
3023
All Children
St. Louis
S65
2017
20
7668
3542
2268
1421
1047
838
All Children
St. Louis
S70
2015
10
76632
44731
31691
24579
19861
16574
All Children
St. Louis
S70
2015
15
24351
12704
8587
6356
4963
3843
All Children
St. Louis
S70
2015
20
10309
4826
3078
2158
1557
1175
All Children
St. Louis
S70
2016
10
89017
54612
40087
31582
25517
21382
All Children
St. Louis
S70
2016
15
32356
17175
11729
8587
6739
5282
All Children
St. Louis
S70
2016
20
14507
6775
4180
2732
2013
1475
All Children
St. Louis
S70
2017
10
87368
53055
38393
30352
24742
20390
All Children
St. Louis
S70
2017
15
28303
15508
10409
7932
6047
4918
All Children
St. Louis
S70
2017
20
11738
5783
3907
2623
1994
1548
All Children
St. Louis
S75
2015
10
88406
53228
38621
30452
24697
20526
All Children
St. Louis
S75
2015
15
31172
16610
11319
8515
6739
5282
All Children
St. Louis
S75
2015
20
13724
6848
4462
3087
2176
1694
All Children
St. Louis
S75
2016
10
102931
65203
48420
38430
31627
26600
All Children
St. Louis
S75
2016
15
40724
22557
15417
11638
9143
7422
All Children
St. Louis
S75
2016
20
19433
9744
6338
4380
3151
2486
All Children
St. Louis
S75
2017
10
101192
63691
46890
37055
30625
25726
All Children
St. Louis
S75
2017
15
36080
20289
13833
10536
8405
6739
All Children
St. Louis
S75
2017
20
15845
8269
5409
3779
2941
2359
Asthma Children
Atlanta
S65
2015
10
15415
8030
5306
3975
2905
2058
Asthma Children
Atlanta
S65
2015
15
3712
1574
1009
625
424
282
Asthma Children
Atlanta
S65
2015
20
1271
363
222
182
161
101
Asthma Children
Atlanta
S65
2016
10
18966
10835
7869
6154
5044
4298
Asthma Children
Atlanta
S65
2016
15
5811
2825
1937
1513
1130
948
Asthma Children
Atlanta
S65
2016
20
2300
1009
545
343
282
202
Asthma Children
Atlanta
S65
2017
10
13619
7465
5165
3773
3127
2522
Asthma Children
Atlanta
S65
2017
15
3612
1796
1150
807
605
343
Asthma Children
Atlanta
S65
2017
20
1190
464
262
202
202
101
3D-Attachment4-68
-------
Study Group
Study Area
AQ
Scenario
Year
FEV,
(percent)
Number of People at or above FEVt Decrement
>1 Day
> 2 Days
> 3 Days
> 4 Days
> 5 Days
> 6 Days
Asthma Children
Atlanta
S70
2015
10
19834
11218
7607
5750
4661
3652
Asthma Children
Atlanta
S70
2015
15
5831
2583
1776
1291
948
706
Asthma Children
Atlanta
S70
2015
20
2199
807
484
242
161
141
Asthma Children
Atlanta
S70
2016
10
23929
14608
10673
8636
7203
5972
Asthma Children
Atlanta
S70
2016
15
8434
4681
3127
2421
1876
1614
Asthma Children
Atlanta
S70
2016
20
3733
1715
1130
847
625
484
Asthma Children
Atlanta
S70
2017
10
17776
9725
7183
5710
4580
3672
Asthma Children
Atlanta
S70
2017
15
4802
2361
1836
1352
1049
747
Asthma Children
Atlanta
S70
2017
20
2078
847
484
262
242
222
Asthma Children
Atlanta
S75
2015
10
24393
14446
10593
8272
6658
5468
Asthma Children
Atlanta
S75
2015
15
8615
4298
2845
2058
1634
1291
Asthma Children
Atlanta
S75
2015
20
3390
1453
928
686
444
262
Asthma Children
Atlanta
S75
2016
10
28994
18764
14083
11501
9584
8071
Asthma Children
Atlanta
S75
2016
15
11723
6638
4661
3672
2986
2583
Asthma Children
Atlanta
S75
2016
20
5448
2724
1876
1392
1090
888
Asthma Children
Atlanta
S75
2017
10
21730
12994
9241
7445
6134
5044
Asthma Children
Atlanta
S75
2017
15
6820
3672
2562
1997
1574
1251
Asthma Children
Atlanta
S75
2017
20
3087
1554
928
565
424
282
Asthma Children
Boston
S65
2015
10
18499
10353
7418
5438
4323
3663
Asthma Children
Boston
S65
2015
15
5416
2366
1707
1069
728
592
Asthma Children
Boston
S65
2015
20
2184
910
432
341
250
159
Asthma Children
Boston
S65
2016
10
19751
10217
7463
5939
4665
3823
Asthma Children
Boston
S65
2016
15
5643
3095
2162
1502
1001
683
Asthma Children
Boston
S65
2016
20
2298
956
523
319
182
137
Asthma Children
Boston
S65
2017
10
20047
11286
7463
5552
4278
3322
Asthma Children
Boston
S65
2017
15
6963
2822
1707
1320
887
592
Asthma Children
Boston
S65
2017
20
2457
1024
569
319
296
228
Asthma Children
Boston
S70
2015
10
21366
12788
9125
6849
5598
4665
Asthma Children
Boston
S70
2015
15
7259
3390
2298
1525
1251
933
Asthma Children
Boston
S70
2015
20
2844
1434
796
432
341
228
Asthma Children
Boston
S70
2016
10
24165
13243
9375
7350
6075
4847
Asthma Children
Boston
S70
2016
15
7691
4210
2913
2230
1684
1229
Asthma Children
Boston
S70
2016
20
3368
1570
887
501
364
228
Asthma Children
Boston
S70
2017
10
24529
14267
9898
7372
5734
4619
Asthma Children
Boston
S70
2017
15
9056
3937
2366
1707
1365
1092
Asthma Children
Boston
S70
2017
20
3959
1547
887
592
410
296
Asthma Children
Boston
S75
2015
10
23483
13835
10331
7759
6235
5347
Asthma Children
Boston
S75
2015
15
8146
4119
2640
1934
1525
1092
Asthma Children
Boston
S75
2015
20
3413
1661
1001
546
455
319
Asthma Children
Boston
S75
2016
10
26827
14972
10626
8920
6986
5552
Asthma Children
Boston
S75
2016
15
9284
5006
3527
2708
2116
1570
Asthma Children
Boston
S75
2016
20
4187
2002
1229
751
523
364
Asthma Children
Boston
S75
2017
10
27965
16429
11332
8669
6599
5438
Asthma Children
Boston
S75
2017
15
10626
5165
3004
2071
1616
1388
Asthma Children
Boston
S75
2017
20
5256
2048
1229
774
478
341
3D-Attachment4-69
-------
Study Group
Study Area
AQ
Scenario
Year
FEV,
(percent)
Number of People at or above FEVt Decrement
>1 Day
> 2 Days
> 3 Days
> 4 Days
> 5 Days
> 6 Days
Asthma Children
Dallas
S65
2015
10
18869
11042
8441
6361
5155
4398
Asthma Children
Dallas
S65
2015
15
6030
3050
2010
1348
1111
969
Asthma Children
Dallas
S65
2015
20
2530
1111
686
426
378
307
Asthma Children
Dallas
S65
2016
10
15393
8985
6881
5249
4280
3452
Asthma Children
Dallas
S65
2016
15
4753
2435
1537
1040
828
567
Asthma Children
Dallas
S65
2016
20
1750
615
284
236
142
95
Asthma Children
Dallas
S65
2017
10
18443
11539
8087
6361
5533
4540
Asthma Children
Dallas
S65
2017
15
5911
3405
2270
1726
1419
1159
Asthma Children
Dallas
S65
2017
20
2719
1301
828
520
355
307
Asthma Children
Dallas
S70
2015
10
22298
13785
10309
8158
6715
5509
Asthma Children
Dallas
S70
2015
15
7779
4587
2979
1986
1655
1301
Asthma Children
Dallas
S70
2015
20
3689
1679
1040
851
567
473
Asthma Children
Dallas
S70
2016
10
17781
10570
7921
6432
5438
4374
Asthma Children
Dallas
S70
2016
15
5911
3121
2104
1442
1111
875
Asthma Children
Dallas
S70
2016
20
2388
1017
615
426
260
142
Asthma Children
Dallas
S70
2017
10
21376
13241
9647
7779
6597
5557
Asthma Children
Dallas
S70
2017
15
7590
4445
2932
2128
1797
1513
Asthma Children
Dallas
S70
2017
20
3547
1773
1230
828
567
473
Asthma Children
Dallas
S75
2015
10
25915
16221
12438
9624
8110
6857
Asthma Children
Dallas
S75
2015
15
9931
5651
4138
3192
2365
1821
Asthma Children
Dallas
S75
2015
20
4682
2317
1466
993
851
733
Asthma Children
Dallas
S75
2016
10
20406
12319
9080
7212
6242
5202
Asthma Children
Dallas
S75
2016
15
7354
3902
2696
2104
1537
1230
Asthma Children
Dallas
S75
2016
20
3168
1395
875
615
402
260
Asthma Children
Dallas
S75
2017
10
23835
15015
11184
9056
7567
6479
Asthma Children
Dallas
S75
2017
15
8914
5084
3783
2956
2246
1892
Asthma Children
Dallas
S75
2017
20
4398
2341
1537
1111
899
686
Asthma Children
Detroit
S65
2015
10
14811
8741
6417
4856
3642
3087
Asthma Children
Detroit
S65
2015
15
4908
2237
1492
1075
798
590
Asthma Children
Detroit
S65
2015
20
1700
780
451
260
173
87
Asthma Children
Detroit
S65
2016
10
17985
10649
7475
5567
4422
3850
Asthma Children
Detroit
S65
2016
15
6122
3295
2046
1353
1058
798
Asthma Children
Detroit
S65
2016
20
2567
1197
746
382
243
191
Asthma Children
Detroit
S65
2017
10
16441
9782
6868
5376
4301
3469
Asthma Children
Detroit
S65
2017
15
5238
2549
1734
1283
1041
746
Asthma Children
Detroit
S65
2017
20
2133
1058
624
382
277
225
Asthma Children
Detroit
S70
2015
10
17603
10649
8134
6226
4839
3989
Asthma Children
Detroit
S70
2015
15
6521
3295
2046
1596
1127
902
Asthma Children
Detroit
S70
2015
20
2653
1214
798
468
382
156
Asthma Children
Detroit
S70
2016
10
21662
13510
9938
7423
5827
5047
Asthma Children
Detroit
S70
2016
15
8151
4544
3052
2151
1682
1335
Asthma Children
Detroit
S70
2016
20
3902
1821
1145
780
468
382
Asthma Children
Detroit
S70
2017
10
19841
11949
8672
6781
5723
4683
Asthma Children
Detroit
S70
2017
15
7007
3711
2497
1960
1613
1179
Asthma Children
Detroit
S70
2017
20
3035
1509
1006
590
468
347
3D-Attachment4-70
-------
Study Group
Study Area
AQ
Scenario
Year
FEV,
(percent)
Number of People at or above FEVt Decrement
>1 Day
> 2 Days
> 3 Days
> 4 Days
> 5 Days
> 6 Days
Asthma Children
Detroit
S75
2015
10
20083
12088
9157
7128
5706
4613
Asthma Children
Detroit
S75
2015
15
7666
4006
2688
1942
1509
1162
Asthma Children
Detroit
S75
2015
20
3295
1509
989
607
503
347
Asthma Children
Detroit
S75
2016
10
24575
15539
11672
9070
7249
5845
Asthma Children
Detroit
S75
2016
15
10319
5584
4024
3000
2203
1925
Asthma Children
Detroit
S75
2016
20
5064
2740
1630
1075
798
572
Asthma Children
Detroit
S75
2017
10
22355
13857
9972
7995
6799
5567
Asthma Children
Detroit
S75
2017
15
8654
4856
3330
2428
1942
1578
Asthma Children
Detroit
S75
2017
20
3972
1942
1422
989
624
486
Asthma Children
Philadelphia
S65
2015
10
19294
12135
8294
6111
4780
4038
Asthma Children
Philadelphia
S65
2015
15
5478
2706
2008
1528
1135
960
Asthma Children
Philadelphia
S65
2015
20
2139
1069
611
415
284
175
Asthma Children
Philadelphia
S65
2016
10
18683
10804
7181
5195
4169
3558
Asthma Children
Philadelphia
S65
2016
15
4758
2488
1659
1179
851
611
Asthma Children
Philadelphia
S65
2016
20
1877
808
437
262
196
175
Asthma Children
Philadelphia
S65
2017
10
18355
10411
7377
5937
4693
3710
Asthma Children
Philadelphia
S65
2017
15
5195
2706
1659
1244
982
786
Asthma Children
Philadelphia
S65
2017
20
2357
917
567
371
153
109
Asthma Children
Philadelphia
S70
2015
10
23135
14449
10651
8272
6439
5173
Asthma Children
Philadelphia
S70
2015
15
7923
3841
2859
2183
1812
1375
Asthma Children
Philadelphia
S70
2015
20
2903
1375
960
786
589
349
Asthma Children
Philadelphia
S70
2016
10
22175
13816
9669
7093
5435
4649
Asthma Children
Philadelphia
S70
2016
15
6722
3318
2292
1790
1310
1048
Asthma Children
Philadelphia
S70
2016
20
2575
1091
698
458
306
240
Asthma Children
Philadelphia
S70
2017
10
21869
12572
9320
7181
6002
4867
Asthma Children
Philadelphia
S70
2017
15
6701
3820
2488
1746
1484
1157
Asthma Children
Philadelphia
S70
2017
20
3099
1484
720
546
393
327
Asthma Children
Philadelphia
S75
2015
10
27850
17766
13510
10826
8796
7159
Asthma Children
Philadelphia
S75
2015
15
10280
5347
3841
2925
2401
2117
Asthma Children
Philadelphia
S75
2015
20
4343
2226
1484
1069
917
742
Asthma Children
Philadelphia
S75
2016
10
26911
17199
12593
9669
7639
6395
Asthma Children
Philadelphia
S75
2016
15
9712
4823
3427
2575
2052
1637
Asthma Children
Philadelphia
S75
2016
20
4147
1855
1310
873
502
458
Asthma Children
Philadelphia
S75
2017
10
26191
15889
11677
9276
7508
6199
Asthma Children
Philadelphia
S75
2017
15
9516
5282
3470
2510
2183
1746
Asthma Children
Philadelphia
S75
2017
20
4452
2204
1353
939
720
524
Asthma Children
Phoenix
S65
2015
10
15215
9992
7685
6086
5223
4444
Asthma Children
Phoenix
S65
2015
15
4982
2972
2066
1656
1373
1118
Asthma Children
Phoenix
S65
2015
20
1882
1090
722
538
467
368
Asthma Children
Phoenix
S65
2016
10
13785
9058
7119
5817
4982
4119
Asthma Children
Phoenix
S65
2016
15
4331
2717
1840
1514
1189
1005
Asthma Children
Phoenix
S65
2016
20
1727
920
580
396
340
269
Asthma Children
Phoenix
S65
2017
10
16064
10502
8166
6666
5732
4925
Asthma Children
Phoenix
S65
2017
15
5223
3255
2392
1882
1557
1373
Asthma Children
Phoenix
S65
2017
20
2335
1231
977
736
637
538
3D-Attachment4-71
-------
Study Group
Study Area
AQ
Scenario
Year
FEV,
(percent)
Number of People at or above FEVt Decrement
>1 Day
> 2 Days
> 3 Days
> 4 Days
> 5 Days
> 6 Days
Asthma Children
Phoenix
S70
2015
10
18201
12200
9539
7940
6695
5746
Asthma Children
Phoenix
S70
2015
15
6836
4161
2930
2293
1911
1571
Asthma Children
Phoenix
S70
2015
20
2930
1684
1104
878
708
566
Asthma Children
Phoenix
S70
2016
10
16460
11025
8846
7445
6298
5506
Asthma Children
Phoenix
S70
2016
15
6114
3552
2689
2081
1783
1543
Asthma Children
Phoenix
S70
2016
20
2434
1429
991
750
651
538
Asthma Children
Phoenix
S70
2017
10
18994
13049
10346
8534
7331
6341
Asthma Children
Phoenix
S70
2017
15
7119
4331
3354
2703
2307
2024
Asthma Children
Phoenix
S70
2017
20
3199
1996
1429
1033
934
807
Asthma Children
Phoenix
S75
2015
10
20735
14139
11294
9539
8294
7289
Asthma Children
Phoenix
S75
2015
15
8492
5378
3935
3114
2548
2137
Asthma Children
Phoenix
S75
2015
20
4331
2307
1613
1203
934
793
Asthma Children
Phoenix
S75
2016
10
18937
12964
10473
8648
7530
6610
Asthma Children
Phoenix
S75
2016
15
7473
4727
3482
2873
2349
2052
Asthma Children
Phoenix
S75
2016
20
3383
1953
1401
1146
948
764
Asthma Children
Phoenix
S75
2017
10
21513
15328
12214
10219
8874
7813
Asthma Children
Phoenix
S75
2017
15
8931
5506
4444
3666
3057
2604
Asthma Children
Phoenix
S75
2017
20
4303
2802
2052
1656
1288
1132
Asthma Children
Sacramento
S65
2015
10
4891
2725
1988
1630
1328
1126
Asthma Children
Sacramento
S65
2015
15
1250
567
303
248
194
124
Asthma Children
Sacramento
S65
2015
20
334
124
70
31
23
16
Asthma Children
Sacramento
S65
2016
10
5256
3075
2189
1731
1429
1211
Asthma Children
Sacramento
S65
2016
15
1413
745
528
427
365
280
Asthma Children
Sacramento
S65
2016
20
590
272
163
124
78
47
Asthma Children
Sacramento
S65
2017
10
5039
2896
1933
1460
1219
986
Asthma Children
Sacramento
S65
2017
15
1242
551
349
287
248
194
Asthma Children
Sacramento
S65
2017
20
419
194
140
70
54
31
Asthma Children
Sacramento
S70
2015
10
6328
3773
2686
2174
1871
1638
Asthma Children
Sacramento
S70
2015
15
1988
1048
714
520
435
280
Asthma Children
Sacramento
S70
2015
20
699
233
155
140
101
70
Asthma Children
Sacramento
S70
2016
10
6972
4294
3113
2453
2096
1731
Asthma Children
Sacramento
S70
2016
15
2244
1320
831
683
551
466
Asthma Children
Sacramento
S70
2016
20
978
536
373
303
225
163
Asthma Children
Sacramento
S70
2017
10
6576
4053
2873
2244
1794
1506
Asthma Children
Sacramento
S70
2017
15
1941
994
629
481
388
311
Asthma Children
Sacramento
S70
2017
20
745
334
225
179
163
132
Asthma Children
Sacramento
S75
2015
10
7477
4635
3432
2795
2376
2057
Asthma Children
Sacramento
S75
2015
15
2733
1530
1040
792
582
474
Asthma Children
Sacramento
S75
2015
20
1064
551
311
210
155
101
Asthma Children
Sacramento
S75
2016
10
8440
5427
4006
3238
2725
2244
Asthma Children
Sacramento
S75
2016
15
2950
1747
1250
955
784
637
Asthma Children
Sacramento
S75
2016
20
1312
745
520
396
334
303
Asthma Children
Sacramento
S75
2017
10
7772
5047
3758
2927
2368
1980
Asthma Children
Sacramento
S75
2017
15
2686
1522
978
691
543
466
Asthma Children
Sacramento
S75
2017
20
1149
505
349
264
225
194
3D-Attachment4-72
-------
Study Group
Study Area
AQ
Scenario
Year
FEV,
(percent)
Number of People at or above FEVt Decrement
>1 Day
> 2 Days
> 3 Days
> 4 Days
> 5 Days
> 6 Days
Asthma Children
St. Louis
S65
2015
10
6730
3843
2659
2158
1794
1412
Asthma Children
St. Louis
S65
2015
15
1821
993
665
464
346
282
Asthma Children
St. Louis
S65
2015
20
729
337
219
146
109
73
Asthma Children
St. Louis
S65
2016
10
7149
4171
3133
2431
2022
1758
Asthma Children
St. Louis
S65
2016
15
2340
1202
920
610
474
301
Asthma Children
St. Louis
S65
2016
20
1002
410
237
137
100
91
Asthma Children
St. Louis
S65
2017
10
7422
4235
2987
2441
2031
1685
Asthma Children
St. Louis
S65
2017
15
2204
1120
738
528
410
319
Asthma Children
St. Louis
S65
2017
20
838
355
246
164
127
82
Asthma Children
St. Louis
S70
2015
10
8278
4863
3624
2741
2249
1958
Asthma Children
St. Louis
S70
2015
15
2650
1421
956
747
583
455
Asthma Children
St. Louis
S70
2015
20
1157
501
346
228
191
164
Asthma Children
St. Louis
S70
2016
10
9043
5437
4034
3151
2705
2277
Asthma Children
St. Louis
S70
2016
15
3206
1721
1275
965
829
647
Asthma Children
St. Louis
S70
2016
20
1512
747
474
319
200
182
Asthma Children
St. Louis
S70
2017
10
9234
5509
3961
3096
2614
2222
Asthma Children
St. Louis
S70
2017
15
3169
1739
1166
865
610
519
Asthma Children
St. Louis
S70
2017
20
1311
610
446
319
219
173
Asthma Children
St. Louis
S75
2015
10
9371
5628
4289
3315
2705
2349
Asthma Children
St. Louis
S75
2015
15
3442
1921
1320
1029
783
592
Asthma Children
St. Louis
S75
2015
20
1503
783
455
337
246
209
Asthma Children
St. Louis
S75
2016
10
10400
6511
4763
3743
3160
2668
Asthma Children
St. Louis
S75
2016
15
4025
2277
1603
1311
1065
865
Asthma Children
St. Louis
S75
2016
20
1994
1084
738
501
346
264
Asthma Children
St. Louis
S75
2017
10
10591
6602
4836
3770
3151
2759
Asthma Children
St. Louis
S75
2017
15
3934
2186
1585
1211
938
701
Asthma Children
St. Louis
S75
2017
20
1785
956
628
419
328
246
All Adults
Atlanta
S65
2015
10
75293
32047
19440
12960
8663
5846
All Adults
Atlanta
S65
2015
15
19228
6480
2676
1831
1268
1057
All Adults
Atlanta
S65
2015
20
7325
2324
1127
704
423
211
All Adults
Atlanta
S65
2016
10
89662
43387
26835
18665
14016
11340
All Adults
Atlanta
S65
2016
15
24088
10565
6550
4015
3099
2465
All Adults
Atlanta
S65
2016
20
9649
3522
2043
1409
1127
916
All Adults
Atlanta
S65
2017
10
74941
32118
20355
13101
8522
6691
All Adults
Atlanta
S65
2017
15
17608
6198
3029
1902
1127
845
All Adults
Atlanta
S65
2017
20
5564
1550
634
211
141
141
All Adults
Atlanta
S70
2015
10
101142
43528
27046
18947
13594
9861
All Adults
Atlanta
S70
2015
15
27187
10354
5494
3029
1831
1268
All Adults
Atlanta
S70
2015
20
10988
3451
1620
1268
986
704
All Adults
Atlanta
S70
2016
10
117483
58037
37752
26553
20426
16693
All Adults
Atlanta
S70
2016
15
35639
15495
9790
7325
5212
4015
All Adults
Atlanta
S70
2016
20
15284
6198
3592
2395
1550
1338
All Adults
Atlanta
S70
2017
10
97903
44444
27539
20496
13735
10283
All Adults
Atlanta
S70
2017
15
25779
10424
5423
3522
1902
1268
All Adults
Atlanta
S70
2017
20
8804
3029
1338
634
141
141
3D-Attachment4-73
-------
Study Group
Study Area
AQ
Scenario
Year
FEV,
(percent)
Number of People at or above FEVt Decrement
>1 Day
> 2 Days
> 3 Days
> 4 Days
> 5 Days
> 6 Days
All Adults
Atlanta
S75
2015
10
129598
60361
37048
25849
19228
14791
All Adults
Atlanta
S75
2015
15
39513
16200
9156
5635
3522
2113
All Adults
Atlanta
S75
2015
20
16974
5283
2536
1690
1338
986
All Adults
Atlanta
S75
2016
10
153404
76491
49726
36414
27610
22116
All Adults
Atlanta
S75
2016
15
49796
22116
14298
9790
7536
5635
All Adults
Atlanta
S75
2016
20
22046
9790
6128
3733
2606
1761
All Adults
Atlanta
S75
2017
10
121286
57333
35498
26342
19017
14791
All Adults
Atlanta
S75
2017
15
35146
14650
8522
5705
3592
2465
All Adults
Atlanta
S75
2017
20
13805
5212
2606
1550
634
493
All Adults
Boston
S65
2015
10
104195
46374
26611
17904
13599
10958
All Adults
Boston
S65
2015
15
25535
9881
5577
2642
1761
1272
All Adults
Boston
S65
2015
20
9294
2935
1370
587
391
294
All Adults
Boston
S65
2016
10
114272
50287
32286
21426
16828
12523
All Adults
Boston
S65
2016
15
31209
14186
7729
4598
2935
2055
All Adults
Boston
S65
2016
20
13893
4892
1957
978
783
489
All Adults
Boston
S65
2017
10
123175
51951
30818
21426
15849
12229
All Adults
Boston
S65
2017
15
30427
11740
7142
4990
3326
2348
All Adults
Boston
S65
2017
20
12816
4500
2739
1663
881
587
All Adults
Boston
S70
2015
10
121022
55473
34047
22111
16339
12816
All Adults
Boston
S70
2015
15
32286
12621
6946
4403
2544
1957
All Adults
Boston
S70
2015
20
12229
4305
2152
1174
881
587
All Adults
Boston
S70
2016
10
136285
60756
39917
28079
21622
16436
All Adults
Boston
S70
2016
15
39036
17513
10958
6457
4109
2642
All Adults
Boston
S70
2016
20
18295
6164
3131
1859
881
685
All Adults
Boston
S70
2017
10
149395
63495
37275
25731
19763
15752
All Adults
Boston
S70
2017
15
41874
16143
9784
5968
4696
3424
All Adults
Boston
S70
2017
20
17806
5968
3620
2544
1468
881
All Adults
Boston
S75
2015
10
133448
60169
36884
23774
18882
15262
All Adults
Boston
S75
2015
15
37667
14969
8120
5087
3033
2152
All Adults
Boston
S75
2015
20
14773
5381
2544
1370
881
587
All Adults
Boston
S75
2016
10
152036
66430
43830
31405
23676
18882
All Adults
Boston
S75
2016
15
44319
19763
12229
7631
5577
3424
All Adults
Boston
S75
2016
20
21132
7925
4109
2446
1468
587
All Adults
Boston
S75
2017
10
169060
72203
41776
29448
22502
17708
All Adults
Boston
S75
2017
15
49603
18491
11447
7142
5185
4109
All Adults
Boston
S75
2017
20
20741
7435
4305
3326
2055
1272
All Adults
Dallas
S65
2015
10
118845
55008
36177
23832
17346
12033
All Adults
Dallas
S65
2015
15
33677
13752
7110
4766
3360
2266
All Adults
Dallas
S65
2015
20
12346
3751
2032
1094
703
469
All Adults
Dallas
S65
2016
10
101734
46804
29067
18675
14611
10861
All Adults
Dallas
S65
2016
15
26488
10548
6095
4141
3125
2188
All Adults
Dallas
S65
2016
20
10001
3907
2110
938
625
391
All Adults
Dallas
S65
2017
10
118142
52898
33286
22972
16721
12814
All Adults
Dallas
S65
2017
15
31020
11173
6329
4454
2735
1797
All Adults
Dallas
S65
2017
20
11408
3751
1875
1328
781
469
3D-Attachment4-74
-------
Study Group
Study Area
AQ
Scenario
Year
FEV,
(percent)
Number of People at or above FEVt Decrement
>1 Day
> 2 Days
> 3 Days
> 4 Days
> 5 Days
> 6 Days
All Adults
Dallas
S70
2015
10
144083
68916
45397
31801
23363
17424
All Adults
Dallas
S70
2015
15
44694
18675
10627
6563
5235
3672
All Adults
Dallas
S70
2015
20
19222
6720
3360
2422
1406
781
All Adults
Dallas
S70
2016
10
119705
57118
34380
23519
17815
13986
All Adults
Dallas
S70
2016
15
33442
13908
7579
4923
3907
2735
All Adults
Dallas
S70
2016
20
13127
5391
2891
1485
1172
781
All Adults
Dallas
S70
2017
10
138145
64853
41569
29223
20862
16174
All Adults
Dallas
S70
2017
15
40553
15705
8048
5704
3594
2500
All Adults
Dallas
S70
2017
20
14690
5626
3516
1563
1094
781
All Adults
Dallas
S75
2015
10
170650
81652
54617
38912
29457
22347
All Adults
Dallas
S75
2015
15
56024
24457
13752
9064
6642
5079
All Adults
Dallas
S75
2015
20
25394
9142
4844
3047
2188
1250
All Adults
Dallas
S75
2016
10
137364
65635
40631
28520
21956
16956
All Adults
Dallas
S75
2016
15
40475
16721
9376
6095
5001
3125
All Adults
Dallas
S75
2016
20
17190
6954
3594
2032
1485
1094
All Adults
Dallas
S75
2017
10
156663
77511
49617
35474
26410
19768
All Adults
Dallas
S75
2017
15
49226
19768
10939
7657
5001
3438
All Adults
Dallas
S75
2017
20
19065
7110
3907
2266
1719
1250
All Adults
Detroit
S65
2015
10
77732
34344
20580
14222
10487
8717
All Adults
Detroit
S65
2015
15
19662
8127
3998
2491
1639
1376
All Adults
Detroit
S65
2015
20
6751
2491
1311
918
655
459
All Adults
Detroit
S65
2016
10
88481
40767
24644
16582
12387
9372
All Adults
Detroit
S65
2016
15
24119
10159
5505
3736
2097
1639
All Adults
Detroit
S65
2016
20
9569
3343
1966
1049
852
524
All Adults
Detroit
S65
2017
10
85663
38407
24185
16451
11863
9176
All Adults
Detroit
S65
2017
15
22284
9438
5243
3212
2556
1639
All Adults
Detroit
S65
2017
20
8651
3605
1573
786
393
197
All Adults
Detroit
S70
2015
10
97067
44634
26020
17893
13829
11339
All Adults
Detroit
S70
2015
15
26872
11142
5833
3801
2491
1835
All Adults
Detroit
S70
2015
20
10159
3867
1966
1114
852
590
All Adults
Detroit
S70
2016
10
110175
52368
33098
22743
17106
13829
All Adults
Detroit
S70
2016
15
33426
14681
7996
5571
3932
2818
All Adults
Detroit
S70
2016
20
13764
5571
2622
1704
1311
786
All Adults
Detroit
S70
2017
10
105128
48566
30739
21432
16320
12518
All Adults
Detroit
S70
2017
15
30936
13567
7472
5112
3605
2687
All Adults
Detroit
S70
2017
20
12060
4260
2359
1639
852
590
All Adults
Detroit
S75
2015
10
109454
50860
30477
20252
15664
12191
All Adults
Detroit
S75
2015
15
31722
13370
7210
4391
3212
2228
All Adults
Detroit
S75
2015
20
12846
4785
2622
1704
918
786
All Adults
Detroit
S75
2016
10
131476
61806
39915
27986
20383
15599
All Adults
Detroit
S75
2016
15
43192
18286
10290
6816
4916
3932
All Adults
Detroit
S75
2016
20
18745
7603
3605
2097
1639
1114
All Adults
Detroit
S75
2017
10
122300
56431
35917
25168
18745
14878
All Adults
Detroit
S75
2017
15
37162
16451
9045
6358
4260
3212
All Adults
Detroit
S75
2017
20
16582
5505
3080
2163
1180
852
3D-Attachment4-75
-------
Study Group
Study Area
AQ
Scenario
Year
FEV,
(percent)
Number of People at or above FEVt Decrement
>1 Day
> 2 Days
> 3 Days
> 4 Days
> 5 Days
> 6 Days
All Adults
Philadelphia
S65
2015
10
109538
48364
29628
20914
15947
12374
All Adults
Philadelphia
S65
2015
15
27711
10719
6884
4009
3050
1917
All Adults
Philadelphia
S65
2015
20
9760
3573
1917
1220
523
349
All Adults
Philadelphia
S65
2016
10
109712
49671
30326
22221
15773
12461
All Adults
Philadelphia
S65
2016
15
26317
10544
6100
4270
2527
1743
All Adults
Philadelphia
S65
2016
20
8889
3834
1743
1046
959
610
All Adults
Philadelphia
S65
2017
10
103874
49235
30761
21263
15250
11241
All Adults
Philadelphia
S65
2017
15
25620
10370
6361
3573
2266
1656
All Adults
Philadelphia
S65
2017
20
9499
2963
1656
871
436
261
All Adults
Philadelphia
S70
2015
10
135506
61261
37994
26840
20217
16208
All Adults
Philadelphia
S70
2015
15
37123
14640
9586
6100
4531
2963
All Adults
Philadelphia
S70
2015
20
14030
5054
2963
1917
1046
436
All Adults
Philadelphia
S70
2016
10
135506
62830
38691
27363
21524
16557
All Adults
Philadelphia
S70
2016
15
35903
13943
7669
5926
3921
2876
All Adults
Philadelphia
S70
2016
20
14030
5316
2701
1917
1394
1133
All Adults
Philadelphia
S70
2017
10
124004
60128
37820
27450
20653
15250
All Adults
Philadelphia
S70
2017
15
35118
13856
8714
5403
3311
2527
All Adults
Philadelphia
S70
2017
20
14466
4706
2440
1394
784
436
All Adults
Philadelphia
S75
2015
10
172542
78951
49410
35990
26578
20478
All Adults
Philadelphia
S75
2015
15
50194
20740
12723
8801
6013
4531
All Adults
Philadelphia
S75
2015
20
20827
8017
4967
2614
1917
1220
All Adults
Philadelphia
S75
2016
10
165658
80084
50107
35380
27450
22134
All Adults
Philadelphia
S75
2016
15
50804
21263
11677
7843
5926
4444
All Adults
Philadelphia
S75
2016
20
20043
7930
4531
3311
2179
1656
All Adults
Philadelphia
S75
2017
10
151541
74594
48103
34944
26666
20827
All Adults
Philadelphia
S75
2017
15
46883
19694
11241
7669
4706
3311
All Adults
Philadelphia
S75
2017
20
20304
7320
3660
1830
1220
959
All Adults
Phoenix
S65
2015
10
104253
55181
37151
28857
23046
18228
All Adults
Phoenix
S65
2015
15
28758
14255
9089
6407
4867
3377
All Adults
Phoenix
S65
2015
20
11225
4619
2881
1689
993
646
All Adults
Phoenix
S65
2016
10
99584
55429
37549
28708
22897
18774
All Adults
Phoenix
S65
2016
15
29254
13261
8543
6258
4520
3775
All Adults
Phoenix
S65
2016
20
10480
4420
2781
1788
1540
894
All Adults
Phoenix
S65
2017
10
114931
61240
42665
30595
23989
19470
All Adults
Phoenix
S65
2017
15
31042
15645
10530
7500
5811
4271
All Adults
Phoenix
S65
2017
20
12566
5960
3129
2235
1341
1093
All Adults
Phoenix
S70
2015
10
129484
70776
48029
36555
29354
24983
All Adults
Phoenix
S70
2015
15
40529
20960
13212
8791
6506
5066
All Adults
Phoenix
S70
2015
20
16887
7152
4222
2732
1838
1242
All Adults
Phoenix
S70
2016
10
121636
68690
48227
36903
29552
24437
All Adults
Phoenix
S70
2016
15
40132
20513
12814
9040
6655
5414
All Adults
Phoenix
S70
2016
20
16887
7202
4172
3030
2285
1838
All Adults
Phoenix
S70
2017
10
143440
78624
55230
41423
32234
26423
All Adults
Phoenix
S70
2017
15
43658
21655
14205
10828
8046
6308
All Adults
Phoenix
S70
2017
20
18327
8642
5761
4172
2583
1987
3D-Attachment4-76
-------
Study Group
Study Area
AQ
Scenario
Year
FEV,
(percent)
Number of People at or above FEVt Decrement
>1 Day
> 2 Days
> 3 Days
> 4 Days
> 5 Days
> 6 Days
All Adults
Phoenix
S75
2015
10
152530
82945
57267
44552
35016
29453
All Adults
Phoenix
S75
2015
15
50115
26274
17533
12119
9238
7251
All Adults
Phoenix
S75
2015
20
22052
10579
6457
4172
3129
2235
All Adults
Phoenix
S75
2016
10
142298
80462
57118
44204
35214
28311
All Adults
Phoenix
S75
2016
15
49270
25728
16738
11771
8990
7301
All Adults
Phoenix
S75
2016
20
22152
9785
6059
4073
3278
2334
All Adults
Phoenix
S75
2017
10
169218
93574
65611
50462
40579
32979
All Adults
Phoenix
S75
2017
15
56075
28807
18675
13311
10728
8543
All Adults
Phoenix
S75
2017
20
24735
11771
7947
5712
4222
2881
All Adults
Sacramento
S65
2015
10
32672
14692
8775
6803
5231
4116
All Adults
Sacramento
S65
2015
15
7661
3316
1887
1315
1000
800
All Adults
Sacramento
S65
2015
20
2916
1086
715
457
172
114
All Adults
Sacramento
S65
2016
10
32329
14864
9118
6031
4516
3516
All Adults
Sacramento
S65
2016
15
7804
2973
1801
1258
743
515
All Adults
Sacramento
S65
2016
20
2773
858
372
257
143
114
All Adults
Sacramento
S65
2017
10
31843
15235
9747
6660
5174
4002
All Adults
Sacramento
S65
2017
15
7546
3287
2144
1229
829
600
All Adults
Sacramento
S65
2017
20
3001
1058
343
257
200
57
All Adults
Sacramento
S70
2015
10
43734
21152
13263
9547
7375
6117
All Adults
Sacramento
S70
2015
15
11691
5260
3287
2287
1658
1229
All Adults
Sacramento
S70
2015
20
4802
1887
1172
715
515
343
All Adults
Sacramento
S70
2016
10
44306
21038
13863
9919
7232
5260
All Adults
Sacramento
S70
2016
15
12291
5088
3087
1972
1401
1058
All Adults
Sacramento
S70
2016
20
4831
1887
1000
486
314
200
All Adults
Sacramento
S70
2017
10
43820
21067
13578
10119
7175
6003
All Adults
Sacramento
S70
2017
15
10948
5260
3201
2344
1801
1229
All Adults
Sacramento
S70
2017
20
4888
1944
1029
629
457
200
All Adults
Sacramento
S75
2015
10
53196
26641
17008
12034
9204
7718
All Adults
Sacramento
S75
2015
15
15350
7060
4316
2887
2258
1829
All Adults
Sacramento
S75
2015
20
6460
2744
1572
1172
743
572
All Adults
Sacramento
S75
2016
10
54939
26984
17608
12834
9662
7546
All Adults
Sacramento
S75
2016
15
17008
6717
4145
2944
2287
1744
All Adults
Sacramento
S75
2016
20
6975
3001
1486
943
657
486
All Adults
Sacramento
S75
2017
10
54739
26669
17465
12720
9833
7746
All Adults
Sacramento
S75
2017
15
15407
7318
4545
3116
2258
1801
All Adults
Sacramento
S75
2017
20
6460
2830
1601
943
657
457
All Adults
St. Louis
S65
2015
10
38915
16453
9836
6653
4864
4006
All Adults
St. Louis
S65
2015
15
9192
3612
2325
1466
1001
715
All Adults
St. Louis
S65
2015
20
3326
1180
537
322
215
72
All Adults
St. Louis
S65
2016
10
48465
23034
13162
8906
6402
5007
All Adults
St. Louis
S65
2016
15
13127
5079
3004
1860
1431
1037
All Adults
St. Louis
S65
2016
20
5329
1717
930
680
393
179
All Adults
St. Louis
S65
2017
10
47535
22283
14200
9514
7118
5472
All Adults
St. Louis
S65
2017
15
12697
4972
2861
1967
1395
1073
All Adults
St. Louis
S65
2017
20
4578
1824
930
644
501
286
3D-Attachment4-77
-------
Study Group
Study Area
AQ
Scenario
Year
FEV,
(percent)
Number of People at or above FEVt Decrement
>1 Day
> 2 Days
> 3 Days
> 4 Days
> 5 Days
> 6 Days
All Adults
St. Louis
S70
2015
10
48965
21174
12948
9156
6903
5759
All Adults
St. Louis
S70
2015
15
13985
5437
3076
2182
1466
1037
All Adults
St. Louis
S70
2015
20
5508
1896
1073
680
358
179
All Adults
St. Louis
S70
2016
10
62593
30545
19100
12447
9121
6975
All Adults
St. Louis
S70
2016
15
19207
7762
4435
2683
1896
1466
All Adults
St. Louis
S70
2016
20
8226
3040
1574
1073
823
644
All Adults
St. Louis
S70
2017
10
60304
29830
19207
13520
9943
8262
All Adults
St. Louis
S70
2017
15
18313
7583
4471
3076
2075
1610
All Adults
St. Louis
S70
2017
20
7189
2647
1502
1109
715
537
All Adults
St. Louis
S75
2015
10
57442
26074
15273
10766
8620
6832
All Adults
St. Louis
S75
2015
15
17776
7082
3934
2826
1931
1431
All Adults
St. Louis
S75
2015
20
7225
2504
1538
1037
572
286
All Adults
St. Louis
S75
2016
10
72965
36840
23070
15702
11839
9192
All Adults
St. Louis
S75
2016
15
24536
10301
5866
3756
2504
1753
All Adults
St. Louis
S75
2016
20
10659
4149
2361
1574
966
751
All Adults
St. Louis
S75
2017
10
70927
35624
23392
16381
12519
10194
All Adults
St. Louis
S75
2017
15
23893
10229
6188
3612
2826
2039
All Adults
St. Louis
S75
2017
20
9586
3827
2182
1538
1180
894
Asthma Adults
Atlanta
S65
2015
10
5494
2324
1479
845
493
282
Asthma Adults
Atlanta
S65
2015
15
1409
493
70
70
70
0
Asthma Adults
Atlanta
S65
2015
20
282
70
70
0
0
0
Asthma Adults
Atlanta
S65
2016
10
6691
3381
2254
1550
1197
986
Asthma Adults
Atlanta
S65
2016
15
1690
916
775
563
423
282
Asthma Adults
Atlanta
S65
2016
20
775
493
282
211
211
211
Asthma Adults
Atlanta
S65
2017
10
5635
2254
1620
845
563
423
Asthma Adults
Atlanta
S65
2017
15
1831
352
211
70
70
70
Asthma Adults
Atlanta
S65
2017
20
282
141
70
0
0
0
Asthma Adults
Atlanta
S70
2015
10
7677
3381
2113
1620
986
634
Asthma Adults
Atlanta
S70
2015
15
2043
986
282
141
141
0
Asthma Adults
Atlanta
S70
2015
20
634
141
70
70
70
0
Asthma Adults
Atlanta
S70
2016
10
8875
4367
3240
2113
1690
1550
Asthma Adults
Atlanta
S70
2016
15
2395
916
775
704
563
493
Asthma Adults
Atlanta
S70
2016
20
1127
563
493
423
211
211
Asthma Adults
Atlanta
S70
2017
10
7818
3663
2324
1409
986
634
Asthma Adults
Atlanta
S70
2017
15
2395
634
282
282
141
70
Asthma Adults
Atlanta
S70
2017
20
493
211
141
0
0
0
Asthma Adults
Atlanta
S75
2015
10
10142
4296
2747
2113
1479
1057
Asthma Adults
Atlanta
S75
2015
15
2817
1409
493
211
211
70
Asthma Adults
Atlanta
S75
2015
20
1338
423
70
70
70
0
Asthma Adults
Atlanta
S75
2016
10
11903
5494
3733
2958
2113
1902
Asthma Adults
Atlanta
S75
2016
15
3803
1479
986
916
704
563
Asthma Adults
Atlanta
S75
2016
20
1479
775
704
423
352
211
Asthma Adults
Atlanta
S75
2017
10
9649
4226
3099
2183
1620
1197
Asthma Adults
Atlanta
S75
2017
15
2747
916
634
423
282
211
Asthma Adults
Atlanta
S75
2017
20
1057
352
141
141
70
0
3D-Attachment4-78
-------
Study Group
Study Area
AQ
Scenario
Year
FEV,
(percent)
Number of People at or above FEVt Decrement
>1 Day
> 2 Days
> 3 Days
> 4 Days
> 5 Days
> 6 Days
Asthma Adults
Boston
S65
2015
10
9490
3816
1957
1468
1076
978
Asthma Adults
Boston
S65
2015
15
1859
881
587
294
0
0
Asthma Adults
Boston
S65
2015
20
685
98
0
0
0
0
Asthma Adults
Boston
S65
2016
10
9784
4403
3229
1957
1370
1174
Asthma Adults
Boston
S65
2016
15
2642
1370
881
196
196
196
Asthma Adults
Boston
S65
2016
20
978
489
294
98
98
0
Asthma Adults
Boston
S65
2017
10
11838
4892
2837
1761
1468
1076
Asthma Adults
Boston
S65
2017
15
3131
881
391
294
294
196
Asthma Adults
Boston
S65
2017
20
1272
391
196
196
0
0
Asthma Adults
Boston
S70
2015
10
11447
4598
2250
1468
1174
978
Asthma Adults
Boston
S70
2015
15
2348
978
587
294
196
98
Asthma Adults
Boston
S70
2015
20
881
196
0
0
0
0
Asthma Adults
Boston
S70
2016
10
12425
5479
3522
2739
1761
1370
Asthma Adults
Boston
S70
2016
15
3522
1859
881
489
294
196
Asthma Adults
Boston
S70
2016
20
1468
587
294
196
98
98
Asthma Adults
Boston
S70
2017
10
15067
5870
3326
2446
1957
1370
Asthma Adults
Boston
S70
2017
15
3816
1272
489
391
391
294
Asthma Adults
Boston
S70
2017
20
1663
489
196
196
98
98
Asthma Adults
Boston
S75
2015
10
12425
5381
2348
1565
1370
1272
Asthma Adults
Boston
S75
2015
15
2642
978
685
489
196
98
Asthma Adults
Boston
S75
2015
20
783
294
98
0
0
0
Asthma Adults
Boston
S75
2016
10
14186
5870
3620
2935
2152
1468
Asthma Adults
Boston
S75
2016
15
3718
2152
1468
881
587
294
Asthma Adults
Boston
S75
2016
20
1859
783
489
196
98
98
Asthma Adults
Boston
S75
2017
10
16143
6653
3522
2739
2250
1370
Asthma Adults
Boston
S75
2017
15
4403
1565
783
587
391
391
Asthma Adults
Boston
S75
2017
20
2055
489
294
196
98
98
Asthma Adults
Dallas
S65
2015
10
6407
2422
1797
1328
1172
547
Asthma Adults
Dallas
S65
2015
15
1797
859
313
156
78
78
Asthma Adults
Dallas
S65
2015
20
547
0
0
0
0
0
Asthma Adults
Dallas
S65
2016
10
5782
2266
1016
625
391
391
Asthma Adults
Dallas
S65
2016
15
1172
313
313
313
156
156
Asthma Adults
Dallas
S65
2016
20
391
234
234
156
78
78
Asthma Adults
Dallas
S65
2017
10
8439
3516
2657
2110
1719
1485
Asthma Adults
Dallas
S65
2017
15
2813
1250
1016
859
469
313
Asthma Adults
Dallas
S65
2017
20
1641
391
156
156
78
0
Asthma Adults
Dallas
S70
2015
10
8751
3438
2188
1485
1172
1016
Asthma Adults
Dallas
S70
2015
15
2110
859
547
313
234
156
Asthma Adults
Dallas
S70
2015
20
703
156
78
78
78
0
Asthma Adults
Dallas
S70
2016
10
6876
2891
1250
938
469
391
Asthma Adults
Dallas
S70
2016
15
1563
469
313
313
234
156
Asthma Adults
Dallas
S70
2016
20
391
234
234
156
156
156
Asthma Adults
Dallas
S70
2017
10
9611
4532
3282
2657
2110
1953
Asthma Adults
Dallas
S70
2017
15
3360
1485
1094
938
625
469
Asthma Adults
Dallas
S70
2017
20
1641
859
625
234
78
78
3D-Attachment4-79
-------
Study Group
Study Area
AQ
Scenario
Year
FEV,
(percent)
Number of People at or above FEVt Decrement
>1 Day
> 2 Days
> 3 Days
> 4 Days
> 5 Days
> 6 Days
Asthma Adults
Dallas
S75
2015
10
10861
3985
2735
1641
1406
1250
Asthma Adults
Dallas
S75
2015
15
2735
938
625
547
313
156
Asthma Adults
Dallas
S75
2015
20
1250
313
156
78
78
0
Asthma Adults
Dallas
S75
2016
10
8361
3672
1641
1094
781
469
Asthma Adults
Dallas
S75
2016
15
2266
547
313
313
313
156
Asthma Adults
Dallas
S75
2016
20
547
234
234
234
156
156
Asthma Adults
Dallas
S75
2017
10
10705
5313
3829
3125
2344
2110
Asthma Adults
Dallas
S75
2017
15
4141
1797
1328
1250
859
547
Asthma Adults
Dallas
S75
2017
20
1875
1016
703
391
313
234
Asthma Adults
Detroit
S65
2015
10
7603
3146
1507
1114
655
590
Asthma Adults
Detroit
S65
2015
15
1770
721
262
131
131
131
Asthma Adults
Detroit
S65
2015
20
655
262
131
131
0
0
Asthma Adults
Detroit
S65
2016
10
8979
3932
2163
1770
1507
1376
Asthma Adults
Detroit
S65
2016
15
2425
1376
852
721
393
262
Asthma Adults
Detroit
S65
2016
20
1376
655
328
66
66
0
Asthma Adults
Detroit
S65
2017
10
8914
4260
2753
2097
1507
1114
Asthma Adults
Detroit
S65
2017
15
2687
1311
590
393
393
328
Asthma Adults
Detroit
S65
2017
20
1442
393
328
262
66
0
Asthma Adults
Detroit
S70
2015
10
9241
4064
1901
1376
983
852
Asthma Adults
Detroit
S70
2015
15
2359
983
524
197
197
131
Asthma Adults
Detroit
S70
2015
20
852
328
131
131
131
66
Asthma Adults
Detroit
S70
2016
10
11011
5047
2949
2359
1901
1770
Asthma Adults
Detroit
S70
2016
15
3343
1704
1180
852
786
721
Asthma Adults
Detroit
S70
2016
20
1770
852
459
262
262
131
Asthma Adults
Detroit
S70
2017
10
11273
5047
3212
2687
1901
1573
Asthma Adults
Detroit
S70
2017
15
3474
1639
918
590
459
393
Asthma Adults
Detroit
S70
2017
20
1966
459
459
328
262
197
Asthma Adults
Detroit
S75
2015
10
10356
4653
2622
1507
1114
918
Asthma Adults
Detroit
S75
2015
15
2753
1114
590
262
262
131
Asthma Adults
Detroit
S75
2015
20
1180
393
262
197
131
131
Asthma Adults
Detroit
S75
2016
10
12977
6095
3736
2556
2097
1770
Asthma Adults
Detroit
S75
2016
15
4326
1901
1311
918
786
721
Asthma Adults
Detroit
S75
2016
20
1966
1114
655
393
393
328
Asthma Adults
Detroit
S75
2017
10
13239
6161
3998
2949
2163
1901
Asthma Adults
Detroit
S75
2017
15
4195
2097
1180
786
524
459
Asthma Adults
Detroit
S75
2017
20
2359
655
459
393
328
262
Asthma Adults
Philadelphia
S65
2015
10
10370
4619
2701
1917
1220
959
Asthma Adults
Philadelphia
S65
2015
15
2614
1046
784
349
349
261
Asthma Adults
Philadelphia
S65
2015
20
1133
349
174
174
87
87
Asthma Adults
Philadelphia
S65
2016
10
8453
3137
1917
1394
1133
697
Asthma Adults
Philadelphia
S65
2016
15
1394
349
174
0
0
0
Asthma Adults
Philadelphia
S65
2016
20
349
174
0
0
0
0
Asthma Adults
Philadelphia
S65
2017
10
9760
4357
3224
2179
1569
1133
Asthma Adults
Philadelphia
S65
2017
15
2440
1307
436
261
174
87
Asthma Adults
Philadelphia
S65
2017
20
1133
174
0
0
0
0
3D-Attachment4-80
-------
Study Group
Study Area
AQ
Scenario
Year
FEV,
(percent)
Number of People at or above FEVt Decrement
>1 Day
> 2 Days
> 3 Days
> 4 Days
> 5 Days
> 6 Days
Asthma Adults
Philadelphia
S70
2015
10
12026
5141
3050
2701
1830
1220
Asthma Adults
Philadelphia
S70
2015
15
3399
1481
959
610
436
349
Asthma Adults
Philadelphia
S70
2015
20
1569
436
349
349
87
87
Asthma Adults
Philadelphia
S70
2016
10
10631
4357
2701
1656
1394
959
Asthma Adults
Philadelphia
S70
2016
15
1917
610
261
174
174
0
Asthma Adults
Philadelphia
S70
2016
20
436
261
87
0
0
0
Asthma Adults
Philadelphia
S70
2017
10
11067
6013
3660
2789
2091
1656
Asthma Adults
Philadelphia
S70
2017
15
3399
1569
959
523
261
174
Asthma Adults
Philadelphia
S70
2017
20
1656
436
174
0
0
0
Asthma Adults
Philadelphia
S75
2015
10
15773
7233
4183
3311
2353
1743
Asthma Adults
Philadelphia
S75
2015
15
4531
1917
1220
784
523
436
Asthma Adults
Philadelphia
S75
2015
20
1917
784
349
349
174
87
Asthma Adults
Philadelphia
S75
2016
10
13594
5403
3660
2440
2004
1394
Asthma Adults
Philadelphia
S75
2016
15
3921
1569
784
349
174
87
Asthma Adults
Philadelphia
S75
2016
20
697
261
87
87
0
0
Asthma Adults
Philadelphia
S75
2017
10
13507
6710
4706
3573
2527
2266
Asthma Adults
Philadelphia
S75
2017
15
4444
1743
1133
959
436
261
Asthma Adults
Philadelphia
S75
2017
20
1917
610
349
87
0
0
Asthma Adults
Phoenix
S65
2015
10
8444
4321
2583
2036
1639
1291
Asthma Adults
Phoenix
S65
2015
15
2334
1440
844
497
397
248
Asthma Adults
Phoenix
S65
2015
20
844
397
298
199
99
0
Asthma Adults
Phoenix
S65
2016
10
6953
3973
2483
1589
1291
1192
Asthma Adults
Phoenix
S65
2016
15
2086
795
447
298
199
99
Asthma Adults
Phoenix
S65
2016
20
596
149
50
0
0
0
Asthma Adults
Phoenix
S65
2017
10
8046
4172
2732
2086
1738
1391
Asthma Adults
Phoenix
S65
2017
15
2235
1341
993
695
497
497
Asthma Adults
Phoenix
S65
2017
20
894
546
298
248
149
99
Asthma Adults
Phoenix
S70
2015
10
10232
5612
3626
2781
1987
1738
Asthma Adults
Phoenix
S70
2015
15
3328
1788
1142
695
546
447
Asthma Adults
Phoenix
S70
2015
20
1738
646
397
248
149
50
Asthma Adults
Phoenix
S70
2016
10
8543
4818
3526
2285
1788
1440
Asthma Adults
Phoenix
S70
2016
15
2632
1341
795
497
248
248
Asthma Adults
Phoenix
S70
2016
20
1093
497
149
50
50
50
Asthma Adults
Phoenix
S70
2017
10
9785
5414
3924
2732
2036
1788
Asthma Adults
Phoenix
S70
2017
15
2980
1788
1142
844
646
596
Asthma Adults
Phoenix
S70
2017
20
1242
646
447
397
248
199
Asthma Adults
Phoenix
S75
2015
10
11871
6705
4420
3328
2483
2086
Asthma Adults
Phoenix
S75
2015
15
4023
2185
1540
1093
745
596
Asthma Adults
Phoenix
S75
2015
20
1937
1043
546
298
149
149
Asthma Adults
Phoenix
S75
2016
10
10182
5612
4222
2881
2285
1838
Asthma Adults
Phoenix
S75
2016
15
3526
1689
1093
745
546
397
Asthma Adults
Phoenix
S75
2016
20
1589
646
348
99
50
50
Asthma Adults
Phoenix
S75
2017
10
11473
6308
4470
3328
2682
2185
Asthma Adults
Phoenix
S75
2017
15
4073
2136
1391
1142
944
695
Asthma Adults
Phoenix
S75
2017
20
1738
894
646
546
397
298
3D-Attachment4-81
-------
Study Group
Study Area
AQ
Scenario
Year
FEV,
(percent)
Number of People at or above FEVt Decrement
>1 Day
> 2 Days
> 3 Days
> 4 Days
> 5 Days
> 6 Days
Asthma Adults
Sacramento
S65
2015
10
2287
858
515
457
343
343
Asthma Adults
Sacramento
S65
2015
15
543
172
86
86
86
29
Asthma Adults
Sacramento
S65
2015
20
257
29
29
29
0
0
Asthma Adults
Sacramento
S65
2016
10
2172
1143
715
543
400
343
Asthma Adults
Sacramento
S65
2016
15
515
200
143
86
86
57
Asthma Adults
Sacramento
S65
2016
20
172
57
57
57
57
29
Asthma Adults
Sacramento
S65
2017
10
2115
915
486
400
314
314
Asthma Adults
Sacramento
S65
2017
15
457
257
172
57
57
29
Asthma Adults
Sacramento
S65
2017
20
172
86
0
0
0
0
Asthma Adults
Sacramento
S70
2015
10
2887
1286
858
657
543
400
Asthma Adults
Sacramento
S70
2015
15
715
314
286
114
114
29
Asthma Adults
Sacramento
S70
2015
20
343
143
86
29
29
29
Asthma Adults
Sacramento
S70
2016
10
2830
1429
1029
772
543
400
Asthma Adults
Sacramento
S70
2016
15
772
400
343
200
114
86
Asthma Adults
Sacramento
S70
2016
20
343
172
114
86
57
29
Asthma Adults
Sacramento
S70
2017
10
2716
1258
743
515
400
314
Asthma Adults
Sacramento
S70
2017
15
800
257
200
143
86
57
Asthma Adults
Sacramento
S70
2017
20
229
143
86
29
0
0
Asthma Adults
Sacramento
S75
2015
10
3544
1744
1172
800
657
572
Asthma Adults
Sacramento
S75
2015
15
1058
429
343
200
143
86
Asthma Adults
Sacramento
S75
2015
20
400
143
86
86
57
29
Asthma Adults
Sacramento
S75
2016
10
3602
1744
1258
943
657
457
Asthma Adults
Sacramento
S75
2016
15
1229
600
372
372
257
114
Asthma Adults
Sacramento
S75
2016
20
457
229
143
114
86
57
Asthma Adults
Sacramento
S75
2017
10
3430
1544
972
657
486
400
Asthma Adults
Sacramento
S75
2017
15
943
400
314
200
114
86
Asthma Adults
Sacramento
S75
2017
20
314
200
86
86
0
0
Asthma Adults
St. Louis
S65
2015
10
3684
1466
894
537
322
322
Asthma Adults
St. Louis
S65
2015
15
894
250
107
72
36
0
Asthma Adults
St. Louis
S65
2015
20
179
36
0
0
0
0
Asthma Adults
St. Louis
S65
2016
10
4185
1931
1073
715
572
429
Asthma Adults
St. Louis
S65
2016
15
1145
501
358
215
143
143
Asthma Adults
St. Louis
S65
2016
20
537
179
107
72
36
0
Asthma Adults
St. Louis
S65
2017
10
4435
1896
1252
751
537
322
Asthma Adults
St. Louis
S65
2017
15
1001
429
286
143
36
36
Asthma Adults
St. Louis
S65
2017
20
358
72
0
0
0
0
Asthma Adults
St. Louis
S70
2015
10
4435
1860
1145
751
537
465
Asthma Adults
St. Louis
S70
2015
15
1288
501
215
107
72
36
Asthma Adults
St. Louis
S70
2015
20
322
72
36
36
36
0
Asthma Adults
St. Louis
S70
2016
10
5687
2611
1574
1073
787
608
Asthma Adults
St. Louis
S70
2016
15
1610
787
429
250
143
143
Asthma Adults
St. Louis
S70
2016
20
715
286
215
143
107
36
Asthma Adults
St. Louis
S70
2017
10
5651
2468
1610
1109
787
608
Asthma Adults
St. Louis
S70
2017
15
1610
608
358
215
107
72
Asthma Adults
St. Louis
S70
2017
20
608
179
107
36
0
0
3D-Attachment4-82
-------
Study Group
Study Area
AQ
Scenario
Year
FEV,
(percent)
Number of People at or above FEVt Decrement
>1 Day
> 2 Days
> 3 Days
> 4 Days
> 5 Days
> 6 Days
Asthma Adults
St. Louis
S75
2015
10
5079
2253
1431
858
680
537
Asthma Adults
St. Louis
S75
2015
15
1717
608
322
179
107
72
Asthma Adults
St. Louis
S75
2015
20
465
143
36
36
36
0
Asthma Adults
St. Louis
S75
2016
10
6545
3219
2039
1395
1073
787
Asthma Adults
St. Louis
S75
2016
15
1967
1001
465
286
215
143
Asthma Adults
St. Louis
S75
2016
20
1001
358
250
215
143
72
Asthma Adults
St. Louis
S75
2017
10
6724
2933
1931
1109
1073
823
Asthma Adults
St. Louis
S75
2017
15
2146
823
465
215
179
107
Asthma Adults
St. Louis
S75
2017
20
823
250
179
72
36
36
3D-Attachment4-83
-------
APPENDIX 4A
EXPOSURE-RESPONSE FUNCTIONS FOR 11 TREE SPECIES AND
TEN CROPS
TABLE OF CONTENTS
4A. 1 Background 2
4A. 1.1 Tree Species Seedling E-R Functions 3
4A. 1.2 Crop Species E-R Functions 10
4A.1.3 Summary Tables for Tree Species and Crops 11
4A. 2 Tree Seedling RBL Studies 16
4A.3 Analysis of Multiple-Year RBL 20
4A.3.1 Comparison of Predicted and Observed O3 Growth Impacts 20
4A.3.2 Comparison of Estimated Impacts of Constant and Annually Varying Seasonal
Exposure 21
References 32
4A-1
-------
4A.1 BACKGROUND
Air quality criteria documents (AQCDs) for prior ozone (O3) reviews have assessed and
characterized the results of a series of studies on the growth effects of a range of seasonal O3
exposure levels. These studies included research conducted by the National Crop Loss
Assessment Network (NCLAN) on commercial crop species and by the EPA's National Health
and Environmental Effects Laboratory Western Ecology Division (NHEERL/WED) on seedlings
of 11 tree species1. These studies included documentation of hourly concentrations across the
full exposures, and multiple exposure scenarios per experiment, which has resulted in their being
the focus of work to characterize exposure-response (E-R) relationships for growth impacts on
crops and tree species.
A subsequent set of publications analyzed the experimental study results to define a
quantitative model that would well describe the E-R relationships of seasonal O3 exposure and
first impaired tree seedling growth and crop yield2. Those studies, which used several different
metrics to quantify exposure (e.g., SUM06, W126), concluded that a three-parameter Weibull
model form provides the most appropriate model for the response of absolute yield and growth to
O3 exposure because of the interpretability of its parameters, its flexibility (given the small
number of parameters), and its tractability for estimation (2013 ISA, section 9.6.2). This three-
parameter Weibull model is presented in equation 4A-1.
'wm\p
Y=ae n '
Equation 4A-1
where:
Y = total yield or biomass;
W126 = 03 exposure (e.g., 3-month sum of daily cumulative W126 from 8am to 8pm);
and,
r| and p are species-specific variables
With removal of the intercept term, a, the model estimates relative yield or biomass
without any further reparameterization. In order to compare E-R functions and associated
estimated across species, genotypes, or experiments (of same species/genotype) for which
1 These programs and the research conducted under them is described in detail in the 1996 AQCD (sections 5.5 and
5.6), summarized in the 2006 AQCD (section 9.5), 2013 ISA (section 9.6), and the current ISA (Appendix 8,
section 8.13).
2 Examples of these analyses include Lee et al. (1994), Gumpertz and Rawlings (1992), Heck et al. (1984), Hogsett
et al. (1997), Lee and Hogsett (1999), Lee et al. (1987), Lee et al. (1988), Lee et al. (1989), Lesser et al. (1990),
Rawlings and Cure (1985).
4A-2
-------
absolute values of the response may vary greatly, the model is reformulated in terms of relative
annual yield (or biomass) or relative yield (or biomass) loss (yield loss=[l-relative yield]). The
resultant 2-parameter model of relative yield was presented in the 1996 and 2006 AQCDs and
2013 and current ISA as basis for deriving common models for multiple species, multiple
genotypes within species and multiple experimental locations (2013 ISA, section 9.6.2; 2020
ISA, Appendix 8, section 8.13.2). The models presented in the AQCDs were in terms of SUM06
over a 3-month season; those models were updated for 12-hour W126 over a 3-month season in
the 2013 ISA (2013 ISA, section 9.6.2). The 2-parameter model structure, for relative biomass
loss (RBL) or relative yield loss (RYL) as a function of W126 is described in equation 4A-2.
RBL = 1 - exp[-(W126/r|)p] Equation 4A-2
Based on this model structure, functions for estimating RBL from seasonal W126 index,
parameterized for each of eleven tree species, are presented and discussed in section 4A.1.1
below, and RYL functions for the 10 crop species are presented in section 4A. 1.2.
4A.1.1 Tree Species Seedling E-R Functions
The RBL functions for each of 11 tree species were derived as median composite
functions from response estimates based on functions derived for each study or experiment for
which data were collected for that species (Lee and Hogsett, 1996, Tables 12 and 13). The eleven
species-specific (composite median) functions, based on Lee and Hogsett (1996)3 are presented
in Table 4A-1.
Table 4A-1. RBL functions for tree species.
Species
RBL Function
n (ppm)
B
Red Maple (Acerrubrum)
1 - exp[-(W126/n)p]
318.12
1.3756
Sugar Maple {Acer saccharum)
36.35
5.7785
Red Alder (Alnus rubra)
179.06
1.2377
Tulip Poplar {Liriodendron tulipifera)
51.38
2.0889
Ponderosa Pine [Pinusponderosa)
159.63
1.1900
Eastern White Pine (Pinus strobus)
63.23
1.6582
Loblolly Pine [Pinus taeda)
1,021.63
0.9954
Virginia Pine [Pinus virqiniana)
1,714.64
1.0000
Quaking Aspen (Popuius tremuioides), wild
109.81
1.2198
Black Cherry (Prunus serotina)
38.92
0.9921
Douglas Fir (Pseudotsuga menzeiesii)
106.83
5.9631
Source: These functions are those presented in Lee and Hogsett (1996), Table 13 or, for loblolly pine,
as presented in Table 8-24 of Appendix 8 of the ISA.
3 The functions presented in Table 4A-1 reflect the median composite response functions presented in Table 13 of
Lee and Hogsett (1996), with the addition of the response curve for loblolly pine from Table 8 24 of Appendix 8
of the 2020 ISA. The process for deriving the composite functions is described in Lee and Hogsett (1996).
4A-3
-------
Figure 4A-1 presents species-specific E-R functions for the tree seedlings. The figures
illustrate how the values of the two parameters affect the shape of the resulting curves. The value
of t] in the RBL function affects the point of the curve where the slope appreciably changes, and
P affects the steepness of the curve. The response functions with smaller values of p (e.g.,
Virginia Pine) or with q values that are above the range shown for of W126 index values (e.g.,
functions for ponderosa pine and red alder) exhibit smaller slopes that have less change across
this W126 range. These functions describe a more constant rate of change in RBL over the range
of Os exposure shown (e.g., up to 30 ppm-hrs). In contrast, the response functions with larger p
values (e.g., the function for Sugar Maple) exhibit a threshold-like behavior, with large changes
in RBL over a small range of W126 index values and relatively small changes at other index
values. In these cases, the "threshold" is determined by the q parameter of the model.
CD
a:
00
o
CO
d
d
CM
o
—i—
30
40
0 10 20
W126 (ppm-hrs)
Figure 4A-1. RBL functions for seedlings of 11 tree species
50
Red Maple
• Sugar Maple
• Red Alder
Tulip Poplar
Ponderosa Pine
• White Pine
• Loblolly Pine
Virginia Pine
• Aspen
• Black Cherry
Douglas Fir
Red Maple
• Sugar Maple
• Red Alder
Tulip Poplar
Ponderosa Pine
• White Pine
Loblolly Pine
Virginia Pine
• Aspen
• Black Cherry
Douglas Fir
10 15 20
W126 (ppm-hrs)
4A-4
-------
The shape of curves presented in Figure 4A-1 also illustrate how sensitive the RBL value
is to changes in O3. Two species, Loblolly Pine (dark grey line) and Virginia Pine (yellow line)
have E-R functions that approach linearity within the W126 range represented on the x-axis,
meaning that a 1 percent change in W126 produces an equal change in RBL. Black Cherry (blue
line) has an E-R function with exhibits a declining slope with increasing W126, with the
appearance of leveling off (Figure 4A-1), which produces a smaller change in RBL relative to
the change in W126. The functions for the remaining species produce somewhat similar percent
changes in RBL relative to changes in W126.
As mentioned above, the species-specific functions were derived from median estimates
based on the functions from the individual experiments for each species. Figure 4A-2 through
Figure 4A-12 present the species-specific functions along with the functions derived from the
experiments available for that species.4 These figures provide a sense of the across-experiment
variability for each species, where such information is available.
O J
CO
o
CD
O
_l
CD
cr
O
CM
O
q
° '—1 1 1 1 1 1—
0 10 20 30 40 50
W126 (ppm-hrs)
Figure 4A-2. RBL functions for Red Maple (Acer rubrum).
4 For aspen, the dark (red) line shown in Figure 4A-1 is the median composite for wild (vs clonal genotype) studies.
4A-5
-------
o
00
o
_l
m
cc
o
CM
O
O
O
0
10
20
30
40
50
W126 (ppm-hrs)
Figure 4A-3. RBL functions for Sugar Maple {Acer saccharum).
20 30
W126 (ppm-hrs)
Figure 4A-4. RBL functions for Red Alder (Alnus rubra).
W126 (ppm-hrs)
Figure 4A-5. RBL functions for Tulip Poplar (Liriodendron tulipifera).
4A-6
-------
CO
o
CD
O
m
a:
O
CNJ
O
O
O
10
20
30
40
50
W126 (ppm-hrs)
Figure 4A-6. RBL functions for Ponderosa Pine (Pinus ponderosa).
CO
O
CD
O
CD
q:
o
CM
o
o
d
20 30
W126 (ppm-hrs)
Figure 4A-7. RBL functions for White Pine (Pinus strobus).
CD
O
CD
a:
o
CNJ
o
o
d
W126 (ppm-hrs)
Figure 4A-8. RBL functions for Loblolly Pine (Pinus taeda).
4A-7
-------
W126 (ppm-hrs)
Figure 4A-9. RBL functions for Virginia Pine (Pinus virginiana).
O
CO
O
CD
O
CM
O
O
O
0
20
30
50
10
40
W126 (ppm-hrs)
Figure 4A-10.RBL functions for Aspen (Populus tremuloides). Red lines = wild, blac.k=clone.
O
°o
O
CD
O
_l
m
on
d
CNJ
o
o
o
0
10
20
30
40
50
W126 (ppm-hrs)
Figure 4A-11.RBL functions for Black Cherry (Prunus serotina).
4A-8
-------
0 10 20 30 40 50
W126 (ppm-hrs)
Figure 4A-12.RBL functions for Douglas Fir (Pseudotsuga menzeiesii).
In the 2015 review, consideration of the E-R functions for the seedlings of 11 tree species
focused on the median estimate across the 11 species-specific functions. Recognizing the extent
to which experimental variation contributes to uncertainty in the species-specific E-R functions,
a stochastic analysis was performed in the quantitative exposure/risk assessment for the 2015
review as an approach to investigating the impact of uncertainty and variability in the E-R
function dataset; an update of this analysis is presented in Figure 4A-13. This figure illustrates
different approaches to estimating a median E-R function from the functions from the individual
experiments. In this figure, each grey curve is the median across 11 species-specific functions
where the species-specific functions are represented by a random draw from the experiment-
specific functions available for each species.5 The red points are the median across the random
draws at that W126 value and the whiskers extend to the 75th and 25th percentiles of those draws.
For reference, the green line is the median across the 11 species-specific functions, and the red
line is the median across the 51 experiments (regardless of species).6
5 For example, there are seven separate experiment-specific E-R functions for ponderosa pine (Lee ang Hogsett,
1996). In each iteration, one of the seven is drawn. This is performed for all eleven species. Each iteration of
these random draws is represented by a single grey line that plots the median of the 11 RBLs derived from the 11
functions for each W126 index level across the range of W126 presented in the figure. At different parts of the
W126 range, different species' E-R functions will produce the median estimate. As a result, the grey line for each
iteration of the random draws has an area of rapid change over a particular range of W126 levels (when the E-R
function producing the median estimate switches to a different species) and then a smoothing (as the median
estimates are being produced by the same E-R function).. That is, since there are 11 species (i.e. an odd number),
each point on each grey line in the figure comes from the curve for the species' function that predicts the 6th
highest (or lowest) RBL for that W126 index value.
6 Both the green and red lines include two step-like changes along the W126 index range from 8 to26 ppm-hrs.
These steps reflect the influence on the median of the functions of species with inflection points that differ from
the others (that can be seen in Figure 4A-1). For example, on the green curve (for the median across the species-
specific functions), from a W126 index of approximately 8 ppm-hrs to 23 ppm-hrs, the curve largely follows the
response function for red alder (which is somewhat centrally located among the functions over that W126 range
4A-9
-------
o
_l
m
-------
o
00
o
CD
d
i
>
a:
o
(XI
o
o
o
0 10 20 30 40 50
W126 (DDm-hrs)
Figure 4A-14.RYL functions for crop species.
4A. 1.3 Summary Tables for Tree Species and Crops
Table 4A-3 and Table 4A-4 below provide estimates of the relative loss for tree biomass
and crop yield, respectively, at various W126 index values using the composite E-R functions for
each species for each integer W126 index value between 7 ppm-hrs and 30 ppm-hrs. The cross-
species median of the species-specific composite functions is calculated for all 11 tree species.
These tables also provide estimates of the number of species for trees and crops respectively that
would be below various reference values (e.g., 2% RBL for trees) at various W126 index values.
Table 4A-5 summarizes the median values for each integer W126 index value between 7 ppm-
hrs and 23 ppm-hrs.
Barley
• Field Corn
• Cotton
Kidney Bean
Lettuce
• Peanut
• Potato
Grain Sorghum
• Soybean
• Winter Wheat
4 A-11
-------
Table 4A-3. Relative biomass loss for eleven individual tree seedlings and median at various W126 index values.
W126
Douglas
Fir
Loblolly
Virginia
Pine
Red Sugar Red
maple maple Alder
Ponderosa
Pine
Aspen
Tulip
Poplar
Eastern
White
Pine
Black
Cherry
Median
(11
species)
Number
of
Species
<2%
Number
of
Species
<5%
Number
of
Species
<10%
Number
of
Species
<15%
30
0.1%
0.8%,
1.7%,
3.8%,
28.1%,
10.4%,
12.8%,
18.(5%,
27.7%,
25.2%,
53.8%,
12.8%,
0
o
4
4
(5
29
0.0%
0.7%,
1.7%
3.6%
23.7%,
10.0%,
12.3%
17.9%
26.1%
24.0%
52.6%
12.3%
o
v)
4
5
6
28
0.0%
0.7%,
1 .(5%,
3.5%,
19.9%,
9.(5%,
11.8%,
17.2%,
24.5%,
22.8%,
51.4%,
11.8%,
3
4
5
(5
27
0.0%
0.7%,
1 .(5%,
3.3%,
1(5.4%,
9.2%,
11.4%,
1(5.5%,
23.0%,
21.(5%,
50.1%
11.4%,
')
v)
4
5
(5
26
0.0%
0.7%
1.5%,
3.1%,
13.4%,
8.8%,
10.!)%
15.8%
21.4%
20.5%
48.8%
10.9%,
')
4
5
7
25
0.0%
0.(5%
1.4%,
3.0%,
10.9%,
8.4%,
10.4%,
15.2%,
19.9%,
19.3%,
47.5%,
10.4%,
')
v)
4
5
7
24
0.0%
0.6%
1.4%,
2.8%,
8.7%
8.0%,
10.0%
14.5%
18.4%
18.2%
46.2%
8.7%
3
4
7
8
23
0.0%
0.(5%
1.3%,
2.7%,
6.9%,
7.(5%,
9.5%
13.8%,
17.0%,
17.1%,
44.8%,
7.(5%,
')
v)
4
7
8
22
0.0%
0.6%
1.3%,
2.5%,
5.3%
7.2%,
9.0%,
13.1%
15.6%
15.9%
43.3%
7.2%
3
4
7
8
21
0.0%
0.5%,
1.2%,
2.4%,
4.1%,
(5.8%,
8.(5%,
12.4%,
14.3%,
14.9%,
41.9%,
(5.8%,
3
5
7
10
20
0.0%,
0.5%,
1.2%,
2.2%,
') 1 ()/
O. /(>
(5.4%,
8.1%,
11.8%,
13.0%,
13.8%,
40.3%,
(5.4%,
0
o
5
7
10
19
0.0%
0.5%,
1.1%,
2.1%,
2.3%,
(5.0%,
7.(5%,
11.1%,
11.8%,
12.7%,
38.8%,
(5.0%,
')
v)
5
7
10
18
0.0%,
0.5%,
1.0%,
1.9%,
1.7%,
5.7%,
7.2%,
10.4%,
10.(5%,
11.7%,
37.2%,
5.7%,
5
5
7
10
17
0.0%,
0.4%,
1.0%,
1.8%,
1.2%,
5.3%,
(5.7%,
9.8%,
9.4%,
10.7%,
35.(5%
5.3%,
5
5
9
10
16
0.0%,
0.4%,
0.9%,
1 .(5%,
0.9%,
4.9%
(5.3%,
9.1%,
8.4%,
9.7%,
')') no/
Ov)..; /o
4.9%,
5
(5
10
10
15
0.0%,
0.4%,
0.9%,
1.5%,
0.(5%,
4.5%,
5.8%,
8.4%,
7.4%,
8.8%,
32.2%,
4.5%,
5
(5
10
10
14
0.0%
0.4%
0.8%,
1.4%,
0.4%
4.2%,
5.4%,
7.8%,
(5.4%,
7.9%
30.4%
4.2%
5
(5
10
10
13
0.0%,
0.3%,
0.8%,
1.2%,
0.3%,
3.8%,
4.9%,
7.1%,
5.5%,
7.0%,
28.(5%,
') yo/
v).0 /()
5
7
10
10
12
0.0%,
0.3%
0.7%,
1.1%,
0.2%
') K ()/
v). J /()
4.5%,
(5.5%,
4.7%,
(5.2%,
2(5.7%
') no/
v). J /o
5
8
10
10
11
0.0%,
0.3%,
0.(5%,
1.0%,
0.1%,
3.1%,
4.1%,
5.9%
') no/
v)..; /o
5.4%,
24.8%,
3.1%,
5
8
10
10
10
0.0%,
0.3%
0.(5%,
0.9%,
0.1%
2.8%,
3.(5%,
5.2%,
3.2%,
4.(5%,
22.9%
2.8%
5
9
10
10
9
0.0%,
0.2%
0.5%,
0.7%
0.0%
2.4%
3.2%
4.6%
2.6%
3.9%
20.9%,
2.4%
5
10
10
10
8
0.0%,
0.2%,
0.5%,
0.(5%,
0.0%,
2.1%,
2.8%,
4.0%,
2.0%,
3.2%,
18.8%,
2.0%,
5
10
10
10
7
0.0%
0.2%
0.4%
0.5%
0.0%
1.8%
2.4%
3.4%
1.5%
2.6%
16.7%
1.5%
7
10
10
10
4A-12
-------
Table 4A-4. Relative yield loss for ten individual crop species and median at various W126 index values.
.. Number Number Number Number e c
,, „ . Median e e f fc , of Species
is7ioc d i t Field Grain „ . „ .. c , Winter „ . . Kidney ,1ft of of of of Species ''
W126 Barley Lettuce „ „ , Peanut Cotton Soybean ,,,, . Potato „ 7 (10 „ „ „ r'. , > 10%
J Corn Sorghum J Wheat Bean . , Species Species Species > 5% and , ^
O CflAPIACI ill onn <
species; < 5% £ 1Q% < 2Q% < 1Q% dim
30
0.1%
5.1%,
2.9%,
2.3%,
10.4%
1(5.3%,
15.7%,
22.5%,
20.2%,
3(5.1%,
13.0%,
')
v)
4
7
1
'.)
v)
29
0.0%
4.4%,
2.7%,
2.1%,
9.7%,
15.(5%,
15.0%,
21.0%,
19.4%,
34.0%,
12.4%,
4
5
8
1
')
v)
28
0.0%
3.7%,
2.4%,
2.0%,
9.1 %,
14.9%,
14.4%,
19.5%
18.7%,
31.9%,
11.8%,
4
5
i)
1
4
27
0.0%
3.1%,
2.2%,
1.9%,
8.(5%,
14.1%,
13.7%,
18.0%,
18.0%,
29.8%,
11.2%,
4
5
9
1
4
26
0.0%,
2.(5%
1.!)%
1.7%
8.0%,
13.4%,
13.1%
1 (5.(5%
17.2%,
27.8%
10.(5%,
4
5
9
1
4
25
0.0%,
2.1%,
1.7%,
1 .(5%,
7.4%,
12.7%,
12.5%,
15.3%,
1 (5.5%
25.8%,
10.0%,
4
5
9
1
4
24
0.0%,
1.7%,
1.5%
1.5%
6.9%,
12.0%,
11.8%
14.0%
15.7%,
23.!)%
9.4%
4
5
9
1
4
23
0.0%,
1.4%,
1.3%,
1.4%,
(5.4%,
11.3%,
11.2%,
12.7%,
15.0%,
22.0%,
8.8%,
4
5
9
1
4
22
0.0%,
1.1%,
1.2%
1.3%
5.9%,
10.(5%,
10.6%
11.5%
14.2%,
20.1%
8.2%,
4
5
S)
1
4
21
0.0%,
0.9%,
1.0%,
1.1%,
5.4%,
10.0%,
10.0%,
10.4%,
13.5%,
18.4%,
7.7%,
4
7
10
3
20
0.0%,
0.7%,
0.9%,
1.0%,
5.0%,
9.3%,
9.4%,
9.3%,
12.7%,
1 (5.(5%,
7.1%,
5
8
10
2
19
0.0%,
0.6%
0.8%,
0.9%,
4.5%
8.7%,
8.8%,
8.3%
12.0%,
15.0%
6.4%
5
8
10
2
18
0.0%,
0.4%,
0.7%,
0.8%,
4.1%,
8.0%,
8.2%,
7.3%,
11.3%,
13.4%,
5.7%,
n
10
2
17
0.0%,
0.3%,
0.6%
0.8%
3.7%
7.4%
7.6%
6.4%
10.5%
11.9%
5.1%
5
10
o
2
16
0.0%,
0.2%,
0.5%,
0.7%,
3.3%,
(5.8%,
7.0%,
5.(5%,
9.8%,
10.5%,
4.4%,
5
9
10
4
1
15
0.0%,
0.2%,
0.4%,
0.(5%,
2.9%,
(5.2%,
(5.4%,
4.8%,
9.1 %,
9.2%,
3.9%,
(5
10
10
4
0
14
0.0%,
0.1%,
0.3%
0.5%
2.(5%,
5.6%
5.9%,
4.1%,
8.4%
7.9%,
3.3%,
(5
10
10
4
0
13
0.0%,
0.1%,
0.2%,
0.5%,
2.2%,
5.0%,
5.3%,
3.5%,
7.7%,
(5.8%,
2.8%,
(5
10
10
4
0
12
0.0%,
0.1%,
0.2%
0.4%
1.9%,
4.5%
4.8%,
2.!)%,
7.0%
5.7%,
2.4%,
8
10
10
0
11
0.0%,
0.0%,
0.2%,
0.3%,
1 .(5%,
3.9%,
4.3%,
2.3%,
(5.3%,
4.7%,
2.0%,
9
10
10
1
0
10
0.0%,
0.0%,
0.1%
0.3%
1.4%,
3.4%
3.8%,
1.!)%,
5.6%
3.8%,
1 .(5%,
9
10
10
1
0
9
0.0%,
0.0%,
0.1%,
0.2%,
1.1%,
2.9%,
3.3%,
1.5%,
4.9%,
3.0%,
1.3%,
10
10
10
0
0
8
0.0%,
0.0%,
0.1%,
0.2%,
0.9%,
2.5%,
2.8%,
1.1%,
4.3%,
2.4%,
1.0%,
10
10
10
0
0
7
0.0%
0.0%
0.0%
0.1%
0.7%
2.0%
2.3%
0.8%
3.6%
1.8%
0.8%
10
10
10
0
0
4A-13
-------
Table 4A-5. Tree seedling RBL and CYL estimated for seasonal W126 O3 exposure.
W126 index
value
for exposure
period
Tree seedling biomass lossA
Crop yield lossc
Median Value6
Individual Species
Median Value0
Individual Species
23 ppm-hrs
Median species
w. 7.6% loss B
< 2% loss: 3/11 species
<5% loss: 4/11 species
<10% loss: 8/11 species
<15% loss: 10/11 species
>40% loss: 1/11 species
Median species w.
8.8 % loss D
<5% loss: 4/10 species
>5,<10% loss: 1/10 species
>10,<20% loss: 4/10 species
>20:1/10 species
22 ppm-hrs
Median species
w. 7.2% loss B
< 2% loss: 3/11 species
<5% loss: 4/11 species
<10% loss: 7/11 species
<15% loss: 10/11 species
>40% loss: 1/11 species
Median species w.
8.2 % loss D
<5% loss: 4/10 species
>5,<10% loss: 1/10 species
>10,<20% loss: 4/10 species
>20:1/10 species
21 ppm-hrs
Median species
w. 6.8% loss B
< 2% loss: 3/11 species
<5% loss: 4/11 species
<10% loss: 7/11 species
<15% loss: 10/11 species
>40% loss: 1/11 species
Median species w.
7.7 % loss D
<5% loss: 4/10 species
>5,<10% loss: 3/10 species
>10,<20% loss: 3/10 species
20 ppm-hrs
Median species
w. 6.4% loss B
< 2% loss: 3/11 species
<5% loss: 5/11 species
<10% loss: 7/11 species
<15% loss: 10/11 species
>40% loss: 1/11 species
Median species w.
7.1 % lossD
<5% loss: 5/10 species
>5,<10% loss: 3/10 species
>10,<20% loss: 2/10 species
19 ppm-hrs
Median species
w. 6.0% loss B
< 2% loss: 3/11 species
<5% loss: 5/11 species
<10% loss: 7/11 species
<15% loss: 10/11 species
>30% loss: 1/11 species
Median species w.
6.4 % loss D
<5% loss: 5/10 species
>5, <10% loss: 3/10 species
>10,<20% loss: 2/10 species
18 ppm-hrs
Median species
w. 5.7% loss B
< 2% loss: 5/11 species
<5% loss: 5/11 species
<10% loss: 7/11 species
<15% loss: 10/11 species
>30% loss: 1/11 species
Median species w.
5.7 % loss D
<5% loss: 5/10 species
>5,<10% loss: 3/10 species
>10,<20% loss: 2/10 species
17 ppm-hrs
Median species
w. 5.3% loss B
< 2% loss: 5/11 species
<5% loss: 5/11 species
<10% loss: 9/11 species
<15% loss: 10/11 species
>30% loss: 1/11 species
Median species w.
5.1 % lossD
<5% loss: 5/10 species
>5, <10% loss: 3/10 species
>10,<20% loss: 2/10 species
16 ppm-hrs
Median species
w. 4.9% loss B
< 2% loss: 5/11 species
<5% loss: 6/11 species
<10% loss: 10/11 species
>30% loss: 1/11 species
Median species w.
<5% lossD
<5% loss: 5/10 species
>5,<10% loss: 4/10 species
>10,<20% loss: 1/10 species
4A-14
-------
W126 index
value
for exposure
period
Tree seedling biomass lossA
Crop yield lossc
Median Value6
Individual Species
Median Value0
Individual Species
15 ppm-hrs
Median species
w. 4.5% loss B
< 2% loss: 5/11 species
<5% loss: 6/11 species
<10% loss: 10/11 species
>30% loss: 1/11 species
Median species w.
<5% lossD
<5% loss: 6/10 species
>5, <10% loss: 4/10 species
14 ppm-hrs
Median species
w. 4.2% loss B
< 2% loss: 5/11 species
<5% loss: 6/11 species
<10% loss: 10/11 species
>30% loss: 1/11 species
Median species w.
<5% lossD
<5% loss: 6/10 species
>5,<10% loss: 4/10 species
13 ppm-hrs
Median species
w. 3.8% loss B
< 2% loss: 5/11 species
<5% loss: 7/11 species
<10% loss: 10/11 species
>20% loss: 1/11 species
Median species
w.<5% lossD
<5% loss: 6/10 species
>5, <10% loss: 4/10 species
12 ppm-hrs
Median species
w. 3.5% loss B
< 2% loss: 5/11 species
<5% loss: 8/11 species
<10% loss: 10/11 species
>20% loss: 1/11 species
Median species w.
<5% lossD
<5% loss: 8/10 species
>5,<10% loss: 2/10 species
11 ppm-hrs
Median species
w. 3.1% loss B
< 2% loss: 5/11 species
<5% loss: 8/11 species
<10% loss: 10/11 species
>20% loss: 1/11 species
Median species w.
<5% lossD
<5% loss: 9/10 species
>5, <10% loss: 1/10 species
10 ppm-hrs
Median species
w. 2.8% loss B
< 2% loss: 5/11 species
<5% loss: 9/11 species
<10% loss: 10/11 species
>20% loss: 1/11 species
Median species w.
<5% lossD
<5% loss: 9/10 species
>5,<10% loss: 1/10 species
9 ppm-hrs
Median species
w. 2.4% lossB
< 2% loss: 5/11 species
^5% loss: 10/11 species
>20% loss: 1/11 species
Median species w.
<5% lossD
<5% loss: all species
8 ppm-hrs
Median species
w. 2.0% loss B
< 2% loss: 5/11 species
<5% loss: 10/11 species
>15% loss: 1/11 species
Median species w.
<5% lossD
<5% loss: all species
7 ppm-hrs
Median species
w. <2% lossB
< 2% loss: 7/11 species
<5% loss: 10/11 species
>15% loss: 1/11 species
Median species w.
<5% lossD
<5% loss: all species
A Estimates here are based on the 11 E-R functions for tree seedlings described in section 4A.1.
B This median value is the median of the composite E-R functions for 11 tree species in Table 4A-3.
C Estimates here are based on the 10 E-R functions for crops described in section 4A.1.
D This median value is the median of the composite E-R functions for 10 crops in Table 4A-4.
4A-15
-------
4A.2 TREE SEEDLING RBL STUDIES
Table 4A-6 below lists the 51 tree seedling exposure cases for which E-R functions were
derived. Eleven tree species are represented by the 51 cases. The exposures in terms of SUM06
for the cases are presented as available.7 As described in section 4A.1 above, species-specific
(composite) functions were derived for each species, and Table 4A-5 above presents median
RBL estimates from the 11 species-specific functions.
7 Based on the approach that the EPA used in the 2007 Staff Paper, a SUM06 index value of 25 ppm-hrs is estimated
to correspond to a W126 index of approximately 21 ppm-hrs (U.S. EPA, 2007, Appendix 7B, p. 7B-2).
4A-16
-------
Table 4A-6. Individual tree seedling experimental cases for which E-R functions were derived in Lee and Hogsett (1996).
Study
IDA
Species
Site
year
Duration
(days)6
Harvest
c
SUM06
(ppm-hr)D
Study/Source and notes
Per 1996 AQCD, Table 5-28; Lee and Hogsett, 1996, Table 12
Per Hogsett et al 1997, Table 2
1
Aspen - wild
OR
1989
84
1
1
Aspen - wild
OR
1989
84
2
2
Aspen - wild
OR
1991
118
1
2
Aspen - wild
OR
1991
118
2
3
Aspen - wild
OR
1990
112
1
0.2,16.1, 72.1,102.8
Hoqsett (unpublished) cited in Hoqsett et al., 1997 Hoqsett et
al., 1995
3
Aspen - wild
OR
1990
112
2
4
Aspen - 216
Ml
1990
82
1
May be described in Karnosky et al., 1996, who reported
statistically siqnificant total biomass loss for clones 259 (at the
2 hiqhest exposure treatments) and clone 271 (at the hiqhest
treatment).
4
Aspen 253
Ml
1990
82
1
4
Aspen 259
Ml
1990
82
1
4
Aspen - 271
Ml
1990
82
1
5
Aspen - 216
Ml
1991
98
1
Karnosky et al., 1996, who reported statistically siqnificant
total biomass loss at the hiqhest exposure concentrations.
5
Aspen - 259
Ml
1991
98
1
0.0,11.5, 24.5, 32.4, 40.3, 60.5
Karnosky et al (1995, in press) cited in Hoqsett et al., 1995
Hoqsett et al., 1997 corresponds to Karnosky et al., 1996, who
reported statistically siqnificant total biomass loss averaqed
across all clones at the hiqhest exposure treatment.
5
Aspen - 271
Ml
1991
98
1
6
Aspen-wild
Ml
1991
98
1
7
Douqlas Fir
OR
1989-90
113
1
7
Douqlas Fir
OR
1989-90
113
2
7
Douqlas Fir
OR
1989-90
234
3
0.1, 33.4,147.2, 207.2, 261.5
Hoqsett (unpublished) cited in Hoqsett et al., 1995 Hoqsett et
al., 1997
7
Douqlas Fir
OR
1989-90
234
4
8
Douqlas Fir
OR
1991-92
118
1
8
Douqlas Fir
OR
1991-92
118
2
8
Douqlas Fir
OR
1991-92
230
3
0.1, 30.4, 60.6,143.0, 202.9
Hoqsett (unpublished) cited in Hoqsett et al., 1995 Hoqsett et
al., 1997
9
Ponderosa Pine
OR
1989
111
1
9
Ponderosa Pine
OR
1989
111
2
10
Ponderosa Pine
OR
1989-90
113
1
May be described in Andersen et al., 1997
Seedlinqs exposed to O3 for two qrowinq seasons were
statistically siqnificant smaller than CF-exposed seedlinqs
when measured in the dormant condition (SUM00 qreater than
253). Total biomass was reduced 58% at the hiqhest
exposure.
10
Ponderosa Pine
OR
1989-90
113
2
10
Ponderosa Pine
OR
1989-90
234
3
10
Ponderosa Pine
OR
1989-90
234
4
11
Ponderosa Pine
OR
1991-92
118
1
Lee and Hoqsett, 1999, who statistically siqnificant biomass
loss at the 2 hiqhest exposures (12-hr W126 qreater than 59)
11
Ponderosa Pine
OR
1991-92
118
2
4A-17
-------
Study
IDA
Species
Site
year
Duration
(days)6
Harvest
c
SUM06
(ppm-hr)D
Study/Source and notes
Per 1996AQCD, Table 5-28; Lee andHoqsett, 1996, Table 12
Per Hoqsett et al 1997, Table 2
11
Ponderosa Pine
OR
1991-92
230
3
0.1, 30.4, 60.6,143.0, 202.9
Hogsett (unpublished) cited in Hogsett et al., 1995 Hogsett et
al., 1997
12
Ponderosa Pine
OR
1992
140
1
May be described in Andersen and Scagel, 1997
Statistically significant reduction in the biomass of all plant
components after two seasons O3 exposure (1992+1993);
reductions were greater with nutrient restriction+03
13
Ponderosa Pine
OR
1991
84
1
14
Red Alder
OR
1990
121
1
15
Red Alder
OR
1989
113
1
15
Red Alder
OR
1989
113
2
16
Red Alder
OR
1991
118
1
16
Red Alder
OR
1991
118
2
0.0,16.0, 31.8, 73.4,103.6
Hogsett (unpublished) cited in Hogsett et al., 1995 Hogsett et
al., 1997
17
Red Alder
OR
1992
112
1
0.1,14.5, 29.1, 70.1, 99.9
Hogsett (unpublished) cited in Hogsett et al., 1995 Hogsett et
al., 1997
18
Black Cherry
SMNPF
1989
76
1
0.0,1.9,17.1,40.6
Neufeld et al., 1995 cited in Hogsett et al., 1995 Hogsett et al.,
1997
Also described in Neufeld and Renfro, 1993.
Statistically significant reduction in highest treatment group
19
Black Cherry
SMNP
1992
140
1
0.0, 0.0, 0.8,18.1,50.2
Neufeld, personal comm cited in Hogsett et al., 1995 Hogsett
et al., 1997
Described in Neufeld et al., 1995 Neufeld and Renfro, 1993
Statistically significant reduction in highest treatment
20
Red Maple
SMNP
1988
55
1
9.2,12,47,125.4
Neufeld (pers comm) cited in Hogsett et al., 1995 Hogsett et
al., 1997
21
Tulip Poplar
SMNP
1990-91
75
1
0.1,2.1,0.2, 0.9,16.6, 38.8
SUM06 exposures provided by Henry Lee email (8/16/2019)
21
Tulip Poplar
SMNP
1990-91
184
3
0.1,0.5,1.4, 34.5, 88.7
Neufeld (pers comm) cited in Hogsett et al., 1995 Hogsett et
al., 1997
22
Tulip Poplar
SMNP
1992
81
1
23
Loblolly GAKR 15-
23
AL
1988-89
555
3
Qiu et al., 1992and Lefohn et al., 1992 cited by Hogsett et al.,
1995 Hogsett et al., 1997, who reported GARK15-23 was
more tolerant to O3 with no significant biomass losses over the
entire exp period, while the more sensitive GARK 15-23 clone
had statistically significant decreases in foliar biomass and
root sguare area at the highest exposure treatment.
23
Loblolly GAKR 15-
91
AL
1988-89
555
3
4.9, 58.5, 301.5, 507.0
24
Sugar Maple
Ml
1990-91
83
1
4A-18
-------
Study
IDA
Species
Site
year
Duration
(days)6
Harvest
c
SUM06
(ppm-hr)D
Study/Source and notes
Per 1996AQCD, Table 5-28; Lee andHoqsett, 1996, Table 12
Per Hoqsett et al 1997, Table 2
24
Sugar Maple
Ml
1990-91
180
3
0.0, 25.2, 27.8, 49.8, 67.6, 94.4
Karnosky (pers. comm.) cited by Hogsett et al., 1995 Hogsett
et al., 1997
May (?) be described in Rebbeck and Loats, 1997, who
reported no statistically significant treatment effects in any of
the seedlings exposed to O3 between two individual seasons
or after exposure to 304 ppm (SUM00 index) over two growing
seasons (total of 225 days).
25
E. White Pine
Ml
1990-91
83
1
25
E. White Pine
Ml
1990-91
180
3
0.0, 25.2, 27.7, 49.8, 64.2, 94.4
Karnosky (pers. comm.) cited by Hogsett et al., 1995 Hogsett
et al., 1997
May be described in Isebrands et al., 2000 pg 170 which
reported no statistically significant difference in height, stem,
root or current year needle biomass in response to O3
26
Virginia Pine
SMNP
1992
98
1
0.0, 0.0,1.9, 21.7, 51.6
Neufeld (pers. comm.) cited in Hogsett et al., 1995 Hogsett et
al., 1997 may be described in Neufeld et al. (2000), who
reported no statistically significant treatment effects on
biomass.from 152 day duration (SUM06 up to 56.2)
A Study ID as in Lee and Hogsett (1996), Table 12.
B Duration corresponds to length in days of the first year of exposure for Harvests 1 and 2 and to the total length of the first and second years of exposure for Harvest 3.
C Harvest 1 occurs immediately following end of first year of exposure. Harvest 2 occurs in spring following first year of exposures. Harvest 3 occurs immediately following end of second year of
exposures. Harvest 4 occurs in spring following second year of exposures.
D First SUM06 treatment value corresponds to charcoal-filtered exposure (Hogsett et al., 1997 Table 2).
E Based on an approach used in the 2007 Staff Paper (and the associated temporal patterns of O3 concentrations in data available at that time), a SUM06 index value of 25 ppm-hrs would be
estimated to correspond to a W126 index of approximately 21 ppm-hrs (U.S. EPA, 2007, Appendix 7B, p. 7B-2).
F SMNP = Smoky Mountains National Park.
4A-19
-------
4A.3 ANALYSIS OF MULTIPLE-YEAR RBL
The sections below consider two illustrative analyses of the influence of seasonal O3
exposures across multiple years.
4A.3.1 Comparison of Predicted and Observed O3 Growth Impacts
The 2013 and 2020 ISAs present comparisons of aspen stand growth observations from
an Aspen FACE multiyear O3 exposure study and predictions derived through the application of
a median composite E-R function for wild aspen and aspen clones8 to seasonal W126 index
values (2013 ISA, section 9.6.3.2; 2020 ISA, Appendix 8, Figure 8-17). The Aspen FACE study
monitored growth of aspen stands annually from 1997 through 2003 (King et al., 2005).9 Growth
was monitored for stands grown in ambient air and under elevated O3 conditions. The elevated
O3 treatment involved increasing hourly concentrations by approximately 1.5 times over the O3
concentrations occurring in ambient air at the site (King et al., 2005). Hourly O3 measurements
were obtained from the authors (for both the "ambient" and "elevated" treatments) and used to
calculate seasonal W126 index for use in the comparisons presented in the 2013 and 2020 ISAs
(cumulative seasonal average in 2013 ISA and single-year average in 2020 ISA). The values for
"observed" above ground total biomass for the aspen stands were derived from measurements
obtained from the authors and allometric equations (2013 ISA, section 9.6.3.2; King et al., 2005
and associated Corrigendum).10
The 2013 ISA presents a comparison of observed biomass to predicted biomass based on
application of the E-R function to W126 in terms of cumulative (multiyear) seasonal average.11
Based on the results of this analysis (which are presented in the Tables 9-14 and 9-15, and Figure
9-20 of the 2013 ISA), the 2013 ISA concludes "the agreement between predictions ... and
observations was very close" and "the function based on one year of growth was shown to be
applicable to subsequent years." (p 9-135). The 2020 ISA also presents such a comparison, but
this one represents the W126 index in terms of single-year seasonal index (2020 ISA, Appendix
8 The median composite function used in the ISAs "was developed from NHEERL/WED data for 11 studies of wild-
type seedlings of aspen as well as four clonally propagated genotypes" (2013 ISA, p. 9-133).
9 Other studies have involved observations involving the same aspen stand extended out to 2008 (e.g., Talhelm et al.,
2014; Zak et al., 2011). Complications associated with performing similar types of comparisons over this longer
time period relate to variation in both the tree measurements taken over the extended period (e.g., diameter
measurements at varying tree heights), and the O- treatments (e.g., the difference in single-year W126 index
ranged from approximately 20 to 30 ppm-hrs through 2003 and then dropped to 10 ppm-hrs for four of the last
five years), as well as changing growth patterns associated with aging trees.
10 The publication by King et al., (2005) reports on measurements for the years 1997 through 2003.
11 The cumulative seasonal average for each year is calculated as the average of the seasonal W126 index values for
that year and all of the preceding years.
4A-20
-------
8, Figure 8-17). These presentations illustrate the variability inherent in the magnitude of growth
impacts of O3 and in the quantitative relationship of O3 exposure and RBL (2013 ISA, Figure 9-
20; 2020 ISA, Appendix 8, Figure 8-17).
4A.3.2 Comparison of Estimated Impacts of Constant and Annually Varying Seasonal
Exposure
The presentation here considers potential differences in aspen growth of O3 exposures
expressed as a constant annual W126 index value compared to exposures expressed as the same
W126 index value in terms of a 3-year average such that the annual values vary while meeting
the 3-year average limit (Figure 4A-15), with the variation reflecting what is shown to be
common at U.S. monitoring locations (e.g., Appendix 4D, section 4D.3.1.2). This analysis is not
intensive or elaborate; rather, it is intended to provide an illustration of concepts associated with
application of the E-R functions described in section 4A. 1 using data from a study with aspen of
the effects of a six-year exposure on accumulating biomass (King et al., 2005), which is also
utilized in a different type of analysis in the 2013 ISA, that is summarized in the ISA and also in
section 4A.3.1 above (2020 ISA, Appendix 8, Figure 8-17; 2013 ISA, section 9.6.3.2).
Description of Analysis: The analysis presented here is intended to simply illustrate the
application of the tree seedling E-R function for aspen over a multi-year period using two types
of air quality scenarios: (1) one in which the O3 concentrations are limited such that each year's
W126 is no higher than 17 ppm-hrs, and (2) a second in which the O3 concentrations are allowed
to vary each year as long as the 3-year average is no higher than 17 ppm-hrs. More specifically,
the two scenarios are (1) repeated years of W126=17 ppm-hrs and (2) repeated 3-year cycles of
the same varying W126 (e.g., 10, 17 and 24 ppm-hrs). This analysis was intended to inform
consideration of potential magnitude of an over or under estimation of growth reduction when
the target W126 value was calculated from a 3-year average or for each individual year. In this
analysis, above-ground tree biomass is estimated for each year through a six-year period.12
The example for this analysis uses aspen, beginning with a seedling, and utilizes data on
growth rates (annual biomass increases) for the control treatment13 in a study by King et al.,
12 In order to avoid extrapolation baseline growth beyond that presented in King et al (2005), the analysis is limited
to the six-year time period. While other Aspen FACE studies have followed the same stand for additional years,
there are aspects of the longer dataset (e.g., different height of tree measurements) that would contribute
uncertainties that lead to the decision to limit the analysis to this duration.
13 Yearly growth- specifically above-ground biomass production- for the baseline situation is based on data
presented in Table 3 of King et al (2005) for the treatment labeled "control." This treatment reflects ambient air
concentrations of O3 at the Aspen FACE location (King et al., 2005). These "control" conditions included single-
year seasonal W126 index values estimated in the 2020 ISA to range from approximately 3 to 9 ppm-hrs across
the 6-year period (2020 ISA, Appendix 8, Figure 8 17), and in the 2013 ISA to be approximately 3 ppm-hrs as the
cumulative multiyear seasonal average across the 6-year period (2013 ISA, Tables 9-14 to 9-15; King et al.,
2005).
4 A-21
-------
2005). The O3 growth effect is estimated by applying the established E-R function for aspen. In
our analysis, above ground biomass loss14 was calculated using the estimated growth rate (yearly
biomass production) and the relative biomass loss (RBL) for the pertinent W126 value based on
the aspen E-R function (Table 4A-8). This biomass loss was calculated for the 3-year average
W126 of 17 ppm and for each of the three individual year values of 10, 17 and 24 ppm (Table
4A-7). King et al., 2005) provided initial biomass for the control and yearly biomass production
for each of the subsequent six years of study. In this analysis, the growth rate information
derived from King were applied over six years of growth (using the yearly growth rates from the
study, g/m2/year for the stand). The above ground biomass of the aspen stand in each year was
compared across the two exposure scenarios (Figure 4A-15; Table 14-7). The difference between
the two 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 always less than 2%.
Summary of Analysis Limitations, Assumptions and Uncertainties: 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
assumption, limitations and uncertainties inherent in the analysis.
• This analysis assumes that baseline growth rate, to which the O3 effect in terms of RBL is
applied, is unaffected by the O3 exposure in the prior year. Although the potential for
"carryover" of an effect of particularly high exposure years into subsequent years (e.g.,
through a reduction in carbohydrate storage) has been discussed in assessments for past
reviews (e.g., 2006 AQCD, section AX9.2.8), the specific exposure levels that might
trigger significant "carryover" effects are not established. Studies cited in those
discussions include exposure levels much higher than conditions associated with meeting
the current standard (e.g., 80 ppb O3 over a full growing season, 2006 AQCD, section
AX9.2.8).
• Additionally, while the availability of multi-year experimental data that can be examined
with regard to this issue for the range of exposures investigated here is limited, a multi-
year study available in the last review (King et al., 2005) did not appear to indicate such
impacts (2013 ISA section 9.6.3.2). The multi-year experimental dataset from King et al
(2005) was assessed in the 2013 ISA and is also discussed in the ISA for this review with
regard to growth effects and correspondence of E-R function predictions with study
observations (2020 ISA, Appendix 8, section 8.13.2 and Figure 8-17; 2013 ISA, section
9.6.3.2, Table 9-15, Figure 9-20). The analysis in the 2013 ISA, which focused on the six
years for which the aspen study reported data, compared observed reductions in growth
for each year of a 6-year period to those predicted by applying the established E-R
function for Aspen to cumulative multi-year average W126 index values (2013 ISA,
14 Above-ground growth (foliage and wood) is used consistent with 2013 and 2020 ISA analyses described in
section 4A.3.1 above).
4A-22
-------
section 9.6.3.2).15 One finding of this evaluation was that "the function based on one year
of growth was shown to be applicable to subsequent years" (2013 ISA, p. 9-135),
indicating that the approach employed in the analysis presented here -for the initial six
years- may be reasonable for the circumstances examined here. In the ISA for this
review, an evaluation slightly different from that in the 2013 ISA was performed,
applying the E-R functions to the W126 index for each year rather than the cumulative
multi-year W126 (2020 ISA, Appendix 8, Figure 8-17). This approach, while indicating
just slightly less tight fit in the later year, was similarly concluded to be "exceptionally
close" to the experimental observations (2020 ISA, Appendix 8, p. 8-192), indicating the
aspen E-R functions to generally reliably predict the yearly findings from the six years of
exposures of the Aspen FACE experiment.
• Variables other than O3 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).
• Additionally, this analysis is based on aspen, and the specific pattern of differences
between the two scenarios might be expected to vary for species with different biomass
growth rates (and E-R functions). 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.
15 For example, the growth impact estimate for year 1 used the W126 index for year 1; the estimate for year 2 used
the average of W126 index in year 1 and W126 index in year 2; the estimate for year 3 used the average of W126
index in years 1, 2 and 3; and so on.
4A-23
-------
Comparison of aboveground growth (biomass)
for annual W126 of 17 ppm-hrs and varying annual W126, with 3-year average of 17 ppm-hrs
3000
2500
w 2000
8
ra
g
m
1500
S5
J 1000
<
500
¦ 'ambient' (control in King) biomass
aW126=17,17,17, etc-biomass (gi'm2)
W 126=10, 24,17, etc-biomass (g
-------
Table 4A-7. 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).
"ambient"
(control
in King)
biomass
Growth -
%
increase
W126=17,
biomass
(q/m2)
W126=10,
24,17,
etc -
biomass
(q/m2)
W126=
24,17,
10, etc -
biomass
(q/m2)
W126=
24,10,
17, etc -
biomass
(q/m2)
W126=
10,17,
24 etc -
biomass
(q/m2)
%
difference
in total
tree
biomass
of
W126=10,
24,17, vs
17
%
difference
in total
tree
biomass
of W126=
24,17,10
vs 17
%
difference
in total
tree
biomass
of W126=
24,10,17
vs 17
%
difference
in total
tree
biomass
of W126=
10,17, 24
vs 17
yO-
1997
9.1
9.1
9.1
9.1
9.1
9.1
Y1
280.2
2979.1%
253.6
266.1
240.9
240.9
266.1
4.9%
-5.0%
-5.0%
4.9%
V2
849.5
203.2%
767.1
752.9
754.4
780.6
779.6
-1.9%
-1.7%
1.8%
1.6%
V3
1335.5
57.2%
1205.5
1191.2
1215.1
1219.0
1195.1
-1.2%
0.8%
1.1%
-0.9%
V4
1581.1
18.4%
1427.0
1424.1
1425.1
1428.9
1428.0
-0.2%
-0.1%
0.1%
0.1%
y5
2099.4
32.8%
1894.6
1867.2
1892.6
1920.3
1895.5
-1.4%
-0.1%
1.4%
0.0%
y6-
2003
2695.2
28.4%
2432.0
2404.6
2457.4
2457.7
2404.9
-1.1%
1.0%
1.1%
-1.1%
4A-25
-------
Table 4A-8. Aboveground growth calculations for subset of scenarios.
"ambient"
(control
in King)
annual
WOOD
growth
that year
"ambient"
(control
in King)
annual
FOLIAGE
growth
that year
"ambient"
(control
in King)
annual
growth
that year
(uses yr-
yr delta
for
foliage)
"ambient"
(control in
King)
aboveground
biomass
Growth -
%
increase
o3
W126
(ppm-
hrs)
03-
impact
(RBL)
W126=17
biomass
(g/m2)
03 W126
(ppm-hrs)
- low->hi-
>ave
03-
impact
(RBL)
W126=10,
24,17, etc
- biomass
(g/m2)
% diff
from
constant
3-yr ave
yO
(1997)
7.6
1.5
9.1
9.1
9.1
yi
226.1
46.5
271.1
280.2
2979.12%
17.0
0.098
253.6
10.0
0.052
266.1
y2
476.6
139.2
569.3
849.5
203.18%
17.0
0.098
767.1
24.0
0.145
752.9
y3
390.5
234.7
486
1335.5
57.21%
17.0
0.098
1205.5
17.0
0.098
1191.2
-1.2%
V4
209
271.3
245.6
1581.1
18.39%
17.0
0.098
1427.0
10.0
0.052
1424.1
-0.2%
y5
434.5
355.1
518.3
2099.4
32.78%
17.0
0.098
1894.6
24.0
0.145
1867.2
-1.4%
ye
(2003)
500.1
450.8
595.8
2695.2
28.38%
17.0
0.098
2432.0
17.0
0.098
2404.6
-1.1%
4A-26
-------
Attachment 1 to Appendix A
4A-27
-------
Derivation of Composite Median Equations (parameterized models)
in Lee and Hogsett (1996)
The following describes the methodology used to produce the sets of parameters in Tables 2, 12,
and 13 of Lee and Hogsett (1996), which have been used in some form in AQCDs and IS As since
1996. "Regression", "parameter estimation", "model estimation" and "model fitting" all refer
to the same statistical procedure of using nonlinear ordinary least square regression to obtain
values for model parameters from a dataset.
1) Tables 12, 13, and 2 in Lee and Hogsett (1996) primarily summarize parameter values
estimated through regression from 51 controlled exposure studies of tree seedlings conducted
by NHEERL/WED. In those studies, 11 species of trees were exposed to a set of ozone
concentrations for durations varying from 55 to 234 days or up to 555 days (in the case of
one species).
2) The model fitted to the data from each of the 51 individual studies (in Table 12) is a three-
parameter Weibull model with the following parameterization: Predicted Biomass = A exp(-
[exposure/B]').u> When removing the intercept A, this model gives biomass relative to no
exposure, and the resulting two-parameter equations all have the same 0-1 range of relative
biomass response and can thus be compared or aggregated across studies with different
ranges of absolute biomass. Predicted Relative Biomass = exp(-[exposure/B]c), and
Predicted Relative Biomass Loss= 1- exp(-[exposure/B]c). When estimating each set of three
parameters for each separate study, the ozone exposure was quantified using the 12-hour
daytime W126 index, summed over the duration of each study.
a. Table 12 gives parameter values for 51 models, one per study, that reflect W126
over each study duration. This table also presents the W126 index estimated for a
92-day duration for RBLs of 10% and 20%.
b. Table 13 presents parameter values for the 51 models given in Table 12, as well
as parameter values for composite models for the 11 tree species included in those
51 studies, two sets of values per species (one for the median and the second for
the 75th percentile). This table also presents W126 index estimates for RBLs of
10%, 20% and 30%. These three estimates for the composite models are estimates
for a 92-day duration.
c. Table 2 presents values for composite models for all experiments for all species
aggregated.
3) The median composite models, one per species (table 13) are derived as follows. For each of
the studies for a given species, the predicted relative biomass loss is first generated at six
values of exposure: 10, 20, 30, 40, 50, and 60 ppm-hr, using the study-specific two-parameter
equation. This is done in a way to obtain six values of exposure for 12-hour daytime
exposures summed over 92 days.17 All but the median of the relative biomass loss estimates
17 Since the W126 index is cumulative, and the duration of exposure varied between studies, the calculated values of
exposure at which some given percent loss is expected were prorated to 92 days using simple linear scaling. For
example, the duration of the first ponderosa pine study in Table 12 is 111 days. To derive the 92day RBL for 10
4A-28
-------
at each value of exposure are then discarded, and the two-parameter model for relative
biomass loss is fitted to the remaining six median points. For example, Ponderosa Pine was
the subject of 11 studies; 11 sets of parameters were estimated through regression (see item 2
above); 66 values of predicted relative biomass loss were then computed, 11 at each of the six
exposure values. All but the median of those 11 relative biomass loss estimates were
discarded at each of the six levels of exposure, and the two-parameter model fitted to the six
remaining points, giving the Ponderosa Pine median composite equation for a 92-day
exposure.
4) The all-species, median composite models (table 2) were estimated using the same
aggregation method, but applied to all 51 studies at once. The 51 equations in Table 12 were
used to compute 306 values of relative biomass loss, six values each for the 51 sets of B and C
parameters, with those six values of exposure generated in a way to obtain 12-hour daytime
exposures summed over 92 days. The two-parameter model was then fitted to the 75th
percentile and the median as in item 3 above. Table 2 also includes the results of the same
method using other exposure indices besides the 12-hour W126 index.
5) For every equation in the tables, values of exposure at which some given percent loss is
expected relative to no exposure, or any other exposure, can be back-calculated using:
Exposure = B * (-In (1-predicted relative biomass loss))1/c. Some of those expected values of
exposure are presented for various loss percentages in tables 2, 12, and 13 for the all-species
median composite model, the 51 studies, and the 11 species-level median composite,
respectively. In the case of single-study calculations in Table 12, the value of exposure for a
given loss percentage was first calculated based on the respective duration of each study, then
simply prorated. For example, the duration of Study 1 Harvest 1 in table 12 was 84 days and
the exposure at which 10% loss is expected over that duration is 13.71 ppm-hr. The prorated
exposure value for a 10% loss over 92 days is calculated as 13.71*92/84 = 15.01 ppm-hr.
Median models for species in Table 13 or for all species in Table 2 were parameterized with
92-day durations and the exposure values for the various loss percentages did not therefore
require prorating.
ppm-hrs, a factor of 92/111 is applied to 10 ppm-hrs before it is input to the experiment-specific equation to
derive an RBL estimate for 10 ppm-hrs over a 92-day exposure.
4A-29
-------
Attachment 2 to Appendix A
As reference, the following table presents calculations for a simple case of multi-year exposure,
application of E-R function estimates, and tree biomass accumulating over that period.
4A-30
-------
Example calculations for simple case of multi-year exposure application of E-R function estimates.
CONCEPTUAL EXAMPLE (RBL =fraction reduction from control tree growth that year)
Growth rate
(fraction of
biomass at end
of prior year)
control
03
RBL
0.3
Y0
1
Y1
1.3
0.138
Y2
1.69
0.118
Y3
2.197
0.098
0.2
Y0
1
Y1
1.2
0.138
Y2
1.44
0.118
Y3
1.728
0.098
0.5
Y0
1
Y1
1.5
0.138
Y2
2.25
0.118
Y3
3.375
0.098
0.1
Y0
1
Y1
1.1
0.138
Y2
1.21
0.118
Y3
1.331
0.098
03 tree
1
1.259
1.603
2.060
0.115
<- RBL over full period
1
1.172
1.384
1.644
0.116 <- RBL over full period
1
1.431
2.093
3.107
0.113 <- RBL over full period
1
1.086
1.183
1.292
0.117 <- RBL over full period
Growth rate
(fraction of
biomass at end
of prior year)
0.3
0.2
0.5
0.1
03
03
control
RBL
tree
Y0
1
1
Y1
1.3
0.118
1.265
Y2
1.69
0.118
1.609
Y3
2.197
0.118
2.056
0.118
Y0
1
1
Y1
1.2
0.118
1.176
Y2
1.44
0.118
1.388
Y3
1.728
0.118
1.642
0.118
Y0
1
1
Y1
1.5
0.118
1.441
Y2
2.25
0.118
2.103
Y3
3.375
0.118
3.095
0.118
Y0
1
1
Y1
1.1
0.118
1.088
Y2
1.21
0.118
1.185
Y3
1.331
0.118
1.292
0.118
<- RBL over full period
4A-31
-------
REFERENCES
Andersen, CP and Scagel, CF (1997). Nutrient availability alters belowground respiration of
ozone-exposed ponderosa pine. Tree Physiology 17(6): 377-387.
Andersen, CP, Wilson, R, Plocher, M and Hogsett, WE (1997). Carry-over effects of ozone on
root growth and carbohydrate concentrations of ponderosa pine seedlings. Tree
Physiology 17(12): 805-811.
Gumpertz, ML and Rawlings, JO (1992). Nonlinear regression with variance components:
Modeling effects of ozone on crop yield. Crop Science 32(1): 219-224.
Heck, WW, Cure, WW, Rawlings, JO, Zaragoza, LJ, Heagle, AS, Heggestad, HE, Kohut, RJ,
Kress, LW and Temple, PJ (1984). Assessing impacts of ozone on agricultural crops: II.
Crop yield functions and alternative exposure statistics. Journal of the Air Pollution
Control Association 34(8): 810-817.
Hogsett, WE, Herstrom, AA, Laurence, JA, Lee, EH, Weber, JE and Tingey, DT, Eds. (1995).
Risk characterization of tropospheric ozone to forests. Air & Waste Management
Association Pittsburgh, PA.
Hogsett, WE, Weber, JE, Tingey, D, Herstrom, A, Lee, EH and Laurence, JA (1997).
Environmental auditing: An approach for characterizing tropospheric ozone risk to
forests. Journal of Environmental Management 21(1): 105-120.
Isebrands, JG, Dickson, RE, Rebbeck, J and Karnosky, DF, Eds. (2000). Interacting effects of
multiple stresses on growth and physiological processes in northern forest trees. Springer-
Verlag New York, NY.
Karnosky, DF, Gagnon, ZE, Dickson, RE, Coleman, MD, Lee, EH and Isebrands, JG (1996).
Changes in growth, leaf abscission, biomass associated with seasonal tropospheric ozone
exposures of Populus tremuloides clones and seedlings. Can J For Res 26(1): 23-37.
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. New Phytol 168(3): 623-635.
Lee, EH and Hogsett, WE (1996). methodology for calculating inputs for ozone secondary
standard benefits analysis part II. Office of Air Quality Planning and Standards. Research
Triangle Park, NC.
Lee, EH and Hogsett, WE (1999). Role of concentrations and time of day in developing ozone
exposure indices for a secondary standard. J Air Waste Manage Assoc 49(6): 669-681.
Lee, EH, Hogsett, WE and Tingey, DT (1994). Attainment and effects issues regarding
alternative secondary ozone air quality standards. Journal of Environmental Quality
23(6): 1129-1140.
4A-32
-------
Lee, EH, Tingey, DT and Hogsett, WE (1987). Selection of the best exposure-response model
using various 7-hour ozone exposure statistics. U.S. Environmental Protection Agency.
Research Triangle Park, NC.
Lee, EH, Tingey, DT and Hogsett, WE (1988). Evaluation of ozone exposure indices in
exposure-response modeling. Environmental Pollution 53(1-4): 43-62.
Lee, EH, Tingey, DT and Hogsett, WE (1989). Interrelation of experimental exposure and
ambient air quality data for comparison of ozone exposure indices and estimating
agricultural losses. EPA/600/3-89/047. U.S. Environmental Protection Agency. Corvallis,
OR.
Lefohn, A, Shadwick, D, Somerville, M, Chappelka, A, Lockaby, B and Meldahl, R (1992). The
characterization and comparison of ozone exposure indices used in assessing the response
of loblolly pine to ozone. Atmospheric Environment, Part A: General Topics 26(2): 287-
298.
Lesser, VM, Rawlings, JO, Spruill, SE and Somerville, MC (1990). Ozone effects on agricultural
crops: Statistical methodologies and estimated dose-response relationships. Crop Science
30(1): 148-155.
Neufeld, HS, Lee, EH, Renfro, JR, Hacker, WD and Yu, BH (1995). Sensitivity of seedlings of
black cherry (Prunus serotina Ehrh) to ozone in Great Smoky Mountains National Park I
Exposure-response curves for biomass. New Phytol 130(3): 447-459.
Neufeld, HS and Renfro, JR (1993). Sensitivity of black cherry seedlings (Prunus serotina Ehrh.)
to ozone in Great Smoky Mountains National Park: The 1989 seedling set. NPS/NRTR-
93/112. U.S. Department of the Interior; National Park Service. Washington, DC.
Qiu, Z, Chappelka, AH, Somers, GL, Lockaby, BG and Meldahl, RS (1992). Effects of ozone
and simulated acidic precipitation on above- and below-ground growth of loblolly pine
(Pinus taeda). Can J For Res 22(4): 582-587.
Rawlings, JO and Cure, WW (1985). The Weibull function as a dose-response model to describe
ozone effects on crop yields. Crop Science 25(5): 807-814.
Rebbeck, J and Loats, K (1997). Ozone effects on seedling sugar maple (Acer saccharum) and
yellow-poplar (Liriodendron tulipifera): gas exchange. Can J For Res 27(10): 1595-1605.
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. Global Change Biol 20(8): 2492-2504.
U.S. EPA (2007). Review of the National Ambient Air Quality Standards for Ozone: Policy
Assessment of Scientific and Technical Information: OAQPS Staff Paper. Office of Air
Quality Planning and Standards. Research Triangle Park, NC. U.S. EPA. EPA-452/R-07-
4A-33
-------
003. January 2007. Available at:
https.Y/nepis. epa.gov/Exe/ZyPURL. cgi?Dockey=Pl0083VX. txt.
Zak, DR, Pregitzer, KS, Kubiske, ME and Burton, AJ (2011). Forest productivity under elevated
C02 and 03: positive feedbacks to soil N cycling sustain decade-long net primary
productivity enhancement by C02. Ecol Lett 14(12): 1220-1226.
4A-34
-------
APPENDIX 4B
U.S. DISTRIBUTION OF 11 TREE SPECIES
4B-1
-------
4B.1. DESCRIPTION
This appendix presents maps of the distribution across the U.S. of 11 tree species for
which there are established exposure-response (E-R) functions, as described in Appendix 4A.
Historical ranges were based on Little (1971, 1976, 1977, and 1978) and basal area of each
species was taken from Wilson et. al (2013) raster data to show present range and estimated
density. Basal area is computed at the stand level as the sum of the basal area values for each
individual tree (in sq. ft.), which is summed across all of the basal area per tree in the
hectare. The map construction consists of tree species abundance, distribution, and basal area at
a 250-meter (m) pixel size for the contiguous United States (Wilson 2013).
4B-2
-------
uaking aspegj^Hopulus tremuloides) \
Range (Little, 1971)
~ Historic range
Basal area (Wilson et al. 2013) o
179.82
0.0003 square meters/hectare
Table 4B-1. Distribution of quaking aspen (Populus tremuloules) in the continental U.S.
4B-3
-------
Red maple (Acer rubrum)
Range (Little, 1971)
Historic range
Basal area (Wilson et al. 2013)
111.04
*- < 0.0001 square meters/hectare
Table 4B-2. Distribution of red maple (Acer rubrum) in the continental U.S.
4B-4
-------
Sugar maple (Acer saccharum)
Range (Little, 1971)
[=~ Historic range
Basal area (Wilson et al. 2013)
¦MT 12321
< 0.0001 square meters/hectare
Table 4B-3. Distribution of sugar maple (Acer saccharum) in the continental U.S.
4B-5
-------
Red alder (Alnus rubra)
Range (Little, 1971)
~ Historic range
Basal area (Wilson et al. 2013)
139 96
-0016 square meters,hectare
Table 4B-4. Distribution of red alder {Alnus rubra) in the continental U.S.
4B-6
-------
Yellow-poplar (Liriodendron tulipifera)
Range (Little, 1971)
~ Historic range
Basal area (Wilson et al. 2013)
v 103 62
' ' * 0 0005 square meters/hectare
Table 4B-5. Distribution of tulip poplar (Liriodendrun tulipifera) in the continental U.S.
4B-7
-------
Ponderosa pine (Pinus ponderosa)
"¦ 1' O^OOOB square meters/hectare
Range (Little, 1971)
~ Historic range
Basal area (Wilson et
. i, 159.77
e>
al. 2013)
Table 4B-6. Distribution of ponderosa pine (Pinus ponderosa) in the continental U.S.
4B-8
-------
Eastern white pine (Pinus strobus)
Range (Little, 1971)
~ Historic range
Basal area (Wilson et al. 2013)
113.7
0.0004 square meters/hectare
Table 4B-7. Distribution of eastern white pine (Pinus strobus) in the continental U.S.
4B-9
-------
Range (Little, 1971)
Historic range
Basal area (Wilson et al. 2013)
¦Mr 14273
0 0004 square meters/hectare
Loblolly pine (Pinus taeda)
Table 4B-8. Distribution of loblolly pine (Pinus taeda) in the continental U.S.
4B-10
-------
Range (Little, 1971)
Historic range
Virginia pine (Pinus virginiana)
Basal area (Wilson et al. 2013)
104-23
0 000B square meters/hectare
Table 4B-9. Distribution of Virginia pine (Pinus virginiana) in the continental U.S.
4B-11
-------
Black cherry (Prunus serotina)
Range (Little, 1971)
(=~ Historic range
Basal area (Wilson et afo2013)
r 85-39
' 0.0002 square meters/hectare
Table 4B-10. Distribution of black cherry (Prunus serotina) in the continental U.S.
4B-12
-------
Douglas fir (Pseudostuga menziesii)
o
Range (Little, 1971)
~ Historic range
Basal area (Wilson et al. 2013)
wmw 374.66 ^
L 0.0008 square meters/hectare
Table 4B-11. Distribution of Douglas fir (Pseudotsuga menziesii) in the continental U.S.
4B-13
-------
REFERENCES
Little, E.L., Jr., 1971, Atlas of United States trees, volume 1, conifers and important hardwoods:
U.S. Department of Agriculture Miscellaneous Publication 1146, 9 p., 200 maps.
Little, E.L., Jr., 1976, Atlas of United States trees, volume 3, minor Western hardwoods: U.S.
Department of Agriculture Miscellaneous Publication 1314, 13 p., 290 maps.
Little, E.L., Jr., 1977, Atlas of United States trees, volume 4, minor Eastern hardwoods: U.S.
Department of Agriculture Miscellaneous Publication 1342, 17 p., 230 maps.
Little, E.L., Jr. 1978, Atlas of United States trees, volume 5, Florida: U.S. Department of
Agriculture Miscellaneous Publication 1361, 262 maps.
Wilson, Barry Tyler; Lister, Andrew J.; Riemann, Rachel I.; Griffith, Douglas M. 2013. Live tree
species basal area of the contiguous United States (2000-2009). Newtown Square, PA:
USDA Forest Service, Rocky Mountain Research Station. https://doi.org/10.2737/RDS-
2013-0013
4B-14
-------
APPENDIX 4C
VISIBLE FOLIAR INJURY SCORES AT U.S. FOREST SERVICE
BIOSITES (2006-2010)
TABLE OF CONTENTS
4C.1 Introduction 2
4C.2 Dataset Preparation 2
4C.3 Dataset Characteristics 5
4C.4 Relationships of Biosite Index Scores with W126 Estimates and Soil Moisture
Categories 6
4C.4.1 Relationships Examined in Full Dataset 6
4C.4.2Examination of Relationships in Dataset Stratified by Soil Moisture Category 9
4C.5 Limitations and Uncertainties 19
4C.6 Summary and Key Observations 20
References 22
4C-1
-------
4C.1 INTRODUCTION
It has long been recognized that elevated ozone (O3) can cause visible foliar injury in
some plants (ISA, Appendix 8, section 8.2). As discussed in the current and past ISAs as well as
past Air Quality Criteria Documents, the severity and extent of visible foliar injury can vary with
a variety of environmental variables (e.g., climatic variables as well as pollutant exposure) as
well as variation in genetic factors within the same plant population (ISA, Appendix 8, section
8.2). Visible foliar injury "occurs only when sensitive plants are exposed to elevated O3
concentrations in a predisposing environment," and "a major modifying factor is the amount of
soil moisture available to a plant during the year when assessed" (U.S. EPA, 2013 [2013 ISA], p.
9-39).
In recognition of the long-standing evidence regarding O3 and visible foliar injury in
susceptible species, the U.S. Forest Service (USFS) and U.S. Park Service have used plant
species with this susceptibility in their biomonitoring programs. A number of publications have
focused on findings from biomonitoring surveys in the USFS-Forest Health Monitoring (FHM)
and Forest Inventory and Analyses (FIA) programs. From the mid 1990s through 2010, this
survey work included collecting information on the presence of visible foliar injury at the
biomonitoring sites (biosites). Data on visible foliar injury incidence and severity data were
collected each year at biosites in forested areas at states across the U.S. and summarized in terms
of a biosite index (BI). The BI is a measure of the severity of 03-induced visible foliar injury
observed at each biosite.
Data from the multi-year USFS survey were used in analyses developed in the 2015 O3
NAAQS review (80 FR 65292, October 26, 2015). These analyses utilized a dataset that had
been developed by merging biosite data collected as part of the USFS FHM/FIA Network during
the years 2006 through 2010, with NOAA soil moisture index values (as a surrogate for soil
moisture measurements) and W126 estimates of seasonal O3 exposure for those sites based on
ambient air monitoring data for those 5 years (Smith and Murphy, 2015; U.S. EPA, 2014 [2014
WREA]) The resultant combined dataset included a BI score, soil moisture index value and a
W126 index estimate each for 5,284 records at locations in 37 states for 1 or more of the years in
the 5-year period from 2006-2010. This appendix brings forward key presentations developed
from the combined dataset for the 2015 O3 review and also includes additional presentations of
key aspects of the dataset and the variables represented within it.
4C.2 DATASET PREPARATION
The combined dataset was developed from three datasets: (1) the national-scale
FIA/FHM dataset of BI scores, (2) the NOAA's National Climatic Data Center national dataset
of monthly drought indices and (3) national surfaces of estimated seasonal W126 index
4C-2
-------
developed by the EPA for the WREA in the last O3 NAAQS review and further analyzed in a
subsequent technical memo (Smith and Murphy, 2015). These individual datasets and how they
were used to create the combined dataset, are described below.
Biosite Index: The USFS O3 biomonitoring program has developed a national-scale data
set focused on visible foliar injury and that includes BI scores at biosites in U.S. forests (Smith,
2012). The field methods, sampling procedures, and analytical techniques are consistent across
biosites and years. The BI is calculated from species-specific scores based on a combination of
the proportion of leaves affected on individual bioindicator plants and the severity of symptoms
on injured foliage using an established scale (Horsfall and Cowling, 1978; Smith, 2012). Each
site is sampled until 30 plants of at least two species have been evaluated (Smith et al., 2007).
The site BI is the average score for each species averaged across all species on the biosite
multiplied by 1,000 (Smith, 2012). The BI score ranges from zero to greater than 25, with a score
of zero indicating no presence of foliar injury symptoms and scores increasingly greater than
zero indicating increasingly greater severity of symptoms (Smith, 2012). Categories that have
been used in publications include little or very light injury (BI greater than 0 up to 5), light injury
(BI greater than 5 up to 15), moderate (BI greater than 15 up to 25) and heavy/severe (BI above
25) (Smith, 2012; Coulston etal., 2003).
The biosite data (BI scores) were obtained from the USFS for the years 2006 to 2010.
While including most states in the contiguous U.S., the data obtained did not include records for
most of the western states (Montana, Idaho, Wyoming, Nevada, Utah, Colorado, Arizona, New
Mexico, Oklahoma, and portions of Texas) because biosite data were not available for those
states during the 2006-2010 period (Smith et al., 2012).
Soil Moisture Index: The NOAA Palmer Z drought index is a monthly moisture anomaly
index that is derived from measurements such as precipitation and temperature. This index
represents the difference between monthly soil moisture and long-term average soil moisture
(Palmer, 1965). The Palmer Z index is derived each month for each of 344 climate region
divisions within the contiguous U.S. by the National Climatic Data Center (NCDC).1 The index
values typically range from -4 to +4, with positive values representing more wetness than normal
and negative values representing more dryness than normal. For the combined dataset, index
values for April through August in the years 2006-2010 were obtained from the NCDC website
(NOAA, 2012). These monthly values were then averaged to create a single growing season
index for each year in each division. Moisture categories were then assigned consistent with
1 There are 344 climate divisions in the continental U.S. For each climate division, monthly station temperature and
precipitation values are computed from the daily observations as described on the website for the National
Climatic Data Center of the U.S. National Atmospheric and Oceanic Administration:
https://www.ncdc.noaa.gov/monitoring-references/maps/us-climate-divisions.php
4C-3
-------
NOAA's Palmer Z drought index, with index values less than -1.25 identified as "dry", values
greater than or equal to 1 identified as "wet", and index values between -1.25 and 1 identified as
"normal." Values beyond the range from -2.75 to +3.5 could be interpreted as extreme drought
and extremely moist, respectively (NCDC, 2012c). The NCDC climate divisions with Palmer Z
data are shown in Figure 4C-1.
Figure 4C-1. Climate divisions for which there are Palmer Z soil moisture index values.
W126 Index Estimates: Estimates of seasonal VV126 exposure index for the years 2006
through 2010 were developed for 12 kilometer (km) by 12 km grid cells in a national-scale
spatial surface. The estimates at this scale were derived from applying a spatial interpolation
technique to annual W126 values derived from O3 measurements at ambient air monitoring
locations. Specifically, the Voronoi Neighbor Averaging (VNA) spatial interpolation technique
was applied to the monitor-location W126 index values to derive an W126 index estimates for
each grid cell (U.S. EPA, 2014, Appendix 4A).2
Combined Dataset: To create the dataset that relates the grid cells with W126 index
estimates to grid cells with B1 scores, the EPA provided a file with the national-scale surface of
grid cells (a "shape" file) to USFS staff, who assigned the BI scores (with sampling year
specified) to grid cells for all but three states. Having this step performed by the USFS ensured
2 The VNA application step used to estimate W126 indices at the centroid of every 12 km x 12 km grid cell, rather
than only at each monitor location (described in Appendix 4 A of the WREA), can result in a lowering of the
highest values in each region (80 PR 65374-65375; October 26, 2015).
4C-4
-------
that the precise and accurate geographic coordinates for each biosite were used in this step,
which allowed the most accurate matching of Palmer Z and W126 index values as possible with
these datasets.3 For three states (California, Oregon, and Washington) the EPA downloaded
biosite indices from the public website and assigned them to the grid cells in which the biosite
was located based on the publicly available geographic coordinates.4 The EPA overlaid the
Palmer Z dataset for each year on the national surface of W126 index estimates for that year to
assign a Palmer Z index to each grid cell in each year's national surface. The completed dataset
(Smith and Murphy, 2015, Appendix) includes the following variables: identifier, year, W126
index, BI score, Palmer Z index, state and soil moisture category (dry, wet, normal)5.
4C.3 DATASET CHARACTERISTICS
The dataset for the analyses included 5,284 biosite records distributed across the 37
different states and the five years from 2006 - 2010 (Smith and Murphy, 2015, Appendix).
Figure 4C-2, reprinted from 2014 WREA, indicates the distribution of sites across the
continental U.S. Table 4C-2 summarizes the biosite index values for each year. The "Damage"
categories used follow the USFS risk categories with the exception of including a separate
category for a biosite index of zero (Smith, 2008, 2012). The zero category was defined and used
as a measure of the presence or absence of any level of visible foliar injury. Across all of the
sites, over 81 percent of the observations recorded no foliar injury. This percentage was similar
across all of the years, with a low value of 78 percent and a high value of 85 percent. Across the
5,284 records in the dataset, only 998 had BI scores greater than zero.
3 This step was taken because the publicly available USFS BI dataset includes location coordinates that have been
slightly altered to avoid specifying the true biosite location for privacy considerations of some property owners.
4 As a result, there is a potential for the biosites for these states to be matched with the W126 index estimate for an
adjacent grid cell rather than the one in which the biosite is truly located.
5 As described earlier in the section on "Soil Moisture Index," all index values less than -1.25 were categorized as
"dry" and all index values greater than or equal to 1 were categorized as "wet."
4C-5
-------
'£ >
FHM Biosites
2010
2009
2008
2006
Figure 4C-2. USFS biomonitoring sites for visible foliar injury ("Biosites").
Table 4C-1. Summary of biosite index scores for 2006 to 2010 USFS biomonitoring sites.
Biosite
Index
Damage
2006
2007
2008
2009
2010
Total
0
None
744
769
796
902
1,075
4,286
<5
Very Liqht
139
131
98
135
183
686
5 to 15
Light
41
29
29
61
65
225
15 to 25
Moderate
15
6
8
6
12
47
>25
Heavy
12
4
4
8
12
40
Total
951
939
935
1,112
1,347
5,284
4C.4 RELATIONSHIPS OF BIOSITE INDEX SCORES WITH W126
ESTIMATES AND SOIL MOISTURE CATEGORIES
4C.4.1 Relationships Examined in Full Dataset
Scatterplots of the full dataset show no clear relationship between Cb and biosite index
(Figure 4C-3), as well as no clear relationship between Os and the Palmer Z drought index,
measured as an average value of the months from April to August (Figure 4C 4). The lack of a
clear relationship is partly due to the high number of observations with no foliar injury (see
Table 4C-1 above and also the distribution of records by soil moisture category and W126
summarized in section 4C.4.2 below) and may also reflect, in part, differing spatial resolutions of
4C-6
-------
the O3 exposure surface, NCDC climate divisions, and the biosites. To investigate the strength of
any relationship in light of the high percentage of zero values, a censored regression was
conducted using a threshold of zero (i.e., including only the non-zero observations). The results
of the regression (Table 4C-2) are consistent with the evaluation of the evidence in the ISA (and
prior ISA and AQCDs), indicating a significant relationship between foliar injury and both O3
and moisture (as measured by Palmer Z), and also a significant interaction between O3 and
moisture. The censored regression does not provide a "goodness of fit" statistic as easily
interpreted as the r-squared value associated with a standard regression, so the results are more
difficult to interpret. Thus, while higher O3 corresponds to higher BI score, the parameters
describing such a relationship in predictive quantitative terms are unresolved.
0 10 20 30 40 50 60
W126 (ppm-hrs)
Figure 4C-3. Scatter plot of biosite index score versus W126 index (ppm-hrs).
4C-7
-------
o
LO —
o
o —
X T—
CD
"O
n.
B
"(7)
o
CD
O _
LO
-2 0 2 4
Average Palmer Z (April to August)
Figure 4C-4. Scatter plot of biosite index score versus Palmer Z (April to August).
Table 4C-2. Statistics from censored regression.
Coefficient
Intercept Estimate
Standard Error
t-value
P
Intercept
-22.5967
0.8934
-25.293
< 0.0001
W126
0.7307
0.0613
11.919
<0.0001
Palmer Z (Apr-Aug)
1.8357
0.4850
3.785
0.0002
W126: Palmer Z
0.1357
0.0437
3.104
0.0019
Marginal Effect
W126
0.1178
0.0099
11.918
<0.0001
Palmer Z (Apr-Aug)
0.2960
0.0777
3.812
0.0001
W126: Palmer Z
0.0219
0.0070
3.093
0.0020
An exploration (in the 2014 WREA) of the use of regression coefficients to calculate
estimated biosite index values did not accurately predict the observed values, likely due in part to
the large number of non-injury observations. It is also of note that the W126 index bin with the
largest percentage of records of each category of BI score (e.g., all, zero, above zero, above 5,
above 15) is that for the lowest W126 index values (0).
4C-8
-------
Table 4C-3. Cumulative percentage of records with specified BI score.
<7
ppm-hrs
>7 -9
ppm-hrs
>9-11
ppm-hrs
>11 -13
ppm-hrs
>13-15
ppm-hrs
>15-17
ppm-hrs
>17-19
ppm-hrs
>19-25
ppm-hrs
>25
ppm-hrs
Cumulative Percentage of Records (percent of records in bin plus all bins to its left)
Of All Records
42%
59% 73%
82% 88%
92%
96%
98%
100%
Of Records with Bl=0 (total in dataset =4286)
43%
60% 73%
83% 88%
93%
97% 99%
100%
Of Record
s with Bl>0 (total in dataset =998)
37%
53% 69%
78%
84%
87%
93%
96%
100%
Of Record
s with Bl>5 (total in dataset =310)
36%
49%
64%
73% 78%
82%
88%
91%
100%
Of Record
s with BI >
5 (total in dataset =85)
33%
45%
49%
59%
64%
69%
78%
86%
100%
4C.4.2 Examination of Relationships in Dataset Stratified by Soil Moisture Category
The following tables and figures describe the data in this dataset with a focus on
consideration of potential trends with W126 index for the different soil moisture categories. The
W126 index estimates were rounded to integer values for consistency with Appendix 4D
analyses and associated clarity in binning of the values.6 Additionally, consistent with USFS
publications (e.g., Campbell et al., 2007), the BI scores7 are rounded to one decimal place. Table
4C-4 presents the counts of records in total and stratified by soil moisture category and W126
index bin. Table 4C-5 presents average BI scores by soil moisture category and W126 bin, and
Table 4C-6 presents the fraction of records with BI scores of differing severity levels
(corresponding to the USFS severity scheme), in the full dataset and also in the subsets by soil
moisture category.
The distribution of records across W126 bins are presented in Table 4C-4 and Figure 4C-
5, and the distribution of scores per bin is presented in Figure 4C-6 through Figure 4C-11. These
figures show that even the lowest W126 index bin (for estimates below 7 ppm-hrs) includes
scores well above 5, and several above 15. Further, zero scores comprise more than half the dry
and normal soil category record scores in eveiy bin, including the highest bin (>25 ppm hrs), as
6 The presentations here are not precise statistical analyses. Rather, they are intended to generally inform
conclusions regarding ability of available datasets to discern air quality conditions contributing to visible foliar
injury occurrences of potential concern. In this light, binning was used to explore the potential for clear
differences in BI scores among sites with differing W126 estimates across the range of interest while also
maintaining reasonable sample sizes.
7 Two records with estimated W126 index below 7 ppm-hrs and BI scores just over 150 are omitted from
presentations in this section as the next highest BI score in this dataset for any W126 index was below 100.
4C-9
-------
seen by the median lines merged with the zero line in Figure 4C-6 and Figure 4C-8. This is also
the case for all but the two highest bins for the wet soil moisture category records, which,
however, contain just a total of 9 records, limiting the extent to which they provide a basis for
interpretation of patterns across W126 bins. The wet soil moisture records have quite limited
sample sizes for the higher W126 index bins, e.g., the number of samples in bins for W126 index
estimates above 13 ppm-hrs represent no more than 1 percent of the total number of wet soil
moisture records (Figure 4C-10).
Focusing on the distribution of scores for records in the normal soil moisture category, it
can be seen that scores are noticeably increased in the highest W126 bin, index estimates greater
than 25 ppm-hrs, over those for the lower bins (Figure 4C-6 and Figure 4C-7). This is also for
the average BI scores per bin, where the highest W126 bin (>25 ppm-hrs) has an average BI
appreciably higher than the others (Table 4C-5). The average BI in this highest bin is 7.9 versus
averages of 1.6 (for W126 >19 to 25 ppm-hrs) and 2.3 (for W126 >17 to 19 ppm-hrs) in the next
lower bins and varying from 0.8 to 1.2 in all the others. Among the records with nonzero scores,
the highest average BI is also in the highest W126 index bin (>25 ppm-hrs); in this case the BI is
approximately 15, more than double the next highest average BI scores for any of the other
W126 index bins (for which no other trend is exhibited). The incidence of records with BI scores
categorized by the USFS as "moderate" or "severe" injury (BI score above 15) is also greatest in
the bin for the highest W126 index estimates (> 25 ppm-hrs), with 20% of those records in this
bin having such a BI score compared to only 2 to 4% of the records in each of the lower bins. A
similar pattern holds for the records with BI scores above 5, while there is much more variability
across the bins for records with any nonzero score (Table 4C-6).
With regard to the dry soil moisture category, there is a suggestion of an increased
incidence of the highest severity scores in the highest two W126 bins. For example, the
proportion of diy soil moisture category records with BI scores categorized by the USFS as
"moderate" or "severe" injury (BI score above 15), is 7 and 8% in the bin for the two highest
W126 index estimates (>19 to 25 and > 25 ppm-hrs, respectively), compared to 0 to 3% in each
of the lower bins. It is noteworthy, however, that the percentages of 7 and 8% reflect no more
than 4 or 5 individual records with this severity score.
As noted above, sample size for the wet soil moisture category is particularly limited for
the W126 index bins above 13 ppm-hrs. In the lower W126 bins, the proportion of such records
with BI scores above 15 varies from 1 to 2%. For BI scores above 5 or above 0, there is a
suggestion of an increased incidence in the relatively higher versus lower W126 index bins,
although it is not known if the significant reduction in sample size that also occurs in comparing
across these bins (see Table 4C-4) and associated variability is playing a role (Table 4C-6).
4C-10
-------
Table 4C-4. Number of biosite records in different W126 index bins.
<7
ppm-hrs
>7 -9
ppm-hrs
>9-11
ppm-hrs
>11 -13
ppm-hrs
>13-15
ppm-hrs
>15-17
ppm-hrs
>17-19
ppm-hrs
>19-25
ppm-hrs
>25
ppm-hrs
All Records (n=5282 A)
Dry (n=866)
155
117
116
76
83
97
99
73
50
Normal (n=3227)
1181
613
522
360
222
147
92
49
41
Wet (n=1189)
868
179
81
43
9B
7B
2b
0B
0B
All
2204
909
719
479
314
251
193
122
91
Records with Bl > 15 (total in dataset =85)
Dry
3
0
0
0
0
0
3
5
4
Normal
20
7
3
7
4
3
3
2
8
Wet
5
3
1
1
0B
2B
1 B
0B
0B
All
28
10
4
8
4
5
7
7
12
Records with Bl>5 (total in dataset =310)
Dry
6
3
5
3
4
1
5
8
10
Normal
56
30
28
18
11
8
12
3
17
Wet
49
9
13
6
1 B
2b
2b
0B
0B
All
111
42
46
27
16
11
19
11
27
Records with Bl>0 (total in dataset =998)
Dry
10
13
9
6
6
9
15
15
23
Normal
158
117
109
68
52
20
35
15
21
Wet
197
36
34
17
5B
4B
2B
0B
0B
All
365
166
152
91
63
33
52
30
44
Records with Bl=0 (total in dataset =4286)
Dry
145
104
107
70
77
88
84
58
27
Normal
1023
496
413
292
170
127
57
34
20
Wet
671
143
47
26
4B
3B
0B
0B
0B
All
1839
743
567
388
251
218
141
92
47
A As noted in the beginning of section 4C.1.2, this count reflects the omission of two outlier values.
B Sample size for this W126 bin is below 1 % of all samples assigned this soil moisture category.
4C-11
-------
All records (# records per W126 bin)
Bl>0 (# records per W126 bin)
2500
2000
1500
1000
500
0
400
350
300
Dry (n=866) i Normal (n=3227) >Wet(n=1189) i All (n=5282) 25Q
200
150
100
50
0
¦ill j.l .l-l .1 I .1 I .. I
Dry (n=106) ¦ Normal (n=595) Wet(n=284) All (n=996)
<7 >7-9 >9-11 >11-13 >13-15 >15-17 >17 -19 >19 - 25 >25 <7 >7-9 >9-11 >11-13 >13-15 >15-17 >17 -19 >19 - 25 >25
Bl>0 (# records per W126 bin)
Bl>15 (# records per W126 bin)
400
350
300
250
200
150
100
50
0
.11
30
25
Dry (n=106) I Normal (n=595) Wet(n=284) All (n=996) 20
15
10
Jl -I.I -lt-iL.il ¦¦ I ¦¦ I o
Dry (n=15) ¦ Normal (n=57) iWet(n=10) ¦ All (n=85)
1
¦
1
III 1.1
1.1 1
II ll
III
II
1.1
III
<7 >7-9 >9-11 >11-13 >13-15 >15-17 >17-19 >19-25 >25 <7 >7-9 >9-11 >11-13 >13-15 >15-17 >17-19 >19-25 >25
Figure 4C-5. Distribution of biosite records by W126 bin and soil moisture type.
4C-12
-------
100
90 -
80
70
60
0
v
m
50 -
at
40 "
30
20
10
o L
<7
>7 -9
>9-11 >11-13 >13-15 >15-17 >17-19 >19-25
>25
Key: The boxes denote the 25th, 50th and 75th percentiles, the x's the mean and the whiskers denote the value equal to the 75th
percentile plus 1.5 times the interquartile range (75th minus 25th percentile). Circles show scores higher than that.
Figure 4C-6. Distribution of BI scores (including zeros) at USFS biosites (normal soil
moisture) grouped by W126 index values.
o
u
<7
>7-9 >9-11 >11-13
>13-15 >15-17 >17-19
W126 Index Bin
>19 - 25 >25
Key: The boxes denote the 25th, 50th and 75th percentiles, the x's the mean and the whiskers denote the value equal to the 75th
percentile plus 1.5 times the interquartile range (75th minus 25th percentile). Circles show scores higher than that.
Figure 4C-7. Distribution of nonzero BI scores at USFS biosites (normal soil moisture)
grouped by W126 index values.
4C-13
-------
30
70
60
50
40
30
20
1©
0
<7 >7-9 >9-11 >11-13 >13-15 >15-17 >17-19 >19-25
W126 Index Vtlue
x
—£_
>25
Key: The boxes denote the 25th, 50th and 75th percentiles, the x's the mean and the whiskers denote the value equal to the 75th
percentile plus 1.5 times the interquartile range (75th minus 25th percentile). Circles show scores higher than that.
Figure 4C-8. Distribution of BI scores (including zeros) at USFS biosites (dry soil
moisture) grouped by W126 index values.
SO
70
SO
so
I 40
m
30
20
10
0
<7
& ty a
csbess
1
>7-9 >9-11 >11-13
>13-15 >15-17 >17-19 >19-25
W126 Index Bin
>25
Key: The boxes denote the 25th, 50th and 75th percentiles, the x's the mean and the whiskers denote the value equal to the 75th
percentile plus (or 25th percentile minus) 1.5 times the interquartile range (75th minus 25th percentile). Circles show still higher scores
Figure 4C-9. Distribution of nonzero BI scores at USFS biosites (dry soil moisture)
grouped by W126 index values.
4C-14
-------
50
45
40
15
30
25
20
15
10
I
0
<7
X
JL
i
-smallsample size-
's.
>7-9 >9-11 >11-13 >13-15
W126 index Bin
>15 -17
>17 -19
Key: The boxes denote the 25th, 50th and 75th percentiles, the x's the mean and the whiskers denote the value equal to the 75th
percentile plus 1.5 times the interquartile range (75th minus 25th percentile). Circles show scores higher than that.
Figure 4C-10. Distribution of BI scores (including zeros) at USFS biosites (wet soil
moisture) grouped by W126 index values.
50
45
40
35
m
30
3W
Q
2
25
m
20
15
10
5
0
< 7
-small sample size--
x
>7-9 >9-11 >11-13 >13-15
W126 Index Bin
>15-17
>17 -19
Key: The boxes denote the 25th, 50th and 75th percentiles, the x's the mean and the whiskers denote the value equal to the 75th
percentile plus (and the 25th minus) 1.5 times the interquartile range (75th minus 25th percentile). Circles show scores above that.
Figure 4C-11. Distribution of nonzero BI scores at USFS biosites (wet soil moisture)
grouped by W126 index values.
4C-15
-------
Table 4C-5. Average BI scores of the records in each W126 index bin.
Soil
Moisture
<7
ppm-hrs
>7 -9
ppm-hrs
>9-11
ppm-hrs
>11 -13
ppm-hrs
>13-15
ppm-hrs
>15-17
ppm-hrs
>17-19
ppm-hrs
>19-25
ppm-hrs
>25
ppm-hrs
Average BI (all records)
Dry
0.9
0.3
0.4
0.4
0.4
0.2
2.3
2.1
3.5
Normal
0.9
0.9
0.8
1.2
1.2
0.9
2.3
1.6
7.9
Wetc
0.9
0.9
1.9
1.9
[2.2]
[6.7]
[13.9]
-
-
All
0.9
0.8
0.8
1.1
1.0
0.8
2.4
1.9
5.5
Average BI (records with BI > 0)
Dry
14.2
3.0
5.1
4.2
5.1
2.6
15.0
10.40
7.60
Normal
6.8
4.7
3.7
6.3
5.2
6.9
6.0
5.19
15.42
Wetc
3.8
4.3
4.6
4.9
[4.0]
[11.8]
[13.9]
-
-
All
5.4
4.4
4.0
6.0
5.1
6.3
9.0
7.8
11.3
Average BI (records with BI >5)
Dry
23.6
6.9
8.1
6.6
7.1
6.3
41.1
18.4
14.2
Normal
17.0
14.3
10.5
19.0
19.7
15.1
13.8
18.3
18.5
Wetc
11.4
14.2
9.7
10.4
[12.5]
[20.2]
[13.9]
-
-
All
14.9
13.7
10.0
15.7
16.1
15.2
21.0
18.4
16.9
Average BI (records with BI >15)
Dry
39
-
-
-
-
-
60.3
24.1
22.9
Normal
32.0
31.2
22.2
36.0
38.0
27.5
30.5
24.4
27.9
Wetc
34.4
25.8
15.2
16.7
-
[20.2]
[17.4]
-
-
All
33.2
29.6
20.4
33.6
38.0
24.6
41.4
24.2
26.3
A Brackets indicate bins in which total sample size for that bin is below 1 % of all for that soil moisture category (i.e., 0 to 9 samples).
4C-16
-------
Table 4C-6. Proportion of records in each W126 index bin with specified BI score.
Soil
Moisture
<7
ppm-hrs
>7 -9
ppm-hrs
>9-11
ppm-hrs
>11 -13
ppm-hrs
>13-15
ppm-hrs
>15-17
ppm-hrs
>17-19
ppm-hrs
>19-25
ppm-hrs
>25
ppm-hrs
Proportion of Records with BI >15 (USFS categories of "moderate"and "severe")
Dry
0.02
0.00
0.00
0.00
0.00
0.00
0.03
0.07
0.08
Normal
0.02
0.01
0.01
0.02
0.02
0.02
0.03
0.04
0.20
WetA
0.01
0.02
0.01
0.02
[0.00]
[0.29 (2)1
[0.50 (1)1
[0.00]
[0.00]
All
0.01
0.01
0.01
0.02
0.01
0.02
0.04
0.06
0.13
Proportion of Records with BI >5 (USFS categories of "low," "moderate"and "severe")
Dry
0.04
0.03
0.04
0.04
0.05
0.01
0.05
0.11
0.20
Normal
0.05
0.05
0.05
0.05
0.05
0.05
0.13
0.06
0.41
WetA
0.06
0.05
0.16
0.14
[0.11 (1)]
[0.29 (2)]
[1.00 (2)]
[0.00]
[0.00]
All
0.05
0.05
0.06
0.06
0.05
0.04
0.10
0.09
0.30
Proportion of Records with BI >5 & <15 (USFS category of "low")
Dry
0.02
0.03
0.04
0.04
0.05
0.01
0.02
0.04
0.12
Normal
0.03
0.04
0.05
0.03
0.03
0.03
0.10
0.02
0.22
WetA
0.05
0.03
0.15
0.12
[0.11 (1)]
[0.00]
[0.50 (1)]
[0.00]
[0.00]
All
0.04
0.04
0.06
0.04
0.04
0.02
0.06
0.03
0.16
Proportion of Records with BI >0 & <5 (USFS category of We")
Dry
0.03
0.09
0.03
0.04
0.02
0.08
0.10
0.10
0.26
Normal
0.09
0.14
0.16
0.14
0.18
0.08
0.25
0.24
0.10
WetA
0.17
0.15
0.26
0.26
[0.44 (4)]
[0.29 (2)]
0.00
[0.00]
0.00
All
0.12
0.14
0.15
0.13
0.15
0.09
0.17
0.16
0.19
Proportion of Records with BI >0
(USFS categories of "Uttie," "low," "moderate"and "severe")
Dry
0.
0.11
0.08
0.08
0.07
0.09
0.15
0.21
0.46
Normal
0.13
0.19
0.21
0.19
0.23
0.14
0.38
0.31
0.51
WetA
0.23
0.20
0.42
0.40
[0.56 (5)]
[0.57 (4)]
[1.00 (2)]
[0.00]
[0.00]
All
0.17
0.18
0.21
0.19
0.20
0.13
0.27
0.25
0.48
Proportion of Records with BI =0 (USFS category of no injury)
Dry
0.94
0.89
0.92
0.92
0.93
0.91
0.85
0.79
0.54
Normal
0.87
0.81
0.79
0.81
0.77
0.86
0.62
0.69
0.49
WetA
0.77
0.80
0.58
0.60
[0.44 (4)]
[0.43 (3)]
[0.00]
[0.00]
[0.00]
All
0.83
0.82
0.79
0.81
0.80
0.87
0.73
0.75
0.52
A Brackets indicate bins in which total sample size for that bin is below 1% of all for that soil moisture category (i.e., 0 to 9 samples).
Additionally, for these entries the value in parenthesis is the number of records in specified BI bin.
4C-17
-------
The observations of visible foliar injury for the highest W126 bin compared to the others
is generally consistent with the evidence regarding visible foliar injury as an indicator of O3
exposure (e.g., ISA, Appendix 8, section 8.2; 2013 ISA, section 9.4.2; U.S. EPA, 2006 [2006
AQCD], p. AX9-22). The evidence indicates a generally greater extent and severity of visible
foliar injury with higher O3 exposure levels and an influence for soil moisture conditions (ISA,
Appendix 8, Section 8.2). Further, consistent with this evidence, the censored regression of the
USFS dataset described in section 4C.1.1 above found a significant relationship between visible
foliar injury and both O3 and moisture, as measured by Palmer Z.
A study cited in the current and 2013 IS As, which analyzed trends in the incidence and
severity of foliar injury, observed a declining trend in the incidence of foliar injury as peak O3
concentrations declined (2013 ISA, p. 9-40; Smith, 2012). Another study, also available in the
last review, that focused on 03-induced visible foliar injury in west coast forests observed that
both percentage of biosites with injury and average BI were higher for sites with average
cumulative O3 concentrations above 25 ppm-hrs in terms of SUM068 as compared to groups of
biosites with lower average cumulative exposure concentrations, with much less clear differences
between the two lower exposure groups (Campbell et al., 2007, Figures 27 and 28 and p. 30). A
similar finding was reported in the 2007 Staff Paper which reported on an analysis that showed a
smaller percentage of biosites with injury among the group of biosites with O3 exposures aot or
below a SUM06 metric of 15 ppm-hrs or a 4th high metric of 74 ppb as compared to larger
groups that also included biosites with SUM06 values up to 25 ppm-hrs or 4th high metric up to
84 ppb, respectively (U.S. EPA, 2007 [2007 Staff Paper], pp. 7-63 to 7-64).
The observations described here have a general consistency with the extensive evidence
base on foliar injury, which indicates that visible foliar injury prevalence and severity are
generally higher at higher (compared to lower) O3 concentrations. As the FIA/FHM biosites vary
in the type of vegetation and species that are present and the vegetation types and species vary in
sensitivity, BI scores would be expected to differ even between two biosites identical in all
environmental characteristics when there are different species present. Therefore, limitations in
the biosite dataset can affect patterns and relationships observed in the BI scores. Additionally,
various environmental and genetic factors influence the exposure-response relationship, with the
most well understood being soil moisture conditions (ISA, Appendix 8, Section 8.2). Our
understanding of specific aspects of these influences on the relationship between O3 exposures,
the most appropriate exposure metrics, and the occurrence or severity of visible foliar injury is,
however, still incomplete.
8 Based on an approach used in the 2007 Staff Paper (and the associated temporal patterns of O3 concentrations in
data available at that time), a SUM06 index value of 2 5 ppm-hrs would be estimated to correspond to a W126
index of approximately 21 ppm-hrs (2007 Staff Paper, Appendix 7B, p. 7B-2).
4C-18
-------
4C.5 LIMITATIONS AND UNCERTAINTIES
The purpose of the analyses and presentations summarized above was to investigate the
potential relationship between BI scores at USFS biosites and O3 in terms of the seasonal W126
index. The lack of a clear relationship (across W126 bins below 25 ppm-hrs) in the presentations
above may relate to inherent limitations and uncertainties in the different aspects of the dataset.
The limitations and uncertainties associated with aspects of the dataset developed for the 2014
WREA, and further investigated above, are presented here. In summarizing these below, they are
grouped into four areas: 1) biosite scores, 2) soil moisture categorization, 3) W126 index
estimates, and 4) combining of datasets.
Biosite data: Site selection, availability, and species presence also contribute to
uncertainty within the dataset and analysis. Data are lacking from many western states including
Montana, Idaho, Wyoming, Nevada, Utah, Colorado, Arizona, New Mexico, Oklahoma, and
portions of Texas. Furthermore, in certain states (California, Washington, and Oregon) exact
locations of sampled sites were not available, and these sites were assigned to the grid based on
publicly available geographic coordinates, increasing the level of uncertainty. Because the grid
sizes are relatively small, limiting the geographic skew of estimated location (7 km in any
direction), it is likely that these locations were at least assigned to adjacent grid cells. While the
extent of such differences and magnitude of any effect on the resultant dataset are unknown, it
may have relatively small difference and low magnitude of influence on the dataset (2014
WREA, p. 7-60).
Soil moisture categories: The use of the Palmer Z soil moisture index contributes
uncertainty of unknown directionality and magnitude. Short-term estimates of soil moisture can
be highly variable from month to month within a single year. Using averages contributes to a
potential temporal mismatch between soil moisture and injury. Soil moisture is also substantially
spatially variable, and the soil moisture data can be hundreds of miles wide in climate regions.
There is much diversity within regions, and some vegetation, such as that along riverbanks, may
experience sufficient soil moisture during periods of drought to exhibit foliar injury. All of these
factors contribute uncertainty to this categorization (2014 WREA, p. 7-61).
W126 index estimates: Ambient air quality measurements have some inherent
uncertainties (considered low [2014 WREA, p. 4-39]) associated with them. These uncertainties
relate to monitoring network design, O3 monitoring seasons, monitor malfunctions, wildlife and
wildfire/smoke impacts, and interpolations of missing data. There is likely somewhat greater
uncertainty associated with the assignment of W126 index estimates to all biosites due to the
need for interpolating between monitor sites to estimate concentrations in unmonitored areas
4C-19
-------
(2014 WREA, sections 4A.2.1).9 Accordingly, there is relatively greater uncertainty associated
with sites at some distance from monitoring sites and lesser uncertainty in densely monitored
areas (2014 WREA, p. 4-40). Unfortunately, which sites are which is unknown.
Combining datasets: Uncertainty is associated with the combination of data types of
different spatial resolution. For example, the biosite scores are available at a much finer spatial
resolution than the W126 index estimates, which represent a much small spatial area than that
represented by the soil moisture categorization. Yet, as recognized above, soil moisture may vary
on much finer scales. To avoid losing resolution of the finest-scale dataset (the biosite scores),
the finest spatial resolution available was used (e.g., rather than averaging the BI scores across
the grids for which W126 index was estimated or across the climate regions for which the soil
moisture scores area available), although this approach contributes its own uncertainty.
There is also uncertainty in the combination step associated with the differing temporal
scales or time-of-year represented by the three types of data.
Overall, we recognize a number of limitations and uncertainties that may be affecting our
ability to identify a relationship between O3, as quantified by seasonal W126 index, and visible
foliar injury at USFS locations (based on BI scores), particularly at sites with W126 index
estimates at or below 25 ppm-hrs.
4C.6 SUMMARY AND KEY OBSERVATIONS
The following are key observations concerning the dataset presented in this appendix,
which includes the subset of USFS biosite data for the years 2006 through 2010, and for which
limitations and uncertainties are recognized in section 4C.5 above.
Full Dataset:
• The combined dataset includes more than 5,000 records, each of which documents a biotic
index scores, soil moisture index value and W126 index estimate, for USFS biosites in 37
states in one or more years from 2006 to 2010.
• The majority of the records are for W126 index estimates at or below 9 ppm-hrs, with
fewer than 10% of records assigned W126 index estimates above 15 ppm-hrs.
• The BI scores (in all soil moisture categories) are quite variable, with at least half the
scores in nearly all bins being zero, and even the bin for the lowest W126 index estimates
(below 7 ppm-hrs) having at least one scores above 5 and 15.
• With regard to soil moisture conditions most of the dataset (61% of all records) are for
soil moisture conditions categorized as normal. The remainder include somewhat more
9 Evaluations of the VNA interpolation technique describe correlations with monitoring data and indicate more
accurate prediction of monitoring data by the VNA method than use of an air quality model (2014 WREA,
section 4.A.3.1).
4C-20
-------
records for wet soil moisture conditions than dry, with 23% of all records categorized as
wet soil moisture conditions and 16% as dry soil moisture.
Records in Wet Soil Moisture Category:
• The wet soil moisture records are concentrated in the two lower W126 index bins which
contain nearly 90% of all records for this soil moisture category.
- Accordingly, interpretations of patterns across W126 bins for this soil moisture
category are limited by small sample size across the bins. For example, the number of
records in each of the W126 bins above 13 ppm-hrs (ranging from zero to 9)
comprise less than 1% of the records in this soil moisture category.
Records in Normal Soil Moisture Category:
• Among records in the normal soil moisture category, BI scores are noticeably increased in
the highest W126 index bin (index estimates above 25 ppm-hrs), averaging 7.9.
- The percentages of records in this W126 bin with scores above 15 or above 5 are
more than three times greater than percentages for these score magnitudes in any of
the lower W126 index bins.
- The average BI score for records in the highest W126 bin is also appreciably greater
than scores for records in the other bins. The average scores in the next two highest
W126 bins are 1.6 and 2.3, respectively, which are only slightly higher than average
scores for the rest of the bins, which vary from 0.8 to 1.2 without a clear relationship
to estimated W126 index.
- Among the records in this category with nonzero scores, the highest average BI is
also in the highest W126 index bin (>25 ppm-hrs); in this case the BI is
approximately 15, more than double the next highest average scores across the other
W126 index bins (for which no other trend is exhibited). The proportion of records
with any injury is also highest in the highest W126 index bin; it is also slightly
increased in the next lower W126 bins compared to the rest (the bins at or below 17
ppm-hrs) across which there is little evident pattern.
Records in Dry Soil Moisture Category:
• Dry soil moisture category records in the two highest W126 bins (>19 and > 25 ppm-hrs)
exhibit the greatest percentages of records with BI above 15 and above 5. For scores
above 15, the percentages are 7 and 8% compared to 0 to 3 % in the other bins, and for
scores above 5, the scores are 11 and 20% compared to 1 to 5% in the other bins.
In summary, the observations described here are generally consistent with the extensive
evidence base on foliar injury and O3, which indicates that foliar injury prevalence and severity
are generally higher at higher (compared to lower) O3 concentrations. The presentations here of
USFS data do not indicate clear trends in BI across the full range of W126 index estimates.
Rather, they indicate increased BI for the highest estimates, with the increase in both incidence
of higher scores and in average score being most clear for W126 index estimates above 25 ppm-
hrs, with a suggestion of slight increase for some records with W126 index estimates above 17 or
4C-21
-------
19 ppm-hrs (dry soil moisture category). Variability as well as sample size limitations contribute
to the lack of more precise conclusions. Additionally, as indicated in the evidence summarized in
the ISA and prior scientific assessments, various environmental and genetic factors influence the
exposure-response relationship. Our understanding of specific aspects of these influences on the
relationship between O3 exposures, the most appropriate exposure metrics, and the occurrence or
severity of visible foliar injury is, however, still incomplete.
REFERENCES
Campbell, SJ, Wanek, R and Coulston, JW (2007). Ozone injury in west coast forests: 6 years of
monitoring - Introduction. U.S. Department of Agriculture. Portland, OR.
Coulston, JW, Smith, GC and Smith, WD (2003). Regional assessment of ozone sensitive tree
species using bioindicator plants. Environmental Monitoring and Assessment 83(2): 113-
127.
Horsfall, J and Cowling, E (1978). Plant disease, an advanced treaties. Academic Press. New
York, NY.
Palmer, WC (1965). Meteorological drought. U.S. Department of Commerce. Washington, DC.
https://www.ncdc.noaa.gov/temp-and-precip/drought/docs/palmer.pdf.
Smith, G (2012). Ambient ozone injury to forest plants in Northeast and North Central USA: 16
years of biomonitoring. Environmental Monitoring and Assessment 184(7): 4049-4065.
Smith, GC, Morin, RS and McCaskill, GL (2012). Ozone injury to forests across the Northeast
and North Central United States, 1994-2010. General Technical Report NRS-103. United
States Department of Agriculture, US Forest Service, Northern Research Station.
Smith J. T.; Murphy, D. (2015). Memorandum to Ozone NAAQS Review Docket (EPA-HQ-
OAR-2008-0699). Additional Observations from WREA Datasets for Visible Foliar
Injury. September 24, 2015.. Docket ID No. EPA-HQ-OAR-2008-0699. Office of Air
Quality Planning and Standards Research Triangle Park, NC. Available at:
https://www.regulations.gov/contentStreamer?documentld=EPA-HQ-OAR-2008-0699-
4250&contentType=pdf.
U.S. EPA (2006). Air Quality Criteria for Ozone and Related Photochemical Oxidants (Volume I
- III). Office of Research and Development U.S. EPA. EPA-600/R-05-004aF, EPA-
600/R-05-004bF, EPA-600/R-05-004cF February 2006. Available at:
https.Y/cfpub. epa.gov/ncea/risk/recordisplay. cfm ?deid=l49923.
U.S. EPA (2007). Review of the National Ambient Air Quality Standards for Ozone: Policy
Assessment of Scientific and Technical Information: OAQPS Staff Paper. Office of Air
Quality Planning and Standards. Research Triangle Park, NC. U.S. EPA. EPA-452/R-07-
4C-22
-------
003. January 2007. Available at:
https.Y/nepis. epa.gov/Exe/ZyPURL. cgi?Dockey=Pl0083VX. txt.
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=P100KETF.txt.
U.S. EPA (2014). Welfare Risk and Exposure Assessment for Ozone (Final).. Office of Air
Quality Planning and Standards. Research Triangle Park, NC. U.S. EPA. EPA-452/P-14-
005a August 2014. Available at:
https.Y/nepis. epa.gov/Exe/ZyPURL. cgi?Dockey=Pl 00KB9D. txt.
4C-23
-------
APPENDIX 4D
ANALYSIS OF THE W126 03 EXPOSURE INDEX AT U.S. AMBIENT AIR
MONITORING SITES
Table of Contents
4D.1 Overview 4D-2
4D.2 Data Handling 4D-2
4D.2.1 Data Retrieval and Preparation 4D-2
4D.2.2 Derivation of the 4th Max and W126 Metrics 4D-2
4D.2.3 Derivation of Temporal Trends 4D-4
4D.2.4 Identification of O3 Monitoring Sites in Federal Class I Areas 4D-5
4D.2.5 Assignment of Monitoring Sites to NOAA Climate Regions 4D-5
4D.3 Results 4D-6
4D.3.1 National Analysis Using Recent Air Quality Data 4D-6
4D.3.1.1 Comparison of the 4th Max and W126 Metrics 4D-8
4D.3.1.2 Relationships Between Metrics and the Annual W126 Index 4D-10
4D.3.2 National Analysis Using Historical Air Quality Data 4D-15
4D.3.2.1 Comparison of the 4th Max and W126 Metrics 4D-16
4D.3.2.2 Trends in W126 Metric 4D-18
4D.3.2.3 Comparison of Trends in the 4th Max and W126 Metrics 4D-20
4D.3.2.4 W126 Metric Values in Federal Class I Areas 4D-25
4D.4 Key Limitations and Uncertainties 4D-33
4D.5 Summary 4D-33
4D.6 References 4D-36
4D-1
-------
4D.1 OVERVIEW
This appendix presents various analyses of ambient air monitoring data for ozone (O3)
concentrations in the U.S. relating to the W126-based cumulative exposure index. These
analyses focus on the annual maximum 3-month sum of daytime hourly weighted O3
concentrations, averaged over 3 consecutive years, hereafter referred to as the "W126 metric,"
calculated as described in section 2 below. These analyses examine spatial and temporal patterns
in the W126 metric using monitoring data from 2000 to 2018 and make various comparisons
between the W126 metric and design values for the current O3 standard (the annual 4th highest
daily maximum 8-hour O3 concentration, averaged over 3 consecutive years; hereafter referred to
as the "4th max metric"). Additional analyses assess the relative variability between the W126
metric and its constituent annual index values and the magnitude of W126 index values at
monitoring sites in or near federally protected ecosystems known as Class I areas. These
analyses are largely parallel to analyses that were completed for the last review of the O3
NAAQS (79 FR 75331, December 17, 2014; 80 FR 65385, October 26, 2015; U.S. EPA, 2014a,
Wells, 2014, Wells, 2015).
4D.2 DATA HANDLING
4D.2.1 Data Retrieval and Preparation
Hourly O3 concentration data were retrieved from the EPA's Air Quality System (AQS,
https://www.epa.gov/aqs) database for 1,981 ambient air monitoring sites which operated
between 2000 and 2018. These data were used to calculate W126 and 4th max metric values for
each 3-year period from 2000-2002 to 2016-2018. Before calculating these metrics, some initial
processing was done on the hourly data. First, data collected using monitoring methods other
than federal reference or equivalent methods, and data collected at monitoring sites not meeting
EPA's quality assurance or other criteria in 40 CFR part 58 were removed from the analysis.
Second, data collected by multiple monitoring instruments operating at the same location were
combined according to Appendix U to 40 CFR Part 50. Finally, data were combined across 95
pairs of monitoring sites approved for such combination by the EPA Regional Offices. The final
hourly O3 concentration dataset contained 1,788 monitoring sites.
4D.2.2 Derivation of the 4th Max and W126 Metrics
The 4th max metric values were calculated according to the data handling procedures in
Appendix U to 40 CFR part 50. First, moving 8-hour averages were calculated from the hourly
O3 concentration data for each site. For each 8-hour period, an 8-hour average value was
calculated if there were at least 6 hourly O3 concentrations available. Each 8-hour average was
stored in the first hour of the period (e.g., the 8-hour average from 12:00 PM to 8:00 PM is
4D-2
-------
stored in the 12:00 PM hour). Daily maximum 8-hour average values were found using the 8-
hour periods beginning from 7:00 AM to 11:00 PM each day. These daily maximum values were
used if at least 13 of the 17 possible 8-hour averages were available, or if the daily maximum
value was greater than 70 parts per billion (ppb). Finally, the annual 4th highest daily maximum
value was found for each year, then averaged across each consecutive 3-year period to obtain the
final set of 4th max metric values in units of ppb. Any decimal digits in these values were
truncated for applications requiring direct comparison to a 4th max level (e.g., Table 4D-2),
otherwise, all decimal digits were retained. The 4th max metric values were considered valid if
daily maximum values were available for at least 90% of the days in the O3 monitoring season
(defined in Appendix D to 40 CFR part 58) on average across the three years, with a minimum of
75% of the days in the O3 monitoring season in any calendar year. In addition, 4th max metric
values were considered valid if they were greater than the 4th max levels to which they were
being compared.
The W126 metric values were calculated using the hourly O3 concentration data in parts
per million (80 FR 65374, October 26, 2015). For daytime hours (defined as the 12-hour period
from 8:00 AM to 8:00 PM Local Standard Time each day), the hourly concentration values at
each O3 monitoring site were weighted using the following equation:
Weighted 03 = 03 / (1 + 4403*exp(-126 * 03)).
These weighted values were summed over each calendar month, then adjusted for
missing data (e.g.; if 80% of the daytime hourly concentrations were available, the sum would be
multiplied by 1/0.8 = 1.25) to obtain the monthly W126 index values. Monthly W126 index
values were not calculated for months where fewer than 75% of the possible daytime hourly
concentrations were available. Next, moving 3-month sums were calculated from the monthly
index values, and the highest of these 3-month sums was determined to be the annual W126
index. Three-month periods spanning multiple years (e.g., November to January, December to
February) were not considered in these calculations. The annual W126 index values were
averaged across each consecutive 3-year period to obtain the final W126 metric values, with
units in parts per million-hours (ppm-hrs). The W126 metric values were rounded to the nearest
unit ppm-hr for applications requiring direct comparison to a W126 level (e.g., Table 4D-3),
otherwise, all decimal digits were retained. For consistency with the 4th max metric calculations,
the W126 metric values were considered valid if hourly O3 concentration values were available
for at least 90% of the daytime hours during the O3 monitoring season on average across the
three years, with a minimum of 75% of the daytime hours during the O3 monitoring season in
any calendar year. Also for consistency with the 4th max metric calculations, the W126 metric
4D-3
-------
values were considered valid if they were greater than the W126 levels to which they were being
compared.
In the final dataset, 1,557 of the 1,788 O3 monitoring sites had sufficient data to calculate
valid 4th max and W126 metric values for at least one 3-year period between 2000-2002 and
2016-2018. The number of sites with valid 4th max and W126 metric values ranged from a low
of 992 in 2000-2002 to a high of 1,119 in 2015-2017, and 543 sites had valid 4th max and W126
metric values for all seventeen 3-year periods.
4D.2.3 Derivation of Temporal Trends
Site-level trends for the W126 metric and annual W126 index values were computed in a
similar manner to the site-level trends for the 4th max metric presented in Chapter 2. Specifically,
for the annual W126 index, a site must have at least 75% annual data completeness for at least 15
of the 19 years, with no more than two consecutive years having less than 75% data
completeness in order to be included in the analysis. For the W126 metric, a site must have a
valid W126 metric value (according to the data completeness criteria presented in the previous
section) in at least 13 of the 17 3-year periods, and no more than two consecutive 3-year periods
that do not havevalid W126 metric values. There were 852 sites meeting these criteria for the
annual W126 index and 713 sites meeting these criteria for the W126 metric. The national
median, 10th percentile, and 90th percentile values of these site-level trends are presented in
Figure 4D-9.
Other analyses presented in Section 4D.3.2.2 use trends in the 4th max and W126 metrics
as well as the annual W126 index calculated with non-parametric regression methods. These
trends were computed using the Theil-Sen estimator (Sen, 1968; Theil, 1950), a type of
regression method that chooses the median slope among all lines crossing through each possible
pair of sample points1. These trends are reported in units of ppb/yr for the 4th max metric or ppm-
hr/yr for the W126 metric and annual W126 index. The data completeness criteria described in
the previous paragraph were also applied to site for which these trends were calculated.2
Statistical tests for significance of the Theil-Sen estimator were computed using the non-
parametric Mann-Kendall test (Kendall, 1948; Mann; 1945).
1 For example, if applying this method to a dataset with W126 metric values for four consecutive years (e.g., W126i,
WI262, W1263, and W1264), the trend would be the median of the per-year changes observed in the six possible
pairs of values (e.g., the median of [W1264-W1263]/l, [W1263-W1262]/l, [W1262-W126i]/1, [W1264-W1262]/2,
[W1263- W126J/2, and [W1264- W126J/3).
2 For the 4th max metric, the data completeness criteria used were valid 4th max metric values (as defined in section
4D.2.2) in 13 of the 17 3-year periods, and no more than two consecutive periods that do not have valid 4th max
metric values. There were 629 sites meeting these criteria, and all of these sites also met the data completeness
criteria for the W126 metric and the annual W126 index.
4D-4
-------
4D.2.4 Identification of O3 Monitoring Sites in Federal Class I Areas
The Clean Air Act (section 162) designated certain federally areas as Class I areas. These
areas are federally mandated to preserve certain air quality values. Class I designation allows the
least amount of deterioration of existing air quality. Areas designated as Class I include all
international parks, national wilderness areas which exceed 5,000 acres in size, national
memorial parks which exceed 5,000 acres in size, and national parks which exceed 6,000 acres in
size, provided the park or wilderness area was in existence on August 7, 1977. There are 158
such areas (e.g., 44 FR 69122, November 30, 1979). Other areas may, and have been,
subsequently designated as Class I consistent with the CAA (section 162). As of July 2019, six
Class II areas on Tribal lands have been re-designated as Class I.3
To identify which O3 monitoring sites represented air quality in federal Class I areas,
shapefiles (i.e., files that specify area boundaries) for all 158 mandated federal Class I areas4
were downloaded from EPA's Environmental Dataset Gateway (EDG; https://edg.epa.govJ) and
augmented with the six tribal areas redesignated as Class I. These boundaries were matched to
the 1,788 O3 monitoring sites in the hourly O3 concentration dataset described in section 4D.2.1.
Since Class I areas include federally designated wilderness areas in which permanent structures
such as air monitoring trailers are prohibited, if there was no monitor located within the area
boundary, the matching was expanded to include the nearest monitoring site within 15 km of the
boundary. For each Class I area and 3-year period, if a 4th max or W126 metric value was not
available for the nearest monitor, the values from the next nearest monitor within 15 km were
used, where applicable. In addition, if a Class I area had multiple monitors inside the boundary,
we used the monitor with the highest 4th max metric value in each 3-year period. These monitors
were extracted from the 4th max and W126 dataset described in section 4D.2.2, yielding a final
Class I areas dataset with a total of 860 records that had valid 4th max and W126 metric values at
79 O3 monitoring sites representing 65 Class I areas (out of 164 total Class I areas).
4D.2.5 Assignment of Monitoring Sites to NOAA Climate Regions
In order to examine regional differences, many of the further analyses were stratified into
the nine NOAA climate regions (Karl and Koss, 1984), which are shown in Figure 4D-1. Since
the NOAA climate regions only cover the contiguous U.S., Alaska was added to the Northwest
region, Hawaii was added to the West region, and Puerto Rico was added to the Southeast
region.
3 The Class I areas on Tribal lands as of December 2018 are listed at:
https://www.nps.gov/subjects/air/tribalclassl.htm. Since then, one additional area has been designated Class I on
Tribal lands (84 FR 34306, July 18, 2019).
4 The set of Class I areas identified in 1977 are referred to here as "mandated."
4D-5
-------
~ NorthWest
¦ West
¦ WestNorthCentral ~
~ SouthWest ~
EastNorthCentral
South
Central
SouthEast
E3 NorthEast
Figure 4D-1. Map of the nine NOAA climate regions.
4D.3 RESULTS
ID.3.1 National Analysis Using Recent Air Quality Data
This section presents various results based on the 4th max and W126 metrics for the 2016-
2018 period. Figure 4D-2 shows a map of the observed W126 metric values based on 2016-2018
data. From this figure, it is apparent that W126 metric values are generally at or below 13 ppm-
hrs in the eastern and northwestern U.S. In the U.S. as a whole, about 60% of all monitoring sites
recorded VV126 metric values at or below 7 ppm-hrs, and over 90% of all monitoring sites
recorded W126 metric values at or below 17 ppm-hrs. The highest W126 metric values occur in
the southwestern U.S. where there are numerous monitoring sites with W126 metric values
above 17 ppm-hrs, however, only one of these sites (near Las Vegas, NV), with a W126 metric
value of 18 ppm-hrs, meets the current standard. Table 4D-1 shows the number of sites in each
NOAA climate region that have a valid 2016-2018 design value meeting the current standard and
the number of sites in each region that have a 2016-2018 design value not meeting the current
standard.
4D-6
-------
• 1-7 ppm-hrs (663 sites) G 14-15 ppm-hrs (25 sites) • 18-62 ppm-hrs (92 sites)
• 8-13 ppm-hrs (312 sites) Q 16-17 ppm-hrs (30 sites) A 4th Max Value > 70 ppb
Figure 4D-2. Map of W126 metric values at U.S. O3 monitoring sites based on 2016-2018
data. Circles indicate monitoring sites with 4th max metric values less than or
equal to 70 ppb. while triangles indicate monitoring sites with 4'Jl max metric
values greater than 70 ppb.
Table 41)1. Number of O3 monitoring sites with valid 2016-2018 design values in each
NO A A climate region
NOAA Climate Region
Total # of
Sites
# of Sites with Design
Value < 70 ppb
# of Sites with Design
Value > 70 ppb
Central
202
169
33 |
EastNorthCentral
86
66
20 I
North East
179
130
49
Northwest
26
22
4
South
132
105
27
South East
166
164
2 I
Southwest
109
65
44 |
West
176
82
94
WestNorthCentral
46
46
0
National
1122
849
273 I
4D-7
-------
4D.3.1.1 Comparison of the 4th Max and W126 Metrics
The following analyses make several comparisons between the 4th max and W126 metric
values based on 2016-2018 data. Table 4D-2 shows the number of sites with 4th max metric
values greater than each 4th max level, and the number of sites with 4th max metric values less
than or equal to each 4th max level. Table 4D-3 shows the number of sites with W126 metric
values greater than each W126 level, and the number of sites with W126 metric values less than
or equal to each W126 level.
The 4th max and W126 metric values were also compared to each combination of 4th max
and W126 levels based on 2016-2018 data. Table 4D-4 shows the number of sites with 4th max
metric values greater than each 4th max level, and W126 metric values less than or equal to each
W126 level (e.g., 147 sites had 4th max metric values greater than 70 ppb and W126 metric
values less than or equal to 13 ppm-hrs). Table 4D-5 shows the number of sites with 4th max
metric values less than or equal to each 4th max level, and W126 metric values greater than each
W126 level (e.g., 27 sites with a 4th max metric value at or below 70 ppb had a W126 metric
value greater than 13 ppm-hrs). Finally, Table 4D-6 shows the number of sites with 4th max
metric values greater than each 4th max level, and W126 metric values greater than each W126
level.
Table 4D-2. Number of sites with 4th max metric values greater than various 4th max levels
based on 2016-2018 data.
4th Max Level (ppb)
75
70
65
# of Sites > Level
10!)
2/3
632
# of Sites < Level
I.OOfi
;M!)
:>03
Total # of SitesA
1.11!.
1.122
1. IS!)
A For each 4"1 max level, a site with a 4"1 max metric value less than or equal to the level is
counted only if it meets the data completeness criteria described in section 4D.2.2, whereas a site
with a 4lh max metric value greater than the level is counted regardless of data completeness.
Therefore, the total number of sites may differ among the columns.
Table 4D-3. Number of sites with W126 metric values greater than various W126 levels
based on 2016-2018 data.
W126 Level (ppm-hrs)
19
17
15
13
11
9
7
# of Sites > Level
78
92
122
147
204
283
471
# of Sites < Level
1,040
1,027
998
975
923
844
663
Total # of SitesA
1.118
1.119
1.120
1.122
1.127
1.127
1.134
A For each W126 level, a site with a W126 metric value less than or equa
completeness criteria described in section 4D.2.2, whereas a site with a \
regardless of data completeness. Therefore, the total number of sites ma
to the level is counted only if it meets the data
¥126 metric value greater than the level is counted
y differ among the columns.
4D-8
-------
Table 4D-4. Number of sites with 4th max metric values greater than various 4th max levels
and W126 metric values less than or equal to various W126 levels based on
2016-2018 data.
# Sites > 4th Max Level
AND < W126 Level
W126 Level (ppm-hrs)
19
17
15
13
11
9
7
4th Max
Level (ppb)
75
33
30
26
23
16
7
1
70
185
173
155
147
127
77
26
65
533
520
491
468
420
350
186
Table 4D-5. Number of sites with 4th max metric values less than or equal to various 4th
max levels and W126 metric values greater than various W126 levels based on
2016-2018 data.
# Sites < 4th Max Level
AND > W126 Level
W126 Level (ppm-hrs)
19
17
15
13
11
9
7
4th Max
Level (ppb)
75
6
15
40
60
105
175
351
70
0
1
12
27
59
88
219
65
0
0
0
0
4
13
30
Table 4D-6. Number of sites with 4th max metric values greater than various 4th max levels
and W126 metric values greater than various W126 levels based on 2016-2018
data.
# Sites > 4th Max Level
AND > W126 Level
W126 Level (ppm-hrs)
19
17
15
13
11
9
7
4th Max
Level (ppb)
75
72
76
81
84
92
101
107
70
78
91
110
119
142
192
245
65
78
92
122
147
200
270
438
According to Table 4D-2, 10% of U.S. O3 monitoring sites had 2016-2018 4th max metric
values greater than 75 ppb, 24% of sites had 4th max metric values greater than 70 ppb, and 56%
of sites had 4th max metric values greater than 65 ppb. According to Table 4D-3, 8% of U.S. O3
monitoring sites had 2016-2018 W126 metric values greater than 17 ppm-hrs, 13% of sites had
W126 metric values greater than 13 ppm-hrs, and 41% of sites had W126 metric values greater
than 7 ppm-hrs. According to Table 4D-5, there were no monitoring sites with a 4th max metric
value less than or equal to 70 ppb and a W126 metric value greater than 19 ppm-hrs, only one
monitoring site with a 4th max less than or equal to 70 ppb and a W126 greater than 17 ppm-hrs.
4D-9
-------
4D.3.1.2 Relationships Between Metrics and the Annual W126 Index
Figure 4D-3 shows a scatter plot comparing the 4th max (x-axis) and W126 (y-axis)
metric values based on 2016-2018 data, with points colored by NOAA climate region. This
figure indicates that there is a strong, positive, non-linear relationship between the 4th max and
W126 metrics. The amount of variability in the relationship between the 4th max and W126
metrics appears to increase as the metric values themselves increase. The relationship between
the 4th max and W126 metrics also appears to vary across regions. In particular, the Southwest
and West regions (i.e., the southwestern U.S.) appear to have higher W126 metric values relative
to their respective 4th max metric values than the rest of the U.S.
Figure 4D-4 shows the same information as Figure 4D-3, but only for monitoring sites
meeting the current standard. This figure shows that all monitoring sites meeting the current
standard have W126 metric values of 18 ppm-hrs or less, and all sites outside the Southwest and
West climate regions have W126 metric values of 12 ppm-hrs or less.
Finally, Figure 4D-5 shows a scatter plot comparing the 4th max metric values (x-axis) to
the annual W126 index values (y-axis) based on 2016-2018 data, with points colored by NOAA
climate region. This figure shows that the annual W126 index values have a similar positive,
non-linear relationship with the 4th max metric values as the W126 metric values. As might be
expected, there is generally more variability in the relationship between the annual W126 index
values and the 4th max metric values than between the W126 metric values and the 4th max
metric values.
Figure 4D-6 shows a scatter plot of the deviations in the 2016, 2017, and 2018 annual
W126 index values (y-axis) from the 2016-2018 average W126 metric values (x-axis). This
figure shows that the magnitude of the annual W126 index deviations from the 3-year average
tend to increase as the W126 metric value increases. About 55% of the annual W126 index
values are within +/- 1 ppm-hr of the 3-year average value, about 81% are within +/- 2 ppm-hrs
of the 3-year average value, and about 98% are within +/- 5 ppm-hrs of the 3-year average value.
Figure 4D-7 also presents the deviations in the 2016, 2017, and 2018 annual W126 index values
from their respective 2016-2018 averages, but with the 2016-2018 average 4th max metric values
plotted on the x-axis instead of the 2016-2018 average W126 metric values. From this figure it
can be seen that lower 4th max metric values generally correspond to smaller inter-annual
variation within W126 metric values, especially for sites meeting the current standard.
4D-10
-------
® Central
o EastNorthCerttral
o NorthEast
O NorthWest
o South
o South East
o SouthWest
• West
• WestNorthCentral
oaM
% •
\ • '
o® O |%
• & •
40
50
60 70 80 90
4th Max Metric Value (ppb)
100
110
Figure 4D-3. Scatter plot of W126 metric values versus 4th max metric values (design
values) based on 2016-2018 monitoring data.
4D-11
-------
® Central
o EastNorthCentral
o North East
© NorthWest
Q South
o SouthEast
o SouthWest
• West
• WestNorthCentral
CM -
O cO(
40
45
r
50 55 60
4th Max Metric Value (ppb)
Figure 4D-4. Scatter plot of W126 metric values versus 4th max metric values (design
values) at monitoring sites meeting the current standard based on 2016-2018
monitoring data.
4D-12
-------
o
CD
O
IO
. o
£ *¦
Q_
Q_
X
0)
T3
C
CD
CM
O
CO
-------
© Central
o EastNorthCentral
o NorthEast
o Northwest
o South
O SouthEast
o Southwest
• West
• WestNorthCentral
A 2016
¦ 2017
• 2018
M • \
Iji • V
f. V A I *
A
10
20 30 40
W126 Metric Value (ppm-hrs)
50
60
Figure 4D-6. Deviation in annual W126 index values from their respective 3-year averages
for all U.S. monitoring sites in 2016-2018.
4D-14
-------
o
Q Central o NorthWest O SouthWest A. 2016
o EastNorthCentral © South ® West ¦ 2017
© NorthEast O SouthEast • WestNorthCentral • 2018
40 50 60 70 80 90 100 110
4th Max Metric Value (ppb)
Figure 4D-7. Deviation in annual W126 index values from their respective 3-year averages
(y-axis) compared to their 4th max metric values (x-axis) in 2016-2018.
4D.3.2 National Analysis Using Historical Air Quality Data
This section presents various results based on the 4th max and W126 metrics for the full
19-year period spanning years 2000 to 2018. Comparisons similar to those shown in section
4D.3.1 are shown in section 41).3.2.1, trends in W126 are shown in section 4D.3.2.2, and several
comparisons of the trends in the 4th max and VV126 metrics are shown in section 4D.3.2.3.
4D-15
-------
4D.3.2.1 Comparison of the 4th Max and W126 Metrics
Table 4D-7 to Table 4D-11 present similar information to Table 4D-2 to Table 4D-6,
respectively, except that the values shown in each cell contain the number of occurrences
summed over all 17 consecutive 3-year periods (2000-2002 to 2016-2018) instead of just the
2016-2018 period. For example, Table 4D-10 shows that over all 17 consecutive 3-year periods,
there were 243 occurrences where sites had 4th max metric values less than or equal to 70 ppb
and W126 metric values greater than 13 ppm-hrs. In general, the relative magnitudes of the
numbers shown in Table 4D-7 to Table 4D-11 compare well to their respective counterparts in
Table 4D-2 to Table 4D-6. According to Table 4D-10, there have been no occurrences over the
entire 19-year period where a site has had a 4th max metric value less than or equal to 70 ppb and
a W126 metric value greater than 19 ppm-hrs.5
Figure 4D-8 shows the distribution of annual W126 index values observed at sites during
3-year periods with different 4th max metric values. These distributions are illustrated by box-
and-whisker plots with boxes showing the 25th, 50th, and 75th percentile of the annual W126
index values occurring with 4th max metric values within each bin, whiskers extending to the 1st
and 99th percentiles of the annual W126 index values, and points occurring outside the 1st and
99th percentiles represented by dots. This figure shows that for the bin with the highest 4th max
metric values meeting the current standard, 66-70 ppb, the 99th percentile of the annual W126
index values was about 19 ppm-hours, or in other words, for sites meeting the current standard,
annual W126 index values were less than or equal to 19 ppm-hrs well over 99% of the time.
5 There was a single occurrence of a site with a 4th max of 70 ppb and a W126 that when rounded, just equaled 19
ppm-hrs.
4D-16
-------
Table 4D-7. Total number of 4th max metric values greater than various 4th max levels
based on all 17 consecutive 3-year periods (2000-2002 to 2016-2018).
4th Max Level (ppb)
75
70
65
Values > Level
1(),(i!)!.
H.9'12
Values < Level
lX.()(i()
8.292
'1.396
Total # of ValuesA
18./31
1(i.!)(i /
19.338
A For each 4"1 max level, a site with a 4"1 max metric value less than or equal to the level is
counted only if it meets the data completeness criteria described in section 4D.2.2, whereas a site
with a 4lh max metric value greater than the level is counted regardless of data completeness.
Therefore, the total number of values may differ among the columns.
Table 4D-8. Total number of W126 metric values greater than various W126 levels based
on all 17 consecutive 3-year periods (2000-2002 to 2016-2018).
W126 Level (ppm-hrs)
19
17
15
13
11
9
7
Values > Level
2,317
3,186
4,397
6,027
7,974
10,347
12,952
Values < Level
16,244
15,414
14,244
12,683
10,840
8,586
6,153
Total # of ValuesA
18.561
18,600
18.641
18.710
18.824
18,933
19,105
A For each W126 level, a site with a W126 metric value less than or egua
completeness criteria described in section 4D.2.2, whereas a site with a \
regardless of data completeness. Therefore, the total number of values rr
to the level is counted only if it meets the data
VI26 metric value greater than the level is counted
lay differ among the columns.
Table 4D-9. Total number of 4th max metric values greater than various 4th max levels and
W126 metric values less than or equal to various W126 levels based on all 17
consecutive 3-year periods (2000-2002 to 2016-2018).
Values > 4th Max Level
AND < W126 Level
W126 Level (ppm-hrs)
19
17
15
13
11
9
7
4th Max
Level (ppb)
75
'1,22/
3,55'l
2,680
1,6/9
832
26/
'10
70
7,920
7,099
5,980
4,609
3,062
1,414
375
65
11,822
10,992
9,822
8,261
6,433
4,292
2,079
Table 4D-10. Total number of 4th max metric values less than or equal to various 4th max
levels and W126 metric values greater than various W126 levels based on all
17 consecutive 3-year periods (2000-2002 to 2016-2018).
Values < 4th Max Level
AND > W126 Level
W126 Level (ppm-hrs)
19
17
15
13
11
9
7
4th Max
Level (ppb)
75
84
236
530
1,085
2,076
3,748
5,947
70
0
8
58
243
536
1,135
2,529
65
0
0
0
0
15
126
343
4D-17
-------
Table 4D-11. Total number of 4th max metric values greater than various 4th max levels and
W126 metric values greater than various W126 levels based on all 17
consecutive 3-year periods (2000-2002 to 2016-2018).
Values > 4th Max Level
AND > W126 Level
W126 Level (ppm-hrs)
19
17
15
13
11
9
7
4th Max
Level (ppb)
75
2,223
2,923
3,820
4,855
5,747
6,341
6,583
70
2,317
3,174
4,328
5,747
7,369
9,088
10,193
65
2,317
3,186
4,397
6,027
7,957
10,201
12,556
O
E
Q_
X
d)
TJ
C
CD
CNJ
|
03
3
C
C
<
O
o
CO
o
o
<60 61-65 66-70 71-75 76-80
4th Max Metric Value (ppb)
> 80
Figure 4D-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 1st
percentile or above the 99th percentile are represented by dots.
4D.3.2.2 Trends in W126 Metric
Figure 4D-9 below shows national trends in both the annual VV126 index and the 3-year
W126 metric based on the monitoring sites reporting data for the full period. Most notably, the
4D-18
-------
figure shows decreasing trends in W126 metric values, with the median value decreasing by
about 60% from 2002 to 2018. The annual W126 index shows considerable year-to-year
variability, with the median value sometimes increasing or decreasing by up to a factor of two
from one year to the next, while the 3 year average is less impacted by this inter-annual
variability, resulting in a smoother trend line.
CO
10th/90th Percentile Annual Value
Median Annual Value
10th/90th Percentile 3-year Average
— Median 3-year Average
O
CO
LO
CL CM
CL
CD
CM
O
LO
o
o^-CNjco^t-Lotor^cocDO-*— cMco^tmtor^co
OOOOOOOO O O ¦*— tt— t— t— t— ¦?— t-
ooooooooooooooooooo
CMCMCMCMCMCMCMCMCNJCNJCMCMCMCNCMCMCMCMCM
Figure 4D-9. National trends in annual W126 index values (2000-2018) and W126 metric
values (2002-2018).
Figure 4D-10 shows a map of the site-level trends in the W126 metric values from 2000-
2002 to 2016-2018. According to Figure 41) 10, nearly 90% of U.S. monitoring sites
experienced significant decreases in W126 over this period, especially in the eastern U.S. and
California where many Os monitoring sites saw decreases of 1 ppm-hr/yr or more. Many
locations in the western U.S. experienced little or no change over this period. Only seven
monitors in disparate locations showed significant increasing trends in the W126 metric during
the 2002-2018 period.
4D-19
-------
»?-
~ Decreasing > 1 ppm-hr/yr (102 sites) 0 No Significant Trend (66 sites)
v Decreasing < 1 ppm-hr/yr (463 sites) A Increasing < 1 ppm-hr/yr (7 sites)
Figure 4D-10.Map of trends in W126 metric values at U.S. O3 monitoring sites from 2000
2002 to 2016-2018.
4D.3.2.3 Comparison of Trends in the 4th Max and W126 Metrics
Figure 4D-11 shows a scatter plot comparing the trends in the 4t!l max metric values (x-
axis, ppb/yr) to the trends in the W126 metric values (y- axis, ppm-hr/yr). These trends are
calculated using the Thiel-Sen estimator as in Figure 4D-1Q. The relationship between the trends
in the two metrics was linear and positive (Pearson correlation coefficient R = 0.82), meaning a
decrease in the 4th max metric is usually accompanied by a decrease in the VV126 metric. The
slope of the regression line shown in Table 41) 12 indicates that, on average, there was a change
of approximately 0.62 ppm-hr in the W126 metric values per unit ppb change in the 4th max
metric values.
Figure 4D-12 shows scatter plots comparing the trends in the 4th max metric values fx-
axis, ppb/yr) to the trends in the W126 metric values (y axis, ppm-hr/yr) in each NOAA climate
region and the associated regression lines fit using the sites within each region. Table 4D-12
provides some summary statistics based on the regional trends comparisons. Figure 4D-12 and
Table 4D-12 show that the positive, linear relationship between the trends in the 4th max metric
values and the trends in the W126 metric values persists within each region, with Pearson
correlation coefficients ranging from 0.53 to 0.94. The regression lines shown in Figure 41) 12
4D-20
-------
with slopes listed in Table 4D-12 indicate that the Southwest region, which had the greatest
potential for sites having higher VV126 metric values relative to their 4th max metric values, also
exhibited the greatest response in the W126 metric values per unit change in the 4tb max metric
values. In Figure 4D-11 and Figure 4D-13 (as well as the West region panels in Figure 4D-12
and Figure 4D-14), there appear to be three sites in the West region with an increasing trend in
the W126 metric (slope > 0) and a decreasing trend in the 4th max metric (slope < -0.5). These
three sites are all located downwind of Los Angeles, CA and generally have 4th max metric
values of 100 ppb or greater, along with W126 metric values in the 30-50 ppm-hr range.
Figure 4D-13, Figure 41) 14 and Table 4D-13 present information similar to that shown
in Figure 4D-11, Figure 4D-12 and Table 4D-12, respectively, except that trends in annual W126
index values are presented instead of W126 metric values. The figures show that the same
general pattern occurs when comparing annual W126 index values to the 4th max metric values
as was seen for the W126 metric values. There is slightly more variability in the relationship, as
can be seen from the slight increase in scatter in the figures and the slightly lower correlation
values shown in Table 4D-12 as compared to Table 4D-11.
• •
o
d)
CO
>
o
CD
CM
i
c
••
¦O
c
0)
I-
• Central
o EastNorthCentral
o NorthEast
o Northwest
© South
o SouthEast
o Southwest
• West
• WestNorthCentral
CO
¦3
¦2
0
1
1
i rend In 4th Max Metric Value (ppb/yr)
Figure 4D-11. Scatter plot comparing trends in 4th max metric values (x-axis) to trends in
W126 metric values (y-axis).
4D-21
-------
NorthWest WestNorthCentral EastNorthCentral
o
CO
¦2
0
•3
¦2
0
West Central SouthEast
o
o
(N
CO
CO
¦3
-2
-1
SouthWest
o
-3
¦2
-1
South
0
NorthEast
O
O
CM -
-2
1
0
Figure 4D-12. Scatter plots comparing the trends in 4th max metric values (x-axis, ppb) and
W126 metric values (y-axis, ppni-hrs) based on O3 monitoring sites within
each of the nine NOAA climate regions.
4D-22
-------
Table 4D-12. Summary statistics based on regional comparisons of trends in 4th max metric
values to trends in W126 metric values.
Mean Trend in
Mean Trend in
Pearson
Number of 03
4th Max Metric
W126 Metric
Regression
Correlation
NOAA Climate Region
Sites
Value (ppb/yr)
Value (ppm-hr/yr)
Slope
Coefficient
Central
152
-1.16
-0.73
0.66
0,89
East North Central
43
-0.95
-0.41
0.40
0.73
Northeast
98
-1.41
-0.76
0.86
0,83
Northwest
11
-0.20
-0.06
0.32
0.81
South
73
1.11
-0.57
0.32
0,53
Southeast
105
-1.29
-0.73
0.83
0.94
Southwest
36
-0.37
-0.24
0.93
0,83
West
105
-0.70
-0.49
0.80
0.75
West North Central
6
-0.08
-0.18
0.43
0,83
National
629
-1.05
-0.61
0.62
0.82
® Central O SouthEast
o EastNorthCentral o SouthWest
o NorthEast • West
o Northwest • WestNorthCentral
o South
? -1 1 !
-3 -2-10 1
Trend in 4th Max Metric Value (ppb/yr)
Figure 4D-13. Scatter plot comparing trends in 4th max metric values (x-axis) and trends in
annual W126 index values (y-axis).
4D-23
-------
V 1 1 1
-3 -2-10 1
SouthWest
V -I 1 1 V -1 1 1
-3 -2-10 1 -3 -2-10 1
Figure 4D-14. Scatter plots comparing trends in 4th max metric values (x-axis, ppb) to
trends in annual W126 index values (y-axis, ppm-hrs) based on O3 monitoring
sites within each of the nine NO A A climate regions.
NorthWest
WestNorthCentral
1 ~i 1 i
-3 -2-10 1
Central
South
EastNorthCentral
CO -| ! (
' -3 -2 -1 0 1
SouthEast
-3 -2 -1 0
NorthEast
4D-24
-------
Table 4D-13. Summary statistics based on regional comparisons of trends in 4th max metric
values and trends in annual W126 index values.
Mean Trend in
Mean Trend in
Annual W126
Pearson
Number of O3
4th Max Metric
Index Value
Regression
Correlation
NOAA Climate Region
Sites
Value (ppb/yr)
(ppm-hr/yr)
Slope
Coefficient
Central
152
-1.16
-0.59
0.61
0.80
East North Central
43
-0.95
-0.29
0.30
0.63
Northeast
98
-1.41
-0.62
0.72
0.83
Northwest
11
-0.20
-0.06
0.26
0.84
South
73
-1.11
-0.57
0.33
0.53
Southeast
105
-1.29
-0.64
0.69
0.93
Southwest
36
-0.37
-0.22
0.85
0.78
West
105
-0.70
-0.45
0.71
0.74
West North Central
6
-0.08
-0.11
0.46
0.77
National
629
-1.05
-0.52
0.54
0.79
4D.3.2.4 W126 Metric Values in Federal Class I Areas
Table 4D-14 below lists the 65 federal Class I areas for which we have monitoring data
available for at least one 3-year period within the 2000-2018 period from a monitor located
either within the area boundaries or within 15 km of the boundary. This summary table indicates
the number of three-year periods for which the two metrics are available, the number of periods
where 4th max metric values were at or below 70 ppb and the range of the W126 metric values
during those periods. In total, the table is summarizing the 860 combinations of Class I area and
3-year period of which 498 have a 4th max metric value at or below 70 ppb and 362 have a 4th
max metric value above 70 ppb. In the most recent period (2016-2018), of the 58 areas for which
we have monitors, 47 sites have 4th max metric values at or below 70 ppb.
Table 4D-15 lists the Class I areas with the highest W126 metric values when the 4th max
metric value is at or below 70 ppb. Among areas with a 4th max metric value at or below 70 ppb
during any of the 3-year periods from 2000 to 2018, five areas (all located in the Southwest
region) had one or more W126 metric values above 17 ppm-hrs, with the highest W126 metric
values equal to 19 ppm-hrs and the highest annual W126 index values equal to 23 ppm-hrs, when
rounded. All seven instances where a Class I area observed a 4th max metric value at or below 70
ppb and a W126 metric value above 17 ppm-hrs occurred prior to 2011. This contrasts with the
much higher values observed in Class I areas when the current standard is not met (Table 4D-
17). In the 2016-2018 period, the W126 metric values range up to 47 ppm-hrs at sites in Class I
areas when the standard is not met, with higher values in the historical Class I dataset.
Figure 4D-15 shows the distribution of annual W126 index values in Class I areas during
3-year periods with different 4th max metric values. The full distribution of annual W126 index
values, including the minimum and maximums, increase with increasing 4th max metric values.
4D-25
-------
For example, the 99th percentile increases from about 20 ppm-hrs or lower to higher than 25
ppm-hrs for 4th max metric values at and below 70 ppb compared to 4th max metric values above
70 ppb. As indicated by Table 4D-15, the 3-year periods with the highest W126 metric values
occurring for 4th max metric values at or below 70 ppb occurred in the earlier years of the dataset
(2000-2010).
Table 4D-16 summarizes the occurrence of relatively higher annual W126 index values
in Class I areas during 3-year periods when the 4th max metric value is at or below 70 ppb. This
figure summarizes the W126 metric (i.e., 3-year average of annual W126 index values), as well
as maximum annual W126 index values in each 3-year period meeting the current standard. For
all instances of an area and 3-year period with a maximum annual W126 index value above 19
ppm-hrs, Figure 4D-16 illustrates the variation in among the annual W126 index values and the
extent to which they differ from the 3-year average.
Finally, Table 4D-17 further documents the ranges of W126 metric values occurring
during periods when the 4th max metric value was above 70 ppb, indicating the extent to which
the current standard appears to be controlling the W126 metric.
4D-26
-------
Table 4D-14. W126 metric values in Class I areas with 4th max metric values at or below 70
ppb (2000-2018).
NOAA Region
Number
Number
of 3-
Range of
W126
(number of Class
of 3-
year
Metric
1 areas1,
number of
year
periods
periods
with 4th
Values
when 4th
states1 with an
with
max<
max < 70
area in region)
State
Area Name2
data
70 ppb
ppb
Kentucky
Mammoth Cave National Park
17
7
5-10
Central
(7, 4)
Tennessee
Great Smoky Mountains National Park3 sw YP'LP'vp'RM'
BC.WP
17
6
7-10
West Virginia
Otter Creek Wilderness vp.Yp RM> *BC.Lp.wp
16
12
5-8
Michigan
Seney Wilderness Area*QA'RW *wp
15
7
4-6
EastNorthCentral
Boundary Waters Canoe Area Wilderness Area*sw
2-4
(6, 3)
Minnesota
QA, WP
Voyageurs National Park QA RM WP
13
13
2-6
North East
(6, 4)
Maine
Acadia National Park RW QA'sw wp
17
6
4-5
New Hampshire
Great Gulf Wilderness Area*wp
14
14
3-8
New Jersey
Brigantine Wilderness Area*BC
16
5
6-8
Idaho
Craters of the Moon Wilderness Area*DF'QA
11
11
6-13
Alpine Lakes Wilderness*DF'pp
17
15
2-6
Northwest
Washington
Mount Rainer National Park,DF
15
14
2-6
(29, 4)
North Cascades National Park*pp'DF'm
3
3
1-2
Olympic National ParkDF'm
7
7
1-2
Alaska
Denali National Park QA (Formerly Mt. McKinley Nat Pk)
17
17
2-4
South
(6, 4)
Arkansas
Caney Creek Wilderness Area*
12
6
4-7
Upper Buffalo Wilderness Area*SM
17
11
3-8
Texas
Big Bend National Park QA'DF'pp
16
15
7-13
Alabama
Sipsey Wilderness*wp. RM> * Yp.Lp.vp
6
1
11
Florida
St. Marks Wilderness Area*
15
10
4-11
Georgia
Cohutta Wilderness Area*wp vp'YP
17
12
5-6
Great Smoky Mountains National Park*sw YP'LP'vp'RM'
See Tennessee for monitor with
South East
(16, 6)
BC, WP
highest design va
ue (4th max)
North Carolina
Linville Gorge Wilderness Area* vp>wp'RM'YP
17
13
5-11
Shining Rock Wilderness Area*
15
8
6-10
South Carolina
Cape Romain Wilderness*
17
10
3-9
Virginia
James River Face Wilderness*wp
17
13
3-8
Shenandoah National Park wp> vp>QA'RW sw YP
17
6
6-11
SouthWest
(38. 4)
Chiricahua National MonumentDF PP
17
9
13-17
Arizona
Grand Canyon National Park DF pp QA
17
9
11-19
Mazatzal Wilderness Area DF'pp
17
2
15
4D-27
-------
NOAA Region
(number of Class
1 areas1,
number of
states1 with an
area in region)
State
Area Name2
Number
of 3-
year
periods
with
data
Number
of 3-
year
periods
with 4th
max<
70 ppb
Range of
W126
Metric
Values
when 4th
max < 70
ppb
Petrified Forest National Park
9
9
11-17
Saguaro Wilderness Area*2 DF'pp
17
5
13-15
Superstition Wilderness Area*pp
17
0
-
Yavapai Reservation*QA' pp>DF
2
2
14-15
Maroon Bells-Snowmass Wilderness. Area*QA DF
12
12
11-19
Colorado
Mesa Verde National Park*pp DF
17
15
12-18
Rocky Mountain National Park*DF' pp>QA
17
4
13-15
Weminuche Wilderness Area*DF'pp
8
3
13-18
New Mexico
San Pedro Parks Wilderness*pp DF
3
3
10-18
Utah
Canyonlands National Parkpp'DF
16
11
10-17
Zion National Park* df.pp.qa
11
6
11-18
Agua Tibia Wilderness*DF
4
0
-
Cucamonga Wilderness Area*DF'pp
17
0
-
Desolation Wilderness Area*pp
8
8-13
Joshua Tree Wilderness Area*
17
0
-
Kaiser Wilderness Area*
1
0
-
Lassen Volcanic National ParkDF'pp
17
13
7-14
West
(32, 3)
California
Pinnacles Wilderness Area*
17
8
8-10
San Gabriel Wilderness Area*DF'pp
17
0
-
San Gorgonio Wilderness Area*pp QA
13
0
-
San Jacinto Wilderness Area*pp
17
0
-
San Rafael Wilderness Area*
17
9
5-9
Sequoia National Parkpp QA'DF
17
0
-
Ventana Wilderness Area*
17
17
2-4
Yosemite National ParkDF'pp QA
17
0
-
Hawaii
Hawaii Volcanoes National Park
2
2
0
Gates of the Mountain Wilderness Area*
6
6
3-5
Montana
Glacier National Park QA'pp'DF
17
17
2-3
WestNorthCentral
(26, 4)
Northern Cheyenne Reservation*
7
7
3-5
North Dakota
Lostwood Wilderness*
13
13
4-5
Theodore Roosevelt National Park2 pp
17
17
5-7
South Dakota
Badlands Wilderness*
11
11
3-12
Wind Cave National Parkpp
10
10
5-15
4D-28
-------
NOAA Region
(number of Class
1 areas1,
number of
states1 with an
area in region)
State
Area Name2
Number
of 3-
year
periods
with
data
Number
of 3-
year
periods
with 4th
max<
70 ppb
Range of
W126
Metric
Values
when 4th
max < 70
ppb
Bridger Wilderness*
13
12
9-16
Wyoming
Grand Teton National ParkDF QA
5
5
5-8
Yellowstone National ParkDF'QA
17
17
6-11
"The monitoring site is outside of the area but within 15 km of the area boundary.
1 Areas are counted in all regions and states with a Class I area.
2 The 2-letter superscripts associated with some area names are abbreviations for species documented to be present in the area for which
there is an established exposure-response function described in Appendix 4A: QA=Quaking Aspen, BC=Black Cherry, C=Cottonwood,
DF=Douglas Fir, LP=Loblolly Pine, PP=Ponderosa Pine, RM=Red Maple, SM=Sugar Maple, VP=Virginia Pine, YP=Yellow (Tulip) Poplar.
Sources include www.NPS.gov,www.inaturalist.org/guides,www.fs.usda.gov,www.msjnha.org/trees,www.wiiderness.net
3 This area has two monitors; it is represented by the one with consistently higher values.
Table 4D-15. Highest W126 metric values occurring in Class I areas when the 4th max
metric value is at or below 70 ppb (2000-2018).
Class I Area
State/County
4th max
Range
(ppb)
3-year Periods
W126 Metric
Range
(ppm-hrs)
Areas with W126 metric values above 17
Grand Canyon National Park
AZ/Coconino
70
2006-2008
19
Maroon Bells-Snowmass Wilderness
CO/Gunnison
70
2000-2002, 2001-2003,
2002-2004
18-19
Mesa Verde National Park
CO/Montezuma
70
2006-2008
18
Weminuche Wilderness A
CO/LaPlata
70
2006-2008
18
Zion National ParkB
UT/Washington
70
2008-2010
18
Areas with W126 metric values at or below 17 and above 15
Bridger Wilderness
WY/Sublette
70
2001-2003
16
Canyonlands National Park
UT/San Juan
70
2006-2008
17
Chiricahua National Monument
AZ/Cochise
69
2006-2008
17
Maroon Bells-Snowmass Wilderness
CO/Gunnison
68
2003-2005, 2004-2006,
2005-2007
16
Mesa Verde National Park
CO/Montezuma
67-70
2000-2002, 2001-2003
2002-2004, 2003-2005
2011-2013
16-17
Petrified Forest National Park
AZ/Navajo
70
2011-2013, 2012-2014
16-17
Zion National Park
UT/Washington
70
2007-2009, 2009-2011
2012-2014
16-17
A Monitoring site is 15.0 km from area.
B Monitoring site is 3.4 km from area.
4D-29
-------
o
CD
O
IO
§-2
S
X
a>
"O
c
<£> o
C M ro
ro
3
!E o
< CJ
<60 61-65 66-70 71-75 76-80
4th Max Metric Value (ppb)
> 80
Figure 4D-15. Range of annual W126 index values in ppm-hrs observed at monitoring sites
in Class I areas based on 2000-2018 monitoring data. Values are binned
according to 4th max metric values in ppb. Boxes show 25th, 50th, and 75th
percentiles, whiskers extend to 1st and 99th percentiles, and points below the 1st
percentile or above the 99th percentile are represented by dots.
Table 4D-16. Summary of Class I area W126 index values when 4th max is at/below 70 ppb.
Time period
Total number of
Area-DVs in time
period
(Number of areas)
Among areas with design values (DVs) < 70 ppb
Number of area-DVs with
W126 metric..,
(number of areas)
\lumber of Area-DVs with
naximum annual W126 index...
number of areas)
>19
>17
<17
>19
>17
<17
2016-2018
58 (58)
0
0
47 (47)
0
3 (3)A
44 (44)
2000-2018
860 (58)
0
7 (5) a
492 (56)
15 (10)B
39 (19)?
460 (56)
A These areas are all in the Southwest Region.
B All but two of these areas are in Southwest Region; the other two are in West and West North Central Regions. The highest maximum
annual W126 index value in dataset is 23 ppm-hrs of which there are four occurrences, all from prior to 2012 in SW, The most recent
maximum annual W126 index value above 19 ppm-hrs is during 2012-2014 period (in 2012) when there are three (20, 20 and 21 ppm-hrs).
c All but eight of these areas are in Southwest Region; the others are in West, South, Central and West North Central Regions.
4D-30
-------
lO
CM
O
CM
cl in
Q_ r-
X
a>
"D
C
CD
CM
(0
3
C
c
<
~
~
a • i
s •
~ ~
~ ~
~
~
Chiricahua NM Wilderness
Desolation Wilderness
Grand Canyon NP
Maroon Bells-Snowmass Wilderness
Mazatzal Wilderness
Mesa Verde NP
Saguaro Wilderness
Weminuche Wilderness
Wind Cave National Park
Zion NP
Figure 4D-16. Range of annual W126 index values observed in each 3-year period where a
site in a Class I area had a design value meeting the current standard and had
at least one annual W126 index value greater than 19 ppm-lirs. Dots show
annual W126 index values and squares show the W126 metric value.
4D-31
-------
Table 4D-17. W126 values in Class I areas with 4th max metric values above 70 ppb (2000-2018).
NOAA Region
W126 metric >19
W126 metric >17
W126 metric >15
W126 metric <15
Number
of areas
W126
metric
range
(ppm-hrs)
Annual
W126
index
range
(ppm-hrs)
Number
of areas
W126
metric
range
(ppm-hrs)
Annual
W126
index
range
(ppm-hrs)
Number
of areas
W126
metric
range
(ppm-hrs)
Annual
W126
index
range
(ppm-hrs)
Number
of areas
W126
metric
range
(ppm-hrs)
Annual
W126
index
range
(ppm-hrs)
2016-2018
Central
0
-
-
0
-
-
0
-
-
0
-
-
EastNorthCentral
0
-
-
0
-
-
0
-
-
0
-
-
NorthEast
0
-
-
0
-
-
0
-
-
0
-
-
NorthWest
0
-
-
0
-
-
0
-
-
0
-
-
South
0
-
-
0
-
-
0
-
-
0
-
-
SouthEast
0
-
0
-
-
0
-
-
0
-
-
SouthWest
-
-
-
1
21-
19-26
3
16-19
12-21
0
-
-
West
8
21-47
17-56
8
21-47
17-56
8
21-47
17-56
0
-
-
WestNorthCentral
0
-
-
0
-
-
0
-
-
0
-
-
2000-2018
Central
1
20-31
9-37
2
18-31
9-37
2
16-31
9-37
3
9-15
6-22
EastNorthCentral
0
-
-
0
-
-
0
-
-
1
6-8
4-11
NorthEast
1
20
18-24
1
20
18-24
1
17-20
9-24
2
5-14
4-18
NorthWest
0
-
-
0
-
-
0
-
-
2
6-7
4-10
South
0
-
-
0
-
-
0
-
-
3
7-14
5-18
SouthEast
1
22
18-25
1
22
18-25
3
16-22
10-25
8
7-15
5-22
SouthWest
8
20-33
10-39
10
18-33
10-39
10
16-37
10-34
5
11-15
7-24
West
12
20-61
16-74
13
18-61
12-74
13
16-61
12-74
4
10-15
8-20
WestNorthCentral
0
-
-
0
-
-
1
17
14-19
0
-
-
4D-32
-------
4D.4 KEY LIMITATIONS AND UNCERTAINTIES
This section summarizes key limitations and uncertainties associated with aspects of the
datasets analyzed in the preceding sections. The first section summarizes key limitations and
uncertainties associated with complete dataset monitoring sites in all U.S. locations (urban and
rural), which focus on patterns and relationships across all monitoring sites. The second section
concentrates on the Class I area sites. Overall, we recognize that while the datasets analyzed are
quite extensive (e.g., more than 1,100 sites covering all 50 states in the most recent 3-year
period), there are limitations and uncertainties associated with the spatial representation of O3
monitoring sites in rural areas and Class I areas, specifically.
Analyses of data for all U.S. monitoring sites: Given that there has been a longstanding
emphasis on urban areas in the EPA's monitoring regulations, urban areas are generally well
represented in the U.S. dataset, with the effect being that the current dataset is more
representative of locations where people live than of complete spatial coverage for all areas in
the U.S., (i.e., the current dataset is more population weighted than geographically weighted). As
O3 precursor sources are also generally more associated with urban areas, one impact of this may
be a greater representation of relatively higher concentration sites. One method that has been
suggested to create a more geographically representative dataset is the use of photochemical air
quality modeling to estimate concentrations. However, this approach has been found to present
its own uncertainties with regard to estimating annual W126 index values (U.S. EPA, 2014b,
Appendix 4A), making it less useful for the current analyses.
Dataset for Class I monitoring sites: A limitation of this dataset is that it includes sites
in only 65 of the 164 Class I areas in the U.S. The representation of states containing Class I
areas is somewhat greater, with monitoring sites in Class I areas in 29 of the 36 states that have
such an area. All nine NOAA climate regions are represented. As can be seen from Figure 4D-2,
sites outside of Class I areas in the states not represented (LA, MO, NV, OK, OR, VT, WI) have
W126 metric values at or below 13 ppm-hrs during the recent 3-year period (2016-2018). Across
the states represented in the dataset, the fraction of a given state's Class I areas included in the
dataset generally ranges from about a third to 100%. An exception to that is New Mexico, for
which a monitoring site is in or near only one of the nine Class I areas in the state. This contrasts
with neighboring Arizona, also in the Southwest region and for which more than half the Class I
areas are represented in the dataset.
4D.5 SUMMARY
The preceding sections present analyses based on 19 years of O3 concentration data
reported at monitoring sites across the U.S. These analyses, intended to inform the review of the
4D-33
-------
current O3 secondary standard, investigate spatial and temporal patterns in the W126 metric
using monitoring data from 2000 to 2018 and the extent of relationships between the W126
metric, annual W126 index values and design values for the current secondary O3 standard (i.e.,
the 4th max metric). Further analyses of O3 concentrations in or near federally protected
ecosystems known as Class I areas focus on examining the levels and distributions of levels of
the W126 metric and the annual W126 index occurring in such areas when the current secondary
standard is met and also when the current secondary standard is not met.
The analyses based on recent (2016-2018) data showed that about one quarter of U.S.
sites had 4th max metric values greater than the current standard level of 70 ppb. By contrast,
only about 1 in 12 U.S. sites had W126 metric values greater than 17 ppm-hrs, and about 1 in 8
U.S. sites had W126 metric values greater than 13 ppm-hrs. There were O3 monitors exceeding
the current standard level of 70 ppb in 8 of 9 climate regions, while two regions, the West and
Southwest, had O3 monitors with W126 metric values exceeding 13 ppm-hrs.
When examining the 4th max and W126 metrics in combination, the 2016-2018 data
showed that there were many U.S. O3 sites with 4th max metric values exceeding the current
standard that had W126 metric values less than or equal to 17 ppm-hrs (173) and 13 ppm-hrs
(147). By contrast, there were relatively few sites meeting the current standard that had W126
metric values greater than 13 ppm-hrs (27); and there was a single site that had a W126 metric
value above 17 ppm-hrs. The 27 sites that met the current standard and had W126 metric values
greater than 13 ppm-hrs were located exclusively in the Southwest and West climate regions,
whereas the 147 sites that exceeded the current standard and had W126 metric values less than or
equal to 13 ppm-hrs had a much broader geographic distribution.
Among O3 monitoring sites in Federal Class I areas, few areas since 2000 have had 4th
max metric values meeting the current standard and W126 metric values above 17 ppm-hrs, the
most recent of which occurred during the 2012-2014 period. These instances are all in or near
Class I areas in the Southwest region, with the highest (19 ppm-hrs) occurring during the 2006-
2008 period.
The analysis of inter-annual variability shows that the distribution of annual W126 index
deviations from their respective 3-year averages generally increase with increasing W126 metric
values. For sites with W126 metric values below 20 ppm-hrs (e.g., focusing on W126 metric
values that have occurred with 4th max metric values at or below 70 ppb), the annual deviation
was generally within 5 ppm-hrs. Additionally, well over 99% of 4th max metric values meeting
the current standard were associated with annual W126 index values of less than or equal to 19
ppm-hrs.
The trends analysis showed that both W126 metric values and annual W126 index values
have generally decreased since 2000, with U.S. median W126 metric values decreasing by over
4D-34
-------
60%, from nearly 17 ppm-hrs in 2002 to less than 7 ppm-hrs in 2018. A substantial number of
U.S. sites have experienced decreases of over 10 ppm-hrs in the past decade, particularly in the
eastern U.S.
The analysis comparing trends in the 4th max metric values and to trends in the W126
metric values based on data from 2000-2018 showed that there was a positive, linear relationship
between the per-year changes in the 4th max and W126 metrics. Nationally, the W126 metric
values decreased by approximately 0.6 ppm-hr per unit ppb decrease in the 4th max metric
values. This relationship varied across the NOAA climate regions. The Southwest and West
regions which showed the greatest potential for exceeding only the W126 levels of interest also
showed the greatest improvement in the W126 metric values per unit decrease in 4th max metric
values. This analysis indicates that W126 metric values in those areas not meeting the current
standard would be expected to decline as the 4th max metric values are reduced to meet the
current standard, consistent with the relationship shown in Figure 4D-11.
4D-35
-------
4D.6 REFERENCES
Karl, T and Koss, WJ (1984). Regional and national monthly, seasonal, and annual temperature
weighted by area, 1895-1983. 4-3. National Environmental Satellite and Data
Information Service (NESDIS). Asheville, NC.
U.S. EPA (2014a). Policy Assessment for the Review of National Ambient Air Quality
Standards for Ozone (Final Report). Office of Air Quality Planning and Standards, Health
and Environmental Impacts Divison. Research Triangle Park, NC. U.S. EPA. EPA-
452/R-14-006 August 2014. Available at:
https://nepis. epa.gov/Exe/ZyPDF. cgi?Dockey=Pl 00KCZ5. txt.
U.S. EPA (2014b). Welfare Risk and Exposure Assessment for Ozone (Final) with Executive
Summary and Appendices. Office of Air Quality Planning and Standards. Research
Triangle Park, NC. U.S. EPA. EPA-452/P-14-005a, EPA-452/R-14-005b and EPA-
452/R-14-005c February 2014. Available at:
https.Y/nepis. epa.gov/Exe/ZyPURL. cgi?Dockey=Pl OOKQLJ. txt.
Wells, B. (2014). Memorandum to Ozone NAAQS Review Docket (EPA-HQ-OAR-2008-0699).
Comparison of Ozone Metrics Considered in Current NAAQS Review.. November 20,
2014.. Docket ID No. EPA-HQ-OAR-2008-0699. Office of Air Quality Planning and
Standards Research Triangle Park, NC. Available at:
https://www.regulations.gov/contentStreamer?documentId=EPA-HQ-OAR-2008-0699-
0155&contentType=pdf.
Wells, B. (2015). Memorandum to Ozone NAAQS Review Docket (EPA-HQ-OAR-2008-0699).
Expanded Comparison of Ozone Metrics Considered in the Current NAAQS Review.
September 28, 2015.. Docket ID No. EPA-HQ-OAR-2008-0699. Office of Air Quality
Planning and Standards Research Triangle Park, NC. Available at:
https://www.regulations.gov/contentStreamer?documentId=EPA-HQ-OAR-2008-0699-
4325&contentType=pdf.
4D-36
-------
APPENDIX 4E
OZONE WELFARE EFFECTS AND RELATED ECOSYSTEM
SERVICES AND PUBLIC WELFARE ASPECTS
-------
Table 4E-1. Ecosystem services and aspects of public welfare potentially affected by the
different types of O3 welfare effects.
03 Effect*
Aspect of Public Welfare Potentially Affected (Examples)6
Ecosystem
Services c
Visible foliar injury
• Appearance and scenic beauty of forests wilderness areas, including
federal, tribal, state, municipally protected areas
• Quality of specific agricultural crops, plant leaf products
• Appearance of plants in residential/commercial areas (ornamentals)
Cultural
Recreation
Provisioning
Reduced vegetation growth
• Food, raw material, and unique biological material and product
production
• Shade provision
• Quality of plants of cultural importance to Native American Tribes
• Changes to national yield and prices
Cultural
Provisioning
Reduced plant reproduction
Reduced yield and quality
of agricultural crops
Reduced productivity in
terrestrial ecosystems
Reduced carbon
sequestration in terrestrial
systems
• Regulation/control of climatological features and meteorological
phenomena
• Changes in pollution removal in urban areas
Regulating
Supporting
Increased tree mortality
• Regulation/control of wildfires
• Regulation of erosion and soil stability
• Decline of ecosystem services provided by trees (see Table 4E-2)
Regulating
Cultural
Supporting
Provisioning
Alteration of terrestrial
community composition
• Intrinsic value of areas specially protected from anthropogenic
degradation
• Production of preferred species of timber
• Preservation of unique or endangered ecosystems or species
• Species diversity in protected areas
Cultural
Provisioning
Supporting
Alteration of belowground
biogeochemical cycles
• Soil quality
• Soil nutrient cycling, decomposition, and availability
• Carbon storage
• Regulation of soil fauna and microbial communities
• Water quality and resource management
• Regulation of hydraulic flow
Supporting
Regulating
Alteration of ecosystem
water cycling
• Water quality and resource management
• Regulation of hydraulic flow
Provisioning
Regulating
Supporting
Altered of herbivore growth
and reproduction
• Food sources, habitat, and protection for native fauna
Supporting
Regulating
Alteration of plant insect
signaling
• Plant-pollinator interactions
• Timber and agricultural plant resistance to insect pest damage
Supporting
Provisioning
Radiative forcing and
related climate effects
• Regulation/control of meteorological phenomena
Regulating
NOTE: Sources include ISA (Appendix 8. Fiaure 8-1 and Table 8-1) and 2014 WREA (Section 5).
A Effects identified as causally or likely causally related to O3 (draft ISA, Appendices 8 and 9).
B Examples provided in Costanza et al., 2017) and 2014 WREA, Section 5 (U.S. EPA, 2014)
c Description of Ecosystem Services in 2013 ISA, Section 9.4.1.2 and in the 2014 WREA, Section 5.1:
• Regulating: Services of importance for human society such as carbon sequestration, climate and water regulation, protection from
natural hazards such as floods, avalanches, or rock-fall, water and air purification, and disease and pest regulation.
• Supporting: The services needed by all the other ecosystem services, either indirectly or directly, such biomass production, production
of atmospheric 02, soil formation and retention, nutrient cycling, water cycling, biodiversity, and provisioning of habitat.
• Provisioning: Services that include market goods, such as food, water, fiber, and medicinal and cosmetic products
• Cultural, services that satisfy human spiritual and aesthetic appreciation of ecosystems and their components including recreational and
other nonmaterial benefits
4E-1
-------
Table 4E-2. Ecosystem services and specific uses of the 11 tree species with robust E-R
functions for reduced growth.
Tree Species
O3 Effect
Role in Ecosystems and Public Uses
Black Cherry
Prunus serotina
Biomass loss,
Visible foliar injury
Cabinets, furniture, paneling, veneers, crafts, toys; Cough remedy, tonic, sedative;
Flavor for rum and brandy; Wine making and jellies; Food and habitat for song birds,
game birds, and mammals
Eastern White Pine
Pinus strobus
Biomass loss
Commercial timber, furniture, woodworking, and Christmas trees; Medicinal uses as
expectorant and antiseptic; Food and habitat for song birds and mammals; Used to
stabilize strip mine soils
Quaking Aspen
Populus tremuloides
Biomass loss,
Visible foliar injury
Commercial logging for pulp, flake-board, pallets, boxes, and plywood; Products
including matchsticks, tongue depressors, and ice cream sticks; Valued for its white
bark and brilliant fall color; Important as a fire break Habitat for variety of wildlife;
Traditional native American use as a food source
Yellow (Tulip) Poplar
Liriodendron tulipifera
Biomass loss,
Visible foliar injury
Furniture stock, veneer, and pulpwood; Street, shade, or ornamental tree - unusual
flowers; Food and habitat for wildlife; Rapid growth for reforestation projects
Ponderosa Pine
Pinus ponderosa
Biomass loss,
Visible foliar injury
Lumber for cabinets and construction; Ornamental and erosion control use;
Recreation areas; Food and habitat for many bird species, including the red-winged
blackbird, chickadee, finches, and nuthatches
Red Alder
Alnus rubra
Biomass loss,
Visible foliar injury
Commercial use in products such as furniture, cabinets, and millwork; Preferred for
smoked salmon; Dyes for baskets, hides, moccasins; Medicinal use for rheumatic
pain, diarrhea, stomach cramps - the bark contains salicin, a chemical similar to
aspirin; Roots used for baskets; Food and habitat for mammals and birds - dam and
lodge construction for beavers; Conservation and erosion control
Red MapleA
Acer rubrum
Biomass loss
One of the most abundant and widespread trees in eastern U.S. Used for
revegetation, especially in riparian buffers and landscaping, where it is valued for its
brilliant fall foliage, some lumber and syrup production; Important wildlife browse
food, especially for elk and white-tailed deer in winter, also leaves are important food
source for some species of butterflies and moths.
Virginia Pine
Pinus virginiana
Biomass loss,
Visible foliar injury
Pulpwood, stabilization of strip mine spoil banks and severely eroded soils; Nesting
for woodpeckers, food and habitat for songbirds and small mammals
Sugar Maple
Acer saccharum
Biomass loss
Commercial syrup production; Native Americans used sap as a candy, beverage -
fresh or fermented into beer, soured into vinegar and used to cook meat; Valued for
its fall foliage and as an ornamental; Commercial logging for furniture, flooring,
paneling, and veneer; Woodenware, musical instruments; Food and habitat for many
birds and mammals
Loblolly Pine*
Biomass loss,
visible foliar injury
Most important and widely cultivated timber species in the southern U.S.; Furniture,
pulpwood, plywood, composite boards, posts, poles, pilings, crates, boxes, pallets.
Also planted to stabilize eroded or damaged soils. It can be used for shade or
ornamental trees, as well as bark mulch; Provides habitat, food and cover for white-
tailed deer, gray squirrel, fox squirrel, bobwhite quail and wild turkey, red-cockaded
woodpeckers, and a variety of other birds and small mammals. Standing dead trees
are frequently used for cavity nests by woodpeckers.
Douglas Fir
Pseudotsuga menziesii
Biomass loss
Commercial timber and used for Christmas trees; Medicinal uses, spiritual and
cultural uses for several Native American tribes; Spotted owl habitat; Food and
habitat for mammals including antelope and mountain sheep
Sources: 2014 WREA, USDA-NRCS (2013); Burns and Honkala, 1990).
ARed maple information from https://www.srs.fs.usda.gov/pubs/misc/ag_654/volume_2/silvics_v2.pdf
"Loblolly pine use information from
https://projects.ncsu.edu/project/dendrology/index/plantae/vascular/seedplants/gymnosperms/conifers/pine/pinus/australes/loblolly/ioblollypine.html.
4E-2
-------
REFERENCES
Burns, RM and Honkala, BH, Eds. (1990). Volume 1: Conifers: Abies balsamea (L.) mill.
Balsam fir. Agriculture Handbook 654. U.S. Department of Agriculture, U.S. Forest
Service Washington, DC.
Costanza, R, De Groot, R, Braat, L, Kubiszewski, I, Fioramonti, L, Sutton, P, Farber, S and
Grasso, M (2017). Twenty years of ecosystem services: How far have we come and how
far do we still need to go? Ecosystem Services 28: 1-16.
U.S. EPA (2014). Welfare Risk and Exposure Assessment for Ozone (Final). . Office of Air
Quality Planning and Standards. Research Triangle Park, NC. U.S. EPA. EPA-452/P-14-
005a August 2014. Available at:
https://nepis. epa.gov/Exe/ZyPURL. cgi?Dockey=P100KB9D. txt.
4E-3
-------
>WC
CTQ 3 3
2 ^ —
3
<¦5
i-i
O
3
3
CD
3
r-K
rD
o
r-K
o"
3
fD
a-
oo
r-K
&3
r-K
a>
go
50
fD
C/5
fD
D3
i-l
o
3-
H
i-i
S"
3
CTQ
I
fD
D3
o
fD
>
TJ
&3
1-1
J77*
o
&3
3
a-
m
< o
tl" c
o EL
3^'
fD HH
3 2
¦ D3
3
3
3 3'
"0
D3 pj
S 3
w a-
a £
S. 3
00 Q^
£* P
CD
o-
c/3
TJ
e
cr
cnj o
o w
00
o o
D
o
m
TJ
>
on
fO
fO
o
o
o
------- |